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.
447 Simple audio dynamic range compression/expansion filter.
449 The filter accepts the following options:
453 Set contrast. Default is 33. Allowed range is between 0 and 100.
458 Copy the input audio source unchanged to the output. This is mainly useful for
463 Apply cross fade from one input audio stream to another input audio stream.
464 The cross fade is applied for specified duration near the end of first stream.
466 The filter accepts the following options:
470 Specify the number of samples for which the cross fade effect has to last.
471 At the end of the cross fade effect the first input audio will be completely
472 silent. Default is 44100.
475 Specify the duration of the cross fade effect. See
476 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
477 for the accepted syntax.
478 By default the duration is determined by @var{nb_samples}.
479 If set this option is used instead of @var{nb_samples}.
482 Should first stream end overlap with second stream start. Default is enabled.
485 Set curve for cross fade transition for first stream.
488 Set curve for cross fade transition for second stream.
490 For description of available curve types see @ref{afade} filter description.
497 Cross fade from one input to another:
499 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
503 Cross fade from one input to another but without overlapping:
505 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
510 Split audio stream into several bands.
512 This filter splits audio stream into two or more frequency ranges.
513 Summing all streams back will give flat output.
515 The filter accepts the following options:
519 Set split frequencies. Those must be positive and increasing.
522 Set filter order, can be @var{2nd}, @var{4th} or @var{8th}.
523 Default is @var{4th}.
528 Reduce audio bit resolution.
530 This filter is bit crusher with enhanced functionality. A bit crusher
531 is used to audibly reduce number of bits an audio signal is sampled
532 with. This doesn't change the bit depth at all, it just produces the
533 effect. Material reduced in bit depth sounds more harsh and "digital".
534 This filter is able to even round to continuous values instead of discrete
536 Additionally it has a D/C offset which results in different crushing of
537 the lower and the upper half of the signal.
538 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
540 Another feature of this filter is the logarithmic mode.
541 This setting switches from linear distances between bits to logarithmic ones.
542 The result is a much more "natural" sounding crusher which doesn't gate low
543 signals for example. The human ear has a logarithmic perception,
544 so this kind of crushing is much more pleasant.
545 Logarithmic crushing is also able to get anti-aliased.
547 The filter accepts the following options:
563 Can be linear: @code{lin} or logarithmic: @code{log}.
572 Set sample reduction.
575 Enable LFO. By default disabled.
586 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
590 Remove impulsive noise from input audio.
592 Samples detected as impulsive noise are replaced by interpolated samples using
593 autoregressive modelling.
597 Set window size, in milliseconds. Allowed range is from @code{10} to
598 @code{100}. Default value is @code{55} milliseconds.
599 This sets size of window which will be processed at once.
602 Set window overlap, in percentage of window size. Allowed range is from
603 @code{50} to @code{95}. Default value is @code{75} percent.
604 Setting this to a very high value increases impulsive noise removal but makes
605 whole process much slower.
608 Set autoregression order, in percentage of window size. Allowed range is from
609 @code{0} to @code{25}. Default value is @code{2} percent. This option also
610 controls quality of interpolated samples using neighbour good samples.
613 Set threshold value. Allowed range is from @code{1} to @code{100}.
614 Default value is @code{2}.
615 This controls the strength of impulsive noise which is going to be removed.
616 The lower value, the more samples will be detected as impulsive noise.
619 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
620 @code{10}. Default value is @code{2}.
621 If any two samples detected as noise are spaced less than this value then any
622 sample between those two samples will be also detected as noise.
627 It accepts the following values:
630 Select overlap-add method. Even not interpolated samples are slightly
631 changed with this method.
634 Select overlap-save method. Not interpolated samples remain unchanged.
637 Default value is @code{a}.
641 Remove clipped samples from input audio.
643 Samples detected as clipped are replaced by interpolated samples using
644 autoregressive modelling.
648 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
649 Default value is @code{55} milliseconds.
650 This sets size of window which will be processed at once.
653 Set window overlap, in percentage of window size. Allowed range is from @code{50}
654 to @code{95}. Default value is @code{75} percent.
657 Set autoregression order, in percentage of window size. Allowed range is from
658 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
659 quality of interpolated samples using neighbour good samples.
662 Set threshold value. Allowed range is from @code{1} to @code{100}.
663 Default value is @code{10}. Higher values make clip detection less aggressive.
666 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
667 Default value is @code{1000}. Higher values make clip detection less aggressive.
672 It accepts the following values:
675 Select overlap-add method. Even not interpolated samples are slightly changed
679 Select overlap-save method. Not interpolated samples remain unchanged.
682 Default value is @code{a}.
687 Delay one or more audio channels.
689 Samples in delayed channel are filled with silence.
691 The filter accepts the following option:
695 Set list of delays in milliseconds for each channel separated by '|'.
696 Unused delays will be silently ignored. If number of given delays is
697 smaller than number of channels all remaining channels will not be delayed.
698 If you want to delay exact number of samples, append 'S' to number.
699 If you want instead to delay in seconds, append 's' to number.
702 Use last set delay for all remaining channels. By default is disabled.
703 This option if enabled changes how option @code{delays} is interpreted.
710 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
711 the second channel (and any other channels that may be present) unchanged.
717 Delay second channel by 500 samples, the third channel by 700 samples and leave
718 the first channel (and any other channels that may be present) unchanged.
724 Delay all channels by same number of samples:
726 adelay=delays=64S:all=1
730 @section aderivative, aintegral
732 Compute derivative/integral of audio stream.
734 Applying both filters one after another produces original audio.
738 Apply echoing to the input audio.
740 Echoes are reflected sound and can occur naturally amongst mountains
741 (and sometimes large buildings) when talking or shouting; digital echo
742 effects emulate this behaviour and are often used to help fill out the
743 sound of a single instrument or vocal. The time difference between the
744 original signal and the reflection is the @code{delay}, and the
745 loudness of the reflected signal is the @code{decay}.
746 Multiple echoes can have different delays and decays.
748 A description of the accepted parameters follows.
752 Set input gain of reflected signal. Default is @code{0.6}.
755 Set output gain of reflected signal. Default is @code{0.3}.
758 Set list of time intervals in milliseconds between original signal and reflections
759 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
760 Default is @code{1000}.
763 Set list of loudness of reflected signals separated by '|'.
764 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
765 Default is @code{0.5}.
772 Make it sound as if there are twice as many instruments as are actually playing:
774 aecho=0.8:0.88:60:0.4
778 If delay is very short, then it sounds like a (metallic) robot playing music:
784 A longer delay will sound like an open air concert in the mountains:
786 aecho=0.8:0.9:1000:0.3
790 Same as above but with one more mountain:
792 aecho=0.8:0.9:1000|1800:0.3|0.25
797 Audio emphasis filter creates or restores material directly taken from LPs or
798 emphased CDs with different filter curves. E.g. to store music on vinyl the
799 signal has to be altered by a filter first to even out the disadvantages of
800 this recording medium.
801 Once the material is played back the inverse filter has to be applied to
802 restore the distortion of the frequency response.
804 The filter accepts the following options:
814 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
815 use @code{production} mode. Default is @code{reproduction} mode.
818 Set filter type. Selects medium. Can be one of the following:
830 select Compact Disc (CD).
836 select 50µs (FM-KF).
838 select 75µs (FM-KF).
844 Modify an audio signal according to the specified expressions.
846 This filter accepts one or more expressions (one for each channel),
847 which are evaluated and used to modify a corresponding audio signal.
849 It accepts the following parameters:
853 Set the '|'-separated expressions list for each separate channel. If
854 the number of input channels is greater than the number of
855 expressions, the last specified expression is used for the remaining
858 @item channel_layout, c
859 Set output channel layout. If not specified, the channel layout is
860 specified by the number of expressions. If set to @samp{same}, it will
861 use by default the same input channel layout.
864 Each expression in @var{exprs} can contain the following constants and functions:
868 channel number of the current expression
871 number of the evaluated sample, starting from 0
877 time of the evaluated sample expressed in seconds
880 @item nb_out_channels
881 input and output number of channels
884 the value of input channel with number @var{CH}
887 Note: this filter is slow. For faster processing you should use a
896 aeval=val(ch)/2:c=same
900 Invert phase of the second channel:
909 Apply fade-in/out effect to input audio.
911 A description of the accepted parameters follows.
915 Specify the effect type, can be either @code{in} for fade-in, or
916 @code{out} for a fade-out effect. Default is @code{in}.
918 @item start_sample, ss
919 Specify the number of the start sample for starting to apply the fade
920 effect. Default is 0.
923 Specify the number of samples for which the fade effect has to last. At
924 the end of the fade-in effect the output audio will have the same
925 volume as the input audio, at the end of the fade-out transition
926 the output audio will be silence. Default is 44100.
929 Specify the start time of the fade effect. Default is 0.
930 The value must be specified as a time duration; see
931 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
932 for the accepted syntax.
933 If set this option is used instead of @var{start_sample}.
936 Specify the duration of the fade effect. See
937 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
938 for the accepted syntax.
939 At the end of the fade-in effect the output audio will have the same
940 volume as the input audio, at the end of the fade-out transition
941 the output audio will be silence.
942 By default the duration is determined by @var{nb_samples}.
943 If set this option is used instead of @var{nb_samples}.
946 Set curve for fade transition.
948 It accepts the following values:
951 select triangular, linear slope (default)
953 select quarter of sine wave
955 select half of sine wave
957 select exponential sine wave
961 select inverted parabola
975 select inverted quarter of sine wave
977 select inverted half of sine wave
979 select double-exponential seat
981 select double-exponential sigmoid
983 select logistic sigmoid
993 Fade in first 15 seconds of audio:
999 Fade out last 25 seconds of a 900 seconds audio:
1001 afade=t=out:st=875:d=25
1006 Denoise audio samples with FFT.
1008 A description of the accepted parameters follows.
1012 Set the noise reduction in dB, allowed range is 0.01 to 97.
1013 Default value is 12 dB.
1016 Set the noise floor in dB, allowed range is -80 to -20.
1017 Default value is -50 dB.
1022 It accepts the following values:
1031 Select shellac noise.
1034 Select custom noise, defined in @code{bn} option.
1036 Default value is white noise.
1040 Set custom band noise for every one of 15 bands.
1041 Bands are separated by ' ' or '|'.
1044 Set the residual floor in dB, allowed range is -80 to -20.
1045 Default value is -38 dB.
1048 Enable noise tracking. By default is disabled.
1049 With this enabled, noise floor is automatically adjusted.
1052 Enable residual tracking. By default is disabled.
1055 Set the output mode.
1057 It accepts the following values:
1060 Pass input unchanged.
1063 Pass noise filtered out.
1068 Default value is @var{o}.
1072 @subsection Commands
1074 This filter supports the following commands:
1076 @item sample_noise, sn
1077 Start or stop measuring noise profile.
1078 Syntax for the command is : "start" or "stop" string.
1079 After measuring noise profile is stopped it will be
1080 automatically applied in filtering.
1082 @item noise_reduction, nr
1083 Change noise reduction. Argument is single float number.
1084 Syntax for the command is : "@var{noise_reduction}"
1086 @item noise_floor, nf
1087 Change noise floor. Argument is single float number.
1088 Syntax for the command is : "@var{noise_floor}"
1090 @item output_mode, om
1091 Change output mode operation.
1092 Syntax for the command is : "i", "o" or "n" string.
1096 Apply arbitrary expressions to samples in frequency domain.
1100 Set frequency domain real expression for each separate channel separated
1101 by '|'. Default is "re".
1102 If the number of input channels is greater than the number of
1103 expressions, the last specified expression is used for the remaining
1107 Set frequency domain imaginary expression for each separate channel
1108 separated by '|'. Default is "im".
1110 Each expression in @var{real} and @var{imag} can contain the following
1111 constants and functions:
1118 current frequency bin number
1121 number of available bins
1124 channel number of the current expression
1133 current real part of frequency bin of current channel
1136 current imaginary part of frequency bin of current channel
1139 Return the value of real part of frequency bin at location (@var{bin},@var{channel})
1142 Return the value of imaginary part of frequency bin at location (@var{bin},@var{channel})
1146 Set window size. Allowed range is from 16 to 131072.
1147 Default is @code{4096}
1150 Set window function. Default is @code{hann}.
1153 Set window overlap. If set to 1, the recommended overlap for selected
1154 window function will be picked. Default is @code{0.75}.
1157 @subsection Examples
1161 Leave almost only low frequencies in audio:
1163 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1167 Apply robotize effect:
1169 afftfilt="real='hypot(re,im)*sin(0)':imag='hypot(re,im)*cos(0)':win_size=512:overlap=0.75"
1173 Apply whisper effect:
1175 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"
1182 Apply an arbitrary Frequency Impulse Response filter.
1184 This filter is designed for applying long FIR filters,
1185 up to 60 seconds long.
1187 It can be used as component for digital crossover filters,
1188 room equalization, cross talk cancellation, wavefield synthesis,
1189 auralization, ambiophonics, ambisonics and spatialization.
1191 This filter uses the second stream as FIR coefficients.
1192 If the second stream holds a single channel, it will be used
1193 for all input channels in the first stream, otherwise
1194 the number of channels in the second stream must be same as
1195 the number of channels in the first stream.
1197 It accepts the following parameters:
1201 Set dry gain. This sets input gain.
1204 Set wet gain. This sets final output gain.
1207 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1210 Enable applying gain measured from power of IR.
1212 Set which approach to use for auto gain measurement.
1216 Do not apply any gain.
1219 select peak gain, very conservative approach. This is default value.
1222 select DC gain, limited application.
1225 select gain to noise approach, this is most popular one.
1229 Set gain to be applied to IR coefficients before filtering.
1230 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1233 Set format of IR stream. Can be @code{mono} or @code{input}.
1234 Default is @code{input}.
1237 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1238 Allowed range is 0.1 to 60 seconds.
1241 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1242 By default it is disabled.
1245 Set for which IR channel to display frequency response. By default is first channel
1246 displayed. This option is used only when @var{response} is enabled.
1249 Set video stream size. This option is used only when @var{response} is enabled.
1252 Set video stream frame rate. This option is used only when @var{response} is enabled.
1255 Set minimal partition size used for convolution. Default is @var{8192}.
1256 Allowed range is from @var{8} to @var{32768}.
1257 Lower values decreases latency at cost of higher CPU usage.
1260 Set maximal partition size used for convolution. Default is @var{8192}.
1261 Allowed range is from @var{8} to @var{32768}.
1262 Lower values may increase CPU usage.
1265 @subsection Examples
1269 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1271 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1278 Set output format constraints for the input audio. The framework will
1279 negotiate the most appropriate format to minimize conversions.
1281 It accepts the following parameters:
1285 A '|'-separated list of requested sample formats.
1288 A '|'-separated list of requested sample rates.
1290 @item channel_layouts
1291 A '|'-separated list of requested channel layouts.
1293 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1294 for the required syntax.
1297 If a parameter is omitted, all values are allowed.
1299 Force the output to either unsigned 8-bit or signed 16-bit stereo
1301 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1306 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1307 processing reduces disturbing noise between useful signals.
1309 Gating is done by detecting the volume below a chosen level @var{threshold}
1310 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1311 floor is set via @var{range}. Because an exact manipulation of the signal
1312 would cause distortion of the waveform the reduction can be levelled over
1313 time. This is done by setting @var{attack} and @var{release}.
1315 @var{attack} determines how long the signal has to fall below the threshold
1316 before any reduction will occur and @var{release} sets the time the signal
1317 has to rise above the threshold to reduce the reduction again.
1318 Shorter signals than the chosen attack time will be left untouched.
1322 Set input level before filtering.
1323 Default is 1. Allowed range is from 0.015625 to 64.
1326 Set the mode of operation. Can be @code{upward} or @code{downward}.
1327 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
1328 will be amplified, expanding dynamic range in upward direction.
1329 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
1332 Set the level of gain reduction when the signal is below the threshold.
1333 Default is 0.06125. Allowed range is from 0 to 1.
1334 Setting this to 0 disables reduction and then filter behaves like expander.
1337 If a signal rises above this level the gain reduction is released.
1338 Default is 0.125. Allowed range is from 0 to 1.
1341 Set a ratio by which the signal is reduced.
1342 Default is 2. Allowed range is from 1 to 9000.
1345 Amount of milliseconds the signal has to rise above the threshold before gain
1347 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1350 Amount of milliseconds the signal has to fall below the threshold before the
1351 reduction is increased again. Default is 250 milliseconds.
1352 Allowed range is from 0.01 to 9000.
1355 Set amount of amplification of signal after processing.
1356 Default is 1. Allowed range is from 1 to 64.
1359 Curve the sharp knee around the threshold to enter gain reduction more softly.
1360 Default is 2.828427125. Allowed range is from 1 to 8.
1363 Choose if exact signal should be taken for detection or an RMS like one.
1364 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1367 Choose if the average level between all channels or the louder channel affects
1369 Default is @code{average}. Can be @code{average} or @code{maximum}.
1374 Apply an arbitrary Infinite Impulse Response filter.
1376 It accepts the following parameters:
1380 Set numerator/zeros coefficients.
1383 Set denominator/poles coefficients.
1395 Set coefficients format.
1401 Z-plane zeros/poles, cartesian (default)
1403 Z-plane zeros/poles, polar radians
1405 Z-plane zeros/poles, polar degrees
1409 Set kind of processing.
1410 Can be @code{d} - direct or @code{s} - serial cascading. Default is @code{s}.
1413 Set filtering precision.
1417 double-precision floating-point (default)
1419 single-precision floating-point
1427 How much to use filtered signal in output. Default is 1.
1428 Range is between 0 and 1.
1431 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1432 By default it is disabled.
1435 Set for which IR channel to display frequency response. By default is first channel
1436 displayed. This option is used only when @var{response} is enabled.
1439 Set video stream size. This option is used only when @var{response} is enabled.
1442 Coefficients in @code{tf} format are separated by spaces and are in ascending
1445 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1446 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1449 Different coefficients and gains can be provided for every channel, in such case
1450 use '|' to separate coefficients or gains. Last provided coefficients will be
1451 used for all remaining channels.
1453 @subsection Examples
1457 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1459 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
1463 Same as above but in @code{zp} format:
1465 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
1471 The limiter prevents an input signal from rising over a desired threshold.
1472 This limiter uses lookahead technology to prevent your signal from distorting.
1473 It means that there is a small delay after the signal is processed. Keep in mind
1474 that the delay it produces is the attack time you set.
1476 The filter accepts the following options:
1480 Set input gain. Default is 1.
1483 Set output gain. Default is 1.
1486 Don't let signals above this level pass the limiter. Default is 1.
1489 The limiter will reach its attenuation level in this amount of time in
1490 milliseconds. Default is 5 milliseconds.
1493 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1494 Default is 50 milliseconds.
1497 When gain reduction is always needed ASC takes care of releasing to an
1498 average reduction level rather than reaching a reduction of 0 in the release
1502 Select how much the release time is affected by ASC, 0 means nearly no changes
1503 in release time while 1 produces higher release times.
1506 Auto level output signal. Default is enabled.
1507 This normalizes audio back to 0dB if enabled.
1510 Depending on picked setting it is recommended to upsample input 2x or 4x times
1511 with @ref{aresample} before applying this filter.
1515 Apply a two-pole all-pass filter with central frequency (in Hz)
1516 @var{frequency}, and filter-width @var{width}.
1517 An all-pass filter changes the audio's frequency to phase relationship
1518 without changing its frequency to amplitude relationship.
1520 The filter accepts the following options:
1524 Set frequency in Hz.
1527 Set method to specify band-width of filter.
1542 Specify the band-width of a filter in width_type units.
1545 How much to use filtered signal in output. Default is 1.
1546 Range is between 0 and 1.
1549 Specify which channels to filter, by default all available are filtered.
1552 @subsection Commands
1554 This filter supports the following commands:
1557 Change allpass frequency.
1558 Syntax for the command is : "@var{frequency}"
1561 Change allpass width_type.
1562 Syntax for the command is : "@var{width_type}"
1565 Change allpass width.
1566 Syntax for the command is : "@var{width}"
1570 Syntax for the command is : "@var{mix}"
1577 The filter accepts the following options:
1581 Set the number of loops. Setting this value to -1 will result in infinite loops.
1585 Set maximal number of samples. Default is 0.
1588 Set first sample of loop. Default is 0.
1594 Merge two or more audio streams into a single multi-channel stream.
1596 The filter accepts the following options:
1601 Set the number of inputs. Default is 2.
1605 If the channel layouts of the inputs are disjoint, and therefore compatible,
1606 the channel layout of the output will be set accordingly and the channels
1607 will be reordered as necessary. If the channel layouts of the inputs are not
1608 disjoint, the output will have all the channels of the first input then all
1609 the channels of the second input, in that order, and the channel layout of
1610 the output will be the default value corresponding to the total number of
1613 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1614 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1615 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1616 first input, b1 is the first channel of the second input).
1618 On the other hand, if both input are in stereo, the output channels will be
1619 in the default order: a1, a2, b1, b2, and the channel layout will be
1620 arbitrarily set to 4.0, which may or may not be the expected value.
1622 All inputs must have the same sample rate, and format.
1624 If inputs do not have the same duration, the output will stop with the
1627 @subsection Examples
1631 Merge two mono files into a stereo stream:
1633 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1637 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1639 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
1645 Mixes multiple audio inputs into a single output.
1647 Note that this filter only supports float samples (the @var{amerge}
1648 and @var{pan} audio filters support many formats). If the @var{amix}
1649 input has integer samples then @ref{aresample} will be automatically
1650 inserted to perform the conversion to float samples.
1654 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1656 will mix 3 input audio streams to a single output with the same duration as the
1657 first input and a dropout transition time of 3 seconds.
1659 It accepts the following parameters:
1663 The number of inputs. If unspecified, it defaults to 2.
1666 How to determine the end-of-stream.
1670 The duration of the longest input. (default)
1673 The duration of the shortest input.
1676 The duration of the first input.
1680 @item dropout_transition
1681 The transition time, in seconds, for volume renormalization when an input
1682 stream ends. The default value is 2 seconds.
1685 Specify weight of each input audio stream as sequence.
1686 Each weight is separated by space. By default all inputs have same weight.
1691 Multiply first audio stream with second audio stream and store result
1692 in output audio stream. Multiplication is done by multiplying each
1693 sample from first stream with sample at same position from second stream.
1695 With this element-wise multiplication one can create amplitude fades and
1696 amplitude modulations.
1698 @section anequalizer
1700 High-order parametric multiband equalizer for each channel.
1702 It accepts the following parameters:
1706 This option string is in format:
1707 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1708 Each equalizer band is separated by '|'.
1712 Set channel number to which equalization will be applied.
1713 If input doesn't have that channel the entry is ignored.
1716 Set central frequency for band.
1717 If input doesn't have that frequency the entry is ignored.
1720 Set band width in hertz.
1723 Set band gain in dB.
1726 Set filter type for band, optional, can be:
1730 Butterworth, this is default.
1741 With this option activated frequency response of anequalizer is displayed
1745 Set video stream size. Only useful if curves option is activated.
1748 Set max gain that will be displayed. Only useful if curves option is activated.
1749 Setting this to a reasonable value makes it possible to display gain which is derived from
1750 neighbour bands which are too close to each other and thus produce higher gain
1751 when both are activated.
1754 Set frequency scale used to draw frequency response in video output.
1755 Can be linear or logarithmic. Default is logarithmic.
1758 Set color for each channel curve which is going to be displayed in video stream.
1759 This is list of color names separated by space or by '|'.
1760 Unrecognised or missing colors will be replaced by white color.
1763 @subsection Examples
1767 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1768 for first 2 channels using Chebyshev type 1 filter:
1770 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1774 @subsection Commands
1776 This filter supports the following commands:
1779 Alter existing filter parameters.
1780 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1782 @var{fN} is existing filter number, starting from 0, if no such filter is available
1784 @var{freq} set new frequency parameter.
1785 @var{width} set new width parameter in herz.
1786 @var{gain} set new gain parameter in dB.
1788 Full filter invocation with asendcmd may look like this:
1789 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1794 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1796 Each sample is adjusted by looking for other samples with similar contexts. This
1797 context similarity is defined by comparing their surrounding patches of size
1798 @option{p}. Patches are searched in an area of @option{r} around the sample.
1800 The filter accepts the following options:
1804 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
1807 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
1808 Default value is 2 milliseconds.
1811 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
1812 Default value is 6 milliseconds.
1815 Set the output mode.
1817 It accepts the following values:
1820 Pass input unchanged.
1823 Pass noise filtered out.
1828 Default value is @var{o}.
1832 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
1835 @subsection Commands
1837 This filter supports the following commands:
1840 Change denoise strength. Argument is single float number.
1841 Syntax for the command is : "@var{s}"
1845 Syntax for the command is : "i", "o" or "n" string.
1849 Apply Normalized Least-Mean-Squares algorithm to the first audio stream using the second audio stream.
1851 This adaptive filter is used to mimic a desired filter by finding the filter coefficients that
1852 relate to producing the least mean square of the error signal (difference between the desired,
1853 2nd input audio stream and the actual signal, the 1st input audio stream).
1855 A description of the accepted options follows.
1868 Set the filter leakage.
1871 It accepts the following values:
1880 Pass filtered samples.
1883 Pass difference between desired and filtered samples.
1885 Default value is @var{o}.
1889 @subsection Examples
1893 One of many usages of this filter is noise reduction, input audio is filtered
1894 with same samples that are delayed by fixed amount, one such example for stereo audio is:
1896 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
1900 @subsection Commands
1902 This filter supports the same commands as options, excluding option @code{order}.
1906 Pass the audio source unchanged to the output.
1910 Pad the end of an audio stream with silence.
1912 This can be used together with @command{ffmpeg} @option{-shortest} to
1913 extend audio streams to the same length as the video stream.
1915 A description of the accepted options follows.
1919 Set silence packet size. Default value is 4096.
1922 Set the number of samples of silence to add to the end. After the
1923 value is reached, the stream is terminated. This option is mutually
1924 exclusive with @option{whole_len}.
1927 Set the minimum total number of samples in the output audio stream. If
1928 the value is longer than the input audio length, silence is added to
1929 the end, until the value is reached. This option is mutually exclusive
1930 with @option{pad_len}.
1933 Specify the duration of samples of silence to add. See
1934 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1935 for the accepted syntax. Used only if set to non-zero value.
1938 Specify the minimum total duration in the output audio stream. See
1939 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1940 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
1941 the input audio length, silence is added to the end, until the value is reached.
1942 This option is mutually exclusive with @option{pad_dur}
1945 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
1946 nor @option{whole_dur} option is set, the filter will add silence to the end of
1947 the input stream indefinitely.
1949 @subsection Examples
1953 Add 1024 samples of silence to the end of the input:
1959 Make sure the audio output will contain at least 10000 samples, pad
1960 the input with silence if required:
1962 apad=whole_len=10000
1966 Use @command{ffmpeg} to pad the audio input with silence, so that the
1967 video stream will always result the shortest and will be converted
1968 until the end in the output file when using the @option{shortest}
1971 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
1976 Add a phasing effect to the input audio.
1978 A phaser filter creates series of peaks and troughs in the frequency spectrum.
1979 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
1981 A description of the accepted parameters follows.
1985 Set input gain. Default is 0.4.
1988 Set output gain. Default is 0.74
1991 Set delay in milliseconds. Default is 3.0.
1994 Set decay. Default is 0.4.
1997 Set modulation speed in Hz. Default is 0.5.
2000 Set modulation type. Default is triangular.
2002 It accepts the following values:
2011 Audio pulsator is something between an autopanner and a tremolo.
2012 But it can produce funny stereo effects as well. Pulsator changes the volume
2013 of the left and right channel based on a LFO (low frequency oscillator) with
2014 different waveforms and shifted phases.
2015 This filter have the ability to define an offset between left and right
2016 channel. An offset of 0 means that both LFO shapes match each other.
2017 The left and right channel are altered equally - a conventional tremolo.
2018 An offset of 50% means that the shape of the right channel is exactly shifted
2019 in phase (or moved backwards about half of the frequency) - pulsator acts as
2020 an autopanner. At 1 both curves match again. Every setting in between moves the
2021 phase shift gapless between all stages and produces some "bypassing" sounds with
2022 sine and triangle waveforms. The more you set the offset near 1 (starting from
2023 the 0.5) the faster the signal passes from the left to the right speaker.
2025 The filter accepts the following options:
2029 Set input gain. By default it is 1. Range is [0.015625 - 64].
2032 Set output gain. By default it is 1. Range is [0.015625 - 64].
2035 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
2036 sawup or sawdown. Default is sine.
2039 Set modulation. Define how much of original signal is affected by the LFO.
2042 Set left channel offset. Default is 0. Allowed range is [0 - 1].
2045 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
2048 Set pulse width. Default is 1. Allowed range is [0 - 2].
2051 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
2054 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
2058 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
2062 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
2063 if timing is set to hz.
2069 Resample the input audio to the specified parameters, using the
2070 libswresample library. If none are specified then the filter will
2071 automatically convert between its input and output.
2073 This filter is also able to stretch/squeeze the audio data to make it match
2074 the timestamps or to inject silence / cut out audio to make it match the
2075 timestamps, do a combination of both or do neither.
2077 The filter accepts the syntax
2078 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
2079 expresses a sample rate and @var{resampler_options} is a list of
2080 @var{key}=@var{value} pairs, separated by ":". See the
2081 @ref{Resampler Options,,"Resampler Options" section in the
2082 ffmpeg-resampler(1) manual,ffmpeg-resampler}
2083 for the complete list of supported options.
2085 @subsection Examples
2089 Resample the input audio to 44100Hz:
2095 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2096 samples per second compensation:
2098 aresample=async=1000
2104 Reverse an audio clip.
2106 Warning: This filter requires memory to buffer the entire clip, so trimming
2109 @subsection Examples
2113 Take the first 5 seconds of a clip, and reverse it.
2115 atrim=end=5,areverse
2121 Reduce noise from speech using Recurrent Neural Networks.
2123 This filter accepts the following options:
2127 Set train model file to load. This option is always required.
2130 @section asetnsamples
2132 Set the number of samples per each output audio frame.
2134 The last output packet may contain a different number of samples, as
2135 the filter will flush all the remaining samples when the input audio
2138 The filter accepts the following options:
2142 @item nb_out_samples, n
2143 Set the number of frames per each output audio frame. The number is
2144 intended as the number of samples @emph{per each channel}.
2145 Default value is 1024.
2148 If set to 1, the filter will pad the last audio frame with zeroes, so
2149 that the last frame will contain the same number of samples as the
2150 previous ones. Default value is 1.
2153 For example, to set the number of per-frame samples to 1234 and
2154 disable padding for the last frame, use:
2156 asetnsamples=n=1234:p=0
2161 Set the sample rate without altering the PCM data.
2162 This will result in a change of speed and pitch.
2164 The filter accepts the following options:
2167 @item sample_rate, r
2168 Set the output sample rate. Default is 44100 Hz.
2173 Show a line containing various information for each input audio frame.
2174 The input audio is not modified.
2176 The shown line contains a sequence of key/value pairs of the form
2177 @var{key}:@var{value}.
2179 The following values are shown in the output:
2183 The (sequential) number of the input frame, starting from 0.
2186 The presentation timestamp of the input frame, in time base units; the time base
2187 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2190 The presentation timestamp of the input frame in seconds.
2193 position of the frame in the input stream, -1 if this information in
2194 unavailable and/or meaningless (for example in case of synthetic audio)
2203 The sample rate for the audio frame.
2206 The number of samples (per channel) in the frame.
2209 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2210 audio, the data is treated as if all the planes were concatenated.
2212 @item plane_checksums
2213 A list of Adler-32 checksums for each data plane.
2217 Apply audio soft clipping.
2219 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2220 along a smooth curve, rather than the abrupt shape of hard-clipping.
2222 This filter accepts the following options:
2226 Set type of soft-clipping.
2228 It accepts the following values:
2240 Set additional parameter which controls sigmoid function.
2244 Automatic Speech Recognition
2246 This filter uses PocketSphinx for speech recognition. To enable
2247 compilation of this filter, you need to configure FFmpeg with
2248 @code{--enable-pocketsphinx}.
2250 It accepts the following options:
2254 Set sampling rate of input audio. Defaults is @code{16000}.
2255 This need to match speech models, otherwise one will get poor results.
2258 Set dictionary containing acoustic model files.
2261 Set pronunciation dictionary.
2264 Set language model file.
2267 Set language model set.
2270 Set which language model to use.
2273 Set output for log messages.
2276 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2281 Display time domain statistical information about the audio channels.
2282 Statistics are calculated and displayed for each audio channel and,
2283 where applicable, an overall figure is also given.
2285 It accepts the following option:
2288 Short window length in seconds, used for peak and trough RMS measurement.
2289 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2293 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2294 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2297 Available keys for each channel are:
2339 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2340 this @code{lavfi.astats.Overall.Peak_count}.
2342 For description what each key means read below.
2345 Set number of frame after which stats are going to be recalculated.
2346 Default is disabled.
2348 @item measure_perchannel
2349 Select the entries which need to be measured per channel. The metadata keys can
2350 be used as flags, default is @option{all} which measures everything.
2351 @option{none} disables all per channel measurement.
2353 @item measure_overall
2354 Select the entries which need to be measured overall. The metadata keys can
2355 be used as flags, default is @option{all} which measures everything.
2356 @option{none} disables all overall measurement.
2360 A description of each shown parameter follows:
2364 Mean amplitude displacement from zero.
2367 Minimal sample level.
2370 Maximal sample level.
2372 @item Min difference
2373 Minimal difference between two consecutive samples.
2375 @item Max difference
2376 Maximal difference between two consecutive samples.
2378 @item Mean difference
2379 Mean difference between two consecutive samples.
2380 The average of each difference between two consecutive samples.
2382 @item RMS difference
2383 Root Mean Square difference between two consecutive samples.
2387 Standard peak and RMS level measured in dBFS.
2391 Peak and trough values for RMS level measured over a short window.
2394 Standard ratio of peak to RMS level (note: not in dB).
2397 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2398 (i.e. either @var{Min level} or @var{Max level}).
2401 Number of occasions (not the number of samples) that the signal attained either
2402 @var{Min level} or @var{Max level}.
2405 Overall bit depth of audio. Number of bits used for each sample.
2408 Measured dynamic range of audio in dB.
2410 @item Zero crossings
2411 Number of points where the waveform crosses the zero level axis.
2413 @item Zero crossings rate
2414 Rate of Zero crossings and number of audio samples.
2421 The filter accepts exactly one parameter, the audio tempo. If not
2422 specified then the filter will assume nominal 1.0 tempo. Tempo must
2423 be in the [0.5, 100.0] range.
2425 Note that tempo greater than 2 will skip some samples rather than
2426 blend them in. If for any reason this is a concern it is always
2427 possible to daisy-chain several instances of atempo to achieve the
2428 desired product tempo.
2430 @subsection Examples
2434 Slow down audio to 80% tempo:
2440 To speed up audio to 300% tempo:
2446 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2448 atempo=sqrt(3),atempo=sqrt(3)
2452 @subsection Commands
2454 This filter supports the following commands:
2457 Change filter tempo scale factor.
2458 Syntax for the command is : "@var{tempo}"
2463 Trim the input so that the output contains one continuous subpart of the input.
2465 It accepts the following parameters:
2468 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2469 sample with the timestamp @var{start} will be the first sample in the output.
2472 Specify time of the first audio sample that will be dropped, i.e. the
2473 audio sample immediately preceding the one with the timestamp @var{end} will be
2474 the last sample in the output.
2477 Same as @var{start}, except this option sets the start timestamp in samples
2481 Same as @var{end}, except this option sets the end timestamp in samples instead
2485 The maximum duration of the output in seconds.
2488 The number of the first sample that should be output.
2491 The number of the first sample that should be dropped.
2494 @option{start}, @option{end}, and @option{duration} are expressed as time
2495 duration specifications; see
2496 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2498 Note that the first two sets of the start/end options and the @option{duration}
2499 option look at the frame timestamp, while the _sample options simply count the
2500 samples that pass through the filter. So start/end_pts and start/end_sample will
2501 give different results when the timestamps are wrong, inexact or do not start at
2502 zero. Also note that this filter does not modify the timestamps. If you wish
2503 to have the output timestamps start at zero, insert the asetpts filter after the
2506 If multiple start or end options are set, this filter tries to be greedy and
2507 keep all samples that match at least one of the specified constraints. To keep
2508 only the part that matches all the constraints at once, chain multiple atrim
2511 The defaults are such that all the input is kept. So it is possible to set e.g.
2512 just the end values to keep everything before the specified time.
2517 Drop everything except the second minute of input:
2519 ffmpeg -i INPUT -af atrim=60:120
2523 Keep only the first 1000 samples:
2525 ffmpeg -i INPUT -af atrim=end_sample=1000
2532 Apply a two-pole Butterworth band-pass filter with central
2533 frequency @var{frequency}, and (3dB-point) band-width width.
2534 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2535 instead of the default: constant 0dB peak gain.
2536 The filter roll off at 6dB per octave (20dB per decade).
2538 The filter accepts the following options:
2542 Set the filter's central frequency. Default is @code{3000}.
2545 Constant skirt gain if set to 1. Defaults to 0.
2548 Set method to specify band-width of filter.
2563 Specify the band-width of a filter in width_type units.
2566 How much to use filtered signal in output. Default is 1.
2567 Range is between 0 and 1.
2570 Specify which channels to filter, by default all available are filtered.
2573 @subsection Commands
2575 This filter supports the following commands:
2578 Change bandpass frequency.
2579 Syntax for the command is : "@var{frequency}"
2582 Change bandpass width_type.
2583 Syntax for the command is : "@var{width_type}"
2586 Change bandpass width.
2587 Syntax for the command is : "@var{width}"
2590 Change bandpass mix.
2591 Syntax for the command is : "@var{mix}"
2596 Apply a two-pole Butterworth band-reject filter with central
2597 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
2598 The filter roll off at 6dB per octave (20dB per decade).
2600 The filter accepts the following options:
2604 Set the filter's central frequency. Default is @code{3000}.
2607 Set method to specify band-width of filter.
2622 Specify the band-width of a filter in width_type units.
2625 How much to use filtered signal in output. Default is 1.
2626 Range is between 0 and 1.
2629 Specify which channels to filter, by default all available are filtered.
2632 @subsection Commands
2634 This filter supports the following commands:
2637 Change bandreject frequency.
2638 Syntax for the command is : "@var{frequency}"
2641 Change bandreject width_type.
2642 Syntax for the command is : "@var{width_type}"
2645 Change bandreject width.
2646 Syntax for the command is : "@var{width}"
2649 Change bandreject mix.
2650 Syntax for the command is : "@var{mix}"
2653 @section bass, lowshelf
2655 Boost or cut the bass (lower) frequencies of the audio using a two-pole
2656 shelving filter with a response similar to that of a standard
2657 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
2659 The filter accepts the following options:
2663 Give the gain at 0 Hz. Its useful range is about -20
2664 (for a large cut) to +20 (for a large boost).
2665 Beware of clipping when using a positive gain.
2668 Set the filter's central frequency and so can be used
2669 to extend or reduce the frequency range to be boosted or cut.
2670 The default value is @code{100} Hz.
2673 Set method to specify band-width of filter.
2688 Determine how steep is the filter's shelf transition.
2691 How much to use filtered signal in output. Default is 1.
2692 Range is between 0 and 1.
2695 Specify which channels to filter, by default all available are filtered.
2698 @subsection Commands
2700 This filter supports the following commands:
2703 Change bass frequency.
2704 Syntax for the command is : "@var{frequency}"
2707 Change bass width_type.
2708 Syntax for the command is : "@var{width_type}"
2712 Syntax for the command is : "@var{width}"
2716 Syntax for the command is : "@var{gain}"
2720 Syntax for the command is : "@var{mix}"
2725 Apply a biquad IIR filter with the given coefficients.
2726 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
2727 are the numerator and denominator coefficients respectively.
2728 and @var{channels}, @var{c} specify which channels to filter, by default all
2729 available are filtered.
2731 @subsection Commands
2733 This filter supports the following commands:
2741 Change biquad parameter.
2742 Syntax for the command is : "@var{value}"
2745 How much to use filtered signal in output. Default is 1.
2746 Range is between 0 and 1.
2750 Bauer stereo to binaural transformation, which improves headphone listening of
2751 stereo audio records.
2753 To enable compilation of this filter you need to configure FFmpeg with
2754 @code{--enable-libbs2b}.
2756 It accepts the following parameters:
2760 Pre-defined crossfeed level.
2764 Default level (fcut=700, feed=50).
2767 Chu Moy circuit (fcut=700, feed=60).
2770 Jan Meier circuit (fcut=650, feed=95).
2775 Cut frequency (in Hz).
2784 Remap input channels to new locations.
2786 It accepts the following parameters:
2789 Map channels from input to output. The argument is a '|'-separated list of
2790 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
2791 @var{in_channel} form. @var{in_channel} can be either the name of the input
2792 channel (e.g. FL for front left) or its index in the input channel layout.
2793 @var{out_channel} is the name of the output channel or its index in the output
2794 channel layout. If @var{out_channel} is not given then it is implicitly an
2795 index, starting with zero and increasing by one for each mapping.
2797 @item channel_layout
2798 The channel layout of the output stream.
2801 If no mapping is present, the filter will implicitly map input channels to
2802 output channels, preserving indices.
2804 @subsection Examples
2808 For example, assuming a 5.1+downmix input MOV file,
2810 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
2812 will create an output WAV file tagged as stereo from the downmix channels of
2816 To fix a 5.1 WAV improperly encoded in AAC's native channel order
2818 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
2822 @section channelsplit
2824 Split each channel from an input audio stream into a separate output stream.
2826 It accepts the following parameters:
2828 @item channel_layout
2829 The channel layout of the input stream. The default is "stereo".
2831 A channel layout describing the channels to be extracted as separate output streams
2832 or "all" to extract each input channel as a separate stream. The default is "all".
2834 Choosing channels not present in channel layout in the input will result in an error.
2837 @subsection Examples
2841 For example, assuming a stereo input MP3 file,
2843 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
2845 will create an output Matroska file with two audio streams, one containing only
2846 the left channel and the other the right channel.
2849 Split a 5.1 WAV file into per-channel files:
2851 ffmpeg -i in.wav -filter_complex
2852 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
2853 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
2854 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
2859 Extract only LFE from a 5.1 WAV file:
2861 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
2862 -map '[LFE]' lfe.wav
2867 Add a chorus effect to the audio.
2869 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
2871 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
2872 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
2873 The modulation depth defines the range the modulated delay is played before or after
2874 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
2875 sound tuned around the original one, like in a chorus where some vocals are slightly
2878 It accepts the following parameters:
2881 Set input gain. Default is 0.4.
2884 Set output gain. Default is 0.4.
2887 Set delays. A typical delay is around 40ms to 60ms.
2899 @subsection Examples
2905 chorus=0.7:0.9:55:0.4:0.25:2
2911 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
2915 Fuller sounding chorus with three delays:
2917 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
2922 Compress or expand the audio's dynamic range.
2924 It accepts the following parameters:
2930 A list of times in seconds for each channel over which the instantaneous level
2931 of the input signal is averaged to determine its volume. @var{attacks} refers to
2932 increase of volume and @var{decays} refers to decrease of volume. For most
2933 situations, the attack time (response to the audio getting louder) should be
2934 shorter than the decay time, because the human ear is more sensitive to sudden
2935 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
2936 a typical value for decay is 0.8 seconds.
2937 If specified number of attacks & decays is lower than number of channels, the last
2938 set attack/decay will be used for all remaining channels.
2941 A list of points for the transfer function, specified in dB relative to the
2942 maximum possible signal amplitude. Each key points list must be defined using
2943 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
2944 @code{x0/y0 x1/y1 x2/y2 ....}
2946 The input values must be in strictly increasing order but the transfer function
2947 does not have to be monotonically rising. The point @code{0/0} is assumed but
2948 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
2949 function are @code{-70/-70|-60/-20|1/0}.
2952 Set the curve radius in dB for all joints. It defaults to 0.01.
2955 Set the additional gain in dB to be applied at all points on the transfer
2956 function. This allows for easy adjustment of the overall gain.
2960 Set an initial volume, in dB, to be assumed for each channel when filtering
2961 starts. This permits the user to supply a nominal level initially, so that, for
2962 example, a very large gain is not applied to initial signal levels before the
2963 companding has begun to operate. A typical value for audio which is initially
2964 quiet is -90 dB. It defaults to 0.
2967 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
2968 delayed before being fed to the volume adjuster. Specifying a delay
2969 approximately equal to the attack/decay times allows the filter to effectively
2970 operate in predictive rather than reactive mode. It defaults to 0.
2974 @subsection Examples
2978 Make music with both quiet and loud passages suitable for listening to in a
2981 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
2984 Another example for audio with whisper and explosion parts:
2986 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
2990 A noise gate for when the noise is at a lower level than the signal:
2992 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
2996 Here is another noise gate, this time for when the noise is at a higher level
2997 than the signal (making it, in some ways, similar to squelch):
2999 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3003 2:1 compression starting at -6dB:
3005 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3009 2:1 compression starting at -9dB:
3011 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3015 2:1 compression starting at -12dB:
3017 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3021 2:1 compression starting at -18dB:
3023 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3027 3:1 compression starting at -15dB:
3029 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3035 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3041 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
3045 Hard limiter at -6dB:
3047 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3051 Hard limiter at -12dB:
3053 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3057 Hard noise gate at -35 dB:
3059 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3065 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3069 @section compensationdelay
3071 Compensation Delay Line is a metric based delay to compensate differing
3072 positions of microphones or speakers.
3074 For example, you have recorded guitar with two microphones placed in
3075 different locations. Because the front of sound wave has fixed speed in
3076 normal conditions, the phasing of microphones can vary and depends on
3077 their location and interposition. The best sound mix can be achieved when
3078 these microphones are in phase (synchronized). Note that a distance of
3079 ~30 cm between microphones makes one microphone capture the signal in
3080 antiphase to the other microphone. That makes the final mix sound moody.
3081 This filter helps to solve phasing problems by adding different delays
3082 to each microphone track and make them synchronized.
3084 The best result can be reached when you take one track as base and
3085 synchronize other tracks one by one with it.
3086 Remember that synchronization/delay tolerance depends on sample rate, too.
3087 Higher sample rates will give more tolerance.
3089 The filter accepts the following parameters:
3093 Set millimeters distance. This is compensation distance for fine tuning.
3097 Set cm distance. This is compensation distance for tightening distance setup.
3101 Set meters distance. This is compensation distance for hard distance setup.
3105 Set dry amount. Amount of unprocessed (dry) signal.
3109 Set wet amount. Amount of processed (wet) signal.
3113 Set temperature in degrees Celsius. This is the temperature of the environment.
3118 Apply headphone crossfeed filter.
3120 Crossfeed is the process of blending the left and right channels of stereo
3122 It is mainly used to reduce extreme stereo separation of low frequencies.
3124 The intent is to produce more speaker like sound to the listener.
3126 The filter accepts the following options:
3130 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3131 This sets gain of low shelf filter for side part of stereo image.
3132 Default is -6dB. Max allowed is -30db when strength is set to 1.
3135 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3136 This sets cut off frequency of low shelf filter. Default is cut off near
3137 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3140 Set input gain. Default is 0.9.
3143 Set output gain. Default is 1.
3146 @section crystalizer
3147 Simple algorithm to expand audio dynamic range.
3149 The filter accepts the following options:
3153 Sets the intensity of effect (default: 2.0). Must be in range between 0.0
3154 (unchanged sound) to 10.0 (maximum effect).
3157 Enable clipping. By default is enabled.
3161 Apply a DC shift to the audio.
3163 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3164 in the recording chain) from the audio. The effect of a DC offset is reduced
3165 headroom and hence volume. The @ref{astats} filter can be used to determine if
3166 a signal has a DC offset.
3170 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3174 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3175 used to prevent clipping.
3180 Apply de-essing to the audio samples.
3184 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3188 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3192 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3196 Set the output mode.
3198 It accepts the following values:
3201 Pass input unchanged.
3204 Pass ess filtered out.
3209 Default value is @var{o}.
3215 Measure audio dynamic range.
3217 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3218 is found in transition material. And anything less that 8 have very poor dynamics
3219 and is very compressed.
3221 The filter accepts the following options:
3225 Set window length in seconds used to split audio into segments of equal length.
3226 Default is 3 seconds.
3230 Dynamic Audio Normalizer.
3232 This filter applies a certain amount of gain to the input audio in order
3233 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3234 contrast to more "simple" normalization algorithms, the Dynamic Audio
3235 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3236 This allows for applying extra gain to the "quiet" sections of the audio
3237 while avoiding distortions or clipping the "loud" sections. In other words:
3238 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3239 sections, in the sense that the volume of each section is brought to the
3240 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3241 this goal *without* applying "dynamic range compressing". It will retain 100%
3242 of the dynamic range *within* each section of the audio file.
3246 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3247 Default is 500 milliseconds.
3248 The Dynamic Audio Normalizer processes the input audio in small chunks,
3249 referred to as frames. This is required, because a peak magnitude has no
3250 meaning for just a single sample value. Instead, we need to determine the
3251 peak magnitude for a contiguous sequence of sample values. While a "standard"
3252 normalizer would simply use the peak magnitude of the complete file, the
3253 Dynamic Audio Normalizer determines the peak magnitude individually for each
3254 frame. The length of a frame is specified in milliseconds. By default, the
3255 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3256 been found to give good results with most files.
3257 Note that the exact frame length, in number of samples, will be determined
3258 automatically, based on the sampling rate of the individual input audio file.
3261 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3262 number. Default is 31.
3263 Probably the most important parameter of the Dynamic Audio Normalizer is the
3264 @code{window size} of the Gaussian smoothing filter. The filter's window size
3265 is specified in frames, centered around the current frame. For the sake of
3266 simplicity, this must be an odd number. Consequently, the default value of 31
3267 takes into account the current frame, as well as the 15 preceding frames and
3268 the 15 subsequent frames. Using a larger window results in a stronger
3269 smoothing effect and thus in less gain variation, i.e. slower gain
3270 adaptation. Conversely, using a smaller window results in a weaker smoothing
3271 effect and thus in more gain variation, i.e. faster gain adaptation.
3272 In other words, the more you increase this value, the more the Dynamic Audio
3273 Normalizer will behave like a "traditional" normalization filter. On the
3274 contrary, the more you decrease this value, the more the Dynamic Audio
3275 Normalizer will behave like a dynamic range compressor.
3278 Set the target peak value. This specifies the highest permissible magnitude
3279 level for the normalized audio input. This filter will try to approach the
3280 target peak magnitude as closely as possible, but at the same time it also
3281 makes sure that the normalized signal will never exceed the peak magnitude.
3282 A frame's maximum local gain factor is imposed directly by the target peak
3283 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3284 It is not recommended to go above this value.
3287 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3288 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3289 factor for each input frame, i.e. the maximum gain factor that does not
3290 result in clipping or distortion. The maximum gain factor is determined by
3291 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3292 additionally bounds the frame's maximum gain factor by a predetermined
3293 (global) maximum gain factor. This is done in order to avoid excessive gain
3294 factors in "silent" or almost silent frames. By default, the maximum gain
3295 factor is 10.0, For most inputs the default value should be sufficient and
3296 it usually is not recommended to increase this value. Though, for input
3297 with an extremely low overall volume level, it may be necessary to allow even
3298 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3299 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3300 Instead, a "sigmoid" threshold function will be applied. This way, the
3301 gain factors will smoothly approach the threshold value, but never exceed that
3305 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3306 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3307 This means that the maximum local gain factor for each frame is defined
3308 (only) by the frame's highest magnitude sample. This way, the samples can
3309 be amplified as much as possible without exceeding the maximum signal
3310 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3311 Normalizer can also take into account the frame's root mean square,
3312 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3313 determine the power of a time-varying signal. It is therefore considered
3314 that the RMS is a better approximation of the "perceived loudness" than
3315 just looking at the signal's peak magnitude. Consequently, by adjusting all
3316 frames to a constant RMS value, a uniform "perceived loudness" can be
3317 established. If a target RMS value has been specified, a frame's local gain
3318 factor is defined as the factor that would result in exactly that RMS value.
3319 Note, however, that the maximum local gain factor is still restricted by the
3320 frame's highest magnitude sample, in order to prevent clipping.
3323 Enable channels coupling. By default is enabled.
3324 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3325 amount. This means the same gain factor will be applied to all channels, i.e.
3326 the maximum possible gain factor is determined by the "loudest" channel.
3327 However, in some recordings, it may happen that the volume of the different
3328 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3329 In this case, this option can be used to disable the channel coupling. This way,
3330 the gain factor will be determined independently for each channel, depending
3331 only on the individual channel's highest magnitude sample. This allows for
3332 harmonizing the volume of the different channels.
3335 Enable DC bias correction. By default is disabled.
3336 An audio signal (in the time domain) is a sequence of sample values.
3337 In the Dynamic Audio Normalizer these sample values are represented in the
3338 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3339 audio signal, or "waveform", should be centered around the zero point.
3340 That means if we calculate the mean value of all samples in a file, or in a
3341 single frame, then the result should be 0.0 or at least very close to that
3342 value. If, however, there is a significant deviation of the mean value from
3343 0.0, in either positive or negative direction, this is referred to as a
3344 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3345 Audio Normalizer provides optional DC bias correction.
3346 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3347 the mean value, or "DC correction" offset, of each input frame and subtract
3348 that value from all of the frame's sample values which ensures those samples
3349 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3350 boundaries, the DC correction offset values will be interpolated smoothly
3351 between neighbouring frames.
3353 @item altboundary, b
3354 Enable alternative boundary mode. By default is disabled.
3355 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3356 around each frame. This includes the preceding frames as well as the
3357 subsequent frames. However, for the "boundary" frames, located at the very
3358 beginning and at the very end of the audio file, not all neighbouring
3359 frames are available. In particular, for the first few frames in the audio
3360 file, the preceding frames are not known. And, similarly, for the last few
3361 frames in the audio file, the subsequent frames are not known. Thus, the
3362 question arises which gain factors should be assumed for the missing frames
3363 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3364 to deal with this situation. The default boundary mode assumes a gain factor
3365 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3366 "fade out" at the beginning and at the end of the input, respectively.
3369 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3370 By default, the Dynamic Audio Normalizer does not apply "traditional"
3371 compression. This means that signal peaks will not be pruned and thus the
3372 full dynamic range will be retained within each local neighbourhood. However,
3373 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3374 normalization algorithm with a more "traditional" compression.
3375 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3376 (thresholding) function. If (and only if) the compression feature is enabled,
3377 all input frames will be processed by a soft knee thresholding function prior
3378 to the actual normalization process. Put simply, the thresholding function is
3379 going to prune all samples whose magnitude exceeds a certain threshold value.
3380 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3381 value. Instead, the threshold value will be adjusted for each individual
3383 In general, smaller parameters result in stronger compression, and vice versa.
3384 Values below 3.0 are not recommended, because audible distortion may appear.
3389 Make audio easier to listen to on headphones.
3391 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3392 so that when listened to on headphones the stereo image is moved from
3393 inside your head (standard for headphones) to outside and in front of
3394 the listener (standard for speakers).
3400 Apply a two-pole peaking equalisation (EQ) filter. With this
3401 filter, the signal-level at and around a selected frequency can
3402 be increased or decreased, whilst (unlike bandpass and bandreject
3403 filters) that at all other frequencies is unchanged.
3405 In order to produce complex equalisation curves, this filter can
3406 be given several times, each with a different central frequency.
3408 The filter accepts the following options:
3412 Set the filter's central frequency in Hz.
3415 Set method to specify band-width of filter.
3430 Specify the band-width of a filter in width_type units.
3433 Set the required gain or attenuation in dB.
3434 Beware of clipping when using a positive gain.
3437 How much to use filtered signal in output. Default is 1.
3438 Range is between 0 and 1.
3441 Specify which channels to filter, by default all available are filtered.
3444 @subsection Examples
3447 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
3449 equalizer=f=1000:t=h:width=200:g=-10
3453 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
3455 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
3459 @subsection Commands
3461 This filter supports the following commands:
3464 Change equalizer frequency.
3465 Syntax for the command is : "@var{frequency}"
3468 Change equalizer width_type.
3469 Syntax for the command is : "@var{width_type}"
3472 Change equalizer width.
3473 Syntax for the command is : "@var{width}"
3476 Change equalizer gain.
3477 Syntax for the command is : "@var{gain}"
3480 Change equalizer mix.
3481 Syntax for the command is : "@var{mix}"
3484 @section extrastereo
3486 Linearly increases the difference between left and right channels which
3487 adds some sort of "live" effect to playback.
3489 The filter accepts the following options:
3493 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
3494 (average of both channels), with 1.0 sound will be unchanged, with
3495 -1.0 left and right channels will be swapped.
3498 Enable clipping. By default is enabled.
3501 @section firequalizer
3502 Apply FIR Equalization using arbitrary frequency response.
3504 The filter accepts the following option:
3508 Set gain curve equation (in dB). The expression can contain variables:
3511 the evaluated frequency
3515 channel number, set to 0 when multichannels evaluation is disabled
3517 channel id, see libavutil/channel_layout.h, set to the first channel id when
3518 multichannels evaluation is disabled
3522 channel_layout, see libavutil/channel_layout.h
3527 @item gain_interpolate(f)
3528 interpolate gain on frequency f based on gain_entry
3529 @item cubic_interpolate(f)
3530 same as gain_interpolate, but smoother
3532 This option is also available as command. Default is @code{gain_interpolate(f)}.
3535 Set gain entry for gain_interpolate function. The expression can
3539 store gain entry at frequency f with value g
3541 This option is also available as command.
3544 Set filter delay in seconds. Higher value means more accurate.
3545 Default is @code{0.01}.
3548 Set filter accuracy in Hz. Lower value means more accurate.
3549 Default is @code{5}.
3552 Set window function. Acceptable values are:
3555 rectangular window, useful when gain curve is already smooth
3557 hann window (default)
3563 3-terms continuous 1st derivative nuttall window
3565 minimum 3-terms discontinuous nuttall window
3567 4-terms continuous 1st derivative nuttall window
3569 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
3571 blackman-harris window
3577 If enabled, use fixed number of audio samples. This improves speed when
3578 filtering with large delay. Default is disabled.
3581 Enable multichannels evaluation on gain. Default is disabled.
3584 Enable zero phase mode by subtracting timestamp to compensate delay.
3585 Default is disabled.
3588 Set scale used by gain. Acceptable values are:
3591 linear frequency, linear gain
3593 linear frequency, logarithmic (in dB) gain (default)
3595 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
3597 logarithmic frequency, logarithmic gain
3601 Set file for dumping, suitable for gnuplot.
3604 Set scale for dumpfile. Acceptable values are same with scale option.
3608 Enable 2-channel convolution using complex FFT. This improves speed significantly.
3609 Default is disabled.
3612 Enable minimum phase impulse response. Default is disabled.
3615 @subsection Examples
3620 firequalizer=gain='if(lt(f,1000), 0, -INF)'
3623 lowpass at 1000 Hz with gain_entry:
3625 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
3628 custom equalization:
3630 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
3633 higher delay with zero phase to compensate delay:
3635 firequalizer=delay=0.1:fixed=on:zero_phase=on
3638 lowpass on left channel, highpass on right channel:
3640 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
3641 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
3646 Apply a flanging effect to the audio.
3648 The filter accepts the following options:
3652 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
3655 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
3658 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
3662 Set percentage of delayed signal mixed with original. Range from 0 to 100.
3663 Default value is 71.
3666 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
3669 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
3670 Default value is @var{sinusoidal}.
3673 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
3674 Default value is 25.
3677 Set delay-line interpolation, @var{linear} or @var{quadratic}.
3678 Default is @var{linear}.
3682 Apply Haas effect to audio.
3684 Note that this makes most sense to apply on mono signals.
3685 With this filter applied to mono signals it give some directionality and
3686 stretches its stereo image.
3688 The filter accepts the following options:
3692 Set input level. By default is @var{1}, or 0dB
3695 Set output level. By default is @var{1}, or 0dB.
3698 Set gain applied to side part of signal. By default is @var{1}.
3701 Set kind of middle source. Can be one of the following:
3711 Pick middle part signal of stereo image.
3714 Pick side part signal of stereo image.
3718 Change middle phase. By default is disabled.
3721 Set left channel delay. By default is @var{2.05} milliseconds.
3724 Set left channel balance. By default is @var{-1}.
3727 Set left channel gain. By default is @var{1}.
3730 Change left phase. By default is disabled.
3733 Set right channel delay. By defaults is @var{2.12} milliseconds.
3736 Set right channel balance. By default is @var{1}.
3739 Set right channel gain. By default is @var{1}.
3742 Change right phase. By default is enabled.
3747 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
3748 embedded HDCD codes is expanded into a 20-bit PCM stream.
3750 The filter supports the Peak Extend and Low-level Gain Adjustment features
3751 of HDCD, and detects the Transient Filter flag.
3754 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
3757 When using the filter with wav, note the default encoding for wav is 16-bit,
3758 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
3759 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
3761 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
3762 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
3765 The filter accepts the following options:
3768 @item disable_autoconvert
3769 Disable any automatic format conversion or resampling in the filter graph.
3771 @item process_stereo
3772 Process the stereo channels together. If target_gain does not match between
3773 channels, consider it invalid and use the last valid target_gain.
3776 Set the code detect timer period in ms.
3779 Always extend peaks above -3dBFS even if PE isn't signaled.
3782 Replace audio with a solid tone and adjust the amplitude to signal some
3783 specific aspect of the decoding process. The output file can be loaded in
3784 an audio editor alongside the original to aid analysis.
3786 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
3793 Gain adjustment level at each sample
3795 Samples where peak extend occurs
3797 Samples where the code detect timer is active
3799 Samples where the target gain does not match between channels
3805 Apply head-related transfer functions (HRTFs) to create virtual
3806 loudspeakers around the user for binaural listening via headphones.
3807 The HRIRs are provided via additional streams, for each channel
3808 one stereo input stream is needed.
3810 The filter accepts the following options:
3814 Set mapping of input streams for convolution.
3815 The argument is a '|'-separated list of channel names in order as they
3816 are given as additional stream inputs for filter.
3817 This also specify number of input streams. Number of input streams
3818 must be not less than number of channels in first stream plus one.
3821 Set gain applied to audio. Value is in dB. Default is 0.
3824 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
3825 processing audio in time domain which is slow.
3826 @var{freq} is processing audio in frequency domain which is fast.
3827 Default is @var{freq}.
3830 Set custom gain for LFE channels. Value is in dB. Default is 0.
3833 Set size of frame in number of samples which will be processed at once.
3834 Default value is @var{1024}. Allowed range is from 1024 to 96000.
3837 Set format of hrir stream.
3838 Default value is @var{stereo}. Alternative value is @var{multich}.
3839 If value is set to @var{stereo}, number of additional streams should
3840 be greater or equal to number of input channels in first input stream.
3841 Also each additional stream should have stereo number of channels.
3842 If value is set to @var{multich}, number of additional streams should
3843 be exactly one. Also number of input channels of additional stream
3844 should be equal or greater than twice number of channels of first input
3848 @subsection Examples
3852 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3853 each amovie filter use stereo file with IR coefficients as input.
3854 The files give coefficients for each position of virtual loudspeaker:
3857 -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"
3862 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3863 but now in @var{multich} @var{hrir} format.
3865 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"
3872 Apply a high-pass filter with 3dB point frequency.
3873 The filter can be either single-pole, or double-pole (the default).
3874 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
3876 The filter accepts the following options:
3880 Set frequency in Hz. Default is 3000.
3883 Set number of poles. Default is 2.
3886 Set method to specify band-width of filter.
3901 Specify the band-width of a filter in width_type units.
3902 Applies only to double-pole filter.
3903 The default is 0.707q and gives a Butterworth response.
3906 How much to use filtered signal in output. Default is 1.
3907 Range is between 0 and 1.
3910 Specify which channels to filter, by default all available are filtered.
3913 @subsection Commands
3915 This filter supports the following commands:
3918 Change highpass frequency.
3919 Syntax for the command is : "@var{frequency}"
3922 Change highpass width_type.
3923 Syntax for the command is : "@var{width_type}"
3926 Change highpass width.
3927 Syntax for the command is : "@var{width}"
3930 Change highpass mix.
3931 Syntax for the command is : "@var{mix}"
3936 Join multiple input streams into one multi-channel stream.
3938 It accepts the following parameters:
3942 The number of input streams. It defaults to 2.
3944 @item channel_layout
3945 The desired output channel layout. It defaults to stereo.
3948 Map channels from inputs to output. The argument is a '|'-separated list of
3949 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
3950 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
3951 can be either the name of the input channel (e.g. FL for front left) or its
3952 index in the specified input stream. @var{out_channel} is the name of the output
3956 The filter will attempt to guess the mappings when they are not specified
3957 explicitly. It does so by first trying to find an unused matching input channel
3958 and if that fails it picks the first unused input channel.
3960 Join 3 inputs (with properly set channel layouts):
3962 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
3965 Build a 5.1 output from 6 single-channel streams:
3967 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
3968 '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'
3974 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
3976 To enable compilation of this filter you need to configure FFmpeg with
3977 @code{--enable-ladspa}.
3981 Specifies the name of LADSPA plugin library to load. If the environment
3982 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
3983 each one of the directories specified by the colon separated list in
3984 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
3985 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
3986 @file{/usr/lib/ladspa/}.
3989 Specifies the plugin within the library. Some libraries contain only
3990 one plugin, but others contain many of them. If this is not set filter
3991 will list all available plugins within the specified library.
3994 Set the '|' separated list of controls which are zero or more floating point
3995 values that determine the behavior of the loaded plugin (for example delay,
3997 Controls need to be defined using the following syntax:
3998 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
3999 @var{valuei} is the value set on the @var{i}-th control.
4000 Alternatively they can be also defined using the following syntax:
4001 @var{value0}|@var{value1}|@var{value2}|..., where
4002 @var{valuei} is the value set on the @var{i}-th control.
4003 If @option{controls} is set to @code{help}, all available controls and
4004 their valid ranges are printed.
4006 @item sample_rate, s
4007 Specify the sample rate, default to 44100. Only used if plugin have
4011 Set the number of samples per channel per each output frame, default
4012 is 1024. Only used if plugin have zero inputs.
4015 Set the minimum duration of the sourced audio. See
4016 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4017 for the accepted syntax.
4018 Note that the resulting duration may be greater than the specified duration,
4019 as the generated audio is always cut at the end of a complete frame.
4020 If not specified, or the expressed duration is negative, the audio is
4021 supposed to be generated forever.
4022 Only used if plugin have zero inputs.
4026 @subsection Examples
4030 List all available plugins within amp (LADSPA example plugin) library:
4036 List all available controls and their valid ranges for @code{vcf_notch}
4037 plugin from @code{VCF} library:
4039 ladspa=f=vcf:p=vcf_notch:c=help
4043 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4046 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4050 Add reverberation to the audio using TAP-plugins
4051 (Tom's Audio Processing plugins):
4053 ladspa=file=tap_reverb:tap_reverb
4057 Generate white noise, with 0.2 amplitude:
4059 ladspa=file=cmt:noise_source_white:c=c0=.2
4063 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4064 @code{C* Audio Plugin Suite} (CAPS) library:
4066 ladspa=file=caps:Click:c=c1=20'
4070 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4072 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4076 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4077 @code{SWH Plugins} collection:
4079 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4083 Attenuate low frequencies using Multiband EQ from Steve Harris
4084 @code{SWH Plugins} collection:
4086 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4090 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4093 ladspa=caps:Narrower
4097 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4099 ladspa=caps:White:.2
4103 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4105 ladspa=caps:Fractal:c=c1=1
4109 Dynamic volume normalization using @code{VLevel} plugin:
4111 ladspa=vlevel-ladspa:vlevel_mono
4115 @subsection Commands
4117 This filter supports the following commands:
4120 Modify the @var{N}-th control value.
4122 If the specified value is not valid, it is ignored and prior one is kept.
4127 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4128 Support for both single pass (livestreams, files) and double pass (files) modes.
4129 This algorithm can target IL, LRA, and maximum true peak. To accurately detect true peaks,
4130 the audio stream will be upsampled to 192 kHz unless the normalization mode is linear.
4131 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4133 The filter accepts the following options:
4137 Set integrated loudness target.
4138 Range is -70.0 - -5.0. Default value is -24.0.
4141 Set loudness range target.
4142 Range is 1.0 - 20.0. Default value is 7.0.
4145 Set maximum true peak.
4146 Range is -9.0 - +0.0. Default value is -2.0.
4148 @item measured_I, measured_i
4149 Measured IL of input file.
4150 Range is -99.0 - +0.0.
4152 @item measured_LRA, measured_lra
4153 Measured LRA of input file.
4154 Range is 0.0 - 99.0.
4156 @item measured_TP, measured_tp
4157 Measured true peak of input file.
4158 Range is -99.0 - +99.0.
4160 @item measured_thresh
4161 Measured threshold of input file.
4162 Range is -99.0 - +0.0.
4165 Set offset gain. Gain is applied before the true-peak limiter.
4166 Range is -99.0 - +99.0. Default is +0.0.
4169 Normalize linearly if possible.
4170 measured_I, measured_LRA, measured_TP, and measured_thresh must also
4171 to be specified in order to use this mode.
4172 Options are true or false. Default is true.
4175 Treat mono input files as "dual-mono". If a mono file is intended for playback
4176 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4177 If set to @code{true}, this option will compensate for this effect.
4178 Multi-channel input files are not affected by this option.
4179 Options are true or false. Default is false.
4182 Set print format for stats. Options are summary, json, or none.
4183 Default value is none.
4188 Apply a low-pass filter with 3dB point frequency.
4189 The filter can be either single-pole or double-pole (the default).
4190 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4192 The filter accepts the following options:
4196 Set frequency in Hz. Default is 500.
4199 Set number of poles. Default is 2.
4202 Set method to specify band-width of filter.
4217 Specify the band-width of a filter in width_type units.
4218 Applies only to double-pole filter.
4219 The default is 0.707q and gives a Butterworth response.
4222 How much to use filtered signal in output. Default is 1.
4223 Range is between 0 and 1.
4226 Specify which channels to filter, by default all available are filtered.
4229 @subsection Examples
4232 Lowpass only LFE channel, it LFE is not present it does nothing:
4238 @subsection Commands
4240 This filter supports the following commands:
4243 Change lowpass frequency.
4244 Syntax for the command is : "@var{frequency}"
4247 Change lowpass width_type.
4248 Syntax for the command is : "@var{width_type}"
4251 Change lowpass width.
4252 Syntax for the command is : "@var{width}"
4256 Syntax for the command is : "@var{mix}"
4261 Load a LV2 (LADSPA Version 2) plugin.
4263 To enable compilation of this filter you need to configure FFmpeg with
4264 @code{--enable-lv2}.
4268 Specifies the plugin URI. You may need to escape ':'.
4271 Set the '|' separated list of controls which are zero or more floating point
4272 values that determine the behavior of the loaded plugin (for example delay,
4274 If @option{controls} is set to @code{help}, all available controls and
4275 their valid ranges are printed.
4277 @item sample_rate, s
4278 Specify the sample rate, default to 44100. Only used if plugin have
4282 Set the number of samples per channel per each output frame, default
4283 is 1024. Only used if plugin have zero inputs.
4286 Set the minimum duration of the sourced audio. See
4287 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4288 for the accepted syntax.
4289 Note that the resulting duration may be greater than the specified duration,
4290 as the generated audio is always cut at the end of a complete frame.
4291 If not specified, or the expressed duration is negative, the audio is
4292 supposed to be generated forever.
4293 Only used if plugin have zero inputs.
4296 @subsection Examples
4300 Apply bass enhancer plugin from Calf:
4302 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4306 Apply vinyl plugin from Calf:
4308 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4312 Apply bit crusher plugin from ArtyFX:
4314 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4319 Multiband Compress or expand the audio's dynamic range.
4321 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4322 This is akin to the crossover of a loudspeaker, and results in flat frequency
4323 response when absent compander action.
4325 It accepts the following parameters:
4329 This option syntax is:
4330 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4331 For explanation of each item refer to compand filter documentation.
4337 Mix channels with specific gain levels. The filter accepts the output
4338 channel layout followed by a set of channels definitions.
4340 This filter is also designed to efficiently remap the channels of an audio
4343 The filter accepts parameters of the form:
4344 "@var{l}|@var{outdef}|@var{outdef}|..."
4348 output channel layout or number of channels
4351 output channel specification, of the form:
4352 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
4355 output channel to define, either a channel name (FL, FR, etc.) or a channel
4356 number (c0, c1, etc.)
4359 multiplicative coefficient for the channel, 1 leaving the volume unchanged
4362 input channel to use, see out_name for details; it is not possible to mix
4363 named and numbered input channels
4366 If the `=' in a channel specification is replaced by `<', then the gains for
4367 that specification will be renormalized so that the total is 1, thus
4368 avoiding clipping noise.
4370 @subsection Mixing examples
4372 For example, if you want to down-mix from stereo to mono, but with a bigger
4373 factor for the left channel:
4375 pan=1c|c0=0.9*c0+0.1*c1
4378 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
4379 7-channels surround:
4381 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
4384 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
4385 that should be preferred (see "-ac" option) unless you have very specific
4388 @subsection Remapping examples
4390 The channel remapping will be effective if, and only if:
4393 @item gain coefficients are zeroes or ones,
4394 @item only one input per channel output,
4397 If all these conditions are satisfied, the filter will notify the user ("Pure
4398 channel mapping detected"), and use an optimized and lossless method to do the
4401 For example, if you have a 5.1 source and want a stereo audio stream by
4402 dropping the extra channels:
4404 pan="stereo| c0=FL | c1=FR"
4407 Given the same source, you can also switch front left and front right channels
4408 and keep the input channel layout:
4410 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
4413 If the input is a stereo audio stream, you can mute the front left channel (and
4414 still keep the stereo channel layout) with:
4419 Still with a stereo audio stream input, you can copy the right channel in both
4420 front left and right:
4422 pan="stereo| c0=FR | c1=FR"
4427 ReplayGain scanner filter. This filter takes an audio stream as an input and
4428 outputs it unchanged.
4429 At end of filtering it displays @code{track_gain} and @code{track_peak}.
4433 Convert the audio sample format, sample rate and channel layout. It is
4434 not meant to be used directly.
4437 Apply time-stretching and pitch-shifting with librubberband.
4439 To enable compilation of this filter, you need to configure FFmpeg with
4440 @code{--enable-librubberband}.
4442 The filter accepts the following options:
4446 Set tempo scale factor.
4449 Set pitch scale factor.
4452 Set transients detector.
4453 Possible values are:
4462 Possible values are:
4471 Possible values are:
4478 Set processing window size.
4479 Possible values are:
4488 Possible values are:
4495 Enable formant preservation when shift pitching.
4496 Possible values are:
4504 Possible values are:
4513 Possible values are:
4520 @subsection Commands
4522 This filter supports the following commands:
4525 Change filter tempo scale factor.
4526 Syntax for the command is : "@var{tempo}"
4529 Change filter pitch scale factor.
4530 Syntax for the command is : "@var{pitch}"
4533 @section sidechaincompress
4535 This filter acts like normal compressor but has the ability to compress
4536 detected signal using second input signal.
4537 It needs two input streams and returns one output stream.
4538 First input stream will be processed depending on second stream signal.
4539 The filtered signal then can be filtered with other filters in later stages of
4540 processing. See @ref{pan} and @ref{amerge} filter.
4542 The filter accepts the following options:
4546 Set input gain. Default is 1. Range is between 0.015625 and 64.
4549 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
4550 Default is @code{downward}.
4553 If a signal of second stream raises above this level it will affect the gain
4554 reduction of first stream.
4555 By default is 0.125. Range is between 0.00097563 and 1.
4558 Set a ratio about which the signal is reduced. 1:2 means that if the level
4559 raised 4dB above the threshold, it will be only 2dB above after the reduction.
4560 Default is 2. Range is between 1 and 20.
4563 Amount of milliseconds the signal has to rise above the threshold before gain
4564 reduction starts. Default is 20. Range is between 0.01 and 2000.
4567 Amount of milliseconds the signal has to fall below the threshold before
4568 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
4571 Set the amount by how much signal will be amplified after processing.
4572 Default is 1. Range is from 1 to 64.
4575 Curve the sharp knee around the threshold to enter gain reduction more softly.
4576 Default is 2.82843. Range is between 1 and 8.
4579 Choose if the @code{average} level between all channels of side-chain stream
4580 or the louder(@code{maximum}) channel of side-chain stream affects the
4581 reduction. Default is @code{average}.
4584 Should the exact signal be taken in case of @code{peak} or an RMS one in case
4585 of @code{rms}. Default is @code{rms} which is mainly smoother.
4588 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
4591 How much to use compressed signal in output. Default is 1.
4592 Range is between 0 and 1.
4595 @subsection Examples
4599 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
4600 depending on the signal of 2nd input and later compressed signal to be
4601 merged with 2nd input:
4603 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
4607 @section sidechaingate
4609 A sidechain gate acts like a normal (wideband) gate but has the ability to
4610 filter the detected signal before sending it to the gain reduction stage.
4611 Normally a gate uses the full range signal to detect a level above the
4613 For example: If you cut all lower frequencies from your sidechain signal
4614 the gate will decrease the volume of your track only if not enough highs
4615 appear. With this technique you are able to reduce the resonation of a
4616 natural drum or remove "rumbling" of muted strokes from a heavily distorted
4618 It needs two input streams and returns one output stream.
4619 First input stream will be processed depending on second stream signal.
4621 The filter accepts the following options:
4625 Set input level before filtering.
4626 Default is 1. Allowed range is from 0.015625 to 64.
4629 Set the mode of operation. Can be @code{upward} or @code{downward}.
4630 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
4631 will be amplified, expanding dynamic range in upward direction.
4632 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
4635 Set the level of gain reduction when the signal is below the threshold.
4636 Default is 0.06125. Allowed range is from 0 to 1.
4637 Setting this to 0 disables reduction and then filter behaves like expander.
4640 If a signal rises above this level the gain reduction is released.
4641 Default is 0.125. Allowed range is from 0 to 1.
4644 Set a ratio about which the signal is reduced.
4645 Default is 2. Allowed range is from 1 to 9000.
4648 Amount of milliseconds the signal has to rise above the threshold before gain
4650 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
4653 Amount of milliseconds the signal has to fall below the threshold before the
4654 reduction is increased again. Default is 250 milliseconds.
4655 Allowed range is from 0.01 to 9000.
4658 Set amount of amplification of signal after processing.
4659 Default is 1. Allowed range is from 1 to 64.
4662 Curve the sharp knee around the threshold to enter gain reduction more softly.
4663 Default is 2.828427125. Allowed range is from 1 to 8.
4666 Choose if exact signal should be taken for detection or an RMS like one.
4667 Default is rms. Can be peak or rms.
4670 Choose if the average level between all channels or the louder channel affects
4672 Default is average. Can be average or maximum.
4675 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
4678 @section silencedetect
4680 Detect silence in an audio stream.
4682 This filter logs a message when it detects that the input audio volume is less
4683 or equal to a noise tolerance value for a duration greater or equal to the
4684 minimum detected noise duration.
4686 The printed times and duration are expressed in seconds.
4688 The filter accepts the following options:
4692 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
4693 specified value) or amplitude ratio. Default is -60dB, or 0.001.
4696 Set silence duration until notification (default is 2 seconds).
4699 Process each channel separately, instead of combined. By default is disabled.
4702 @subsection Examples
4706 Detect 5 seconds of silence with -50dB noise tolerance:
4708 silencedetect=n=-50dB:d=5
4712 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
4713 tolerance in @file{silence.mp3}:
4715 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
4719 @section silenceremove
4721 Remove silence from the beginning, middle or end of the audio.
4723 The filter accepts the following options:
4727 This value is used to indicate if audio should be trimmed at beginning of
4728 the audio. A value of zero indicates no silence should be trimmed from the
4729 beginning. When specifying a non-zero value, it trims audio up until it
4730 finds non-silence. Normally, when trimming silence from beginning of audio
4731 the @var{start_periods} will be @code{1} but it can be increased to higher
4732 values to trim all audio up to specific count of non-silence periods.
4733 Default value is @code{0}.
4735 @item start_duration
4736 Specify the amount of time that non-silence must be detected before it stops
4737 trimming audio. By increasing the duration, bursts of noises can be treated
4738 as silence and trimmed off. Default value is @code{0}.
4740 @item start_threshold
4741 This indicates what sample value should be treated as silence. For digital
4742 audio, a value of @code{0} may be fine but for audio recorded from analog,
4743 you may wish to increase the value to account for background noise.
4744 Can be specified in dB (in case "dB" is appended to the specified value)
4745 or amplitude ratio. Default value is @code{0}.
4748 Specify max duration of silence at beginning that will be kept after
4749 trimming. Default is 0, which is equal to trimming all samples detected
4753 Specify mode of detection of silence end in start of multi-channel audio.
4754 Can be @var{any} or @var{all}. Default is @var{any}.
4755 With @var{any}, any sample that is detected as non-silence will cause
4756 stopped trimming of silence.
4757 With @var{all}, only if all channels are detected as non-silence will cause
4758 stopped trimming of silence.
4761 Set the count for trimming silence from the end of audio.
4762 To remove silence from the middle of a file, specify a @var{stop_periods}
4763 that is negative. This value is then treated as a positive value and is
4764 used to indicate the effect should restart processing as specified by
4765 @var{start_periods}, making it suitable for removing periods of silence
4766 in the middle of the audio.
4767 Default value is @code{0}.
4770 Specify a duration of silence that must exist before audio is not copied any
4771 more. By specifying a higher duration, silence that is wanted can be left in
4773 Default value is @code{0}.
4775 @item stop_threshold
4776 This is the same as @option{start_threshold} but for trimming silence from
4778 Can be specified in dB (in case "dB" is appended to the specified value)
4779 or amplitude ratio. Default value is @code{0}.
4782 Specify max duration of silence at end that will be kept after
4783 trimming. Default is 0, which is equal to trimming all samples detected
4787 Specify mode of detection of silence start in end of multi-channel audio.
4788 Can be @var{any} or @var{all}. Default is @var{any}.
4789 With @var{any}, any sample that is detected as non-silence will cause
4790 stopped trimming of silence.
4791 With @var{all}, only if all channels are detected as non-silence will cause
4792 stopped trimming of silence.
4795 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
4796 and works better with digital silence which is exactly 0.
4797 Default value is @code{rms}.
4800 Set duration in number of seconds used to calculate size of window in number
4801 of samples for detecting silence.
4802 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
4805 @subsection Examples
4809 The following example shows how this filter can be used to start a recording
4810 that does not contain the delay at the start which usually occurs between
4811 pressing the record button and the start of the performance:
4813 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
4817 Trim all silence encountered from beginning to end where there is more than 1
4818 second of silence in audio:
4820 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
4824 Trim all digital silence samples, using peak detection, from beginning to end
4825 where there is more than 0 samples of digital silence in audio and digital
4826 silence is detected in all channels at same positions in stream:
4828 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
4834 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
4835 loudspeakers around the user for binaural listening via headphones (audio
4836 formats up to 9 channels supported).
4837 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
4838 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
4839 Austrian Academy of Sciences.
4841 To enable compilation of this filter you need to configure FFmpeg with
4842 @code{--enable-libmysofa}.
4844 The filter accepts the following options:
4848 Set the SOFA file used for rendering.
4851 Set gain applied to audio. Value is in dB. Default is 0.
4854 Set rotation of virtual loudspeakers in deg. Default is 0.
4857 Set elevation of virtual speakers in deg. Default is 0.
4860 Set distance in meters between loudspeakers and the listener with near-field
4861 HRTFs. Default is 1.
4864 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4865 processing audio in time domain which is slow.
4866 @var{freq} is processing audio in frequency domain which is fast.
4867 Default is @var{freq}.
4870 Set custom positions of virtual loudspeakers. Syntax for this option is:
4871 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
4872 Each virtual loudspeaker is described with short channel name following with
4873 azimuth and elevation in degrees.
4874 Each virtual loudspeaker description is separated by '|'.
4875 For example to override front left and front right channel positions use:
4876 'speakers=FL 45 15|FR 345 15'.
4877 Descriptions with unrecognised channel names are ignored.
4880 Set custom gain for LFE channels. Value is in dB. Default is 0.
4883 Set custom frame size in number of samples. Default is 1024.
4884 Allowed range is from 1024 to 96000. Only used if option @samp{type}
4885 is set to @var{freq}.
4888 Should all IRs be normalized upon importing SOFA file.
4889 By default is enabled.
4892 Should nearest IRs be interpolated with neighbor IRs if exact position
4893 does not match. By default is disabled.
4896 Minphase all IRs upon loading of SOFA file. By default is disabled.
4899 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
4902 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
4905 @subsection Examples
4909 Using ClubFritz6 sofa file:
4911 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
4915 Using ClubFritz12 sofa file and bigger radius with small rotation:
4917 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
4921 Similar as above but with custom speaker positions for front left, front right, back left and back right
4922 and also with custom gain:
4924 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
4928 @section stereotools
4930 This filter has some handy utilities to manage stereo signals, for converting
4931 M/S stereo recordings to L/R signal while having control over the parameters
4932 or spreading the stereo image of master track.
4934 The filter accepts the following options:
4938 Set input level before filtering for both channels. Defaults is 1.
4939 Allowed range is from 0.015625 to 64.
4942 Set output level after filtering for both channels. Defaults is 1.
4943 Allowed range is from 0.015625 to 64.
4946 Set input balance between both channels. Default is 0.
4947 Allowed range is from -1 to 1.
4950 Set output balance between both channels. Default is 0.
4951 Allowed range is from -1 to 1.
4954 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
4955 clipping. Disabled by default.
4958 Mute the left channel. Disabled by default.
4961 Mute the right channel. Disabled by default.
4964 Change the phase of the left channel. Disabled by default.
4967 Change the phase of the right channel. Disabled by default.
4970 Set stereo mode. Available values are:
4974 Left/Right to Left/Right, this is default.
4977 Left/Right to Mid/Side.
4980 Mid/Side to Left/Right.
4983 Left/Right to Left/Left.
4986 Left/Right to Right/Right.
4989 Left/Right to Left + Right.
4992 Left/Right to Right/Left.
4995 Mid/Side to Left/Left.
4998 Mid/Side to Right/Right.
5002 Set level of side signal. Default is 1.
5003 Allowed range is from 0.015625 to 64.
5006 Set balance of side signal. Default is 0.
5007 Allowed range is from -1 to 1.
5010 Set level of the middle signal. Default is 1.
5011 Allowed range is from 0.015625 to 64.
5014 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5017 Set stereo base between mono and inversed channels. Default is 0.
5018 Allowed range is from -1 to 1.
5021 Set delay in milliseconds how much to delay left from right channel and
5022 vice versa. Default is 0. Allowed range is from -20 to 20.
5025 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5028 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5030 @item bmode_in, bmode_out
5031 Set balance mode for balance_in/balance_out option.
5033 Can be one of the following:
5037 Classic balance mode. Attenuate one channel at time.
5038 Gain is raised up to 1.
5041 Similar as classic mode above but gain is raised up to 2.
5044 Equal power distribution, from -6dB to +6dB range.
5048 @subsection Examples
5052 Apply karaoke like effect:
5054 stereotools=mlev=0.015625
5058 Convert M/S signal to L/R:
5060 "stereotools=mode=ms>lr"
5064 @section stereowiden
5066 This filter enhance the stereo effect by suppressing signal common to both
5067 channels and by delaying the signal of left into right and vice versa,
5068 thereby widening the stereo effect.
5070 The filter accepts the following options:
5074 Time in milliseconds of the delay of left signal into right and vice versa.
5075 Default is 20 milliseconds.
5078 Amount of gain in delayed signal into right and vice versa. Gives a delay
5079 effect of left signal in right output and vice versa which gives widening
5080 effect. Default is 0.3.
5083 Cross feed of left into right with inverted phase. This helps in suppressing
5084 the mono. If the value is 1 it will cancel all the signal common to both
5085 channels. Default is 0.3.
5088 Set level of input signal of original channel. Default is 0.8.
5091 @section superequalizer
5092 Apply 18 band equalizer.
5094 The filter accepts the following options:
5101 Set 131Hz band gain.
5103 Set 185Hz band gain.
5105 Set 262Hz band gain.
5107 Set 370Hz band gain.
5109 Set 523Hz band gain.
5111 Set 740Hz band gain.
5113 Set 1047Hz band gain.
5115 Set 1480Hz band gain.
5117 Set 2093Hz band gain.
5119 Set 2960Hz band gain.
5121 Set 4186Hz band gain.
5123 Set 5920Hz band gain.
5125 Set 8372Hz band gain.
5127 Set 11840Hz band gain.
5129 Set 16744Hz band gain.
5131 Set 20000Hz band gain.
5135 Apply audio surround upmix filter.
5137 This filter allows to produce multichannel output from audio stream.
5139 The filter accepts the following options:
5143 Set output channel layout. By default, this is @var{5.1}.
5145 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5146 for the required syntax.
5149 Set input channel layout. By default, this is @var{stereo}.
5151 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5152 for the required syntax.
5155 Set input volume level. By default, this is @var{1}.
5158 Set output volume level. By default, this is @var{1}.
5161 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5164 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5167 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5170 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5171 In @var{add} mode, LFE channel is created from input audio and added to output.
5172 In @var{sub} mode, LFE channel is created from input audio and added to output but
5173 also all non-LFE output channels are subtracted with output LFE channel.
5176 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5177 Default is @var{90}.
5180 Set front center input volume. By default, this is @var{1}.
5183 Set front center output volume. By default, this is @var{1}.
5186 Set front left input volume. By default, this is @var{1}.
5189 Set front left output volume. By default, this is @var{1}.
5192 Set front right input volume. By default, this is @var{1}.
5195 Set front right output volume. By default, this is @var{1}.
5198 Set side left input volume. By default, this is @var{1}.
5201 Set side left output volume. By default, this is @var{1}.
5204 Set side right input volume. By default, this is @var{1}.
5207 Set side right output volume. By default, this is @var{1}.
5210 Set back left input volume. By default, this is @var{1}.
5213 Set back left output volume. By default, this is @var{1}.
5216 Set back right input volume. By default, this is @var{1}.
5219 Set back right output volume. By default, this is @var{1}.
5222 Set back center input volume. By default, this is @var{1}.
5225 Set back center output volume. By default, this is @var{1}.
5228 Set LFE input volume. By default, this is @var{1}.
5231 Set LFE output volume. By default, this is @var{1}.
5234 Set spread usage of stereo image across X axis for all channels.
5237 Set spread usage of stereo image across Y axis for all channels.
5239 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5240 Set spread usage of stereo image across X axis for each channel.
5242 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5243 Set spread usage of stereo image across Y axis for each channel.
5246 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5249 Set window function.
5251 It accepts the following values:
5274 Default is @code{hann}.
5277 Set window overlap. If set to 1, the recommended overlap for selected
5278 window function will be picked. Default is @code{0.5}.
5281 @section treble, highshelf
5283 Boost or cut treble (upper) frequencies of the audio using a two-pole
5284 shelving filter with a response similar to that of a standard
5285 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
5287 The filter accepts the following options:
5291 Give the gain at whichever is the lower of ~22 kHz and the
5292 Nyquist frequency. Its useful range is about -20 (for a large cut)
5293 to +20 (for a large boost). Beware of clipping when using a positive gain.
5296 Set the filter's central frequency and so can be used
5297 to extend or reduce the frequency range to be boosted or cut.
5298 The default value is @code{3000} Hz.
5301 Set method to specify band-width of filter.
5316 Determine how steep is the filter's shelf transition.
5319 How much to use filtered signal in output. Default is 1.
5320 Range is between 0 and 1.
5323 Specify which channels to filter, by default all available are filtered.
5326 @subsection Commands
5328 This filter supports the following commands:
5331 Change treble frequency.
5332 Syntax for the command is : "@var{frequency}"
5335 Change treble width_type.
5336 Syntax for the command is : "@var{width_type}"
5339 Change treble width.
5340 Syntax for the command is : "@var{width}"
5344 Syntax for the command is : "@var{gain}"
5348 Syntax for the command is : "@var{mix}"
5353 Sinusoidal amplitude modulation.
5355 The filter accepts the following options:
5359 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
5360 (20 Hz or lower) will result in a tremolo effect.
5361 This filter may also be used as a ring modulator by specifying
5362 a modulation frequency higher than 20 Hz.
5363 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5366 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5367 Default value is 0.5.
5372 Sinusoidal phase modulation.
5374 The filter accepts the following options:
5378 Modulation frequency in Hertz.
5379 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5382 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5383 Default value is 0.5.
5388 Adjust the input audio volume.
5390 It accepts the following parameters:
5394 Set audio volume expression.
5396 Output values are clipped to the maximum value.
5398 The output audio volume is given by the relation:
5400 @var{output_volume} = @var{volume} * @var{input_volume}
5403 The default value for @var{volume} is "1.0".
5406 This parameter represents the mathematical precision.
5408 It determines which input sample formats will be allowed, which affects the
5409 precision of the volume scaling.
5413 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
5415 32-bit floating-point; this limits input sample format to FLT. (default)
5417 64-bit floating-point; this limits input sample format to DBL.
5421 Choose the behaviour on encountering ReplayGain side data in input frames.
5425 Remove ReplayGain side data, ignoring its contents (the default).
5428 Ignore ReplayGain side data, but leave it in the frame.
5431 Prefer the track gain, if present.
5434 Prefer the album gain, if present.
5437 @item replaygain_preamp
5438 Pre-amplification gain in dB to apply to the selected replaygain gain.
5440 Default value for @var{replaygain_preamp} is 0.0.
5443 Set when the volume expression is evaluated.
5445 It accepts the following values:
5448 only evaluate expression once during the filter initialization, or
5449 when the @samp{volume} command is sent
5452 evaluate expression for each incoming frame
5455 Default value is @samp{once}.
5458 The volume expression can contain the following parameters.
5462 frame number (starting at zero)
5465 @item nb_consumed_samples
5466 number of samples consumed by the filter
5468 number of samples in the current frame
5470 original frame position in the file
5476 PTS at start of stream
5478 time at start of stream
5484 last set volume value
5487 Note that when @option{eval} is set to @samp{once} only the
5488 @var{sample_rate} and @var{tb} variables are available, all other
5489 variables will evaluate to NAN.
5491 @subsection Commands
5493 This filter supports the following commands:
5496 Modify the volume expression.
5497 The command accepts the same syntax of the corresponding option.
5499 If the specified expression is not valid, it is kept at its current
5501 @item replaygain_noclip
5502 Prevent clipping by limiting the gain applied.
5504 Default value for @var{replaygain_noclip} is 1.
5508 @subsection Examples
5512 Halve the input audio volume:
5516 volume=volume=-6.0206dB
5519 In all the above example the named key for @option{volume} can be
5520 omitted, for example like in:
5526 Increase input audio power by 6 decibels using fixed-point precision:
5528 volume=volume=6dB:precision=fixed
5532 Fade volume after time 10 with an annihilation period of 5 seconds:
5534 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
5538 @section volumedetect
5540 Detect the volume of the input video.
5542 The filter has no parameters. The input is not modified. Statistics about
5543 the volume will be printed in the log when the input stream end is reached.
5545 In particular it will show the mean volume (root mean square), maximum
5546 volume (on a per-sample basis), and the beginning of a histogram of the
5547 registered volume values (from the maximum value to a cumulated 1/1000 of
5550 All volumes are in decibels relative to the maximum PCM value.
5552 @subsection Examples
5554 Here is an excerpt of the output:
5556 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
5557 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
5558 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
5559 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
5560 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
5561 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
5562 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
5563 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
5564 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
5570 The mean square energy is approximately -27 dB, or 10^-2.7.
5572 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
5574 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
5577 In other words, raising the volume by +4 dB does not cause any clipping,
5578 raising it by +5 dB causes clipping for 6 samples, etc.
5580 @c man end AUDIO FILTERS
5582 @chapter Audio Sources
5583 @c man begin AUDIO SOURCES
5585 Below is a description of the currently available audio sources.
5589 Buffer audio frames, and make them available to the filter chain.
5591 This source is mainly intended for a programmatic use, in particular
5592 through the interface defined in @file{libavfilter/asrc_abuffer.h}.
5594 It accepts the following parameters:
5598 The timebase which will be used for timestamps of submitted frames. It must be
5599 either a floating-point number or in @var{numerator}/@var{denominator} form.
5602 The sample rate of the incoming audio buffers.
5605 The sample format of the incoming audio buffers.
5606 Either a sample format name or its corresponding integer representation from
5607 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
5609 @item channel_layout
5610 The channel layout of the incoming audio buffers.
5611 Either a channel layout name from channel_layout_map in
5612 @file{libavutil/channel_layout.c} or its corresponding integer representation
5613 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
5616 The number of channels of the incoming audio buffers.
5617 If both @var{channels} and @var{channel_layout} are specified, then they
5622 @subsection Examples
5625 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
5628 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
5629 Since the sample format with name "s16p" corresponds to the number
5630 6 and the "stereo" channel layout corresponds to the value 0x3, this is
5633 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
5638 Generate an audio signal specified by an expression.
5640 This source accepts in input one or more expressions (one for each
5641 channel), which are evaluated and used to generate a corresponding
5644 This source accepts the following options:
5648 Set the '|'-separated expressions list for each separate channel. In case the
5649 @option{channel_layout} option is not specified, the selected channel layout
5650 depends on the number of provided expressions. Otherwise the last
5651 specified expression is applied to the remaining output channels.
5653 @item channel_layout, c
5654 Set the channel layout. The number of channels in the specified layout
5655 must be equal to the number of specified expressions.
5658 Set the minimum duration of the sourced audio. See
5659 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5660 for the accepted syntax.
5661 Note that the resulting duration may be greater than the specified
5662 duration, as the generated audio is always cut at the end of a
5665 If not specified, or the expressed duration is negative, the audio is
5666 supposed to be generated forever.
5669 Set the number of samples per channel per each output frame,
5672 @item sample_rate, s
5673 Specify the sample rate, default to 44100.
5676 Each expression in @var{exprs} can contain the following constants:
5680 number of the evaluated sample, starting from 0
5683 time of the evaluated sample expressed in seconds, starting from 0
5690 @subsection Examples
5700 Generate a sin signal with frequency of 440 Hz, set sample rate to
5703 aevalsrc="sin(440*2*PI*t):s=8000"
5707 Generate a two channels signal, specify the channel layout (Front
5708 Center + Back Center) explicitly:
5710 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
5714 Generate white noise:
5716 aevalsrc="-2+random(0)"
5720 Generate an amplitude modulated signal:
5722 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
5726 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
5728 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
5735 The null audio source, return unprocessed audio frames. It is mainly useful
5736 as a template and to be employed in analysis / debugging tools, or as
5737 the source for filters which ignore the input data (for example the sox
5740 This source accepts the following options:
5744 @item channel_layout, cl
5746 Specifies the channel layout, and can be either an integer or a string
5747 representing a channel layout. The default value of @var{channel_layout}
5750 Check the channel_layout_map definition in
5751 @file{libavutil/channel_layout.c} for the mapping between strings and
5752 channel layout values.
5754 @item sample_rate, r
5755 Specifies the sample rate, and defaults to 44100.
5758 Set the number of samples per requested frames.
5762 @subsection Examples
5766 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
5768 anullsrc=r=48000:cl=4
5772 Do the same operation with a more obvious syntax:
5774 anullsrc=r=48000:cl=mono
5778 All the parameters need to be explicitly defined.
5782 Synthesize a voice utterance using the libflite library.
5784 To enable compilation of this filter you need to configure FFmpeg with
5785 @code{--enable-libflite}.
5787 Note that versions of the flite library prior to 2.0 are not thread-safe.
5789 The filter accepts the following options:
5794 If set to 1, list the names of the available voices and exit
5795 immediately. Default value is 0.
5798 Set the maximum number of samples per frame. Default value is 512.
5801 Set the filename containing the text to speak.
5804 Set the text to speak.
5807 Set the voice to use for the speech synthesis. Default value is
5808 @code{kal}. See also the @var{list_voices} option.
5811 @subsection Examples
5815 Read from file @file{speech.txt}, and synthesize the text using the
5816 standard flite voice:
5818 flite=textfile=speech.txt
5822 Read the specified text selecting the @code{slt} voice:
5824 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
5828 Input text to ffmpeg:
5830 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
5834 Make @file{ffplay} speak the specified text, using @code{flite} and
5835 the @code{lavfi} device:
5837 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
5841 For more information about libflite, check:
5842 @url{http://www.festvox.org/flite/}
5846 Generate a noise audio signal.
5848 The filter accepts the following options:
5851 @item sample_rate, r
5852 Specify the sample rate. Default value is 48000 Hz.
5855 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
5859 Specify the duration of the generated audio stream. Not specifying this option
5860 results in noise with an infinite length.
5862 @item color, colour, c
5863 Specify the color of noise. Available noise colors are white, pink, brown,
5864 blue and violet. Default color is white.
5867 Specify a value used to seed the PRNG.
5870 Set the number of samples per each output frame, default is 1024.
5873 @subsection Examples
5878 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
5880 anoisesrc=d=60:c=pink:r=44100:a=0.5
5886 Generate odd-tap Hilbert transform FIR coefficients.
5888 The resulting stream can be used with @ref{afir} filter for phase-shifting
5889 the signal by 90 degrees.
5891 This is used in many matrix coding schemes and for analytic signal generation.
5892 The process is often written as a multiplication by i (or j), the imaginary unit.
5894 The filter accepts the following options:
5898 @item sample_rate, s
5899 Set sample rate, default is 44100.
5902 Set length of FIR filter, default is 22051.
5905 Set number of samples per each frame.
5908 Set window function to be used when generating FIR coefficients.
5913 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
5915 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
5917 The filter accepts the following options:
5920 @item sample_rate, r
5921 Set sample rate, default is 44100.
5924 Set number of samples per each frame. Default is 1024.
5927 Set high-pass frequency. Default is 0.
5930 Set low-pass frequency. Default is 0.
5931 If high-pass frequency is lower than low-pass frequency and low-pass frequency
5932 is higher than 0 then filter will create band-pass filter coefficients,
5933 otherwise band-reject filter coefficients.
5936 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
5939 Set Kaiser window beta.
5942 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
5945 Enable rounding, by default is disabled.
5948 Set number of taps for high-pass filter.
5951 Set number of taps for low-pass filter.
5956 Generate an audio signal made of a sine wave with amplitude 1/8.
5958 The audio signal is bit-exact.
5960 The filter accepts the following options:
5965 Set the carrier frequency. Default is 440 Hz.
5967 @item beep_factor, b
5968 Enable a periodic beep every second with frequency @var{beep_factor} times
5969 the carrier frequency. Default is 0, meaning the beep is disabled.
5971 @item sample_rate, r
5972 Specify the sample rate, default is 44100.
5975 Specify the duration of the generated audio stream.
5977 @item samples_per_frame
5978 Set the number of samples per output frame.
5980 The expression can contain the following constants:
5984 The (sequential) number of the output audio frame, starting from 0.
5987 The PTS (Presentation TimeStamp) of the output audio frame,
5988 expressed in @var{TB} units.
5991 The PTS of the output audio frame, expressed in seconds.
5994 The timebase of the output audio frames.
5997 Default is @code{1024}.
6000 @subsection Examples
6005 Generate a simple 440 Hz sine wave:
6011 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6015 sine=frequency=220:beep_factor=4:duration=5
6019 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6022 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6026 @c man end AUDIO SOURCES
6028 @chapter Audio Sinks
6029 @c man begin AUDIO SINKS
6031 Below is a description of the currently available audio sinks.
6033 @section abuffersink
6035 Buffer audio frames, and make them available to the end of filter chain.
6037 This sink is mainly intended for programmatic use, in particular
6038 through the interface defined in @file{libavfilter/buffersink.h}
6039 or the options system.
6041 It accepts a pointer to an AVABufferSinkContext structure, which
6042 defines the incoming buffers' formats, to be passed as the opaque
6043 parameter to @code{avfilter_init_filter} for initialization.
6046 Null audio sink; do absolutely nothing with the input audio. It is
6047 mainly useful as a template and for use in analysis / debugging
6050 @c man end AUDIO SINKS
6052 @chapter Video Filters
6053 @c man begin VIDEO FILTERS
6055 When you configure your FFmpeg build, you can disable any of the
6056 existing filters using @code{--disable-filters}.
6057 The configure output will show the video filters included in your
6060 Below is a description of the currently available video filters.
6064 Mark a region of interest in a video frame.
6066 The frame data is passed through unchanged, but metadata is attached
6067 to the frame indicating regions of interest which can affect the
6068 behaviour of later encoding. Multiple regions can be marked by
6069 applying the filter multiple times.
6073 Region distance in pixels from the left edge of the frame.
6075 Region distance in pixels from the top edge of the frame.
6077 Region width in pixels.
6079 Region height in pixels.
6081 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6082 and may contain the following variables:
6085 Width of the input frame.
6087 Height of the input frame.
6091 Quantisation offset to apply within the region.
6093 This must be a real value in the range -1 to +1. A value of zero
6094 indicates no quality change. A negative value asks for better quality
6095 (less quantisation), while a positive value asks for worse quality
6096 (greater quantisation).
6098 The range is calibrated so that the extreme values indicate the
6099 largest possible offset - if the rest of the frame is encoded with the
6100 worst possible quality, an offset of -1 indicates that this region
6101 should be encoded with the best possible quality anyway. Intermediate
6102 values are then interpolated in some codec-dependent way.
6104 For example, in 10-bit H.264 the quantisation parameter varies between
6105 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6106 this region should be encoded with a QP around one-tenth of the full
6107 range better than the rest of the frame. So, if most of the frame
6108 were to be encoded with a QP of around 30, this region would get a QP
6109 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6110 An extreme value of -1 would indicate that this region should be
6111 encoded with the best possible quality regardless of the treatment of
6112 the rest of the frame - that is, should be encoded at a QP of -12.
6114 If set to true, remove any existing regions of interest marked on the
6115 frame before adding the new one.
6118 @subsection Examples
6122 Mark the centre quarter of the frame as interesting.
6124 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6127 Mark the 100-pixel-wide region on the left edge of the frame as very
6128 uninteresting (to be encoded at much lower quality than the rest of
6131 addroi=0:0:100:ih:+1/5
6135 @section alphaextract
6137 Extract the alpha component from the input as a grayscale video. This
6138 is especially useful with the @var{alphamerge} filter.
6142 Add or replace the alpha component of the primary input with the
6143 grayscale value of a second input. This is intended for use with
6144 @var{alphaextract} to allow the transmission or storage of frame
6145 sequences that have alpha in a format that doesn't support an alpha
6148 For example, to reconstruct full frames from a normal YUV-encoded video
6149 and a separate video created with @var{alphaextract}, you might use:
6151 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6154 Since this filter is designed for reconstruction, it operates on frame
6155 sequences without considering timestamps, and terminates when either
6156 input reaches end of stream. This will cause problems if your encoding
6157 pipeline drops frames. If you're trying to apply an image as an
6158 overlay to a video stream, consider the @var{overlay} filter instead.
6162 Amplify differences between current pixel and pixels of adjacent frames in
6163 same pixel location.
6165 This filter accepts the following options:
6169 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6170 For example radius of 3 will instruct filter to calculate average of 7 frames.
6173 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6176 Set threshold for difference amplification. Any difference greater or equal to
6177 this value will not alter source pixel. Default is 10.
6178 Allowed range is from 0 to 65535.
6181 Set tolerance for difference amplification. Any difference lower to
6182 this value will not alter source pixel. Default is 0.
6183 Allowed range is from 0 to 65535.
6186 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6187 This option controls maximum possible value that will decrease source pixel value.
6190 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6191 This option controls maximum possible value that will increase source pixel value.
6194 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
6197 @subsection Commands
6199 This filter supports the following @ref{commands} that corresponds to option of same name:
6211 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
6212 and libavformat to work. On the other hand, it is limited to ASS (Advanced
6213 Substation Alpha) subtitles files.
6215 This filter accepts the following option in addition to the common options from
6216 the @ref{subtitles} filter:
6220 Set the shaping engine
6222 Available values are:
6225 The default libass shaping engine, which is the best available.
6227 Fast, font-agnostic shaper that can do only substitutions
6229 Slower shaper using OpenType for substitutions and positioning
6232 The default is @code{auto}.
6236 Apply an Adaptive Temporal Averaging Denoiser to the video input.
6238 The filter accepts the following options:
6242 Set threshold A for 1st plane. Default is 0.02.
6243 Valid range is 0 to 0.3.
6246 Set threshold B for 1st plane. Default is 0.04.
6247 Valid range is 0 to 5.
6250 Set threshold A for 2nd plane. Default is 0.02.
6251 Valid range is 0 to 0.3.
6254 Set threshold B for 2nd plane. Default is 0.04.
6255 Valid range is 0 to 5.
6258 Set threshold A for 3rd plane. Default is 0.02.
6259 Valid range is 0 to 0.3.
6262 Set threshold B for 3rd plane. Default is 0.04.
6263 Valid range is 0 to 5.
6265 Threshold A is designed to react on abrupt changes in the input signal and
6266 threshold B is designed to react on continuous changes in the input signal.
6269 Set number of frames filter will use for averaging. Default is 9. Must be odd
6270 number in range [5, 129].
6273 Set what planes of frame filter will use for averaging. Default is all.
6276 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
6277 Alternatively can be set to @code{s} serial.
6279 Parallel can be faster then serial, while other way around is never true.
6280 Parallel will abort early on first change being greater then thresholds, while serial
6281 will continue processing other side of frames if they are equal or bellow thresholds.
6286 Apply average blur filter.
6288 The filter accepts the following options:
6292 Set horizontal radius size.
6295 Set which planes to filter. By default all planes are filtered.
6298 Set vertical radius size, if zero it will be same as @code{sizeX}.
6299 Default is @code{0}.
6302 @subsection Commands
6303 This filter supports same commands as options.
6304 The command accepts the same syntax of the corresponding option.
6306 If the specified expression is not valid, it is kept at its current
6311 Compute the bounding box for the non-black pixels in the input frame
6314 This filter computes the bounding box containing all the pixels with a
6315 luminance value greater than the minimum allowed value.
6316 The parameters describing the bounding box are printed on the filter
6319 The filter accepts the following option:
6323 Set the minimal luminance value. Default is @code{16}.
6326 @section bitplanenoise
6328 Show and measure bit plane noise.
6330 The filter accepts the following options:
6334 Set which plane to analyze. Default is @code{1}.
6337 Filter out noisy pixels from @code{bitplane} set above.
6338 Default is disabled.
6341 @section blackdetect
6343 Detect video intervals that are (almost) completely black. Can be
6344 useful to detect chapter transitions, commercials, or invalid
6345 recordings. Output lines contains the time for the start, end and
6346 duration of the detected black interval expressed in seconds.
6348 In order to display the output lines, you need to set the loglevel at
6349 least to the AV_LOG_INFO value.
6351 The filter accepts the following options:
6354 @item black_min_duration, d
6355 Set the minimum detected black duration expressed in seconds. It must
6356 be a non-negative floating point number.
6358 Default value is 2.0.
6360 @item picture_black_ratio_th, pic_th
6361 Set the threshold for considering a picture "black".
6362 Express the minimum value for the ratio:
6364 @var{nb_black_pixels} / @var{nb_pixels}
6367 for which a picture is considered black.
6368 Default value is 0.98.
6370 @item pixel_black_th, pix_th
6371 Set the threshold for considering a pixel "black".
6373 The threshold expresses the maximum pixel luminance value for which a
6374 pixel is considered "black". The provided value is scaled according to
6375 the following equation:
6377 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
6380 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
6381 the input video format, the range is [0-255] for YUV full-range
6382 formats and [16-235] for YUV non full-range formats.
6384 Default value is 0.10.
6387 The following example sets the maximum pixel threshold to the minimum
6388 value, and detects only black intervals of 2 or more seconds:
6390 blackdetect=d=2:pix_th=0.00
6395 Detect frames that are (almost) completely black. Can be useful to
6396 detect chapter transitions or commercials. Output lines consist of
6397 the frame number of the detected frame, the percentage of blackness,
6398 the position in the file if known or -1 and the timestamp in seconds.
6400 In order to display the output lines, you need to set the loglevel at
6401 least to the AV_LOG_INFO value.
6403 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
6404 The value represents the percentage of pixels in the picture that
6405 are below the threshold value.
6407 It accepts the following parameters:
6412 The percentage of the pixels that have to be below the threshold; it defaults to
6415 @item threshold, thresh
6416 The threshold below which a pixel value is considered black; it defaults to
6421 @section blend, tblend
6423 Blend two video frames into each other.
6425 The @code{blend} filter takes two input streams and outputs one
6426 stream, the first input is the "top" layer and second input is
6427 "bottom" layer. By default, the output terminates when the longest input terminates.
6429 The @code{tblend} (time blend) filter takes two consecutive frames
6430 from one single stream, and outputs the result obtained by blending
6431 the new frame on top of the old frame.
6433 A description of the accepted options follows.
6441 Set blend mode for specific pixel component or all pixel components in case
6442 of @var{all_mode}. Default value is @code{normal}.
6444 Available values for component modes are:
6486 Set blend opacity for specific pixel component or all pixel components in case
6487 of @var{all_opacity}. Only used in combination with pixel component blend modes.
6494 Set blend expression for specific pixel component or all pixel components in case
6495 of @var{all_expr}. Note that related mode options will be ignored if those are set.
6497 The expressions can use the following variables:
6501 The sequential number of the filtered frame, starting from @code{0}.
6505 the coordinates of the current sample
6509 the width and height of currently filtered plane
6513 Width and height scale for the plane being filtered. It is the
6514 ratio between the dimensions of the current plane to the luma plane,
6515 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
6516 the luma plane and @code{0.5,0.5} for the chroma planes.
6519 Time of the current frame, expressed in seconds.
6522 Value of pixel component at current location for first video frame (top layer).
6525 Value of pixel component at current location for second video frame (bottom layer).
6529 The @code{blend} filter also supports the @ref{framesync} options.
6531 @subsection Examples
6535 Apply transition from bottom layer to top layer in first 10 seconds:
6537 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
6541 Apply linear horizontal transition from top layer to bottom layer:
6543 blend=all_expr='A*(X/W)+B*(1-X/W)'
6547 Apply 1x1 checkerboard effect:
6549 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
6553 Apply uncover left effect:
6555 blend=all_expr='if(gte(N*SW+X,W),A,B)'
6559 Apply uncover down effect:
6561 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
6565 Apply uncover up-left effect:
6567 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
6571 Split diagonally video and shows top and bottom layer on each side:
6573 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
6577 Display differences between the current and the previous frame:
6579 tblend=all_mode=grainextract
6585 Denoise frames using Block-Matching 3D algorithm.
6587 The filter accepts the following options.
6591 Set denoising strength. Default value is 1.
6592 Allowed range is from 0 to 999.9.
6593 The denoising algorithm is very sensitive to sigma, so adjust it
6594 according to the source.
6597 Set local patch size. This sets dimensions in 2D.
6600 Set sliding step for processing blocks. Default value is 4.
6601 Allowed range is from 1 to 64.
6602 Smaller values allows processing more reference blocks and is slower.
6605 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
6606 When set to 1, no block matching is done. Larger values allows more blocks
6608 Allowed range is from 1 to 256.
6611 Set radius for search block matching. Default is 9.
6612 Allowed range is from 1 to INT32_MAX.
6615 Set step between two search locations for block matching. Default is 1.
6616 Allowed range is from 1 to 64. Smaller is slower.
6619 Set threshold of mean square error for block matching. Valid range is 0 to
6623 Set thresholding parameter for hard thresholding in 3D transformed domain.
6624 Larger values results in stronger hard-thresholding filtering in frequency
6628 Set filtering estimation mode. Can be @code{basic} or @code{final}.
6629 Default is @code{basic}.
6632 If enabled, filter will use 2nd stream for block matching.
6633 Default is disabled for @code{basic} value of @var{estim} option,
6634 and always enabled if value of @var{estim} is @code{final}.
6637 Set planes to filter. Default is all available except alpha.
6640 @subsection Examples
6644 Basic filtering with bm3d:
6646 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
6650 Same as above, but filtering only luma:
6652 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
6656 Same as above, but with both estimation modes:
6658 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
6662 Same as above, but prefilter with @ref{nlmeans} filter instead:
6664 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
6670 Apply a boxblur algorithm to the input video.
6672 It accepts the following parameters:
6676 @item luma_radius, lr
6677 @item luma_power, lp
6678 @item chroma_radius, cr
6679 @item chroma_power, cp
6680 @item alpha_radius, ar
6681 @item alpha_power, ap
6685 A description of the accepted options follows.
6688 @item luma_radius, lr
6689 @item chroma_radius, cr
6690 @item alpha_radius, ar
6691 Set an expression for the box radius in pixels used for blurring the
6692 corresponding input plane.
6694 The radius value must be a non-negative number, and must not be
6695 greater than the value of the expression @code{min(w,h)/2} for the
6696 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
6699 Default value for @option{luma_radius} is "2". If not specified,
6700 @option{chroma_radius} and @option{alpha_radius} default to the
6701 corresponding value set for @option{luma_radius}.
6703 The expressions can contain the following constants:
6707 The input width and height in pixels.
6711 The input chroma image width and height in pixels.
6715 The horizontal and vertical chroma subsample values. For example, for the
6716 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
6719 @item luma_power, lp
6720 @item chroma_power, cp
6721 @item alpha_power, ap
6722 Specify how many times the boxblur filter is applied to the
6723 corresponding plane.
6725 Default value for @option{luma_power} is 2. If not specified,
6726 @option{chroma_power} and @option{alpha_power} default to the
6727 corresponding value set for @option{luma_power}.
6729 A value of 0 will disable the effect.
6732 @subsection Examples
6736 Apply a boxblur filter with the luma, chroma, and alpha radii
6739 boxblur=luma_radius=2:luma_power=1
6744 Set the luma radius to 2, and alpha and chroma radius to 0:
6746 boxblur=2:1:cr=0:ar=0
6750 Set the luma and chroma radii to a fraction of the video dimension:
6752 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
6758 Deinterlace the input video ("bwdif" stands for "Bob Weaver
6759 Deinterlacing Filter").
6761 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
6762 interpolation algorithms.
6763 It accepts the following parameters:
6767 The interlacing mode to adopt. It accepts one of the following values:
6771 Output one frame for each frame.
6773 Output one frame for each field.
6776 The default value is @code{send_field}.
6779 The picture field parity assumed for the input interlaced video. It accepts one
6780 of the following values:
6784 Assume the top field is first.
6786 Assume the bottom field is first.
6788 Enable automatic detection of field parity.
6791 The default value is @code{auto}.
6792 If the interlacing is unknown or the decoder does not export this information,
6793 top field first will be assumed.
6796 Specify which frames to deinterlace. Accepts one of the following
6801 Deinterlace all frames.
6803 Only deinterlace frames marked as interlaced.
6806 The default value is @code{all}.
6810 Remove all color information for all colors except for certain one.
6812 The filter accepts the following options:
6816 The color which will not be replaced with neutral chroma.
6819 Similarity percentage with the above color.
6820 0.01 matches only the exact key color, while 1.0 matches everything.
6824 0.0 makes pixels either fully gray, or not gray at all.
6825 Higher values result in more preserved color.
6828 Signals that the color passed is already in YUV instead of RGB.
6830 Literal colors like "green" or "red" don't make sense with this enabled anymore.
6831 This can be used to pass exact YUV values as hexadecimal numbers.
6835 YUV colorspace color/chroma keying.
6837 The filter accepts the following options:
6841 The color which will be replaced with transparency.
6844 Similarity percentage with the key color.
6846 0.01 matches only the exact key color, while 1.0 matches everything.
6851 0.0 makes pixels either fully transparent, or not transparent at all.
6853 Higher values result in semi-transparent pixels, with a higher transparency
6854 the more similar the pixels color is to the key color.
6857 Signals that the color passed is already in YUV instead of RGB.
6859 Literal colors like "green" or "red" don't make sense with this enabled anymore.
6860 This can be used to pass exact YUV values as hexadecimal numbers.
6863 @subsection Examples
6867 Make every green pixel in the input image transparent:
6869 ffmpeg -i input.png -vf chromakey=green out.png
6873 Overlay a greenscreen-video on top of a static black background.
6875 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
6879 @section chromashift
6880 Shift chroma pixels horizontally and/or vertically.
6882 The filter accepts the following options:
6885 Set amount to shift chroma-blue horizontally.
6887 Set amount to shift chroma-blue vertically.
6889 Set amount to shift chroma-red horizontally.
6891 Set amount to shift chroma-red vertically.
6893 Set edge mode, can be @var{smear}, default, or @var{warp}.
6898 Display CIE color diagram with pixels overlaid onto it.
6900 The filter accepts the following options:
6915 @item uhdtv, rec2020
6929 Set what gamuts to draw.
6931 See @code{system} option for available values.
6934 Set ciescope size, by default set to 512.
6937 Set intensity used to map input pixel values to CIE diagram.
6940 Set contrast used to draw tongue colors that are out of active color system gamut.
6943 Correct gamma displayed on scope, by default enabled.
6946 Show white point on CIE diagram, by default disabled.
6949 Set input gamma. Used only with XYZ input color space.
6954 Visualize information exported by some codecs.
6956 Some codecs can export information through frames using side-data or other
6957 means. For example, some MPEG based codecs export motion vectors through the
6958 @var{export_mvs} flag in the codec @option{flags2} option.
6960 The filter accepts the following option:
6964 Set motion vectors to visualize.
6966 Available flags for @var{mv} are:
6970 forward predicted MVs of P-frames
6972 forward predicted MVs of B-frames
6974 backward predicted MVs of B-frames
6978 Display quantization parameters using the chroma planes.
6981 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
6983 Available flags for @var{mv_type} are:
6987 forward predicted MVs
6989 backward predicted MVs
6992 @item frame_type, ft
6993 Set frame type to visualize motion vectors of.
6995 Available flags for @var{frame_type} are:
6999 intra-coded frames (I-frames)
7001 predicted frames (P-frames)
7003 bi-directionally predicted frames (B-frames)
7007 @subsection Examples
7011 Visualize forward predicted MVs of all frames using @command{ffplay}:
7013 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7017 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7019 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7023 @section colorbalance
7024 Modify intensity of primary colors (red, green and blue) of input frames.
7026 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7027 regions for the red-cyan, green-magenta or blue-yellow balance.
7029 A positive adjustment value shifts the balance towards the primary color, a negative
7030 value towards the complementary color.
7032 The filter accepts the following options:
7038 Adjust red, green and blue shadows (darkest pixels).
7043 Adjust red, green and blue midtones (medium pixels).
7048 Adjust red, green and blue highlights (brightest pixels).
7050 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7053 @subsection Examples
7057 Add red color cast to shadows:
7063 @section colorchannelmixer
7065 Adjust video input frames by re-mixing color channels.
7067 This filter modifies a color channel by adding the values associated to
7068 the other channels of the same pixels. For example if the value to
7069 modify is red, the output value will be:
7071 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
7074 The filter accepts the following options:
7081 Adjust contribution of input red, green, blue and alpha channels for output red channel.
7082 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
7088 Adjust contribution of input red, green, blue and alpha channels for output green channel.
7089 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
7095 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
7096 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
7102 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
7103 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
7105 Allowed ranges for options are @code{[-2.0, 2.0]}.
7108 @subsection Examples
7112 Convert source to grayscale:
7114 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
7117 Simulate sepia tones:
7119 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
7123 @subsection Commands
7125 This filter supports the all above options as @ref{commands}.
7128 RGB colorspace color keying.
7130 The filter accepts the following options:
7134 The color which will be replaced with transparency.
7137 Similarity percentage with the key color.
7139 0.01 matches only the exact key color, while 1.0 matches everything.
7144 0.0 makes pixels either fully transparent, or not transparent at all.
7146 Higher values result in semi-transparent pixels, with a higher transparency
7147 the more similar the pixels color is to the key color.
7150 @subsection Examples
7154 Make every green pixel in the input image transparent:
7156 ffmpeg -i input.png -vf colorkey=green out.png
7160 Overlay a greenscreen-video on top of a static background image.
7162 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
7167 Remove all color information for all RGB colors except for certain one.
7169 The filter accepts the following options:
7173 The color which will not be replaced with neutral gray.
7176 Similarity percentage with the above color.
7177 0.01 matches only the exact key color, while 1.0 matches everything.
7180 Blend percentage. 0.0 makes pixels fully gray.
7181 Higher values result in more preserved color.
7184 @section colorlevels
7186 Adjust video input frames using levels.
7188 The filter accepts the following options:
7195 Adjust red, green, blue and alpha input black point.
7196 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7202 Adjust red, green, blue and alpha input white point.
7203 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
7205 Input levels are used to lighten highlights (bright tones), darken shadows
7206 (dark tones), change the balance of bright and dark tones.
7212 Adjust red, green, blue and alpha output black point.
7213 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
7219 Adjust red, green, blue and alpha output white point.
7220 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
7222 Output levels allows manual selection of a constrained output level range.
7225 @subsection Examples
7229 Make video output darker:
7231 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
7237 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
7241 Make video output lighter:
7243 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
7247 Increase brightness:
7249 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
7253 @section colormatrix
7255 Convert color matrix.
7257 The filter accepts the following options:
7262 Specify the source and destination color matrix. Both values must be
7265 The accepted values are:
7293 For example to convert from BT.601 to SMPTE-240M, use the command:
7295 colormatrix=bt601:smpte240m
7300 Convert colorspace, transfer characteristics or color primaries.
7301 Input video needs to have an even size.
7303 The filter accepts the following options:
7308 Specify all color properties at once.
7310 The accepted values are:
7340 Specify output colorspace.
7342 The accepted values are:
7351 BT.470BG or BT.601-6 625
7354 SMPTE-170M or BT.601-6 525
7363 BT.2020 with non-constant luminance
7369 Specify output transfer characteristics.
7371 The accepted values are:
7383 Constant gamma of 2.2
7386 Constant gamma of 2.8
7389 SMPTE-170M, BT.601-6 625 or BT.601-6 525
7407 BT.2020 for 10-bits content
7410 BT.2020 for 12-bits content
7416 Specify output color primaries.
7418 The accepted values are:
7427 BT.470BG or BT.601-6 625
7430 SMPTE-170M or BT.601-6 525
7454 Specify output color range.
7456 The accepted values are:
7459 TV (restricted) range
7462 MPEG (restricted) range
7473 Specify output color format.
7475 The accepted values are:
7478 YUV 4:2:0 planar 8-bits
7481 YUV 4:2:0 planar 10-bits
7484 YUV 4:2:0 planar 12-bits
7487 YUV 4:2:2 planar 8-bits
7490 YUV 4:2:2 planar 10-bits
7493 YUV 4:2:2 planar 12-bits
7496 YUV 4:4:4 planar 8-bits
7499 YUV 4:4:4 planar 10-bits
7502 YUV 4:4:4 planar 12-bits
7507 Do a fast conversion, which skips gamma/primary correction. This will take
7508 significantly less CPU, but will be mathematically incorrect. To get output
7509 compatible with that produced by the colormatrix filter, use fast=1.
7512 Specify dithering mode.
7514 The accepted values are:
7520 Floyd-Steinberg dithering
7524 Whitepoint adaptation mode.
7526 The accepted values are:
7529 Bradford whitepoint adaptation
7532 von Kries whitepoint adaptation
7535 identity whitepoint adaptation (i.e. no whitepoint adaptation)
7539 Override all input properties at once. Same accepted values as @ref{all}.
7542 Override input colorspace. Same accepted values as @ref{space}.
7545 Override input color primaries. Same accepted values as @ref{primaries}.
7548 Override input transfer characteristics. Same accepted values as @ref{trc}.
7551 Override input color range. Same accepted values as @ref{range}.
7555 The filter converts the transfer characteristics, color space and color
7556 primaries to the specified user values. The output value, if not specified,
7557 is set to a default value based on the "all" property. If that property is
7558 also not specified, the filter will log an error. The output color range and
7559 format default to the same value as the input color range and format. The
7560 input transfer characteristics, color space, color primaries and color range
7561 should be set on the input data. If any of these are missing, the filter will
7562 log an error and no conversion will take place.
7564 For example to convert the input to SMPTE-240M, use the command:
7566 colorspace=smpte240m
7569 @section convolution
7571 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
7573 The filter accepts the following options:
7580 Set matrix for each plane.
7581 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
7582 and from 1 to 49 odd number of signed integers in @var{row} mode.
7588 Set multiplier for calculated value for each plane.
7589 If unset or 0, it will be sum of all matrix elements.
7595 Set bias for each plane. This value is added to the result of the multiplication.
7596 Useful for making the overall image brighter or darker. Default is 0.0.
7602 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
7603 Default is @var{square}.
7606 @subsection Examples
7612 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"
7618 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"
7624 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"
7630 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"
7634 Apply laplacian edge detector which includes diagonals:
7636 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"
7642 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"
7648 Apply 2D convolution of video stream in frequency domain using second stream
7651 The filter accepts the following options:
7655 Set which planes to process.
7658 Set which impulse video frames will be processed, can be @var{first}
7659 or @var{all}. Default is @var{all}.
7662 The @code{convolve} filter also supports the @ref{framesync} options.
7666 Copy the input video source unchanged to the output. This is mainly useful for
7671 Video filtering on GPU using Apple's CoreImage API on OSX.
7673 Hardware acceleration is based on an OpenGL context. Usually, this means it is
7674 processed by video hardware. However, software-based OpenGL implementations
7675 exist which means there is no guarantee for hardware processing. It depends on
7678 There are many filters and image generators provided by Apple that come with a
7679 large variety of options. The filter has to be referenced by its name along
7682 The coreimage filter accepts the following options:
7685 List all available filters and generators along with all their respective
7686 options as well as possible minimum and maximum values along with the default
7693 Specify all filters by their respective name and options.
7694 Use @var{list_filters} to determine all valid filter names and options.
7695 Numerical options are specified by a float value and are automatically clamped
7696 to their respective value range. Vector and color options have to be specified
7697 by a list of space separated float values. Character escaping has to be done.
7698 A special option name @code{default} is available to use default options for a
7701 It is required to specify either @code{default} or at least one of the filter options.
7702 All omitted options are used with their default values.
7703 The syntax of the filter string is as follows:
7705 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
7709 Specify a rectangle where the output of the filter chain is copied into the
7710 input image. It is given by a list of space separated float values:
7712 output_rect=x\ y\ width\ height
7714 If not given, the output rectangle equals the dimensions of the input image.
7715 The output rectangle is automatically cropped at the borders of the input
7716 image. Negative values are valid for each component.
7718 output_rect=25\ 25\ 100\ 100
7722 Several filters can be chained for successive processing without GPU-HOST
7723 transfers allowing for fast processing of complex filter chains.
7724 Currently, only filters with zero (generators) or exactly one (filters) input
7725 image and one output image are supported. Also, transition filters are not yet
7728 Some filters generate output images with additional padding depending on the
7729 respective filter kernel. The padding is automatically removed to ensure the
7730 filter output has the same size as the input image.
7732 For image generators, the size of the output image is determined by the
7733 previous output image of the filter chain or the input image of the whole
7734 filterchain, respectively. The generators do not use the pixel information of
7735 this image to generate their output. However, the generated output is
7736 blended onto this image, resulting in partial or complete coverage of the
7739 The @ref{coreimagesrc} video source can be used for generating input images
7740 which are directly fed into the filter chain. By using it, providing input
7741 images by another video source or an input video is not required.
7743 @subsection Examples
7748 List all filters available:
7750 coreimage=list_filters=true
7754 Use the CIBoxBlur filter with default options to blur an image:
7756 coreimage=filter=CIBoxBlur@@default
7760 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
7761 its center at 100x100 and a radius of 50 pixels:
7763 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
7767 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
7768 given as complete and escaped command-line for Apple's standard bash shell:
7770 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
7776 Cover a rectangular object
7778 It accepts the following options:
7782 Filepath of the optional cover image, needs to be in yuv420.
7787 It accepts the following values:
7790 cover it by the supplied image
7792 cover it by interpolating the surrounding pixels
7795 Default value is @var{blur}.
7798 @subsection Examples
7802 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
7804 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
7810 Crop the input video to given dimensions.
7812 It accepts the following parameters:
7816 The width of the output video. It defaults to @code{iw}.
7817 This expression is evaluated only once during the filter
7818 configuration, or when the @samp{w} or @samp{out_w} command is sent.
7821 The height of the output video. It defaults to @code{ih}.
7822 This expression is evaluated only once during the filter
7823 configuration, or when the @samp{h} or @samp{out_h} command is sent.
7826 The horizontal position, in the input video, of the left edge of the output
7827 video. It defaults to @code{(in_w-out_w)/2}.
7828 This expression is evaluated per-frame.
7831 The vertical position, in the input video, of the top edge of the output video.
7832 It defaults to @code{(in_h-out_h)/2}.
7833 This expression is evaluated per-frame.
7836 If set to 1 will force the output display aspect ratio
7837 to be the same of the input, by changing the output sample aspect
7838 ratio. It defaults to 0.
7841 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
7842 width/height/x/y as specified and will not be rounded to nearest smaller value.
7846 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
7847 expressions containing the following constants:
7852 The computed values for @var{x} and @var{y}. They are evaluated for
7857 The input width and height.
7861 These are the same as @var{in_w} and @var{in_h}.
7865 The output (cropped) width and height.
7869 These are the same as @var{out_w} and @var{out_h}.
7872 same as @var{iw} / @var{ih}
7875 input sample aspect ratio
7878 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
7882 horizontal and vertical chroma subsample values. For example for the
7883 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
7886 The number of the input frame, starting from 0.
7889 the position in the file of the input frame, NAN if unknown
7892 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
7896 The expression for @var{out_w} may depend on the value of @var{out_h},
7897 and the expression for @var{out_h} may depend on @var{out_w}, but they
7898 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
7899 evaluated after @var{out_w} and @var{out_h}.
7901 The @var{x} and @var{y} parameters specify the expressions for the
7902 position of the top-left corner of the output (non-cropped) area. They
7903 are evaluated for each frame. If the evaluated value is not valid, it
7904 is approximated to the nearest valid value.
7906 The expression for @var{x} may depend on @var{y}, and the expression
7907 for @var{y} may depend on @var{x}.
7909 @subsection Examples
7913 Crop area with size 100x100 at position (12,34).
7918 Using named options, the example above becomes:
7920 crop=w=100:h=100:x=12:y=34
7924 Crop the central input area with size 100x100:
7930 Crop the central input area with size 2/3 of the input video:
7932 crop=2/3*in_w:2/3*in_h
7936 Crop the input video central square:
7943 Delimit the rectangle with the top-left corner placed at position
7944 100:100 and the right-bottom corner corresponding to the right-bottom
7945 corner of the input image.
7947 crop=in_w-100:in_h-100:100:100
7951 Crop 10 pixels from the left and right borders, and 20 pixels from
7952 the top and bottom borders
7954 crop=in_w-2*10:in_h-2*20
7958 Keep only the bottom right quarter of the input image:
7960 crop=in_w/2:in_h/2:in_w/2:in_h/2
7964 Crop height for getting Greek harmony:
7966 crop=in_w:1/PHI*in_w
7970 Apply trembling effect:
7972 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)
7976 Apply erratic camera effect depending on timestamp:
7978 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)"
7982 Set x depending on the value of y:
7984 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
7988 @subsection Commands
7990 This filter supports the following commands:
7996 Set width/height of the output video and the horizontal/vertical position
7998 The command accepts the same syntax of the corresponding option.
8000 If the specified expression is not valid, it is kept at its current
8006 Auto-detect the crop size.
8008 It calculates the necessary cropping parameters and prints the
8009 recommended parameters via the logging system. The detected dimensions
8010 correspond to the non-black area of the input video.
8012 It accepts the following parameters:
8017 Set higher black value threshold, which can be optionally specified
8018 from nothing (0) to everything (255 for 8-bit based formats). An intensity
8019 value greater to the set value is considered non-black. It defaults to 24.
8020 You can also specify a value between 0.0 and 1.0 which will be scaled depending
8021 on the bitdepth of the pixel format.
8024 The value which the width/height should be divisible by. It defaults to
8025 16. The offset is automatically adjusted to center the video. Use 2 to
8026 get only even dimensions (needed for 4:2:2 video). 16 is best when
8027 encoding to most video codecs.
8029 @item reset_count, reset
8030 Set the counter that determines after how many frames cropdetect will
8031 reset the previously detected largest video area and start over to
8032 detect the current optimal crop area. Default value is 0.
8034 This can be useful when channel logos distort the video area. 0
8035 indicates 'never reset', and returns the largest area encountered during
8042 Delay video filtering until a given wallclock timestamp. The filter first
8043 passes on @option{preroll} amount of frames, then it buffers at most
8044 @option{buffer} amount of frames and waits for the cue. After reaching the cue
8045 it forwards the buffered frames and also any subsequent frames coming in its
8048 The filter can be used synchronize the output of multiple ffmpeg processes for
8049 realtime output devices like decklink. By putting the delay in the filtering
8050 chain and pre-buffering frames the process can pass on data to output almost
8051 immediately after the target wallclock timestamp is reached.
8053 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
8059 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
8062 The duration of content to pass on as preroll expressed in seconds. Default is 0.
8065 The maximum duration of content to buffer before waiting for the cue expressed
8066 in seconds. Default is 0.
8073 Apply color adjustments using curves.
8075 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
8076 component (red, green and blue) has its values defined by @var{N} key points
8077 tied from each other using a smooth curve. The x-axis represents the pixel
8078 values from the input frame, and the y-axis the new pixel values to be set for
8081 By default, a component curve is defined by the two points @var{(0;0)} and
8082 @var{(1;1)}. This creates a straight line where each original pixel value is
8083 "adjusted" to its own value, which means no change to the image.
8085 The filter allows you to redefine these two points and add some more. A new
8086 curve (using a natural cubic spline interpolation) will be define to pass
8087 smoothly through all these new coordinates. The new defined points needs to be
8088 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
8089 be in the @var{[0;1]} interval. If the computed curves happened to go outside
8090 the vector spaces, the values will be clipped accordingly.
8092 The filter accepts the following options:
8096 Select one of the available color presets. This option can be used in addition
8097 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
8098 options takes priority on the preset values.
8099 Available presets are:
8102 @item color_negative
8105 @item increase_contrast
8107 @item linear_contrast
8108 @item medium_contrast
8110 @item strong_contrast
8113 Default is @code{none}.
8115 Set the master key points. These points will define a second pass mapping. It
8116 is sometimes called a "luminance" or "value" mapping. It can be used with
8117 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
8118 post-processing LUT.
8120 Set the key points for the red component.
8122 Set the key points for the green component.
8124 Set the key points for the blue component.
8126 Set the key points for all components (not including master).
8127 Can be used in addition to the other key points component
8128 options. In this case, the unset component(s) will fallback on this
8129 @option{all} setting.
8131 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
8133 Save Gnuplot script of the curves in specified file.
8136 To avoid some filtergraph syntax conflicts, each key points list need to be
8137 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
8139 @subsection Examples
8143 Increase slightly the middle level of blue:
8145 curves=blue='0/0 0.5/0.58 1/1'
8151 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'
8153 Here we obtain the following coordinates for each components:
8156 @code{(0;0.11) (0.42;0.51) (1;0.95)}
8158 @code{(0;0) (0.50;0.48) (1;1)}
8160 @code{(0;0.22) (0.49;0.44) (1;0.80)}
8164 The previous example can also be achieved with the associated built-in preset:
8166 curves=preset=vintage
8176 Use a Photoshop preset and redefine the points of the green component:
8178 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
8182 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
8183 and @command{gnuplot}:
8185 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
8186 gnuplot -p /tmp/curves.plt
8192 Video data analysis filter.
8194 This filter shows hexadecimal pixel values of part of video.
8196 The filter accepts the following options:
8200 Set output video size.
8203 Set x offset from where to pick pixels.
8206 Set y offset from where to pick pixels.
8209 Set scope mode, can be one of the following:
8212 Draw hexadecimal pixel values with white color on black background.
8215 Draw hexadecimal pixel values with input video pixel color on black
8219 Draw hexadecimal pixel values on color background picked from input video,
8220 the text color is picked in such way so its always visible.
8224 Draw rows and columns numbers on left and top of video.
8227 Set background opacity.
8232 Denoise frames using 2D DCT (frequency domain filtering).
8234 This filter is not designed for real time.
8236 The filter accepts the following options:
8240 Set the noise sigma constant.
8242 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
8243 coefficient (absolute value) below this threshold with be dropped.
8245 If you need a more advanced filtering, see @option{expr}.
8247 Default is @code{0}.
8250 Set number overlapping pixels for each block. Since the filter can be slow, you
8251 may want to reduce this value, at the cost of a less effective filter and the
8252 risk of various artefacts.
8254 If the overlapping value doesn't permit processing the whole input width or
8255 height, a warning will be displayed and according borders won't be denoised.
8257 Default value is @var{blocksize}-1, which is the best possible setting.
8260 Set the coefficient factor expression.
8262 For each coefficient of a DCT block, this expression will be evaluated as a
8263 multiplier value for the coefficient.
8265 If this is option is set, the @option{sigma} option will be ignored.
8267 The absolute value of the coefficient can be accessed through the @var{c}
8271 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
8272 @var{blocksize}, which is the width and height of the processed blocks.
8274 The default value is @var{3} (8x8) and can be raised to @var{4} for a
8275 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
8276 on the speed processing. Also, a larger block size does not necessarily means a
8280 @subsection Examples
8282 Apply a denoise with a @option{sigma} of @code{4.5}:
8287 The same operation can be achieved using the expression system:
8289 dctdnoiz=e='gte(c, 4.5*3)'
8292 Violent denoise using a block size of @code{16x16}:
8299 Remove banding artifacts from input video.
8300 It works by replacing banded pixels with average value of referenced pixels.
8302 The filter accepts the following options:
8309 Set banding detection threshold for each plane. Default is 0.02.
8310 Valid range is 0.00003 to 0.5.
8311 If difference between current pixel and reference pixel is less than threshold,
8312 it will be considered as banded.
8315 Banding detection range in pixels. Default is 16. If positive, random number
8316 in range 0 to set value will be used. If negative, exact absolute value
8318 The range defines square of four pixels around current pixel.
8321 Set direction in radians from which four pixel will be compared. If positive,
8322 random direction from 0 to set direction will be picked. If negative, exact of
8323 absolute value will be picked. For example direction 0, -PI or -2*PI radians
8324 will pick only pixels on same row and -PI/2 will pick only pixels on same
8328 If enabled, current pixel is compared with average value of all four
8329 surrounding pixels. The default is enabled. If disabled current pixel is
8330 compared with all four surrounding pixels. The pixel is considered banded
8331 if only all four differences with surrounding pixels are less than threshold.
8334 If enabled, current pixel is changed if and only if all pixel components are banded,
8335 e.g. banding detection threshold is triggered for all color components.
8336 The default is disabled.
8341 Remove blocking artifacts from input video.
8343 The filter accepts the following options:
8347 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
8348 This controls what kind of deblocking is applied.
8351 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
8357 Set blocking detection thresholds. Allowed range is 0 to 1.
8358 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
8359 Using higher threshold gives more deblocking strength.
8360 Setting @var{alpha} controls threshold detection at exact edge of block.
8361 Remaining options controls threshold detection near the edge. Each one for
8362 below/above or left/right. Setting any of those to @var{0} disables
8366 Set planes to filter. Default is to filter all available planes.
8369 @subsection Examples
8373 Deblock using weak filter and block size of 4 pixels.
8375 deblock=filter=weak:block=4
8379 Deblock using strong filter, block size of 4 pixels and custom thresholds for
8380 deblocking more edges.
8382 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
8386 Similar as above, but filter only first plane.
8388 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
8392 Similar as above, but filter only second and third plane.
8394 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
8401 Drop duplicated frames at regular intervals.
8403 The filter accepts the following options:
8407 Set the number of frames from which one will be dropped. Setting this to
8408 @var{N} means one frame in every batch of @var{N} frames will be dropped.
8409 Default is @code{5}.
8412 Set the threshold for duplicate detection. If the difference metric for a frame
8413 is less than or equal to this value, then it is declared as duplicate. Default
8417 Set scene change threshold. Default is @code{15}.
8421 Set the size of the x and y-axis blocks used during metric calculations.
8422 Larger blocks give better noise suppression, but also give worse detection of
8423 small movements. Must be a power of two. Default is @code{32}.
8426 Mark main input as a pre-processed input and activate clean source input
8427 stream. This allows the input to be pre-processed with various filters to help
8428 the metrics calculation while keeping the frame selection lossless. When set to
8429 @code{1}, the first stream is for the pre-processed input, and the second
8430 stream is the clean source from where the kept frames are chosen. Default is
8434 Set whether or not chroma is considered in the metric calculations. Default is
8440 Apply 2D deconvolution of video stream in frequency domain using second stream
8443 The filter accepts the following options:
8447 Set which planes to process.
8450 Set which impulse video frames will be processed, can be @var{first}
8451 or @var{all}. Default is @var{all}.
8454 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
8455 and height are not same and not power of 2 or if stream prior to convolving
8459 The @code{deconvolve} filter also supports the @ref{framesync} options.
8463 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
8465 It accepts the following options:
8469 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
8470 @var{rainbows} for cross-color reduction.
8473 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
8476 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
8479 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
8482 Set temporal chroma threshold. Lower values increases reduction of cross-color.
8487 Apply deflate effect to the video.
8489 This filter replaces the pixel by the local(3x3) average by taking into account
8490 only values lower than the pixel.
8492 It accepts the following options:
8499 Limit the maximum change for each plane, default is 65535.
8500 If 0, plane will remain unchanged.
8505 Remove temporal frame luminance variations.
8507 It accepts the following options:
8511 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
8514 Set averaging mode to smooth temporal luminance variations.
8516 Available values are:
8541 Do not actually modify frame. Useful when one only wants metadata.
8546 Remove judder produced by partially interlaced telecined content.
8548 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
8549 source was partially telecined content then the output of @code{pullup,dejudder}
8550 will have a variable frame rate. May change the recorded frame rate of the
8551 container. Aside from that change, this filter will not affect constant frame
8554 The option available in this filter is:
8558 Specify the length of the window over which the judder repeats.
8560 Accepts any integer greater than 1. Useful values are:
8564 If the original was telecined from 24 to 30 fps (Film to NTSC).
8567 If the original was telecined from 25 to 30 fps (PAL to NTSC).
8570 If a mixture of the two.
8573 The default is @samp{4}.
8578 Suppress a TV station logo by a simple interpolation of the surrounding
8579 pixels. Just set a rectangle covering the logo and watch it disappear
8580 (and sometimes something even uglier appear - your mileage may vary).
8582 It accepts the following parameters:
8587 Specify the top left corner coordinates of the logo. They must be
8592 Specify the width and height of the logo to clear. They must be
8596 Specify the thickness of the fuzzy edge of the rectangle (added to
8597 @var{w} and @var{h}). The default value is 1. This option is
8598 deprecated, setting higher values should no longer be necessary and
8602 When set to 1, a green rectangle is drawn on the screen to simplify
8603 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
8604 The default value is 0.
8606 The rectangle is drawn on the outermost pixels which will be (partly)
8607 replaced with interpolated values. The values of the next pixels
8608 immediately outside this rectangle in each direction will be used to
8609 compute the interpolated pixel values inside the rectangle.
8613 @subsection Examples
8617 Set a rectangle covering the area with top left corner coordinates 0,0
8618 and size 100x77, and a band of size 10:
8620 delogo=x=0:y=0:w=100:h=77:band=10
8627 Remove the rain in the input image/video by applying the derain methods based on
8628 convolutional neural networks. Supported models:
8632 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
8633 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
8636 Training as well as model generation scripts are provided in
8637 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
8639 Native model files (.model) can be generated from TensorFlow model
8640 files (.pb) by using tools/python/convert.py
8642 The filter accepts the following options:
8646 Specify which filter to use. This option accepts the following values:
8650 Derain filter. To conduct derain filter, you need to use a derain model.
8653 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
8655 Default value is @samp{derain}.
8658 Specify which DNN backend to use for model loading and execution. This option accepts
8659 the following values:
8663 Native implementation of DNN loading and execution.
8666 TensorFlow backend. To enable this backend you
8667 need to install the TensorFlow for C library (see
8668 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
8669 @code{--enable-libtensorflow}
8671 Default value is @samp{native}.
8674 Set path to model file specifying network architecture and its parameters.
8675 Note that different backends use different file formats. TensorFlow and native
8676 backend can load files for only its format.
8681 Attempt to fix small changes in horizontal and/or vertical shift. This
8682 filter helps remove camera shake from hand-holding a camera, bumping a
8683 tripod, moving on a vehicle, etc.
8685 The filter accepts the following options:
8693 Specify a rectangular area where to limit the search for motion
8695 If desired the search for motion vectors can be limited to a
8696 rectangular area of the frame defined by its top left corner, width
8697 and height. These parameters have the same meaning as the drawbox
8698 filter which can be used to visualise the position of the bounding
8701 This is useful when simultaneous movement of subjects within the frame
8702 might be confused for camera motion by the motion vector search.
8704 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
8705 then the full frame is used. This allows later options to be set
8706 without specifying the bounding box for the motion vector search.
8708 Default - search the whole frame.
8712 Specify the maximum extent of movement in x and y directions in the
8713 range 0-64 pixels. Default 16.
8716 Specify how to generate pixels to fill blanks at the edge of the
8717 frame. Available values are:
8720 Fill zeroes at blank locations
8722 Original image at blank locations
8724 Extruded edge value at blank locations
8726 Mirrored edge at blank locations
8728 Default value is @samp{mirror}.
8731 Specify the blocksize to use for motion search. Range 4-128 pixels,
8735 Specify the contrast threshold for blocks. Only blocks with more than
8736 the specified contrast (difference between darkest and lightest
8737 pixels) will be considered. Range 1-255, default 125.
8740 Specify the search strategy. Available values are:
8743 Set exhaustive search
8745 Set less exhaustive search.
8747 Default value is @samp{exhaustive}.
8750 If set then a detailed log of the motion search is written to the
8757 Remove unwanted contamination of foreground colors, caused by reflected color of
8758 greenscreen or bluescreen.
8760 This filter accepts the following options:
8764 Set what type of despill to use.
8767 Set how spillmap will be generated.
8770 Set how much to get rid of still remaining spill.
8773 Controls amount of red in spill area.
8776 Controls amount of green in spill area.
8777 Should be -1 for greenscreen.
8780 Controls amount of blue in spill area.
8781 Should be -1 for bluescreen.
8784 Controls brightness of spill area, preserving colors.
8787 Modify alpha from generated spillmap.
8792 Apply an exact inverse of the telecine operation. It requires a predefined
8793 pattern specified using the pattern option which must be the same as that passed
8794 to the telecine filter.
8796 This filter accepts the following options:
8805 The default value is @code{top}.
8809 A string of numbers representing the pulldown pattern you wish to apply.
8810 The default value is @code{23}.
8813 A number representing position of the first frame with respect to the telecine
8814 pattern. This is to be used if the stream is cut. The default value is @code{0}.
8819 Apply dilation effect to the video.
8821 This filter replaces the pixel by the local(3x3) maximum.
8823 It accepts the following options:
8830 Limit the maximum change for each plane, default is 65535.
8831 If 0, plane will remain unchanged.
8834 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
8837 Flags to local 3x3 coordinates maps like this:
8846 Displace pixels as indicated by second and third input stream.
8848 It takes three input streams and outputs one stream, the first input is the
8849 source, and second and third input are displacement maps.
8851 The second input specifies how much to displace pixels along the
8852 x-axis, while the third input specifies how much to displace pixels
8854 If one of displacement map streams terminates, last frame from that
8855 displacement map will be used.
8857 Note that once generated, displacements maps can be reused over and over again.
8859 A description of the accepted options follows.
8863 Set displace behavior for pixels that are out of range.
8865 Available values are:
8868 Missing pixels are replaced by black pixels.
8871 Adjacent pixels will spread out to replace missing pixels.
8874 Out of range pixels are wrapped so they point to pixels of other side.
8877 Out of range pixels will be replaced with mirrored pixels.
8879 Default is @samp{smear}.
8883 @subsection Examples
8887 Add ripple effect to rgb input of video size hd720:
8889 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
8893 Add wave effect to rgb input of video size hd720:
8895 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
8901 Draw a colored box on the input image.
8903 It accepts the following parameters:
8908 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
8912 The expressions which specify the width and height of the box; if 0 they are interpreted as
8913 the input width and height. It defaults to 0.
8916 Specify the color of the box to write. For the general syntax of this option,
8917 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
8918 value @code{invert} is used, the box edge color is the same as the
8919 video with inverted luma.
8922 The expression which sets the thickness of the box edge.
8923 A value of @code{fill} will create a filled box. Default value is @code{3}.
8925 See below for the list of accepted constants.
8928 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
8929 will overwrite the video's color and alpha pixels.
8930 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
8933 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
8934 following constants:
8938 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
8942 horizontal and vertical chroma subsample values. For example for the
8943 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8947 The input width and height.
8950 The input sample aspect ratio.
8954 The x and y offset coordinates where the box is drawn.
8958 The width and height of the drawn box.
8961 The thickness of the drawn box.
8963 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
8964 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
8968 @subsection Examples
8972 Draw a black box around the edge of the input image:
8978 Draw a box with color red and an opacity of 50%:
8980 drawbox=10:20:200:60:red@@0.5
8983 The previous example can be specified as:
8985 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
8989 Fill the box with pink color:
8991 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
8995 Draw a 2-pixel red 2.40:1 mask:
8997 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
9001 @subsection Commands
9002 This filter supports same commands as options.
9003 The command accepts the same syntax of the corresponding option.
9005 If the specified expression is not valid, it is kept at its current
9010 Draw a grid on the input image.
9012 It accepts the following parameters:
9017 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
9021 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
9022 input width and height, respectively, minus @code{thickness}, so image gets
9023 framed. Default to 0.
9026 Specify the color of the grid. For the general syntax of this option,
9027 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
9028 value @code{invert} is used, the grid color is the same as the
9029 video with inverted luma.
9032 The expression which sets the thickness of the grid line. Default value is @code{1}.
9034 See below for the list of accepted constants.
9037 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
9038 will overwrite the video's color and alpha pixels.
9039 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
9042 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
9043 following constants:
9047 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
9051 horizontal and vertical chroma subsample values. For example for the
9052 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9056 The input grid cell width and height.
9059 The input sample aspect ratio.
9063 The x and y coordinates of some point of grid intersection (meant to configure offset).
9067 The width and height of the drawn cell.
9070 The thickness of the drawn cell.
9072 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
9073 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
9077 @subsection Examples
9081 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
9083 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
9087 Draw a white 3x3 grid with an opacity of 50%:
9089 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
9093 @subsection Commands
9094 This filter supports same commands as options.
9095 The command accepts the same syntax of the corresponding option.
9097 If the specified expression is not valid, it is kept at its current
9103 Draw a text string or text from a specified file on top of a video, using the
9104 libfreetype library.
9106 To enable compilation of this filter, you need to configure FFmpeg with
9107 @code{--enable-libfreetype}.
9108 To enable default font fallback and the @var{font} option you need to
9109 configure FFmpeg with @code{--enable-libfontconfig}.
9110 To enable the @var{text_shaping} option, you need to configure FFmpeg with
9111 @code{--enable-libfribidi}.
9115 It accepts the following parameters:
9120 Used to draw a box around text using the background color.
9121 The value must be either 1 (enable) or 0 (disable).
9122 The default value of @var{box} is 0.
9125 Set the width of the border to be drawn around the box using @var{boxcolor}.
9126 The default value of @var{boxborderw} is 0.
9129 The color to be used for drawing box around text. For the syntax of this
9130 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9132 The default value of @var{boxcolor} is "white".
9135 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
9136 The default value of @var{line_spacing} is 0.
9139 Set the width of the border to be drawn around the text using @var{bordercolor}.
9140 The default value of @var{borderw} is 0.
9143 Set the color to be used for drawing border around text. For the syntax of this
9144 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9146 The default value of @var{bordercolor} is "black".
9149 Select how the @var{text} is expanded. Can be either @code{none},
9150 @code{strftime} (deprecated) or
9151 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
9155 Set a start time for the count. Value is in microseconds. Only applied
9156 in the deprecated strftime expansion mode. To emulate in normal expansion
9157 mode use the @code{pts} function, supplying the start time (in seconds)
9158 as the second argument.
9161 If true, check and fix text coords to avoid clipping.
9164 The color to be used for drawing fonts. For the syntax of this option, check
9165 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9167 The default value of @var{fontcolor} is "black".
9169 @item fontcolor_expr
9170 String which is expanded the same way as @var{text} to obtain dynamic
9171 @var{fontcolor} value. By default this option has empty value and is not
9172 processed. When this option is set, it overrides @var{fontcolor} option.
9175 The font family to be used for drawing text. By default Sans.
9178 The font file to be used for drawing text. The path must be included.
9179 This parameter is mandatory if the fontconfig support is disabled.
9182 Draw the text applying alpha blending. The value can
9183 be a number between 0.0 and 1.0.
9184 The expression accepts the same variables @var{x, y} as well.
9185 The default value is 1.
9186 Please see @var{fontcolor_expr}.
9189 The font size to be used for drawing text.
9190 The default value of @var{fontsize} is 16.
9193 If set to 1, attempt to shape the text (for example, reverse the order of
9194 right-to-left text and join Arabic characters) before drawing it.
9195 Otherwise, just draw the text exactly as given.
9196 By default 1 (if supported).
9199 The flags to be used for loading the fonts.
9201 The flags map the corresponding flags supported by libfreetype, and are
9202 a combination of the following values:
9209 @item vertical_layout
9210 @item force_autohint
9213 @item ignore_global_advance_width
9215 @item ignore_transform
9221 Default value is "default".
9223 For more information consult the documentation for the FT_LOAD_*
9227 The color to be used for drawing a shadow behind the drawn text. For the
9228 syntax of this option, check the @ref{color syntax,,"Color" section in the
9229 ffmpeg-utils manual,ffmpeg-utils}.
9231 The default value of @var{shadowcolor} is "black".
9235 The x and y offsets for the text shadow position with respect to the
9236 position of the text. They can be either positive or negative
9237 values. The default value for both is "0".
9240 The starting frame number for the n/frame_num variable. The default value
9244 The size in number of spaces to use for rendering the tab.
9248 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
9249 format. It can be used with or without text parameter. @var{timecode_rate}
9250 option must be specified.
9252 @item timecode_rate, rate, r
9253 Set the timecode frame rate (timecode only). Value will be rounded to nearest
9254 integer. Minimum value is "1".
9255 Drop-frame timecode is supported for frame rates 30 & 60.
9258 If set to 1, the output of the timecode option will wrap around at 24 hours.
9259 Default is 0 (disabled).
9262 The text string to be drawn. The text must be a sequence of UTF-8
9264 This parameter is mandatory if no file is specified with the parameter
9268 A text file containing text to be drawn. The text must be a sequence
9269 of UTF-8 encoded characters.
9271 This parameter is mandatory if no text string is specified with the
9272 parameter @var{text}.
9274 If both @var{text} and @var{textfile} are specified, an error is thrown.
9277 If set to 1, the @var{textfile} will be reloaded before each frame.
9278 Be sure to update it atomically, or it may be read partially, or even fail.
9282 The expressions which specify the offsets where text will be drawn
9283 within the video frame. They are relative to the top/left border of the
9286 The default value of @var{x} and @var{y} is "0".
9288 See below for the list of accepted constants and functions.
9291 The parameters for @var{x} and @var{y} are expressions containing the
9292 following constants and functions:
9296 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
9300 horizontal and vertical chroma subsample values. For example for the
9301 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9304 the height of each text line
9312 @item max_glyph_a, ascent
9313 the maximum distance from the baseline to the highest/upper grid
9314 coordinate used to place a glyph outline point, for all the rendered
9316 It is a positive value, due to the grid's orientation with the Y axis
9319 @item max_glyph_d, descent
9320 the maximum distance from the baseline to the lowest grid coordinate
9321 used to place a glyph outline point, for all the rendered glyphs.
9322 This is a negative value, due to the grid's orientation, with the Y axis
9326 maximum glyph height, that is the maximum height for all the glyphs
9327 contained in the rendered text, it is equivalent to @var{ascent} -
9331 maximum glyph width, that is the maximum width for all the glyphs
9332 contained in the rendered text
9335 the number of input frame, starting from 0
9337 @item rand(min, max)
9338 return a random number included between @var{min} and @var{max}
9341 The input sample aspect ratio.
9344 timestamp expressed in seconds, NAN if the input timestamp is unknown
9347 the height of the rendered text
9350 the width of the rendered text
9354 the x and y offset coordinates where the text is drawn.
9356 These parameters allow the @var{x} and @var{y} expressions to refer
9357 to each other, so you can for example specify @code{y=x/dar}.
9360 A one character description of the current frame's picture type.
9363 The current packet's position in the input file or stream
9364 (in bytes, from the start of the input). A value of -1 indicates
9365 this info is not available.
9368 The current packet's duration, in seconds.
9371 The current packet's size (in bytes).
9374 @anchor{drawtext_expansion}
9375 @subsection Text expansion
9377 If @option{expansion} is set to @code{strftime},
9378 the filter recognizes strftime() sequences in the provided text and
9379 expands them accordingly. Check the documentation of strftime(). This
9380 feature is deprecated.
9382 If @option{expansion} is set to @code{none}, the text is printed verbatim.
9384 If @option{expansion} is set to @code{normal} (which is the default),
9385 the following expansion mechanism is used.
9387 The backslash character @samp{\}, followed by any character, always expands to
9388 the second character.
9390 Sequences of the form @code{%@{...@}} are expanded. The text between the
9391 braces is a function name, possibly followed by arguments separated by ':'.
9392 If the arguments contain special characters or delimiters (':' or '@}'),
9393 they should be escaped.
9395 Note that they probably must also be escaped as the value for the
9396 @option{text} option in the filter argument string and as the filter
9397 argument in the filtergraph description, and possibly also for the shell,
9398 that makes up to four levels of escaping; using a text file avoids these
9401 The following functions are available:
9406 The expression evaluation result.
9408 It must take one argument specifying the expression to be evaluated,
9409 which accepts the same constants and functions as the @var{x} and
9410 @var{y} values. Note that not all constants should be used, for
9411 example the text size is not known when evaluating the expression, so
9412 the constants @var{text_w} and @var{text_h} will have an undefined
9415 @item expr_int_format, eif
9416 Evaluate the expression's value and output as formatted integer.
9418 The first argument is the expression to be evaluated, just as for the @var{expr} function.
9419 The second argument specifies the output format. Allowed values are @samp{x},
9420 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
9421 @code{printf} function.
9422 The third parameter is optional and sets the number of positions taken by the output.
9423 It can be used to add padding with zeros from the left.
9426 The time at which the filter is running, expressed in UTC.
9427 It can accept an argument: a strftime() format string.
9430 The time at which the filter is running, expressed in the local time zone.
9431 It can accept an argument: a strftime() format string.
9434 Frame metadata. Takes one or two arguments.
9436 The first argument is mandatory and specifies the metadata key.
9438 The second argument is optional and specifies a default value, used when the
9439 metadata key is not found or empty.
9441 Available metadata can be identified by inspecting entries
9442 starting with TAG included within each frame section
9443 printed by running @code{ffprobe -show_frames}.
9445 String metadata generated in filters leading to
9446 the drawtext filter are also available.
9449 The frame number, starting from 0.
9452 A one character description of the current picture type.
9455 The timestamp of the current frame.
9456 It can take up to three arguments.
9458 The first argument is the format of the timestamp; it defaults to @code{flt}
9459 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
9460 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
9461 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
9462 @code{localtime} stands for the timestamp of the frame formatted as
9463 local time zone time.
9465 The second argument is an offset added to the timestamp.
9467 If the format is set to @code{hms}, a third argument @code{24HH} may be
9468 supplied to present the hour part of the formatted timestamp in 24h format
9471 If the format is set to @code{localtime} or @code{gmtime},
9472 a third argument may be supplied: a strftime() format string.
9473 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
9476 @subsection Commands
9478 This filter supports altering parameters via commands:
9481 Alter existing filter parameters.
9483 Syntax for the argument is the same as for filter invocation, e.g.
9486 fontsize=56:fontcolor=green:text='Hello World'
9489 Full filter invocation with sendcmd would look like this:
9492 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
9496 If the entire argument can't be parsed or applied as valid values then the filter will
9497 continue with its existing parameters.
9499 @subsection Examples
9503 Draw "Test Text" with font FreeSerif, using the default values for the
9504 optional parameters.
9507 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
9511 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
9512 and y=50 (counting from the top-left corner of the screen), text is
9513 yellow with a red box around it. Both the text and the box have an
9517 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
9518 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
9521 Note that the double quotes are not necessary if spaces are not used
9522 within the parameter list.
9525 Show the text at the center of the video frame:
9527 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
9531 Show the text at a random position, switching to a new position every 30 seconds:
9533 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)"
9537 Show a text line sliding from right to left in the last row of the video
9538 frame. The file @file{LONG_LINE} is assumed to contain a single line
9541 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
9545 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
9547 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
9551 Draw a single green letter "g", at the center of the input video.
9552 The glyph baseline is placed at half screen height.
9554 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
9558 Show text for 1 second every 3 seconds:
9560 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
9564 Use fontconfig to set the font. Note that the colons need to be escaped.
9566 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
9570 Print the date of a real-time encoding (see strftime(3)):
9572 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
9576 Show text fading in and out (appearing/disappearing):
9579 DS=1.0 # display start
9580 DE=10.0 # display end
9581 FID=1.5 # fade in duration
9582 FOD=5 # fade out duration
9583 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 @}"
9587 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
9588 and the @option{fontsize} value are included in the @option{y} offset.
9590 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
9591 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
9596 For more information about libfreetype, check:
9597 @url{http://www.freetype.org/}.
9599 For more information about fontconfig, check:
9600 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
9602 For more information about libfribidi, check:
9603 @url{http://fribidi.org/}.
9607 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
9609 The filter accepts the following options:
9614 Set low and high threshold values used by the Canny thresholding
9617 The high threshold selects the "strong" edge pixels, which are then
9618 connected through 8-connectivity with the "weak" edge pixels selected
9619 by the low threshold.
9621 @var{low} and @var{high} threshold values must be chosen in the range
9622 [0,1], and @var{low} should be lesser or equal to @var{high}.
9624 Default value for @var{low} is @code{20/255}, and default value for @var{high}
9628 Define the drawing mode.
9632 Draw white/gray wires on black background.
9635 Mix the colors to create a paint/cartoon effect.
9638 Apply Canny edge detector on all selected planes.
9640 Default value is @var{wires}.
9643 Select planes for filtering. By default all available planes are filtered.
9646 @subsection Examples
9650 Standard edge detection with custom values for the hysteresis thresholding:
9652 edgedetect=low=0.1:high=0.4
9656 Painting effect without thresholding:
9658 edgedetect=mode=colormix:high=0
9664 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
9666 For each input image, the filter will compute the optimal mapping from
9667 the input to the output given the codebook length, that is the number
9668 of distinct output colors.
9670 This filter accepts the following options.
9673 @item codebook_length, l
9674 Set codebook length. The value must be a positive integer, and
9675 represents the number of distinct output colors. Default value is 256.
9678 Set the maximum number of iterations to apply for computing the optimal
9679 mapping. The higher the value the better the result and the higher the
9680 computation time. Default value is 1.
9683 Set a random seed, must be an integer included between 0 and
9684 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
9685 will try to use a good random seed on a best effort basis.
9688 Set pal8 output pixel format. This option does not work with codebook
9689 length greater than 256.
9694 Measure graylevel entropy in histogram of color channels of video frames.
9696 It accepts the following parameters:
9700 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
9702 @var{diff} mode measures entropy of histogram delta values, absolute differences
9703 between neighbour histogram values.
9707 Set brightness, contrast, saturation and approximate gamma adjustment.
9709 The filter accepts the following options:
9713 Set the contrast expression. The value must be a float value in range
9714 @code{-1000.0} to @code{1000.0}. The default value is "1".
9717 Set the brightness expression. The value must be a float value in
9718 range @code{-1.0} to @code{1.0}. The default value is "0".
9721 Set the saturation expression. The value must be a float in
9722 range @code{0.0} to @code{3.0}. The default value is "1".
9725 Set the gamma expression. The value must be a float in range
9726 @code{0.1} to @code{10.0}. The default value is "1".
9729 Set the gamma expression for red. The value must be a float in
9730 range @code{0.1} to @code{10.0}. The default value is "1".
9733 Set the gamma expression for green. The value must be a float in range
9734 @code{0.1} to @code{10.0}. The default value is "1".
9737 Set the gamma expression for blue. The value must be a float in range
9738 @code{0.1} to @code{10.0}. The default value is "1".
9741 Set the gamma weight expression. It can be used to reduce the effect
9742 of a high gamma value on bright image areas, e.g. keep them from
9743 getting overamplified and just plain white. The value must be a float
9744 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
9745 gamma correction all the way down while @code{1.0} leaves it at its
9746 full strength. Default is "1".
9749 Set when the expressions for brightness, contrast, saturation and
9750 gamma expressions are evaluated.
9752 It accepts the following values:
9755 only evaluate expressions once during the filter initialization or
9756 when a command is processed
9759 evaluate expressions for each incoming frame
9762 Default value is @samp{init}.
9765 The expressions accept the following parameters:
9768 frame count of the input frame starting from 0
9771 byte position of the corresponding packet in the input file, NAN if
9775 frame rate of the input video, NAN if the input frame rate is unknown
9778 timestamp expressed in seconds, NAN if the input timestamp is unknown
9781 @subsection Commands
9782 The filter supports the following commands:
9786 Set the contrast expression.
9789 Set the brightness expression.
9792 Set the saturation expression.
9795 Set the gamma expression.
9798 Set the gamma_r expression.
9801 Set gamma_g expression.
9804 Set gamma_b expression.
9807 Set gamma_weight expression.
9809 The command accepts the same syntax of the corresponding option.
9811 If the specified expression is not valid, it is kept at its current
9818 Apply erosion effect to the video.
9820 This filter replaces the pixel by the local(3x3) minimum.
9822 It accepts the following options:
9829 Limit the maximum change for each plane, default is 65535.
9830 If 0, plane will remain unchanged.
9833 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9836 Flags to local 3x3 coordinates maps like this:
9843 @section extractplanes
9845 Extract color channel components from input video stream into
9846 separate grayscale video streams.
9848 The filter accepts the following option:
9852 Set plane(s) to extract.
9854 Available values for planes are:
9865 Choosing planes not available in the input will result in an error.
9866 That means you cannot select @code{r}, @code{g}, @code{b} planes
9867 with @code{y}, @code{u}, @code{v} planes at same time.
9870 @subsection Examples
9874 Extract luma, u and v color channel component from input video frame
9875 into 3 grayscale outputs:
9877 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
9883 Apply a fade-in/out effect to the input video.
9885 It accepts the following parameters:
9889 The effect type can be either "in" for a fade-in, or "out" for a fade-out
9891 Default is @code{in}.
9893 @item start_frame, s
9894 Specify the number of the frame to start applying the fade
9895 effect at. Default is 0.
9898 The number of frames that the fade effect lasts. At the end of the
9899 fade-in effect, the output video will have the same intensity as the input video.
9900 At the end of the fade-out transition, the output video will be filled with the
9901 selected @option{color}.
9905 If set to 1, fade only alpha channel, if one exists on the input.
9908 @item start_time, st
9909 Specify the timestamp (in seconds) of the frame to start to apply the fade
9910 effect. If both start_frame and start_time are specified, the fade will start at
9911 whichever comes last. Default is 0.
9914 The number of seconds for which the fade effect has to last. At the end of the
9915 fade-in effect the output video will have the same intensity as the input video,
9916 at the end of the fade-out transition the output video will be filled with the
9917 selected @option{color}.
9918 If both duration and nb_frames are specified, duration is used. Default is 0
9919 (nb_frames is used by default).
9922 Specify the color of the fade. Default is "black".
9925 @subsection Examples
9929 Fade in the first 30 frames of video:
9934 The command above is equivalent to:
9940 Fade out the last 45 frames of a 200-frame video:
9943 fade=type=out:start_frame=155:nb_frames=45
9947 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
9949 fade=in:0:25, fade=out:975:25
9953 Make the first 5 frames yellow, then fade in from frame 5-24:
9955 fade=in:5:20:color=yellow
9959 Fade in alpha over first 25 frames of video:
9961 fade=in:0:25:alpha=1
9965 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
9967 fade=t=in:st=5.5:d=0.5
9973 Denoise frames using 3D FFT (frequency domain filtering).
9975 The filter accepts the following options:
9979 Set the noise sigma constant. This sets denoising strength.
9980 Default value is 1. Allowed range is from 0 to 30.
9981 Using very high sigma with low overlap may give blocking artifacts.
9984 Set amount of denoising. By default all detected noise is reduced.
9985 Default value is 1. Allowed range is from 0 to 1.
9988 Set size of block, Default is 4, can be 3, 4, 5 or 6.
9989 Actual size of block in pixels is 2 to power of @var{block}, so by default
9990 block size in pixels is 2^4 which is 16.
9993 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
9996 Set number of previous frames to use for denoising. By default is set to 0.
9999 Set number of next frames to to use for denoising. By default is set to 0.
10002 Set planes which will be filtered, by default are all available filtered
10007 Apply arbitrary expressions to samples in frequency domain
10011 Adjust the dc value (gain) of the luma plane of the image. The filter
10012 accepts an integer value in range @code{0} to @code{1000}. The default
10013 value is set to @code{0}.
10016 Adjust the dc value (gain) of the 1st chroma plane of the image. The
10017 filter accepts an integer value in range @code{0} to @code{1000}. The
10018 default value is set to @code{0}.
10021 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
10022 filter accepts an integer value in range @code{0} to @code{1000}. The
10023 default value is set to @code{0}.
10026 Set the frequency domain weight expression for the luma plane.
10029 Set the frequency domain weight expression for the 1st chroma plane.
10032 Set the frequency domain weight expression for the 2nd chroma plane.
10035 Set when the expressions are evaluated.
10037 It accepts the following values:
10040 Only evaluate expressions once during the filter initialization.
10043 Evaluate expressions for each incoming frame.
10046 Default value is @samp{init}.
10048 The filter accepts the following variables:
10051 The coordinates of the current sample.
10055 The width and height of the image.
10058 The number of input frame, starting from 0.
10061 @subsection Examples
10067 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
10073 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
10079 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
10085 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
10092 Extract a single field from an interlaced image using stride
10093 arithmetic to avoid wasting CPU time. The output frames are marked as
10096 The filter accepts the following options:
10100 Specify whether to extract the top (if the value is @code{0} or
10101 @code{top}) or the bottom field (if the value is @code{1} or
10107 Create new frames by copying the top and bottom fields from surrounding frames
10108 supplied as numbers by the hint file.
10112 Set file containing hints: absolute/relative frame numbers.
10114 There must be one line for each frame in a clip. Each line must contain two
10115 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
10116 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
10117 is current frame number for @code{absolute} mode or out of [-1, 1] range
10118 for @code{relative} mode. First number tells from which frame to pick up top
10119 field and second number tells from which frame to pick up bottom field.
10121 If optionally followed by @code{+} output frame will be marked as interlaced,
10122 else if followed by @code{-} output frame will be marked as progressive, else
10123 it will be marked same as input frame.
10124 If optionally followed by @code{t} output frame will use only top field, or in
10125 case of @code{b} it will use only bottom field.
10126 If line starts with @code{#} or @code{;} that line is skipped.
10129 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
10132 Example of first several lines of @code{hint} file for @code{relative} mode:
10134 0,0 - # first frame
10135 1,0 - # second frame, use third's frame top field and second's frame bottom field
10136 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
10151 @section fieldmatch
10153 Field matching filter for inverse telecine. It is meant to reconstruct the
10154 progressive frames from a telecined stream. The filter does not drop duplicated
10155 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
10156 followed by a decimation filter such as @ref{decimate} in the filtergraph.
10158 The separation of the field matching and the decimation is notably motivated by
10159 the possibility of inserting a de-interlacing filter fallback between the two.
10160 If the source has mixed telecined and real interlaced content,
10161 @code{fieldmatch} will not be able to match fields for the interlaced parts.
10162 But these remaining combed frames will be marked as interlaced, and thus can be
10163 de-interlaced by a later filter such as @ref{yadif} before decimation.
10165 In addition to the various configuration options, @code{fieldmatch} can take an
10166 optional second stream, activated through the @option{ppsrc} option. If
10167 enabled, the frames reconstruction will be based on the fields and frames from
10168 this second stream. This allows the first input to be pre-processed in order to
10169 help the various algorithms of the filter, while keeping the output lossless
10170 (assuming the fields are matched properly). Typically, a field-aware denoiser,
10171 or brightness/contrast adjustments can help.
10173 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
10174 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
10175 which @code{fieldmatch} is based on. While the semantic and usage are very
10176 close, some behaviour and options names can differ.
10178 The @ref{decimate} filter currently only works for constant frame rate input.
10179 If your input has mixed telecined (30fps) and progressive content with a lower
10180 framerate like 24fps use the following filterchain to produce the necessary cfr
10181 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
10183 The filter accepts the following options:
10187 Specify the assumed field order of the input stream. Available values are:
10191 Auto detect parity (use FFmpeg's internal parity value).
10193 Assume bottom field first.
10195 Assume top field first.
10198 Note that it is sometimes recommended not to trust the parity announced by the
10201 Default value is @var{auto}.
10204 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
10205 sense that it won't risk creating jerkiness due to duplicate frames when
10206 possible, but if there are bad edits or blended fields it will end up
10207 outputting combed frames when a good match might actually exist. On the other
10208 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
10209 but will almost always find a good frame if there is one. The other values are
10210 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
10211 jerkiness and creating duplicate frames versus finding good matches in sections
10212 with bad edits, orphaned fields, blended fields, etc.
10214 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
10216 Available values are:
10220 2-way matching (p/c)
10222 2-way matching, and trying 3rd match if still combed (p/c + n)
10224 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
10226 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
10227 still combed (p/c + n + u/b)
10229 3-way matching (p/c/n)
10231 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
10232 detected as combed (p/c/n + u/b)
10235 The parenthesis at the end indicate the matches that would be used for that
10236 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
10239 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
10242 Default value is @var{pc_n}.
10245 Mark the main input stream as a pre-processed input, and enable the secondary
10246 input stream as the clean source to pick the fields from. See the filter
10247 introduction for more details. It is similar to the @option{clip2} feature from
10250 Default value is @code{0} (disabled).
10253 Set the field to match from. It is recommended to set this to the same value as
10254 @option{order} unless you experience matching failures with that setting. In
10255 certain circumstances changing the field that is used to match from can have a
10256 large impact on matching performance. Available values are:
10260 Automatic (same value as @option{order}).
10262 Match from the bottom field.
10264 Match from the top field.
10267 Default value is @var{auto}.
10270 Set whether or not chroma is included during the match comparisons. In most
10271 cases it is recommended to leave this enabled. You should set this to @code{0}
10272 only if your clip has bad chroma problems such as heavy rainbowing or other
10273 artifacts. Setting this to @code{0} could also be used to speed things up at
10274 the cost of some accuracy.
10276 Default value is @code{1}.
10280 These define an exclusion band which excludes the lines between @option{y0} and
10281 @option{y1} from being included in the field matching decision. An exclusion
10282 band can be used to ignore subtitles, a logo, or other things that may
10283 interfere with the matching. @option{y0} sets the starting scan line and
10284 @option{y1} sets the ending line; all lines in between @option{y0} and
10285 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
10286 @option{y0} and @option{y1} to the same value will disable the feature.
10287 @option{y0} and @option{y1} defaults to @code{0}.
10290 Set the scene change detection threshold as a percentage of maximum change on
10291 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
10292 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
10293 @option{scthresh} is @code{[0.0, 100.0]}.
10295 Default value is @code{12.0}.
10298 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
10299 account the combed scores of matches when deciding what match to use as the
10300 final match. Available values are:
10304 No final matching based on combed scores.
10306 Combed scores are only used when a scene change is detected.
10308 Use combed scores all the time.
10311 Default is @var{sc}.
10314 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
10315 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
10316 Available values are:
10320 No forced calculation.
10322 Force p/c/n calculations.
10324 Force p/c/n/u/b calculations.
10327 Default value is @var{none}.
10330 This is the area combing threshold used for combed frame detection. This
10331 essentially controls how "strong" or "visible" combing must be to be detected.
10332 Larger values mean combing must be more visible and smaller values mean combing
10333 can be less visible or strong and still be detected. Valid settings are from
10334 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
10335 be detected as combed). This is basically a pixel difference value. A good
10336 range is @code{[8, 12]}.
10338 Default value is @code{9}.
10341 Sets whether or not chroma is considered in the combed frame decision. Only
10342 disable this if your source has chroma problems (rainbowing, etc.) that are
10343 causing problems for the combed frame detection with chroma enabled. Actually,
10344 using @option{chroma}=@var{0} is usually more reliable, except for the case
10345 where there is chroma only combing in the source.
10347 Default value is @code{0}.
10351 Respectively set the x-axis and y-axis size of the window used during combed
10352 frame detection. This has to do with the size of the area in which
10353 @option{combpel} pixels are required to be detected as combed for a frame to be
10354 declared combed. See the @option{combpel} parameter description for more info.
10355 Possible values are any number that is a power of 2 starting at 4 and going up
10358 Default value is @code{16}.
10361 The number of combed pixels inside any of the @option{blocky} by
10362 @option{blockx} size blocks on the frame for the frame to be detected as
10363 combed. While @option{cthresh} controls how "visible" the combing must be, this
10364 setting controls "how much" combing there must be in any localized area (a
10365 window defined by the @option{blockx} and @option{blocky} settings) on the
10366 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
10367 which point no frames will ever be detected as combed). This setting is known
10368 as @option{MI} in TFM/VFM vocabulary.
10370 Default value is @code{80}.
10373 @anchor{p/c/n/u/b meaning}
10374 @subsection p/c/n/u/b meaning
10376 @subsubsection p/c/n
10378 We assume the following telecined stream:
10381 Top fields: 1 2 2 3 4
10382 Bottom fields: 1 2 3 4 4
10385 The numbers correspond to the progressive frame the fields relate to. Here, the
10386 first two frames are progressive, the 3rd and 4th are combed, and so on.
10388 When @code{fieldmatch} is configured to run a matching from bottom
10389 (@option{field}=@var{bottom}) this is how this input stream get transformed:
10394 B 1 2 3 4 4 <-- matching reference
10403 As a result of the field matching, we can see that some frames get duplicated.
10404 To perform a complete inverse telecine, you need to rely on a decimation filter
10405 after this operation. See for instance the @ref{decimate} filter.
10407 The same operation now matching from top fields (@option{field}=@var{top})
10412 T 1 2 2 3 4 <-- matching reference
10422 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
10423 basically, they refer to the frame and field of the opposite parity:
10426 @item @var{p} matches the field of the opposite parity in the previous frame
10427 @item @var{c} matches the field of the opposite parity in the current frame
10428 @item @var{n} matches the field of the opposite parity in the next frame
10433 The @var{u} and @var{b} matching are a bit special in the sense that they match
10434 from the opposite parity flag. In the following examples, we assume that we are
10435 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
10436 'x' is placed above and below each matched fields.
10438 With bottom matching (@option{field}=@var{bottom}):
10443 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
10444 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
10452 With top matching (@option{field}=@var{top}):
10457 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
10458 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
10466 @subsection Examples
10468 Simple IVTC of a top field first telecined stream:
10470 fieldmatch=order=tff:combmatch=none, decimate
10473 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
10475 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
10478 @section fieldorder
10480 Transform the field order of the input video.
10482 It accepts the following parameters:
10487 The output field order. Valid values are @var{tff} for top field first or @var{bff}
10488 for bottom field first.
10491 The default value is @samp{tff}.
10493 The transformation is done by shifting the picture content up or down
10494 by one line, and filling the remaining line with appropriate picture content.
10495 This method is consistent with most broadcast field order converters.
10497 If the input video is not flagged as being interlaced, or it is already
10498 flagged as being of the required output field order, then this filter does
10499 not alter the incoming video.
10501 It is very useful when converting to or from PAL DV material,
10502 which is bottom field first.
10506 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
10509 @section fifo, afifo
10511 Buffer input images and send them when they are requested.
10513 It is mainly useful when auto-inserted by the libavfilter
10516 It does not take parameters.
10518 @section fillborders
10520 Fill borders of the input video, without changing video stream dimensions.
10521 Sometimes video can have garbage at the four edges and you may not want to
10522 crop video input to keep size multiple of some number.
10524 This filter accepts the following options:
10528 Number of pixels to fill from left border.
10531 Number of pixels to fill from right border.
10534 Number of pixels to fill from top border.
10537 Number of pixels to fill from bottom border.
10542 It accepts the following values:
10545 fill pixels using outermost pixels
10548 fill pixels using mirroring
10551 fill pixels with constant value
10554 Default is @var{smear}.
10557 Set color for pixels in fixed mode. Default is @var{black}.
10562 Find a rectangular object
10564 It accepts the following options:
10568 Filepath of the object image, needs to be in gray8.
10571 Detection threshold, default is 0.5.
10574 Number of mipmaps, default is 3.
10576 @item xmin, ymin, xmax, ymax
10577 Specifies the rectangle in which to search.
10580 @subsection Examples
10584 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
10586 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
10592 Flood area with values of same pixel components with another values.
10594 It accepts the following options:
10597 Set pixel x coordinate.
10600 Set pixel y coordinate.
10603 Set source #0 component value.
10606 Set source #1 component value.
10609 Set source #2 component value.
10612 Set source #3 component value.
10615 Set destination #0 component value.
10618 Set destination #1 component value.
10621 Set destination #2 component value.
10624 Set destination #3 component value.
10630 Convert the input video to one of the specified pixel formats.
10631 Libavfilter will try to pick one that is suitable as input to
10634 It accepts the following parameters:
10638 A '|'-separated list of pixel format names, such as
10639 "pix_fmts=yuv420p|monow|rgb24".
10643 @subsection Examples
10647 Convert the input video to the @var{yuv420p} format
10649 format=pix_fmts=yuv420p
10652 Convert the input video to any of the formats in the list
10654 format=pix_fmts=yuv420p|yuv444p|yuv410p
10661 Convert the video to specified constant frame rate by duplicating or dropping
10662 frames as necessary.
10664 It accepts the following parameters:
10668 The desired output frame rate. The default is @code{25}.
10671 Assume the first PTS should be the given value, in seconds. This allows for
10672 padding/trimming at the start of stream. By default, no assumption is made
10673 about the first frame's expected PTS, so no padding or trimming is done.
10674 For example, this could be set to 0 to pad the beginning with duplicates of
10675 the first frame if a video stream starts after the audio stream or to trim any
10676 frames with a negative PTS.
10679 Timestamp (PTS) rounding method.
10681 Possible values are:
10688 round towards -infinity
10690 round towards +infinity
10694 The default is @code{near}.
10697 Action performed when reading the last frame.
10699 Possible values are:
10702 Use same timestamp rounding method as used for other frames.
10704 Pass through last frame if input duration has not been reached yet.
10706 The default is @code{round}.
10710 Alternatively, the options can be specified as a flat string:
10711 @var{fps}[:@var{start_time}[:@var{round}]].
10713 See also the @ref{setpts} filter.
10715 @subsection Examples
10719 A typical usage in order to set the fps to 25:
10725 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
10727 fps=fps=film:round=near
10733 Pack two different video streams into a stereoscopic video, setting proper
10734 metadata on supported codecs. The two views should have the same size and
10735 framerate and processing will stop when the shorter video ends. Please note
10736 that you may conveniently adjust view properties with the @ref{scale} and
10739 It accepts the following parameters:
10743 The desired packing format. Supported values are:
10748 The views are next to each other (default).
10751 The views are on top of each other.
10754 The views are packed by line.
10757 The views are packed by column.
10760 The views are temporally interleaved.
10769 # Convert left and right views into a frame-sequential video
10770 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
10772 # Convert views into a side-by-side video with the same output resolution as the input
10773 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
10778 Change the frame rate by interpolating new video output frames from the source
10781 This filter is not designed to function correctly with interlaced media. If
10782 you wish to change the frame rate of interlaced media then you are required
10783 to deinterlace before this filter and re-interlace after this filter.
10785 A description of the accepted options follows.
10789 Specify the output frames per second. This option can also be specified
10790 as a value alone. The default is @code{50}.
10793 Specify the start of a range where the output frame will be created as a
10794 linear interpolation of two frames. The range is [@code{0}-@code{255}],
10795 the default is @code{15}.
10798 Specify the end of a range where the output frame will be created as a
10799 linear interpolation of two frames. The range is [@code{0}-@code{255}],
10800 the default is @code{240}.
10803 Specify the level at which a scene change is detected as a value between
10804 0 and 100 to indicate a new scene; a low value reflects a low
10805 probability for the current frame to introduce a new scene, while a higher
10806 value means the current frame is more likely to be one.
10807 The default is @code{8.2}.
10810 Specify flags influencing the filter process.
10812 Available value for @var{flags} is:
10815 @item scene_change_detect, scd
10816 Enable scene change detection using the value of the option @var{scene}.
10817 This flag is enabled by default.
10823 Select one frame every N-th frame.
10825 This filter accepts the following option:
10828 Select frame after every @code{step} frames.
10829 Allowed values are positive integers higher than 0. Default value is @code{1}.
10832 @section freezedetect
10834 Detect frozen video.
10836 This filter logs a message and sets frame metadata when it detects that the
10837 input video has no significant change in content during a specified duration.
10838 Video freeze detection calculates the mean average absolute difference of all
10839 the components of video frames and compares it to a noise floor.
10841 The printed times and duration are expressed in seconds. The
10842 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
10843 whose timestamp equals or exceeds the detection duration and it contains the
10844 timestamp of the first frame of the freeze. The
10845 @code{lavfi.freezedetect.freeze_duration} and
10846 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
10849 The filter accepts the following options:
10853 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
10854 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
10858 Set freeze duration until notification (default is 2 seconds).
10864 Apply a frei0r effect to the input video.
10866 To enable the compilation of this filter, you need to install the frei0r
10867 header and configure FFmpeg with @code{--enable-frei0r}.
10869 It accepts the following parameters:
10874 The name of the frei0r effect to load. If the environment variable
10875 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
10876 directories specified by the colon-separated list in @env{FREI0R_PATH}.
10877 Otherwise, the standard frei0r paths are searched, in this order:
10878 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
10879 @file{/usr/lib/frei0r-1/}.
10881 @item filter_params
10882 A '|'-separated list of parameters to pass to the frei0r effect.
10886 A frei0r effect parameter can be a boolean (its value is either
10887 "y" or "n"), a double, a color (specified as
10888 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
10889 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
10890 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
10891 a position (specified as @var{X}/@var{Y}, where
10892 @var{X} and @var{Y} are floating point numbers) and/or a string.
10894 The number and types of parameters depend on the loaded effect. If an
10895 effect parameter is not specified, the default value is set.
10897 @subsection Examples
10901 Apply the distort0r effect, setting the first two double parameters:
10903 frei0r=filter_name=distort0r:filter_params=0.5|0.01
10907 Apply the colordistance effect, taking a color as the first parameter:
10909 frei0r=colordistance:0.2/0.3/0.4
10910 frei0r=colordistance:violet
10911 frei0r=colordistance:0x112233
10915 Apply the perspective effect, specifying the top left and top right image
10918 frei0r=perspective:0.2/0.2|0.8/0.2
10922 For more information, see
10923 @url{http://frei0r.dyne.org}
10927 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
10929 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
10930 processing filter, one of them is performed once per block, not per pixel.
10931 This allows for much higher speed.
10933 The filter accepts the following options:
10937 Set quality. This option defines the number of levels for averaging. It accepts
10938 an integer in the range 4-5. Default value is @code{4}.
10941 Force a constant quantization parameter. It accepts an integer in range 0-63.
10942 If not set, the filter will use the QP from the video stream (if available).
10945 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
10946 more details but also more artifacts, while higher values make the image smoother
10947 but also blurrier. Default value is @code{0} − PSNR optimal.
10949 @item use_bframe_qp
10950 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
10951 option may cause flicker since the B-Frames have often larger QP. Default is
10952 @code{0} (not enabled).
10958 Apply Gaussian blur filter.
10960 The filter accepts the following options:
10964 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
10967 Set number of steps for Gaussian approximation. Default is @code{1}.
10970 Set which planes to filter. By default all planes are filtered.
10973 Set vertical sigma, if negative it will be same as @code{sigma}.
10974 Default is @code{-1}.
10977 @subsection Commands
10978 This filter supports same commands as options.
10979 The command accepts the same syntax of the corresponding option.
10981 If the specified expression is not valid, it is kept at its current
10986 Apply generic equation to each pixel.
10988 The filter accepts the following options:
10991 @item lum_expr, lum
10992 Set the luminance expression.
10994 Set the chrominance blue expression.
10996 Set the chrominance red expression.
10997 @item alpha_expr, a
10998 Set the alpha expression.
11000 Set the red expression.
11001 @item green_expr, g
11002 Set the green expression.
11004 Set the blue expression.
11007 The colorspace is selected according to the specified options. If one
11008 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
11009 options is specified, the filter will automatically select a YCbCr
11010 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
11011 @option{blue_expr} options is specified, it will select an RGB
11014 If one of the chrominance expression is not defined, it falls back on the other
11015 one. If no alpha expression is specified it will evaluate to opaque value.
11016 If none of chrominance expressions are specified, they will evaluate
11017 to the luminance expression.
11019 The expressions can use the following variables and functions:
11023 The sequential number of the filtered frame, starting from @code{0}.
11027 The coordinates of the current sample.
11031 The width and height of the image.
11035 Width and height scale depending on the currently filtered plane. It is the
11036 ratio between the corresponding luma plane number of pixels and the current
11037 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
11038 @code{0.5,0.5} for chroma planes.
11041 Time of the current frame, expressed in seconds.
11044 Return the value of the pixel at location (@var{x},@var{y}) of the current
11048 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
11052 Return the value of the pixel at location (@var{x},@var{y}) of the
11053 blue-difference chroma plane. Return 0 if there is no such plane.
11056 Return the value of the pixel at location (@var{x},@var{y}) of the
11057 red-difference chroma plane. Return 0 if there is no such plane.
11062 Return the value of the pixel at location (@var{x},@var{y}) of the
11063 red/green/blue component. Return 0 if there is no such component.
11066 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
11067 plane. Return 0 if there is no such plane.
11069 @item interpolation
11070 Set one of interpolation methods:
11075 Default is bilinear.
11078 For functions, if @var{x} and @var{y} are outside the area, the value will be
11079 automatically clipped to the closer edge.
11081 @subsection Examples
11085 Flip the image horizontally:
11091 Generate a bidimensional sine wave, with angle @code{PI/3} and a
11092 wavelength of 100 pixels:
11094 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
11098 Generate a fancy enigmatic moving light:
11100 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
11104 Generate a quick emboss effect:
11106 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
11110 Modify RGB components depending on pixel position:
11112 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
11116 Create a radial gradient that is the same size as the input (also see
11117 the @ref{vignette} filter):
11119 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
11125 Fix the banding artifacts that are sometimes introduced into nearly flat
11126 regions by truncation to 8-bit color depth.
11127 Interpolate the gradients that should go where the bands are, and
11130 It is designed for playback only. Do not use it prior to
11131 lossy compression, because compression tends to lose the dither and
11132 bring back the bands.
11134 It accepts the following parameters:
11139 The maximum amount by which the filter will change any one pixel. This is also
11140 the threshold for detecting nearly flat regions. Acceptable values range from
11141 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
11145 The neighborhood to fit the gradient to. A larger radius makes for smoother
11146 gradients, but also prevents the filter from modifying the pixels near detailed
11147 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
11148 values will be clipped to the valid range.
11152 Alternatively, the options can be specified as a flat string:
11153 @var{strength}[:@var{radius}]
11155 @subsection Examples
11159 Apply the filter with a @code{3.5} strength and radius of @code{8}:
11165 Specify radius, omitting the strength (which will fall-back to the default
11173 @section graphmonitor, agraphmonitor
11174 Show various filtergraph stats.
11176 With this filter one can debug complete filtergraph.
11177 Especially issues with links filling with queued frames.
11179 The filter accepts the following options:
11183 Set video output size. Default is @var{hd720}.
11186 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
11189 Set output mode, can be @var{fulll} or @var{compact}.
11190 In @var{compact} mode only filters with some queued frames have displayed stats.
11193 Set flags which enable which stats are shown in video.
11195 Available values for flags are:
11198 Display number of queued frames in each link.
11200 @item frame_count_in
11201 Display number of frames taken from filter.
11203 @item frame_count_out
11204 Display number of frames given out from filter.
11207 Display current filtered frame pts.
11210 Display current filtered frame time.
11213 Display time base for filter link.
11216 Display used format for filter link.
11219 Display video size or number of audio channels in case of audio used by filter link.
11222 Display video frame rate or sample rate in case of audio used by filter link.
11226 Set upper limit for video rate of output stream, Default value is @var{25}.
11227 This guarantee that output video frame rate will not be higher than this value.
11231 A color constancy variation filter which estimates scene illumination via grey edge algorithm
11232 and corrects the scene colors accordingly.
11234 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
11236 The filter accepts the following options:
11240 The order of differentiation to be applied on the scene. Must be chosen in the range
11241 [0,2] and default value is 1.
11244 The Minkowski parameter to be used for calculating the Minkowski distance. Must
11245 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
11246 max value instead of calculating Minkowski distance.
11249 The standard deviation of Gaussian blur to be applied on the scene. Must be
11250 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
11251 can't be equal to 0 if @var{difford} is greater than 0.
11254 @subsection Examples
11260 greyedge=difford=1:minknorm=5:sigma=2
11266 greyedge=difford=1:minknorm=0:sigma=2
11274 Apply a Hald CLUT to a video stream.
11276 First input is the video stream to process, and second one is the Hald CLUT.
11277 The Hald CLUT input can be a simple picture or a complete video stream.
11279 The filter accepts the following options:
11283 Force termination when the shortest input terminates. Default is @code{0}.
11285 Continue applying the last CLUT after the end of the stream. A value of
11286 @code{0} disable the filter after the last frame of the CLUT is reached.
11287 Default is @code{1}.
11290 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
11291 filters share the same internals).
11293 This filter also supports the @ref{framesync} options.
11295 More information about the Hald CLUT can be found on Eskil Steenberg's website
11296 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
11298 @subsection Workflow examples
11300 @subsubsection Hald CLUT video stream
11302 Generate an identity Hald CLUT stream altered with various effects:
11304 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
11307 Note: make sure you use a lossless codec.
11309 Then use it with @code{haldclut} to apply it on some random stream:
11311 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
11314 The Hald CLUT will be applied to the 10 first seconds (duration of
11315 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
11316 to the remaining frames of the @code{mandelbrot} stream.
11318 @subsubsection Hald CLUT with preview
11320 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
11321 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
11322 biggest possible square starting at the top left of the picture. The remaining
11323 padding pixels (bottom or right) will be ignored. This area can be used to add
11324 a preview of the Hald CLUT.
11326 Typically, the following generated Hald CLUT will be supported by the
11327 @code{haldclut} filter:
11330 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
11331 pad=iw+320 [padded_clut];
11332 smptebars=s=320x256, split [a][b];
11333 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
11334 [main][b] overlay=W-320" -frames:v 1 clut.png
11337 It contains the original and a preview of the effect of the CLUT: SMPTE color
11338 bars are displayed on the right-top, and below the same color bars processed by
11341 Then, the effect of this Hald CLUT can be visualized with:
11343 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
11348 Flip the input video horizontally.
11350 For example, to horizontally flip the input video with @command{ffmpeg}:
11352 ffmpeg -i in.avi -vf "hflip" out.avi
11356 This filter applies a global color histogram equalization on a
11359 It can be used to correct video that has a compressed range of pixel
11360 intensities. The filter redistributes the pixel intensities to
11361 equalize their distribution across the intensity range. It may be
11362 viewed as an "automatically adjusting contrast filter". This filter is
11363 useful only for correcting degraded or poorly captured source
11366 The filter accepts the following options:
11370 Determine the amount of equalization to be applied. As the strength
11371 is reduced, the distribution of pixel intensities more-and-more
11372 approaches that of the input frame. The value must be a float number
11373 in the range [0,1] and defaults to 0.200.
11376 Set the maximum intensity that can generated and scale the output
11377 values appropriately. The strength should be set as desired and then
11378 the intensity can be limited if needed to avoid washing-out. The value
11379 must be a float number in the range [0,1] and defaults to 0.210.
11382 Set the antibanding level. If enabled the filter will randomly vary
11383 the luminance of output pixels by a small amount to avoid banding of
11384 the histogram. Possible values are @code{none}, @code{weak} or
11385 @code{strong}. It defaults to @code{none}.
11390 Compute and draw a color distribution histogram for the input video.
11392 The computed histogram is a representation of the color component
11393 distribution in an image.
11395 Standard histogram displays the color components distribution in an image.
11396 Displays color graph for each color component. Shows distribution of
11397 the Y, U, V, A or R, G, B components, depending on input format, in the
11398 current frame. Below each graph a color component scale meter is shown.
11400 The filter accepts the following options:
11404 Set height of level. Default value is @code{200}.
11405 Allowed range is [50, 2048].
11408 Set height of color scale. Default value is @code{12}.
11409 Allowed range is [0, 40].
11413 It accepts the following values:
11416 Per color component graphs are placed below each other.
11419 Per color component graphs are placed side by side.
11422 Presents information identical to that in the @code{parade}, except
11423 that the graphs representing color components are superimposed directly
11426 Default is @code{stack}.
11429 Set mode. Can be either @code{linear}, or @code{logarithmic}.
11430 Default is @code{linear}.
11433 Set what color components to display.
11434 Default is @code{7}.
11437 Set foreground opacity. Default is @code{0.7}.
11440 Set background opacity. Default is @code{0.5}.
11443 @subsection Examples
11448 Calculate and draw histogram:
11450 ffplay -i input -vf histogram
11458 This is a high precision/quality 3d denoise filter. It aims to reduce
11459 image noise, producing smooth images and making still images really
11460 still. It should enhance compressibility.
11462 It accepts the following optional parameters:
11466 A non-negative floating point number which specifies spatial luma strength.
11467 It defaults to 4.0.
11469 @item chroma_spatial
11470 A non-negative floating point number which specifies spatial chroma strength.
11471 It defaults to 3.0*@var{luma_spatial}/4.0.
11474 A floating point number which specifies luma temporal strength. It defaults to
11475 6.0*@var{luma_spatial}/4.0.
11478 A floating point number which specifies chroma temporal strength. It defaults to
11479 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
11482 @anchor{hwdownload}
11483 @section hwdownload
11485 Download hardware frames to system memory.
11487 The input must be in hardware frames, and the output a non-hardware format.
11488 Not all formats will be supported on the output - it may be necessary to insert
11489 an additional @option{format} filter immediately following in the graph to get
11490 the output in a supported format.
11494 Map hardware frames to system memory or to another device.
11496 This filter has several different modes of operation; which one is used depends
11497 on the input and output formats:
11500 Hardware frame input, normal frame output
11502 Map the input frames to system memory and pass them to the output. If the
11503 original hardware frame is later required (for example, after overlaying
11504 something else on part of it), the @option{hwmap} filter can be used again
11505 in the next mode to retrieve it.
11507 Normal frame input, hardware frame output
11509 If the input is actually a software-mapped hardware frame, then unmap it -
11510 that is, return the original hardware frame.
11512 Otherwise, a device must be provided. Create new hardware surfaces on that
11513 device for the output, then map them back to the software format at the input
11514 and give those frames to the preceding filter. This will then act like the
11515 @option{hwupload} filter, but may be able to avoid an additional copy when
11516 the input is already in a compatible format.
11518 Hardware frame input and output
11520 A device must be supplied for the output, either directly or with the
11521 @option{derive_device} option. The input and output devices must be of
11522 different types and compatible - the exact meaning of this is
11523 system-dependent, but typically it means that they must refer to the same
11524 underlying hardware context (for example, refer to the same graphics card).
11526 If the input frames were originally created on the output device, then unmap
11527 to retrieve the original frames.
11529 Otherwise, map the frames to the output device - create new hardware frames
11530 on the output corresponding to the frames on the input.
11533 The following additional parameters are accepted:
11537 Set the frame mapping mode. Some combination of:
11540 The mapped frame should be readable.
11542 The mapped frame should be writeable.
11544 The mapping will always overwrite the entire frame.
11546 This may improve performance in some cases, as the original contents of the
11547 frame need not be loaded.
11549 The mapping must not involve any copying.
11551 Indirect mappings to copies of frames are created in some cases where either
11552 direct mapping is not possible or it would have unexpected properties.
11553 Setting this flag ensures that the mapping is direct and will fail if that is
11556 Defaults to @var{read+write} if not specified.
11558 @item derive_device @var{type}
11559 Rather than using the device supplied at initialisation, instead derive a new
11560 device of type @var{type} from the device the input frames exist on.
11563 In a hardware to hardware mapping, map in reverse - create frames in the sink
11564 and map them back to the source. This may be necessary in some cases where
11565 a mapping in one direction is required but only the opposite direction is
11566 supported by the devices being used.
11568 This option is dangerous - it may break the preceding filter in undefined
11569 ways if there are any additional constraints on that filter's output.
11570 Do not use it without fully understanding the implications of its use.
11576 Upload system memory frames to hardware surfaces.
11578 The device to upload to must be supplied when the filter is initialised. If
11579 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
11582 @anchor{hwupload_cuda}
11583 @section hwupload_cuda
11585 Upload system memory frames to a CUDA device.
11587 It accepts the following optional parameters:
11591 The number of the CUDA device to use
11596 Apply a high-quality magnification filter designed for pixel art. This filter
11597 was originally created by Maxim Stepin.
11599 It accepts the following option:
11603 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
11604 @code{hq3x} and @code{4} for @code{hq4x}.
11605 Default is @code{3}.
11609 Stack input videos horizontally.
11611 All streams must be of same pixel format and of same height.
11613 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
11614 to create same output.
11616 The filter accepts the following option:
11620 Set number of input streams. Default is 2.
11623 If set to 1, force the output to terminate when the shortest input
11624 terminates. Default value is 0.
11629 Modify the hue and/or the saturation of the input.
11631 It accepts the following parameters:
11635 Specify the hue angle as a number of degrees. It accepts an expression,
11636 and defaults to "0".
11639 Specify the saturation in the [-10,10] range. It accepts an expression and
11643 Specify the hue angle as a number of radians. It accepts an
11644 expression, and defaults to "0".
11647 Specify the brightness in the [-10,10] range. It accepts an expression and
11651 @option{h} and @option{H} are mutually exclusive, and can't be
11652 specified at the same time.
11654 The @option{b}, @option{h}, @option{H} and @option{s} option values are
11655 expressions containing the following constants:
11659 frame count of the input frame starting from 0
11662 presentation timestamp of the input frame expressed in time base units
11665 frame rate of the input video, NAN if the input frame rate is unknown
11668 timestamp expressed in seconds, NAN if the input timestamp is unknown
11671 time base of the input video
11674 @subsection Examples
11678 Set the hue to 90 degrees and the saturation to 1.0:
11684 Same command but expressing the hue in radians:
11690 Rotate hue and make the saturation swing between 0
11691 and 2 over a period of 1 second:
11693 hue="H=2*PI*t: s=sin(2*PI*t)+1"
11697 Apply a 3 seconds saturation fade-in effect starting at 0:
11699 hue="s=min(t/3\,1)"
11702 The general fade-in expression can be written as:
11704 hue="s=min(0\, max((t-START)/DURATION\, 1))"
11708 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
11710 hue="s=max(0\, min(1\, (8-t)/3))"
11713 The general fade-out expression can be written as:
11715 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
11720 @subsection Commands
11722 This filter supports the following commands:
11728 Modify the hue and/or the saturation and/or brightness of the input video.
11729 The command accepts the same syntax of the corresponding option.
11731 If the specified expression is not valid, it is kept at its current
11735 @section hysteresis
11737 Grow first stream into second stream by connecting components.
11738 This makes it possible to build more robust edge masks.
11740 This filter accepts the following options:
11744 Set which planes will be processed as bitmap, unprocessed planes will be
11745 copied from first stream.
11746 By default value 0xf, all planes will be processed.
11749 Set threshold which is used in filtering. If pixel component value is higher than
11750 this value filter algorithm for connecting components is activated.
11751 By default value is 0.
11756 Detect video interlacing type.
11758 This filter tries to detect if the input frames are interlaced, progressive,
11759 top or bottom field first. It will also try to detect fields that are
11760 repeated between adjacent frames (a sign of telecine).
11762 Single frame detection considers only immediately adjacent frames when classifying each frame.
11763 Multiple frame detection incorporates the classification history of previous frames.
11765 The filter will log these metadata values:
11768 @item single.current_frame
11769 Detected type of current frame using single-frame detection. One of:
11770 ``tff'' (top field first), ``bff'' (bottom field first),
11771 ``progressive'', or ``undetermined''
11774 Cumulative number of frames detected as top field first using single-frame detection.
11777 Cumulative number of frames detected as top field first using multiple-frame detection.
11780 Cumulative number of frames detected as bottom field first using single-frame detection.
11782 @item multiple.current_frame
11783 Detected type of current frame using multiple-frame detection. One of:
11784 ``tff'' (top field first), ``bff'' (bottom field first),
11785 ``progressive'', or ``undetermined''
11788 Cumulative number of frames detected as bottom field first using multiple-frame detection.
11790 @item single.progressive
11791 Cumulative number of frames detected as progressive using single-frame detection.
11793 @item multiple.progressive
11794 Cumulative number of frames detected as progressive using multiple-frame detection.
11796 @item single.undetermined
11797 Cumulative number of frames that could not be classified using single-frame detection.
11799 @item multiple.undetermined
11800 Cumulative number of frames that could not be classified using multiple-frame detection.
11802 @item repeated.current_frame
11803 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
11805 @item repeated.neither
11806 Cumulative number of frames with no repeated field.
11809 Cumulative number of frames with the top field repeated from the previous frame's top field.
11811 @item repeated.bottom
11812 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
11815 The filter accepts the following options:
11819 Set interlacing threshold.
11821 Set progressive threshold.
11823 Threshold for repeated field detection.
11825 Number of frames after which a given frame's contribution to the
11826 statistics is halved (i.e., it contributes only 0.5 to its
11827 classification). The default of 0 means that all frames seen are given
11828 full weight of 1.0 forever.
11829 @item analyze_interlaced_flag
11830 When this is not 0 then idet will use the specified number of frames to determine
11831 if the interlaced flag is accurate, it will not count undetermined frames.
11832 If the flag is found to be accurate it will be used without any further
11833 computations, if it is found to be inaccurate it will be cleared without any
11834 further computations. This allows inserting the idet filter as a low computational
11835 method to clean up the interlaced flag
11840 Deinterleave or interleave fields.
11842 This filter allows one to process interlaced images fields without
11843 deinterlacing them. Deinterleaving splits the input frame into 2
11844 fields (so called half pictures). Odd lines are moved to the top
11845 half of the output image, even lines to the bottom half.
11846 You can process (filter) them independently and then re-interleave them.
11848 The filter accepts the following options:
11852 @item chroma_mode, c
11853 @item alpha_mode, a
11854 Available values for @var{luma_mode}, @var{chroma_mode} and
11855 @var{alpha_mode} are:
11861 @item deinterleave, d
11862 Deinterleave fields, placing one above the other.
11864 @item interleave, i
11865 Interleave fields. Reverse the effect of deinterleaving.
11867 Default value is @code{none}.
11869 @item luma_swap, ls
11870 @item chroma_swap, cs
11871 @item alpha_swap, as
11872 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
11877 Apply inflate effect to the video.
11879 This filter replaces the pixel by the local(3x3) average by taking into account
11880 only values higher than the pixel.
11882 It accepts the following options:
11889 Limit the maximum change for each plane, default is 65535.
11890 If 0, plane will remain unchanged.
11895 Simple interlacing filter from progressive contents. This interleaves upper (or
11896 lower) lines from odd frames with lower (or upper) lines from even frames,
11897 halving the frame rate and preserving image height.
11900 Original Original New Frame
11901 Frame 'j' Frame 'j+1' (tff)
11902 ========== =========== ==================
11903 Line 0 --------------------> Frame 'j' Line 0
11904 Line 1 Line 1 ----> Frame 'j+1' Line 1
11905 Line 2 ---------------------> Frame 'j' Line 2
11906 Line 3 Line 3 ----> Frame 'j+1' Line 3
11908 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
11911 It accepts the following optional parameters:
11915 This determines whether the interlaced frame is taken from the even
11916 (tff - default) or odd (bff) lines of the progressive frame.
11919 Vertical lowpass filter to avoid twitter interlacing and
11920 reduce moire patterns.
11924 Disable vertical lowpass filter
11927 Enable linear filter (default)
11930 Enable complex filter. This will slightly less reduce twitter and moire
11931 but better retain detail and subjective sharpness impression.
11938 Deinterlace input video by applying Donald Graft's adaptive kernel
11939 deinterling. Work on interlaced parts of a video to produce
11940 progressive frames.
11942 The description of the accepted parameters follows.
11946 Set the threshold which affects the filter's tolerance when
11947 determining if a pixel line must be processed. It must be an integer
11948 in the range [0,255] and defaults to 10. A value of 0 will result in
11949 applying the process on every pixels.
11952 Paint pixels exceeding the threshold value to white if set to 1.
11956 Set the fields order. Swap fields if set to 1, leave fields alone if
11960 Enable additional sharpening if set to 1. Default is 0.
11963 Enable twoway sharpening if set to 1. Default is 0.
11966 @subsection Examples
11970 Apply default values:
11972 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
11976 Enable additional sharpening:
11982 Paint processed pixels in white:
11990 Slowly update darker pixels.
11992 This filter makes short flashes of light appear longer.
11993 This filter accepts the following options:
11997 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
12000 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
12003 @section lenscorrection
12005 Correct radial lens distortion
12007 This filter can be used to correct for radial distortion as can result from the use
12008 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
12009 one can use tools available for example as part of opencv or simply trial-and-error.
12010 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
12011 and extract the k1 and k2 coefficients from the resulting matrix.
12013 Note that effectively the same filter is available in the open-source tools Krita and
12014 Digikam from the KDE project.
12016 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
12017 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
12018 brightness distribution, so you may want to use both filters together in certain
12019 cases, though you will have to take care of ordering, i.e. whether vignetting should
12020 be applied before or after lens correction.
12022 @subsection Options
12024 The filter accepts the following options:
12028 Relative x-coordinate of the focal point of the image, and thereby the center of the
12029 distortion. This value has a range [0,1] and is expressed as fractions of the image
12030 width. Default is 0.5.
12032 Relative y-coordinate of the focal point of the image, and thereby the center of the
12033 distortion. This value has a range [0,1] and is expressed as fractions of the image
12034 height. Default is 0.5.
12036 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
12037 no correction. Default is 0.
12039 Coefficient of the double quadratic correction term. This value has a range [-1,1].
12040 0 means no correction. Default is 0.
12043 The formula that generates the correction is:
12045 @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)
12047 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
12048 distances from the focal point in the source and target images, respectively.
12052 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
12054 The @code{lensfun} filter requires the camera make, camera model, and lens model
12055 to apply the lens correction. The filter will load the lensfun database and
12056 query it to find the corresponding camera and lens entries in the database. As
12057 long as these entries can be found with the given options, the filter can
12058 perform corrections on frames. Note that incomplete strings will result in the
12059 filter choosing the best match with the given options, and the filter will
12060 output the chosen camera and lens models (logged with level "info"). You must
12061 provide the make, camera model, and lens model as they are required.
12063 The filter accepts the following options:
12067 The make of the camera (for example, "Canon"). This option is required.
12070 The model of the camera (for example, "Canon EOS 100D"). This option is
12074 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
12075 option is required.
12078 The type of correction to apply. The following values are valid options:
12082 Enables fixing lens vignetting.
12085 Enables fixing lens geometry. This is the default.
12088 Enables fixing chromatic aberrations.
12091 Enables fixing lens vignetting and lens geometry.
12094 Enables fixing lens vignetting and chromatic aberrations.
12097 Enables fixing both lens geometry and chromatic aberrations.
12100 Enables all possible corrections.
12104 The focal length of the image/video (zoom; expected constant for video). For
12105 example, a 18--55mm lens has focal length range of [18--55], so a value in that
12106 range should be chosen when using that lens. Default 18.
12109 The aperture of the image/video (expected constant for video). Note that
12110 aperture is only used for vignetting correction. Default 3.5.
12112 @item focus_distance
12113 The focus distance of the image/video (expected constant for video). Note that
12114 focus distance is only used for vignetting and only slightly affects the
12115 vignetting correction process. If unknown, leave it at the default value (which
12119 The scale factor which is applied after transformation. After correction the
12120 video is no longer necessarily rectangular. This parameter controls how much of
12121 the resulting image is visible. The value 0 means that a value will be chosen
12122 automatically such that there is little or no unmapped area in the output
12123 image. 1.0 means that no additional scaling is done. Lower values may result
12124 in more of the corrected image being visible, while higher values may avoid
12125 unmapped areas in the output.
12127 @item target_geometry
12128 The target geometry of the output image/video. The following values are valid
12132 @item rectilinear (default)
12135 @item equirectangular
12136 @item fisheye_orthographic
12137 @item fisheye_stereographic
12138 @item fisheye_equisolid
12139 @item fisheye_thoby
12142 Apply the reverse of image correction (instead of correcting distortion, apply
12145 @item interpolation
12146 The type of interpolation used when correcting distortion. The following values
12151 @item linear (default)
12156 @subsection Examples
12160 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
12161 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
12165 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
12169 Apply the same as before, but only for the first 5 seconds of video.
12172 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
12179 Obtain the VMAF (Video Multi-Method Assessment Fusion)
12180 score between two input videos.
12182 The obtained VMAF score is printed through the logging system.
12184 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
12185 After installing the library it can be enabled using:
12186 @code{./configure --enable-libvmaf --enable-version3}.
12187 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
12189 The filter has following options:
12193 Set the model path which is to be used for SVM.
12194 Default value: @code{"vmaf_v0.6.1.pkl"}
12197 Set the file path to be used to store logs.
12200 Set the format of the log file (xml or json).
12202 @item enable_transform
12203 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
12204 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
12205 Default value: @code{false}
12208 Invokes the phone model which will generate VMAF scores higher than in the
12209 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
12212 Enables computing psnr along with vmaf.
12215 Enables computing ssim along with vmaf.
12218 Enables computing ms_ssim along with vmaf.
12221 Set the pool method (mean, min or harmonic mean) to be used for computing vmaf.
12224 Set number of threads to be used when computing vmaf.
12227 Set interval for frame subsampling used when computing vmaf.
12229 @item enable_conf_interval
12230 Enables confidence interval.
12233 This filter also supports the @ref{framesync} options.
12235 On the below examples the input file @file{main.mpg} being processed is
12236 compared with the reference file @file{ref.mpg}.
12239 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
12242 Example with options:
12244 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
12249 Limits the pixel components values to the specified range [min, max].
12251 The filter accepts the following options:
12255 Lower bound. Defaults to the lowest allowed value for the input.
12258 Upper bound. Defaults to the highest allowed value for the input.
12261 Specify which planes will be processed. Defaults to all available.
12268 The filter accepts the following options:
12272 Set the number of loops. Setting this value to -1 will result in infinite loops.
12276 Set maximal size in number of frames. Default is 0.
12279 Set first frame of loop. Default is 0.
12282 @subsection Examples
12286 Loop single first frame infinitely:
12288 loop=loop=-1:size=1:start=0
12292 Loop single first frame 10 times:
12294 loop=loop=10:size=1:start=0
12298 Loop 10 first frames 5 times:
12300 loop=loop=5:size=10:start=0
12306 Apply a 1D LUT to an input video.
12308 The filter accepts the following options:
12312 Set the 1D LUT file name.
12314 Currently supported formats:
12323 Select interpolation mode.
12325 Available values are:
12329 Use values from the nearest defined point.
12331 Interpolate values using the linear interpolation.
12333 Interpolate values using the cosine interpolation.
12335 Interpolate values using the cubic interpolation.
12337 Interpolate values using the spline interpolation.
12344 Apply a 3D LUT to an input video.
12346 The filter accepts the following options:
12350 Set the 3D LUT file name.
12352 Currently supported formats:
12366 Select interpolation mode.
12368 Available values are:
12372 Use values from the nearest defined point.
12374 Interpolate values using the 8 points defining a cube.
12376 Interpolate values using a tetrahedron.
12382 Turn certain luma values into transparency.
12384 The filter accepts the following options:
12388 Set the luma which will be used as base for transparency.
12389 Default value is @code{0}.
12392 Set the range of luma values to be keyed out.
12393 Default value is @code{0}.
12396 Set the range of softness. Default value is @code{0}.
12397 Use this to control gradual transition from zero to full transparency.
12400 @section lut, lutrgb, lutyuv
12402 Compute a look-up table for binding each pixel component input value
12403 to an output value, and apply it to the input video.
12405 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
12406 to an RGB input video.
12408 These filters accept the following parameters:
12411 set first pixel component expression
12413 set second pixel component expression
12415 set third pixel component expression
12417 set fourth pixel component expression, corresponds to the alpha component
12420 set red component expression
12422 set green component expression
12424 set blue component expression
12426 alpha component expression
12429 set Y/luminance component expression
12431 set U/Cb component expression
12433 set V/Cr component expression
12436 Each of them specifies the expression to use for computing the lookup table for
12437 the corresponding pixel component values.
12439 The exact component associated to each of the @var{c*} options depends on the
12442 The @var{lut} filter requires either YUV or RGB pixel formats in input,
12443 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
12445 The expressions can contain the following constants and functions:
12450 The input width and height.
12453 The input value for the pixel component.
12456 The input value, clipped to the @var{minval}-@var{maxval} range.
12459 The maximum value for the pixel component.
12462 The minimum value for the pixel component.
12465 The negated value for the pixel component value, clipped to the
12466 @var{minval}-@var{maxval} range; it corresponds to the expression
12467 "maxval-clipval+minval".
12470 The computed value in @var{val}, clipped to the
12471 @var{minval}-@var{maxval} range.
12473 @item gammaval(gamma)
12474 The computed gamma correction value of the pixel component value,
12475 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
12477 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
12481 All expressions default to "val".
12483 @subsection Examples
12487 Negate input video:
12489 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
12490 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
12493 The above is the same as:
12495 lutrgb="r=negval:g=negval:b=negval"
12496 lutyuv="y=negval:u=negval:v=negval"
12506 Remove chroma components, turning the video into a graytone image:
12508 lutyuv="u=128:v=128"
12512 Apply a luma burning effect:
12518 Remove green and blue components:
12524 Set a constant alpha channel value on input:
12526 format=rgba,lutrgb=a="maxval-minval/2"
12530 Correct luminance gamma by a factor of 0.5:
12532 lutyuv=y=gammaval(0.5)
12536 Discard least significant bits of luma:
12538 lutyuv=y='bitand(val, 128+64+32)'
12542 Technicolor like effect:
12544 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
12548 @section lut2, tlut2
12550 The @code{lut2} filter takes two input streams and outputs one
12553 The @code{tlut2} (time lut2) filter takes two consecutive frames
12554 from one single stream.
12556 This filter accepts the following parameters:
12559 set first pixel component expression
12561 set second pixel component expression
12563 set third pixel component expression
12565 set fourth pixel component expression, corresponds to the alpha component
12568 set output bit depth, only available for @code{lut2} filter. By default is 0,
12569 which means bit depth is automatically picked from first input format.
12572 Each of them specifies the expression to use for computing the lookup table for
12573 the corresponding pixel component values.
12575 The exact component associated to each of the @var{c*} options depends on the
12578 The expressions can contain the following constants:
12583 The input width and height.
12586 The first input value for the pixel component.
12589 The second input value for the pixel component.
12592 The first input video bit depth.
12595 The second input video bit depth.
12598 All expressions default to "x".
12600 @subsection Examples
12604 Highlight differences between two RGB video streams:
12606 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)'
12610 Highlight differences between two YUV video streams:
12612 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)'
12616 Show max difference between two video streams:
12618 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)))'
12622 @section maskedclamp
12624 Clamp the first input stream with the second input and third input stream.
12626 Returns the value of first stream to be between second input
12627 stream - @code{undershoot} and third input stream + @code{overshoot}.
12629 This filter accepts the following options:
12632 Default value is @code{0}.
12635 Default value is @code{0}.
12638 Set which planes will be processed as bitmap, unprocessed planes will be
12639 copied from first stream.
12640 By default value 0xf, all planes will be processed.
12643 @section maskedmerge
12645 Merge the first input stream with the second input stream using per pixel
12646 weights in the third input stream.
12648 A value of 0 in the third stream pixel component means that pixel component
12649 from first stream is returned unchanged, while maximum value (eg. 255 for
12650 8-bit videos) means that pixel component from second stream is returned
12651 unchanged. Intermediate values define the amount of merging between both
12652 input stream's pixel components.
12654 This filter accepts the following options:
12657 Set which planes will be processed as bitmap, unprocessed planes will be
12658 copied from first stream.
12659 By default value 0xf, all planes will be processed.
12663 Create mask from input video.
12665 For example it is useful to create motion masks after @code{tblend} filter.
12667 This filter accepts the following options:
12671 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
12674 Set high threshold. Any pixel component higher than this value will be set to max value
12675 allowed for current pixel format.
12678 Set planes to filter, by default all available planes are filtered.
12681 Fill all frame pixels with this value.
12684 Set max average pixel value for frame. If sum of all pixel components is higher that this
12685 average, output frame will be completely filled with value set by @var{fill} option.
12686 Typically useful for scene changes when used in combination with @code{tblend} filter.
12691 Apply motion-compensation deinterlacing.
12693 It needs one field per frame as input and must thus be used together
12694 with yadif=1/3 or equivalent.
12696 This filter accepts the following options:
12699 Set the deinterlacing mode.
12701 It accepts one of the following values:
12706 use iterative motion estimation
12708 like @samp{slow}, but use multiple reference frames.
12710 Default value is @samp{fast}.
12713 Set the picture field parity assumed for the input video. It must be
12714 one of the following values:
12718 assume top field first
12720 assume bottom field first
12723 Default value is @samp{bff}.
12726 Set per-block quantization parameter (QP) used by the internal
12729 Higher values should result in a smoother motion vector field but less
12730 optimal individual vectors. Default value is 1.
12733 @section mergeplanes
12735 Merge color channel components from several video streams.
12737 The filter accepts up to 4 input streams, and merge selected input
12738 planes to the output video.
12740 This filter accepts the following options:
12743 Set input to output plane mapping. Default is @code{0}.
12745 The mappings is specified as a bitmap. It should be specified as a
12746 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
12747 mapping for the first plane of the output stream. 'A' sets the number of
12748 the input stream to use (from 0 to 3), and 'a' the plane number of the
12749 corresponding input to use (from 0 to 3). The rest of the mappings is
12750 similar, 'Bb' describes the mapping for the output stream second
12751 plane, 'Cc' describes the mapping for the output stream third plane and
12752 'Dd' describes the mapping for the output stream fourth plane.
12755 Set output pixel format. Default is @code{yuva444p}.
12758 @subsection Examples
12762 Merge three gray video streams of same width and height into single video stream:
12764 [a0][a1][a2]mergeplanes=0x001020:yuv444p
12768 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
12770 [a0][a1]mergeplanes=0x00010210:yuva444p
12774 Swap Y and A plane in yuva444p stream:
12776 format=yuva444p,mergeplanes=0x03010200:yuva444p
12780 Swap U and V plane in yuv420p stream:
12782 format=yuv420p,mergeplanes=0x000201:yuv420p
12786 Cast a rgb24 clip to yuv444p:
12788 format=rgb24,mergeplanes=0x000102:yuv444p
12794 Estimate and export motion vectors using block matching algorithms.
12795 Motion vectors are stored in frame side data to be used by other filters.
12797 This filter accepts the following options:
12800 Specify the motion estimation method. Accepts one of the following values:
12804 Exhaustive search algorithm.
12806 Three step search algorithm.
12808 Two dimensional logarithmic search algorithm.
12810 New three step search algorithm.
12812 Four step search algorithm.
12814 Diamond search algorithm.
12816 Hexagon-based search algorithm.
12818 Enhanced predictive zonal search algorithm.
12820 Uneven multi-hexagon search algorithm.
12822 Default value is @samp{esa}.
12825 Macroblock size. Default @code{16}.
12828 Search parameter. Default @code{7}.
12831 @section midequalizer
12833 Apply Midway Image Equalization effect using two video streams.
12835 Midway Image Equalization adjusts a pair of images to have the same
12836 histogram, while maintaining their dynamics as much as possible. It's
12837 useful for e.g. matching exposures from a pair of stereo cameras.
12839 This filter has two inputs and one output, which must be of same pixel format, but
12840 may be of different sizes. The output of filter is first input adjusted with
12841 midway histogram of both inputs.
12843 This filter accepts the following option:
12847 Set which planes to process. Default is @code{15}, which is all available planes.
12850 @section minterpolate
12852 Convert the video to specified frame rate using motion interpolation.
12854 This filter accepts the following options:
12857 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}.
12860 Motion interpolation mode. Following values are accepted:
12863 Duplicate previous or next frame for interpolating new ones.
12865 Blend source frames. Interpolated frame is mean of previous and next frames.
12867 Motion compensated interpolation. Following options are effective when this mode is selected:
12871 Motion compensation mode. Following values are accepted:
12874 Overlapped block motion compensation.
12876 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
12878 Default mode is @samp{obmc}.
12881 Motion estimation mode. Following values are accepted:
12884 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
12886 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
12888 Default mode is @samp{bilat}.
12891 The algorithm to be used for motion estimation. Following values are accepted:
12894 Exhaustive search algorithm.
12896 Three step search algorithm.
12898 Two dimensional logarithmic search algorithm.
12900 New three step search algorithm.
12902 Four step search algorithm.
12904 Diamond search algorithm.
12906 Hexagon-based search algorithm.
12908 Enhanced predictive zonal search algorithm.
12910 Uneven multi-hexagon search algorithm.
12912 Default algorithm is @samp{epzs}.
12915 Macroblock size. Default @code{16}.
12918 Motion estimation search parameter. Default @code{32}.
12921 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).
12926 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:
12929 Disable scene change detection.
12931 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
12933 Default method is @samp{fdiff}.
12935 @item scd_threshold
12936 Scene change detection threshold. Default is @code{5.0}.
12941 Mix several video input streams into one video stream.
12943 A description of the accepted options follows.
12947 The number of inputs. If unspecified, it defaults to 2.
12950 Specify weight of each input video stream as sequence.
12951 Each weight is separated by space. If number of weights
12952 is smaller than number of @var{frames} last specified
12953 weight will be used for all remaining unset weights.
12956 Specify scale, if it is set it will be multiplied with sum
12957 of each weight multiplied with pixel values to give final destination
12958 pixel value. By default @var{scale} is auto scaled to sum of weights.
12961 Specify how end of stream is determined.
12964 The duration of the longest input. (default)
12967 The duration of the shortest input.
12970 The duration of the first input.
12974 @section mpdecimate
12976 Drop frames that do not differ greatly from the previous frame in
12977 order to reduce frame rate.
12979 The main use of this filter is for very-low-bitrate encoding
12980 (e.g. streaming over dialup modem), but it could in theory be used for
12981 fixing movies that were inverse-telecined incorrectly.
12983 A description of the accepted options follows.
12987 Set the maximum number of consecutive frames which can be dropped (if
12988 positive), or the minimum interval between dropped frames (if
12989 negative). If the value is 0, the frame is dropped disregarding the
12990 number of previous sequentially dropped frames.
12992 Default value is 0.
12997 Set the dropping threshold values.
12999 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
13000 represent actual pixel value differences, so a threshold of 64
13001 corresponds to 1 unit of difference for each pixel, or the same spread
13002 out differently over the block.
13004 A frame is a candidate for dropping if no 8x8 blocks differ by more
13005 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
13006 meaning the whole image) differ by more than a threshold of @option{lo}.
13008 Default value for @option{hi} is 64*12, default value for @option{lo} is
13009 64*5, and default value for @option{frac} is 0.33.
13015 Negate (invert) the input video.
13017 It accepts the following option:
13022 With value 1, it negates the alpha component, if present. Default value is 0.
13028 Denoise frames using Non-Local Means algorithm.
13030 Each pixel is adjusted by looking for other pixels with similar contexts. This
13031 context similarity is defined by comparing their surrounding patches of size
13032 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
13035 Note that the research area defines centers for patches, which means some
13036 patches will be made of pixels outside that research area.
13038 The filter accepts the following options.
13042 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
13045 Set patch size. Default is 7. Must be odd number in range [0, 99].
13048 Same as @option{p} but for chroma planes.
13050 The default value is @var{0} and means automatic.
13053 Set research size. Default is 15. Must be odd number in range [0, 99].
13056 Same as @option{r} but for chroma planes.
13058 The default value is @var{0} and means automatic.
13063 Deinterlace video using neural network edge directed interpolation.
13065 This filter accepts the following options:
13069 Mandatory option, without binary file filter can not work.
13070 Currently file can be found here:
13071 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
13074 Set which frames to deinterlace, by default it is @code{all}.
13075 Can be @code{all} or @code{interlaced}.
13078 Set mode of operation.
13080 Can be one of the following:
13084 Use frame flags, both fields.
13086 Use frame flags, single field.
13088 Use top field only.
13090 Use bottom field only.
13092 Use both fields, top first.
13094 Use both fields, bottom first.
13098 Set which planes to process, by default filter process all frames.
13101 Set size of local neighborhood around each pixel, used by the predictor neural
13104 Can be one of the following:
13117 Set the number of neurons in predictor neural network.
13118 Can be one of the following:
13129 Controls the number of different neural network predictions that are blended
13130 together to compute the final output value. Can be @code{fast}, default or
13134 Set which set of weights to use in the predictor.
13135 Can be one of the following:
13139 weights trained to minimize absolute error
13141 weights trained to minimize squared error
13145 Controls whether or not the prescreener neural network is used to decide
13146 which pixels should be processed by the predictor neural network and which
13147 can be handled by simple cubic interpolation.
13148 The prescreener is trained to know whether cubic interpolation will be
13149 sufficient for a pixel or whether it should be predicted by the predictor nn.
13150 The computational complexity of the prescreener nn is much less than that of
13151 the predictor nn. Since most pixels can be handled by cubic interpolation,
13152 using the prescreener generally results in much faster processing.
13153 The prescreener is pretty accurate, so the difference between using it and not
13154 using it is almost always unnoticeable.
13156 Can be one of the following:
13164 Default is @code{new}.
13167 Set various debugging flags.
13172 Force libavfilter not to use any of the specified pixel formats for the
13173 input to the next filter.
13175 It accepts the following parameters:
13179 A '|'-separated list of pixel format names, such as
13180 pix_fmts=yuv420p|monow|rgb24".
13184 @subsection Examples
13188 Force libavfilter to use a format different from @var{yuv420p} for the
13189 input to the vflip filter:
13191 noformat=pix_fmts=yuv420p,vflip
13195 Convert the input video to any of the formats not contained in the list:
13197 noformat=yuv420p|yuv444p|yuv410p
13203 Add noise on video input frame.
13205 The filter accepts the following options:
13213 Set noise seed for specific pixel component or all pixel components in case
13214 of @var{all_seed}. Default value is @code{123457}.
13216 @item all_strength, alls
13217 @item c0_strength, c0s
13218 @item c1_strength, c1s
13219 @item c2_strength, c2s
13220 @item c3_strength, c3s
13221 Set noise strength for specific pixel component or all pixel components in case
13222 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
13224 @item all_flags, allf
13225 @item c0_flags, c0f
13226 @item c1_flags, c1f
13227 @item c2_flags, c2f
13228 @item c3_flags, c3f
13229 Set pixel component flags or set flags for all components if @var{all_flags}.
13230 Available values for component flags are:
13233 averaged temporal noise (smoother)
13235 mix random noise with a (semi)regular pattern
13237 temporal noise (noise pattern changes between frames)
13239 uniform noise (gaussian otherwise)
13243 @subsection Examples
13245 Add temporal and uniform noise to input video:
13247 noise=alls=20:allf=t+u
13252 Normalize RGB video (aka histogram stretching, contrast stretching).
13253 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
13255 For each channel of each frame, the filter computes the input range and maps
13256 it linearly to the user-specified output range. The output range defaults
13257 to the full dynamic range from pure black to pure white.
13259 Temporal smoothing can be used on the input range to reduce flickering (rapid
13260 changes in brightness) caused when small dark or bright objects enter or leave
13261 the scene. This is similar to the auto-exposure (automatic gain control) on a
13262 video camera, and, like a video camera, it may cause a period of over- or
13263 under-exposure of the video.
13265 The R,G,B channels can be normalized independently, which may cause some
13266 color shifting, or linked together as a single channel, which prevents
13267 color shifting. Linked normalization preserves hue. Independent normalization
13268 does not, so it can be used to remove some color casts. Independent and linked
13269 normalization can be combined in any ratio.
13271 The normalize filter accepts the following options:
13276 Colors which define the output range. The minimum input value is mapped to
13277 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
13278 The defaults are black and white respectively. Specifying white for
13279 @var{blackpt} and black for @var{whitept} will give color-inverted,
13280 normalized video. Shades of grey can be used to reduce the dynamic range
13281 (contrast). Specifying saturated colors here can create some interesting
13285 The number of previous frames to use for temporal smoothing. The input range
13286 of each channel is smoothed using a rolling average over the current frame
13287 and the @var{smoothing} previous frames. The default is 0 (no temporal
13291 Controls the ratio of independent (color shifting) channel normalization to
13292 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
13293 independent. Defaults to 1.0 (fully independent).
13296 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
13297 expensive no-op. Defaults to 1.0 (full strength).
13301 @subsection Examples
13303 Stretch video contrast to use the full dynamic range, with no temporal
13304 smoothing; may flicker depending on the source content:
13306 normalize=blackpt=black:whitept=white:smoothing=0
13309 As above, but with 50 frames of temporal smoothing; flicker should be
13310 reduced, depending on the source content:
13312 normalize=blackpt=black:whitept=white:smoothing=50
13315 As above, but with hue-preserving linked channel normalization:
13317 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
13320 As above, but with half strength:
13322 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
13325 Map the darkest input color to red, the brightest input color to cyan:
13327 normalize=blackpt=red:whitept=cyan
13332 Pass the video source unchanged to the output.
13335 Optical Character Recognition
13337 This filter uses Tesseract for optical character recognition. To enable
13338 compilation of this filter, you need to configure FFmpeg with
13339 @code{--enable-libtesseract}.
13341 It accepts the following options:
13345 Set datapath to tesseract data. Default is to use whatever was
13346 set at installation.
13349 Set language, default is "eng".
13352 Set character whitelist.
13355 Set character blacklist.
13358 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
13359 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
13363 Apply a video transform using libopencv.
13365 To enable this filter, install the libopencv library and headers and
13366 configure FFmpeg with @code{--enable-libopencv}.
13368 It accepts the following parameters:
13373 The name of the libopencv filter to apply.
13375 @item filter_params
13376 The parameters to pass to the libopencv filter. If not specified, the default
13377 values are assumed.
13381 Refer to the official libopencv documentation for more precise
13383 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
13385 Several libopencv filters are supported; see the following subsections.
13390 Dilate an image by using a specific structuring element.
13391 It corresponds to the libopencv function @code{cvDilate}.
13393 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
13395 @var{struct_el} represents a structuring element, and has the syntax:
13396 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
13398 @var{cols} and @var{rows} represent the number of columns and rows of
13399 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
13400 point, and @var{shape} the shape for the structuring element. @var{shape}
13401 must be "rect", "cross", "ellipse", or "custom".
13403 If the value for @var{shape} is "custom", it must be followed by a
13404 string of the form "=@var{filename}". The file with name
13405 @var{filename} is assumed to represent a binary image, with each
13406 printable character corresponding to a bright pixel. When a custom
13407 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
13408 or columns and rows of the read file are assumed instead.
13410 The default value for @var{struct_el} is "3x3+0x0/rect".
13412 @var{nb_iterations} specifies the number of times the transform is
13413 applied to the image, and defaults to 1.
13417 # Use the default values
13420 # Dilate using a structuring element with a 5x5 cross, iterating two times
13421 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
13423 # Read the shape from the file diamond.shape, iterating two times.
13424 # The file diamond.shape may contain a pattern of characters like this
13430 # The specified columns and rows are ignored
13431 # but the anchor point coordinates are not
13432 ocv=dilate:0x0+2x2/custom=diamond.shape|2
13437 Erode an image by using a specific structuring element.
13438 It corresponds to the libopencv function @code{cvErode}.
13440 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
13441 with the same syntax and semantics as the @ref{dilate} filter.
13445 Smooth the input video.
13447 The filter takes the following parameters:
13448 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
13450 @var{type} is the type of smooth filter to apply, and must be one of
13451 the following values: "blur", "blur_no_scale", "median", "gaussian",
13452 or "bilateral". The default value is "gaussian".
13454 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
13455 depends on the smooth type. @var{param1} and
13456 @var{param2} accept integer positive values or 0. @var{param3} and
13457 @var{param4} accept floating point values.
13459 The default value for @var{param1} is 3. The default value for the
13460 other parameters is 0.
13462 These parameters correspond to the parameters assigned to the
13463 libopencv function @code{cvSmooth}.
13465 @section oscilloscope
13467 2D Video Oscilloscope.
13469 Useful to measure spatial impulse, step responses, chroma delays, etc.
13471 It accepts the following parameters:
13475 Set scope center x position.
13478 Set scope center y position.
13481 Set scope size, relative to frame diagonal.
13484 Set scope tilt/rotation.
13490 Set trace center x position.
13493 Set trace center y position.
13496 Set trace width, relative to width of frame.
13499 Set trace height, relative to height of frame.
13502 Set which components to trace. By default it traces first three components.
13505 Draw trace grid. By default is enabled.
13508 Draw some statistics. By default is enabled.
13511 Draw scope. By default is enabled.
13514 @subsection Examples
13518 Inspect full first row of video frame.
13520 oscilloscope=x=0.5:y=0:s=1
13524 Inspect full last row of video frame.
13526 oscilloscope=x=0.5:y=1:s=1
13530 Inspect full 5th line of video frame of height 1080.
13532 oscilloscope=x=0.5:y=5/1080:s=1
13536 Inspect full last column of video frame.
13538 oscilloscope=x=1:y=0.5:s=1:t=1
13546 Overlay one video on top of another.
13548 It takes two inputs and has one output. The first input is the "main"
13549 video on which the second input is overlaid.
13551 It accepts the following parameters:
13553 A description of the accepted options follows.
13558 Set the expression for the x and y coordinates of the overlaid video
13559 on the main video. Default value is "0" for both expressions. In case
13560 the expression is invalid, it is set to a huge value (meaning that the
13561 overlay will not be displayed within the output visible area).
13564 See @ref{framesync}.
13567 Set when the expressions for @option{x}, and @option{y} are evaluated.
13569 It accepts the following values:
13572 only evaluate expressions once during the filter initialization or
13573 when a command is processed
13576 evaluate expressions for each incoming frame
13579 Default value is @samp{frame}.
13582 See @ref{framesync}.
13585 Set the format for the output video.
13587 It accepts the following values:
13590 force YUV420 output
13593 force YUV422 output
13596 force YUV444 output
13599 force packed RGB output
13602 force planar RGB output
13605 automatically pick format
13608 Default value is @samp{yuv420}.
13611 See @ref{framesync}.
13614 Set format of alpha of the overlaid video, it can be @var{straight} or
13615 @var{premultiplied}. Default is @var{straight}.
13618 The @option{x}, and @option{y} expressions can contain the following
13624 The main input width and height.
13628 The overlay input width and height.
13632 The computed values for @var{x} and @var{y}. They are evaluated for
13637 horizontal and vertical chroma subsample values of the output
13638 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
13642 the number of input frame, starting from 0
13645 the position in the file of the input frame, NAN if unknown
13648 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
13652 This filter also supports the @ref{framesync} options.
13654 Note that the @var{n}, @var{pos}, @var{t} variables are available only
13655 when evaluation is done @emph{per frame}, and will evaluate to NAN
13656 when @option{eval} is set to @samp{init}.
13658 Be aware that frames are taken from each input video in timestamp
13659 order, hence, if their initial timestamps differ, it is a good idea
13660 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
13661 have them begin in the same zero timestamp, as the example for
13662 the @var{movie} filter does.
13664 You can chain together more overlays but you should test the
13665 efficiency of such approach.
13667 @subsection Commands
13669 This filter supports the following commands:
13673 Modify the x and y of the overlay input.
13674 The command accepts the same syntax of the corresponding option.
13676 If the specified expression is not valid, it is kept at its current
13680 @subsection Examples
13684 Draw the overlay at 10 pixels from the bottom right corner of the main
13687 overlay=main_w-overlay_w-10:main_h-overlay_h-10
13690 Using named options the example above becomes:
13692 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
13696 Insert a transparent PNG logo in the bottom left corner of the input,
13697 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
13699 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
13703 Insert 2 different transparent PNG logos (second logo on bottom
13704 right corner) using the @command{ffmpeg} tool:
13706 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
13710 Add a transparent color layer on top of the main video; @code{WxH}
13711 must specify the size of the main input to the overlay filter:
13713 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
13717 Play an original video and a filtered version (here with the deshake
13718 filter) side by side using the @command{ffplay} tool:
13720 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
13723 The above command is the same as:
13725 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
13729 Make a sliding overlay appearing from the left to the right top part of the
13730 screen starting since time 2:
13732 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
13736 Compose output by putting two input videos side to side:
13738 ffmpeg -i left.avi -i right.avi -filter_complex "
13739 nullsrc=size=200x100 [background];
13740 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
13741 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
13742 [background][left] overlay=shortest=1 [background+left];
13743 [background+left][right] overlay=shortest=1:x=100 [left+right]
13748 Mask 10-20 seconds of a video by applying the delogo filter to a section
13750 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
13751 -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]'
13756 Chain several overlays in cascade:
13758 nullsrc=s=200x200 [bg];
13759 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
13760 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
13761 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
13762 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
13763 [in3] null, [mid2] overlay=100:100 [out0]
13770 Apply Overcomplete Wavelet denoiser.
13772 The filter accepts the following options:
13778 Larger depth values will denoise lower frequency components more, but
13779 slow down filtering.
13781 Must be an int in the range 8-16, default is @code{8}.
13783 @item luma_strength, ls
13786 Must be a double value in the range 0-1000, default is @code{1.0}.
13788 @item chroma_strength, cs
13789 Set chroma strength.
13791 Must be a double value in the range 0-1000, default is @code{1.0}.
13797 Add paddings to the input image, and place the original input at the
13798 provided @var{x}, @var{y} coordinates.
13800 It accepts the following parameters:
13805 Specify an expression for the size of the output image with the
13806 paddings added. If the value for @var{width} or @var{height} is 0, the
13807 corresponding input size is used for the output.
13809 The @var{width} expression can reference the value set by the
13810 @var{height} expression, and vice versa.
13812 The default value of @var{width} and @var{height} is 0.
13816 Specify the offsets to place the input image at within the padded area,
13817 with respect to the top/left border of the output image.
13819 The @var{x} expression can reference the value set by the @var{y}
13820 expression, and vice versa.
13822 The default value of @var{x} and @var{y} is 0.
13824 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
13825 so the input image is centered on the padded area.
13828 Specify the color of the padded area. For the syntax of this option,
13829 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
13830 manual,ffmpeg-utils}.
13832 The default value of @var{color} is "black".
13835 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
13837 It accepts the following values:
13841 Only evaluate expressions once during the filter initialization or when
13842 a command is processed.
13845 Evaluate expressions for each incoming frame.
13849 Default value is @samp{init}.
13852 Pad to aspect instead to a resolution.
13856 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
13857 options are expressions containing the following constants:
13862 The input video width and height.
13866 These are the same as @var{in_w} and @var{in_h}.
13870 The output width and height (the size of the padded area), as
13871 specified by the @var{width} and @var{height} expressions.
13875 These are the same as @var{out_w} and @var{out_h}.
13879 The x and y offsets as specified by the @var{x} and @var{y}
13880 expressions, or NAN if not yet specified.
13883 same as @var{iw} / @var{ih}
13886 input sample aspect ratio
13889 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
13893 The horizontal and vertical chroma subsample values. For example for the
13894 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
13897 @subsection Examples
13901 Add paddings with the color "violet" to the input video. The output video
13902 size is 640x480, and the top-left corner of the input video is placed at
13905 pad=640:480:0:40:violet
13908 The example above is equivalent to the following command:
13910 pad=width=640:height=480:x=0:y=40:color=violet
13914 Pad the input to get an output with dimensions increased by 3/2,
13915 and put the input video at the center of the padded area:
13917 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
13921 Pad the input to get a squared output with size equal to the maximum
13922 value between the input width and height, and put the input video at
13923 the center of the padded area:
13925 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
13929 Pad the input to get a final w/h ratio of 16:9:
13931 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
13935 In case of anamorphic video, in order to set the output display aspect
13936 correctly, it is necessary to use @var{sar} in the expression,
13937 according to the relation:
13939 (ih * X / ih) * sar = output_dar
13940 X = output_dar / sar
13943 Thus the previous example needs to be modified to:
13945 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
13949 Double the output size and put the input video in the bottom-right
13950 corner of the output padded area:
13952 pad="2*iw:2*ih:ow-iw:oh-ih"
13956 @anchor{palettegen}
13957 @section palettegen
13959 Generate one palette for a whole video stream.
13961 It accepts the following options:
13965 Set the maximum number of colors to quantize in the palette.
13966 Note: the palette will still contain 256 colors; the unused palette entries
13969 @item reserve_transparent
13970 Create a palette of 255 colors maximum and reserve the last one for
13971 transparency. Reserving the transparency color is useful for GIF optimization.
13972 If not set, the maximum of colors in the palette will be 256. You probably want
13973 to disable this option for a standalone image.
13976 @item transparency_color
13977 Set the color that will be used as background for transparency.
13980 Set statistics mode.
13982 It accepts the following values:
13985 Compute full frame histograms.
13987 Compute histograms only for the part that differs from previous frame. This
13988 might be relevant to give more importance to the moving part of your input if
13989 the background is static.
13991 Compute new histogram for each frame.
13994 Default value is @var{full}.
13997 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
13998 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
13999 color quantization of the palette. This information is also visible at
14000 @var{info} logging level.
14002 @subsection Examples
14006 Generate a representative palette of a given video using @command{ffmpeg}:
14008 ffmpeg -i input.mkv -vf palettegen palette.png
14012 @section paletteuse
14014 Use a palette to downsample an input video stream.
14016 The filter takes two inputs: one video stream and a palette. The palette must
14017 be a 256 pixels image.
14019 It accepts the following options:
14023 Select dithering mode. Available algorithms are:
14026 Ordered 8x8 bayer dithering (deterministic)
14028 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
14029 Note: this dithering is sometimes considered "wrong" and is included as a
14031 @item floyd_steinberg
14032 Floyd and Steingberg dithering (error diffusion)
14034 Frankie Sierra dithering v2 (error diffusion)
14036 Frankie Sierra dithering v2 "Lite" (error diffusion)
14039 Default is @var{sierra2_4a}.
14042 When @var{bayer} dithering is selected, this option defines the scale of the
14043 pattern (how much the crosshatch pattern is visible). A low value means more
14044 visible pattern for less banding, and higher value means less visible pattern
14045 at the cost of more banding.
14047 The option must be an integer value in the range [0,5]. Default is @var{2}.
14050 If set, define the zone to process
14054 Only the changing rectangle will be reprocessed. This is similar to GIF
14055 cropping/offsetting compression mechanism. This option can be useful for speed
14056 if only a part of the image is changing, and has use cases such as limiting the
14057 scope of the error diffusal @option{dither} to the rectangle that bounds the
14058 moving scene (it leads to more deterministic output if the scene doesn't change
14059 much, and as a result less moving noise and better GIF compression).
14062 Default is @var{none}.
14065 Take new palette for each output frame.
14067 @item alpha_threshold
14068 Sets the alpha threshold for transparency. Alpha values above this threshold
14069 will be treated as completely opaque, and values below this threshold will be
14070 treated as completely transparent.
14072 The option must be an integer value in the range [0,255]. Default is @var{128}.
14075 @subsection Examples
14079 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
14080 using @command{ffmpeg}:
14082 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
14086 @section perspective
14088 Correct perspective of video not recorded perpendicular to the screen.
14090 A description of the accepted parameters follows.
14101 Set coordinates expression for top left, top right, bottom left and bottom right corners.
14102 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
14103 If the @code{sense} option is set to @code{source}, then the specified points will be sent
14104 to the corners of the destination. If the @code{sense} option is set to @code{destination},
14105 then the corners of the source will be sent to the specified coordinates.
14107 The expressions can use the following variables:
14112 the width and height of video frame.
14116 Output frame count.
14119 @item interpolation
14120 Set interpolation for perspective correction.
14122 It accepts the following values:
14128 Default value is @samp{linear}.
14131 Set interpretation of coordinate options.
14133 It accepts the following values:
14137 Send point in the source specified by the given coordinates to
14138 the corners of the destination.
14140 @item 1, destination
14142 Send the corners of the source to the point in the destination specified
14143 by the given coordinates.
14145 Default value is @samp{source}.
14149 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
14151 It accepts the following values:
14154 only evaluate expressions once during the filter initialization or
14155 when a command is processed
14158 evaluate expressions for each incoming frame
14161 Default value is @samp{init}.
14166 Delay interlaced video by one field time so that the field order changes.
14168 The intended use is to fix PAL movies that have been captured with the
14169 opposite field order to the film-to-video transfer.
14171 A description of the accepted parameters follows.
14177 It accepts the following values:
14180 Capture field order top-first, transfer bottom-first.
14181 Filter will delay the bottom field.
14184 Capture field order bottom-first, transfer top-first.
14185 Filter will delay the top field.
14188 Capture and transfer with the same field order. This mode only exists
14189 for the documentation of the other options to refer to, but if you
14190 actually select it, the filter will faithfully do nothing.
14193 Capture field order determined automatically by field flags, transfer
14195 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
14196 basis using field flags. If no field information is available,
14197 then this works just like @samp{u}.
14200 Capture unknown or varying, transfer opposite.
14201 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
14202 analyzing the images and selecting the alternative that produces best
14203 match between the fields.
14206 Capture top-first, transfer unknown or varying.
14207 Filter selects among @samp{t} and @samp{p} using image analysis.
14210 Capture bottom-first, transfer unknown or varying.
14211 Filter selects among @samp{b} and @samp{p} using image analysis.
14214 Capture determined by field flags, transfer unknown or varying.
14215 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
14216 image analysis. If no field information is available, then this works just
14217 like @samp{U}. This is the default mode.
14220 Both capture and transfer unknown or varying.
14221 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
14225 @section photosensitivity
14226 Reduce various flashes in video, so to help users with epilepsy.
14228 It accepts the following options:
14231 Set how many frames to use when filtering. Default is 30.
14234 Set detection threshold factor. Default is 1.
14238 Set how many pixels to skip when sampling frames. Defalt is 1.
14239 Allowed range is from 1 to 1024.
14242 Leave frames unchanged. Default is disabled.
14245 @section pixdesctest
14247 Pixel format descriptor test filter, mainly useful for internal
14248 testing. The output video should be equal to the input video.
14252 format=monow, pixdesctest
14255 can be used to test the monowhite pixel format descriptor definition.
14259 Display sample values of color channels. Mainly useful for checking color
14260 and levels. Minimum supported resolution is 640x480.
14262 The filters accept the following options:
14266 Set scope X position, relative offset on X axis.
14269 Set scope Y position, relative offset on Y axis.
14278 Set window opacity. This window also holds statistics about pixel area.
14281 Set window X position, relative offset on X axis.
14284 Set window Y position, relative offset on Y axis.
14289 Enable the specified chain of postprocessing subfilters using libpostproc. This
14290 library should be automatically selected with a GPL build (@code{--enable-gpl}).
14291 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
14292 Each subfilter and some options have a short and a long name that can be used
14293 interchangeably, i.e. dr/dering are the same.
14295 The filters accept the following options:
14299 Set postprocessing subfilters string.
14302 All subfilters share common options to determine their scope:
14306 Honor the quality commands for this subfilter.
14309 Do chrominance filtering, too (default).
14312 Do luminance filtering only (no chrominance).
14315 Do chrominance filtering only (no luminance).
14318 These options can be appended after the subfilter name, separated by a '|'.
14320 Available subfilters are:
14323 @item hb/hdeblock[|difference[|flatness]]
14324 Horizontal deblocking filter
14327 Difference factor where higher values mean more deblocking (default: @code{32}).
14329 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14332 @item vb/vdeblock[|difference[|flatness]]
14333 Vertical deblocking filter
14336 Difference factor where higher values mean more deblocking (default: @code{32}).
14338 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14341 @item ha/hadeblock[|difference[|flatness]]
14342 Accurate horizontal deblocking filter
14345 Difference factor where higher values mean more deblocking (default: @code{32}).
14347 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14350 @item va/vadeblock[|difference[|flatness]]
14351 Accurate vertical deblocking filter
14354 Difference factor where higher values mean more deblocking (default: @code{32}).
14356 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14360 The horizontal and vertical deblocking filters share the difference and
14361 flatness values so you cannot set different horizontal and vertical
14365 @item h1/x1hdeblock
14366 Experimental horizontal deblocking filter
14368 @item v1/x1vdeblock
14369 Experimental vertical deblocking filter
14374 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
14377 larger -> stronger filtering
14379 larger -> stronger filtering
14381 larger -> stronger filtering
14384 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
14387 Stretch luminance to @code{0-255}.
14390 @item lb/linblenddeint
14391 Linear blend deinterlacing filter that deinterlaces the given block by
14392 filtering all lines with a @code{(1 2 1)} filter.
14394 @item li/linipoldeint
14395 Linear interpolating deinterlacing filter that deinterlaces the given block by
14396 linearly interpolating every second line.
14398 @item ci/cubicipoldeint
14399 Cubic interpolating deinterlacing filter deinterlaces the given block by
14400 cubically interpolating every second line.
14402 @item md/mediandeint
14403 Median deinterlacing filter that deinterlaces the given block by applying a
14404 median filter to every second line.
14406 @item fd/ffmpegdeint
14407 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
14408 second line with a @code{(-1 4 2 4 -1)} filter.
14411 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
14412 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
14414 @item fq/forceQuant[|quantizer]
14415 Overrides the quantizer table from the input with the constant quantizer you
14423 Default pp filter combination (@code{hb|a,vb|a,dr|a})
14426 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
14429 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
14432 @subsection Examples
14436 Apply horizontal and vertical deblocking, deringing and automatic
14437 brightness/contrast:
14443 Apply default filters without brightness/contrast correction:
14449 Apply default filters and temporal denoiser:
14451 pp=default/tmpnoise|1|2|3
14455 Apply deblocking on luminance only, and switch vertical deblocking on or off
14456 automatically depending on available CPU time:
14463 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
14464 similar to spp = 6 with 7 point DCT, where only the center sample is
14467 The filter accepts the following options:
14471 Force a constant quantization parameter. It accepts an integer in range
14472 0 to 63. If not set, the filter will use the QP from the video stream
14476 Set thresholding mode. Available modes are:
14480 Set hard thresholding.
14482 Set soft thresholding (better de-ringing effect, but likely blurrier).
14484 Set medium thresholding (good results, default).
14488 @section premultiply
14489 Apply alpha premultiply effect to input video stream using first plane
14490 of second stream as alpha.
14492 Both streams must have same dimensions and same pixel format.
14494 The filter accepts the following option:
14498 Set which planes will be processed, unprocessed planes will be copied.
14499 By default value 0xf, all planes will be processed.
14502 Do not require 2nd input for processing, instead use alpha plane from input stream.
14506 Apply prewitt operator to input video stream.
14508 The filter accepts the following option:
14512 Set which planes will be processed, unprocessed planes will be copied.
14513 By default value 0xf, all planes will be processed.
14516 Set value which will be multiplied with filtered result.
14519 Set value which will be added to filtered result.
14522 @anchor{program_opencl}
14523 @section program_opencl
14525 Filter video using an OpenCL program.
14530 OpenCL program source file.
14533 Kernel name in program.
14536 Number of inputs to the filter. Defaults to 1.
14539 Size of output frames. Defaults to the same as the first input.
14543 The program source file must contain a kernel function with the given name,
14544 which will be run once for each plane of the output. Each run on a plane
14545 gets enqueued as a separate 2D global NDRange with one work-item for each
14546 pixel to be generated. The global ID offset for each work-item is therefore
14547 the coordinates of a pixel in the destination image.
14549 The kernel function needs to take the following arguments:
14552 Destination image, @var{__write_only image2d_t}.
14554 This image will become the output; the kernel should write all of it.
14556 Frame index, @var{unsigned int}.
14558 This is a counter starting from zero and increasing by one for each frame.
14560 Source images, @var{__read_only image2d_t}.
14562 These are the most recent images on each input. The kernel may read from
14563 them to generate the output, but they can't be written to.
14570 Copy the input to the output (output must be the same size as the input).
14572 __kernel void copy(__write_only image2d_t destination,
14573 unsigned int index,
14574 __read_only image2d_t source)
14576 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
14578 int2 location = (int2)(get_global_id(0), get_global_id(1));
14580 float4 value = read_imagef(source, sampler, location);
14582 write_imagef(destination, location, value);
14587 Apply a simple transformation, rotating the input by an amount increasing
14588 with the index counter. Pixel values are linearly interpolated by the
14589 sampler, and the output need not have the same dimensions as the input.
14591 __kernel void rotate_image(__write_only image2d_t dst,
14592 unsigned int index,
14593 __read_only image2d_t src)
14595 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
14596 CLK_FILTER_LINEAR);
14598 float angle = (float)index / 100.0f;
14600 float2 dst_dim = convert_float2(get_image_dim(dst));
14601 float2 src_dim = convert_float2(get_image_dim(src));
14603 float2 dst_cen = dst_dim / 2.0f;
14604 float2 src_cen = src_dim / 2.0f;
14606 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
14608 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
14610 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
14611 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
14613 src_pos = src_pos * src_dim / dst_dim;
14615 float2 src_loc = src_pos + src_cen;
14617 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
14618 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
14619 write_imagef(dst, dst_loc, 0.5f);
14621 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
14626 Blend two inputs together, with the amount of each input used varying
14627 with the index counter.
14629 __kernel void blend_images(__write_only image2d_t dst,
14630 unsigned int index,
14631 __read_only image2d_t src1,
14632 __read_only image2d_t src2)
14634 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
14635 CLK_FILTER_LINEAR);
14637 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
14639 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
14640 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
14641 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
14643 float4 val1 = read_imagef(src1, sampler, src1_loc);
14644 float4 val2 = read_imagef(src2, sampler, src2_loc);
14646 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
14652 @section pseudocolor
14654 Alter frame colors in video with pseudocolors.
14656 This filter accepts the following options:
14660 set pixel first component expression
14663 set pixel second component expression
14666 set pixel third component expression
14669 set pixel fourth component expression, corresponds to the alpha component
14672 set component to use as base for altering colors
14675 Each of them specifies the expression to use for computing the lookup table for
14676 the corresponding pixel component values.
14678 The expressions can contain the following constants and functions:
14683 The input width and height.
14686 The input value for the pixel component.
14688 @item ymin, umin, vmin, amin
14689 The minimum allowed component value.
14691 @item ymax, umax, vmax, amax
14692 The maximum allowed component value.
14695 All expressions default to "val".
14697 @subsection Examples
14701 Change too high luma values to gradient:
14703 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'"
14709 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
14710 Ratio) between two input videos.
14712 This filter takes in input two input videos, the first input is
14713 considered the "main" source and is passed unchanged to the
14714 output. The second input is used as a "reference" video for computing
14717 Both video inputs must have the same resolution and pixel format for
14718 this filter to work correctly. Also it assumes that both inputs
14719 have the same number of frames, which are compared one by one.
14721 The obtained average PSNR is printed through the logging system.
14723 The filter stores the accumulated MSE (mean squared error) of each
14724 frame, and at the end of the processing it is averaged across all frames
14725 equally, and the following formula is applied to obtain the PSNR:
14728 PSNR = 10*log10(MAX^2/MSE)
14731 Where MAX is the average of the maximum values of each component of the
14734 The description of the accepted parameters follows.
14737 @item stats_file, f
14738 If specified the filter will use the named file to save the PSNR of
14739 each individual frame. When filename equals "-" the data is sent to
14742 @item stats_version
14743 Specifies which version of the stats file format to use. Details of
14744 each format are written below.
14745 Default value is 1.
14747 @item stats_add_max
14748 Determines whether the max value is output to the stats log.
14749 Default value is 0.
14750 Requires stats_version >= 2. If this is set and stats_version < 2,
14751 the filter will return an error.
14754 This filter also supports the @ref{framesync} options.
14756 The file printed if @var{stats_file} is selected, contains a sequence of
14757 key/value pairs of the form @var{key}:@var{value} for each compared
14760 If a @var{stats_version} greater than 1 is specified, a header line precedes
14761 the list of per-frame-pair stats, with key value pairs following the frame
14762 format with the following parameters:
14765 @item psnr_log_version
14766 The version of the log file format. Will match @var{stats_version}.
14769 A comma separated list of the per-frame-pair parameters included in
14773 A description of each shown per-frame-pair parameter follows:
14777 sequential number of the input frame, starting from 1
14780 Mean Square Error pixel-by-pixel average difference of the compared
14781 frames, averaged over all the image components.
14783 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
14784 Mean Square Error pixel-by-pixel average difference of the compared
14785 frames for the component specified by the suffix.
14787 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
14788 Peak Signal to Noise ratio of the compared frames for the component
14789 specified by the suffix.
14791 @item max_avg, max_y, max_u, max_v
14792 Maximum allowed value for each channel, and average over all
14798 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
14799 [main][ref] psnr="stats_file=stats.log" [out]
14802 On this example the input file being processed is compared with the
14803 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
14804 is stored in @file{stats.log}.
14809 Pulldown reversal (inverse telecine) filter, capable of handling mixed
14810 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
14813 The pullup filter is designed to take advantage of future context in making
14814 its decisions. This filter is stateless in the sense that it does not lock
14815 onto a pattern to follow, but it instead looks forward to the following
14816 fields in order to identify matches and rebuild progressive frames.
14818 To produce content with an even framerate, insert the fps filter after
14819 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
14820 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
14822 The filter accepts the following options:
14829 These options set the amount of "junk" to ignore at the left, right, top, and
14830 bottom of the image, respectively. Left and right are in units of 8 pixels,
14831 while top and bottom are in units of 2 lines.
14832 The default is 8 pixels on each side.
14835 Set the strict breaks. Setting this option to 1 will reduce the chances of
14836 filter generating an occasional mismatched frame, but it may also cause an
14837 excessive number of frames to be dropped during high motion sequences.
14838 Conversely, setting it to -1 will make filter match fields more easily.
14839 This may help processing of video where there is slight blurring between
14840 the fields, but may also cause there to be interlaced frames in the output.
14841 Default value is @code{0}.
14844 Set the metric plane to use. It accepts the following values:
14850 Use chroma blue plane.
14853 Use chroma red plane.
14856 This option may be set to use chroma plane instead of the default luma plane
14857 for doing filter's computations. This may improve accuracy on very clean
14858 source material, but more likely will decrease accuracy, especially if there
14859 is chroma noise (rainbow effect) or any grayscale video.
14860 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
14861 load and make pullup usable in realtime on slow machines.
14864 For best results (without duplicated frames in the output file) it is
14865 necessary to change the output frame rate. For example, to inverse
14866 telecine NTSC input:
14868 ffmpeg -i input -vf pullup -r 24000/1001 ...
14873 Change video quantization parameters (QP).
14875 The filter accepts the following option:
14879 Set expression for quantization parameter.
14882 The expression is evaluated through the eval API and can contain, among others,
14883 the following constants:
14887 1 if index is not 129, 0 otherwise.
14890 Sequential index starting from -129 to 128.
14893 @subsection Examples
14897 Some equation like:
14905 Flush video frames from internal cache of frames into a random order.
14906 No frame is discarded.
14907 Inspired by @ref{frei0r} nervous filter.
14911 Set size in number of frames of internal cache, in range from @code{2} to
14912 @code{512}. Default is @code{30}.
14915 Set seed for random number generator, must be an integer included between
14916 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
14917 less than @code{0}, the filter will try to use a good random seed on a
14921 @section readeia608
14923 Read closed captioning (EIA-608) information from the top lines of a video frame.
14925 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
14926 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
14927 with EIA-608 data (starting from 0). A description of each metadata value follows:
14930 @item lavfi.readeia608.X.cc
14931 The two bytes stored as EIA-608 data (printed in hexadecimal).
14933 @item lavfi.readeia608.X.line
14934 The number of the line on which the EIA-608 data was identified and read.
14937 This filter accepts the following options:
14941 Set the line to start scanning for EIA-608 data. Default is @code{0}.
14944 Set the line to end scanning for EIA-608 data. Default is @code{29}.
14947 Set minimal acceptable amplitude change for sync codes detection.
14948 Default is @code{0.2}. Allowed range is @code{[0.001 - 1]}.
14951 Set the ratio of width reserved for sync code detection.
14952 Default is @code{0.27}. Allowed range is @code{[0.01 - 0.7]}.
14955 Set the max peaks height difference for sync code detection.
14956 Default is @code{0.1}. Allowed range is @code{[0.0 - 0.5]}.
14959 Set max peaks period difference for sync code detection.
14960 Default is @code{0.1}. Allowed range is @code{[0.0 - 0.5]}.
14963 Set the first two max start code bits differences.
14964 Default is @code{0.02}. Allowed range is @code{[0.0 - 0.5]}.
14967 Set the minimum ratio of bits height compared to 3rd start code bit.
14968 Default is @code{0.75}. Allowed range is @code{[0.01 - 1]}.
14971 Set the white color threshold. Default is @code{0.35}. Allowed range is @code{[0.1 - 1]}.
14974 Set the black color threshold. Default is @code{0.15}. Allowed range is @code{[0.0 - 0.5]}.
14977 Enable checking the parity bit. In the event of a parity error, the filter will output
14978 @code{0x00} for that character. Default is false.
14981 Lowpass lines prior to further processing. Default is disabled.
14984 @subsection Examples
14988 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
14990 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
14996 Read vertical interval timecode (VITC) information from the top lines of a
14999 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
15000 timecode value, if a valid timecode has been detected. Further metadata key
15001 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
15002 timecode data has been found or not.
15004 This filter accepts the following options:
15008 Set the maximum number of lines to scan for VITC data. If the value is set to
15009 @code{-1} the full video frame is scanned. Default is @code{45}.
15012 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
15013 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
15016 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
15017 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
15020 @subsection Examples
15024 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
15025 draw @code{--:--:--:--} as a placeholder:
15027 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
15033 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
15035 Destination pixel at position (X, Y) will be picked from source (x, y) position
15036 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
15037 value for pixel will be used for destination pixel.
15039 Xmap and Ymap input video streams must be of same dimensions. Output video stream
15040 will have Xmap/Ymap video stream dimensions.
15041 Xmap and Ymap input video streams are 16bit depth, single channel.
15045 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
15046 Default is @code{color}.
15049 @section removegrain
15051 The removegrain filter is a spatial denoiser for progressive video.
15055 Set mode for the first plane.
15058 Set mode for the second plane.
15061 Set mode for the third plane.
15064 Set mode for the fourth plane.
15067 Range of mode is from 0 to 24. Description of each mode follows:
15071 Leave input plane unchanged. Default.
15074 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
15077 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
15080 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
15083 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
15084 This is equivalent to a median filter.
15087 Line-sensitive clipping giving the minimal change.
15090 Line-sensitive clipping, intermediate.
15093 Line-sensitive clipping, intermediate.
15096 Line-sensitive clipping, intermediate.
15099 Line-sensitive clipping on a line where the neighbours pixels are the closest.
15102 Replaces the target pixel with the closest neighbour.
15105 [1 2 1] horizontal and vertical kernel blur.
15111 Bob mode, interpolates top field from the line where the neighbours
15112 pixels are the closest.
15115 Bob mode, interpolates bottom field from the line where the neighbours
15116 pixels are the closest.
15119 Bob mode, interpolates top field. Same as 13 but with a more complicated
15120 interpolation formula.
15123 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
15124 interpolation formula.
15127 Clips the pixel with the minimum and maximum of respectively the maximum and
15128 minimum of each pair of opposite neighbour pixels.
15131 Line-sensitive clipping using opposite neighbours whose greatest distance from
15132 the current pixel is minimal.
15135 Replaces the pixel with the average of its 8 neighbours.
15138 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
15141 Clips pixels using the averages of opposite neighbour.
15144 Same as mode 21 but simpler and faster.
15147 Small edge and halo removal, but reputed useless.
15153 @section removelogo
15155 Suppress a TV station logo, using an image file to determine which
15156 pixels comprise the logo. It works by filling in the pixels that
15157 comprise the logo with neighboring pixels.
15159 The filter accepts the following options:
15163 Set the filter bitmap file, which can be any image format supported by
15164 libavformat. The width and height of the image file must match those of the
15165 video stream being processed.
15168 Pixels in the provided bitmap image with a value of zero are not
15169 considered part of the logo, non-zero pixels are considered part of
15170 the logo. If you use white (255) for the logo and black (0) for the
15171 rest, you will be safe. For making the filter bitmap, it is
15172 recommended to take a screen capture of a black frame with the logo
15173 visible, and then using a threshold filter followed by the erode
15174 filter once or twice.
15176 If needed, little splotches can be fixed manually. Remember that if
15177 logo pixels are not covered, the filter quality will be much
15178 reduced. Marking too many pixels as part of the logo does not hurt as
15179 much, but it will increase the amount of blurring needed to cover over
15180 the image and will destroy more information than necessary, and extra
15181 pixels will slow things down on a large logo.
15183 @section repeatfields
15185 This filter uses the repeat_field flag from the Video ES headers and hard repeats
15186 fields based on its value.
15190 Reverse a video clip.
15192 Warning: This filter requires memory to buffer the entire clip, so trimming
15195 @subsection Examples
15199 Take the first 5 seconds of a clip, and reverse it.
15206 Shift R/G/B/A pixels horizontally and/or vertically.
15208 The filter accepts the following options:
15211 Set amount to shift red horizontally.
15213 Set amount to shift red vertically.
15215 Set amount to shift green horizontally.
15217 Set amount to shift green vertically.
15219 Set amount to shift blue horizontally.
15221 Set amount to shift blue vertically.
15223 Set amount to shift alpha horizontally.
15225 Set amount to shift alpha vertically.
15227 Set edge mode, can be @var{smear}, default, or @var{warp}.
15231 Apply roberts cross operator to input video stream.
15233 The filter accepts the following option:
15237 Set which planes will be processed, unprocessed planes will be copied.
15238 By default value 0xf, all planes will be processed.
15241 Set value which will be multiplied with filtered result.
15244 Set value which will be added to filtered result.
15249 Rotate video by an arbitrary angle expressed in radians.
15251 The filter accepts the following options:
15253 A description of the optional parameters follows.
15256 Set an expression for the angle by which to rotate the input video
15257 clockwise, expressed as a number of radians. A negative value will
15258 result in a counter-clockwise rotation. By default it is set to "0".
15260 This expression is evaluated for each frame.
15263 Set the output width expression, default value is "iw".
15264 This expression is evaluated just once during configuration.
15267 Set the output height expression, default value is "ih".
15268 This expression is evaluated just once during configuration.
15271 Enable bilinear interpolation if set to 1, a value of 0 disables
15272 it. Default value is 1.
15275 Set the color used to fill the output area not covered by the rotated
15276 image. For the general syntax of this option, check the
15277 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
15278 If the special value "none" is selected then no
15279 background is printed (useful for example if the background is never shown).
15281 Default value is "black".
15284 The expressions for the angle and the output size can contain the
15285 following constants and functions:
15289 sequential number of the input frame, starting from 0. It is always NAN
15290 before the first frame is filtered.
15293 time in seconds of the input frame, it is set to 0 when the filter is
15294 configured. It is always NAN before the first frame is filtered.
15298 horizontal and vertical chroma subsample values. For example for the
15299 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15303 the input video width and height
15307 the output width and height, that is the size of the padded area as
15308 specified by the @var{width} and @var{height} expressions
15312 the minimal width/height required for completely containing the input
15313 video rotated by @var{a} radians.
15315 These are only available when computing the @option{out_w} and
15316 @option{out_h} expressions.
15319 @subsection Examples
15323 Rotate the input by PI/6 radians clockwise:
15329 Rotate the input by PI/6 radians counter-clockwise:
15335 Rotate the input by 45 degrees clockwise:
15341 Apply a constant rotation with period T, starting from an angle of PI/3:
15343 rotate=PI/3+2*PI*t/T
15347 Make the input video rotation oscillating with a period of T
15348 seconds and an amplitude of A radians:
15350 rotate=A*sin(2*PI/T*t)
15354 Rotate the video, output size is chosen so that the whole rotating
15355 input video is always completely contained in the output:
15357 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
15361 Rotate the video, reduce the output size so that no background is ever
15364 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
15368 @subsection Commands
15370 The filter supports the following commands:
15374 Set the angle expression.
15375 The command accepts the same syntax of the corresponding option.
15377 If the specified expression is not valid, it is kept at its current
15383 Apply Shape Adaptive Blur.
15385 The filter accepts the following options:
15388 @item luma_radius, lr
15389 Set luma blur filter strength, must be a value in range 0.1-4.0, default
15390 value is 1.0. A greater value will result in a more blurred image, and
15391 in slower processing.
15393 @item luma_pre_filter_radius, lpfr
15394 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
15397 @item luma_strength, ls
15398 Set luma maximum difference between pixels to still be considered, must
15399 be a value in the 0.1-100.0 range, default value is 1.0.
15401 @item chroma_radius, cr
15402 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
15403 greater value will result in a more blurred image, and in slower
15406 @item chroma_pre_filter_radius, cpfr
15407 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
15409 @item chroma_strength, cs
15410 Set chroma maximum difference between pixels to still be considered,
15411 must be a value in the -0.9-100.0 range.
15414 Each chroma option value, if not explicitly specified, is set to the
15415 corresponding luma option value.
15420 Scale (resize) the input video, using the libswscale library.
15422 The scale filter forces the output display aspect ratio to be the same
15423 of the input, by changing the output sample aspect ratio.
15425 If the input image format is different from the format requested by
15426 the next filter, the scale filter will convert the input to the
15429 @subsection Options
15430 The filter accepts the following options, or any of the options
15431 supported by the libswscale scaler.
15433 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
15434 the complete list of scaler options.
15439 Set the output video dimension expression. Default value is the input
15442 If the @var{width} or @var{w} value is 0, the input width is used for
15443 the output. If the @var{height} or @var{h} value is 0, the input height
15444 is used for the output.
15446 If one and only one of the values is -n with n >= 1, the scale filter
15447 will use a value that maintains the aspect ratio of the input image,
15448 calculated from the other specified dimension. After that it will,
15449 however, make sure that the calculated dimension is divisible by n and
15450 adjust the value if necessary.
15452 If both values are -n with n >= 1, the behavior will be identical to
15453 both values being set to 0 as previously detailed.
15455 See below for the list of accepted constants for use in the dimension
15459 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
15463 Only evaluate expressions once during the filter initialization or when a command is processed.
15466 Evaluate expressions for each incoming frame.
15470 Default value is @samp{init}.
15474 Set the interlacing mode. It accepts the following values:
15478 Force interlaced aware scaling.
15481 Do not apply interlaced scaling.
15484 Select interlaced aware scaling depending on whether the source frames
15485 are flagged as interlaced or not.
15488 Default value is @samp{0}.
15491 Set libswscale scaling flags. See
15492 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
15493 complete list of values. If not explicitly specified the filter applies
15497 @item param0, param1
15498 Set libswscale input parameters for scaling algorithms that need them. See
15499 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
15500 complete documentation. If not explicitly specified the filter applies
15506 Set the video size. For the syntax of this option, check the
15507 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
15509 @item in_color_matrix
15510 @item out_color_matrix
15511 Set in/output YCbCr color space type.
15513 This allows the autodetected value to be overridden as well as allows forcing
15514 a specific value used for the output and encoder.
15516 If not specified, the color space type depends on the pixel format.
15522 Choose automatically.
15525 Format conforming to International Telecommunication Union (ITU)
15526 Recommendation BT.709.
15529 Set color space conforming to the United States Federal Communications
15530 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
15535 Set color space conforming to:
15539 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
15542 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
15545 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
15550 Set color space conforming to SMPTE ST 240:1999.
15553 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
15558 Set in/output YCbCr sample range.
15560 This allows the autodetected value to be overridden as well as allows forcing
15561 a specific value used for the output and encoder. If not specified, the
15562 range depends on the pixel format. Possible values:
15566 Choose automatically.
15569 Set full range (0-255 in case of 8-bit luma).
15571 @item mpeg/limited/tv
15572 Set "MPEG" range (16-235 in case of 8-bit luma).
15575 @item force_original_aspect_ratio
15576 Enable decreasing or increasing output video width or height if necessary to
15577 keep the original aspect ratio. Possible values:
15581 Scale the video as specified and disable this feature.
15584 The output video dimensions will automatically be decreased if needed.
15587 The output video dimensions will automatically be increased if needed.
15591 One useful instance of this option is that when you know a specific device's
15592 maximum allowed resolution, you can use this to limit the output video to
15593 that, while retaining the aspect ratio. For example, device A allows
15594 1280x720 playback, and your video is 1920x800. Using this option (set it to
15595 decrease) and specifying 1280x720 to the command line makes the output
15598 Please note that this is a different thing than specifying -1 for @option{w}
15599 or @option{h}, you still need to specify the output resolution for this option
15602 @item force_divisible_by
15603 Ensures that both the output dimensions, width and height, are divisible by the
15604 given integer when used together with @option{force_original_aspect_ratio}. This
15605 works similar to using @code{-n} in the @option{w} and @option{h} options.
15607 This option respects the value set for @option{force_original_aspect_ratio},
15608 increasing or decreasing the resolution accordingly. The video's aspect ratio
15609 may be slightly modified.
15611 This option can be handy if you need to have a video fit within or exceed
15612 a defined resolution using @option{force_original_aspect_ratio} but also have
15613 encoder restrictions on width or height divisibility.
15617 The values of the @option{w} and @option{h} options are expressions
15618 containing the following constants:
15623 The input width and height
15627 These are the same as @var{in_w} and @var{in_h}.
15631 The output (scaled) width and height
15635 These are the same as @var{out_w} and @var{out_h}
15638 The same as @var{iw} / @var{ih}
15641 input sample aspect ratio
15644 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
15648 horizontal and vertical input chroma subsample values. For example for the
15649 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15653 horizontal and vertical output chroma subsample values. For example for the
15654 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15657 @subsection Examples
15661 Scale the input video to a size of 200x100
15666 This is equivalent to:
15677 Specify a size abbreviation for the output size:
15682 which can also be written as:
15688 Scale the input to 2x:
15690 scale=w=2*iw:h=2*ih
15694 The above is the same as:
15696 scale=2*in_w:2*in_h
15700 Scale the input to 2x with forced interlaced scaling:
15702 scale=2*iw:2*ih:interl=1
15706 Scale the input to half size:
15708 scale=w=iw/2:h=ih/2
15712 Increase the width, and set the height to the same size:
15718 Seek Greek harmony:
15725 Increase the height, and set the width to 3/2 of the height:
15727 scale=w=3/2*oh:h=3/5*ih
15731 Increase the size, making the size a multiple of the chroma
15734 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
15738 Increase the width to a maximum of 500 pixels,
15739 keeping the same aspect ratio as the input:
15741 scale=w='min(500\, iw*3/2):h=-1'
15745 Make pixels square by combining scale and setsar:
15747 scale='trunc(ih*dar):ih',setsar=1/1
15751 Make pixels square by combining scale and setsar,
15752 making sure the resulting resolution is even (required by some codecs):
15754 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
15758 @subsection Commands
15760 This filter supports the following commands:
15764 Set the output video dimension expression.
15765 The command accepts the same syntax of the corresponding option.
15767 If the specified expression is not valid, it is kept at its current
15773 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
15774 format conversion on CUDA video frames. Setting the output width and height
15775 works in the same way as for the @var{scale} filter.
15777 The following additional options are accepted:
15780 The pixel format of the output CUDA frames. If set to the string "same" (the
15781 default), the input format will be kept. Note that automatic format negotiation
15782 and conversion is not yet supported for hardware frames
15785 The interpolation algorithm used for resizing. One of the following:
15792 @item cubic2p_bspline
15793 2-parameter cubic (B=1, C=0)
15795 @item cubic2p_catmullrom
15796 2-parameter cubic (B=0, C=1/2)
15798 @item cubic2p_b05c03
15799 2-parameter cubic (B=1/2, C=3/10)
15811 Scale (resize) the input video, based on a reference video.
15813 See the scale filter for available options, scale2ref supports the same but
15814 uses the reference video instead of the main input as basis. scale2ref also
15815 supports the following additional constants for the @option{w} and
15816 @option{h} options:
15821 The main input video's width and height
15824 The same as @var{main_w} / @var{main_h}
15827 The main input video's sample aspect ratio
15829 @item main_dar, mdar
15830 The main input video's display aspect ratio. Calculated from
15831 @code{(main_w / main_h) * main_sar}.
15835 The main input video's horizontal and vertical chroma subsample values.
15836 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
15840 @subsection Examples
15844 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
15846 'scale2ref[b][a];[a][b]overlay'
15850 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
15852 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
15857 Scroll input video horizontally and/or vertically by constant speed.
15859 The filter accepts the following options:
15861 @item horizontal, h
15862 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
15863 Negative values changes scrolling direction.
15866 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
15867 Negative values changes scrolling direction.
15870 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
15873 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
15876 @subsection Commands
15878 This filter supports the following @ref{commands}:
15880 @item horizontal, h
15881 Set the horizontal scrolling speed.
15883 Set the vertical scrolling speed.
15886 @anchor{selectivecolor}
15887 @section selectivecolor
15889 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
15890 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
15891 by the "purity" of the color (that is, how saturated it already is).
15893 This filter is similar to the Adobe Photoshop Selective Color tool.
15895 The filter accepts the following options:
15898 @item correction_method
15899 Select color correction method.
15901 Available values are:
15904 Specified adjustments are applied "as-is" (added/subtracted to original pixel
15907 Specified adjustments are relative to the original component value.
15909 Default is @code{absolute}.
15911 Adjustments for red pixels (pixels where the red component is the maximum)
15913 Adjustments for yellow pixels (pixels where the blue component is the minimum)
15915 Adjustments for green pixels (pixels where the green component is the maximum)
15917 Adjustments for cyan pixels (pixels where the red component is the minimum)
15919 Adjustments for blue pixels (pixels where the blue component is the maximum)
15921 Adjustments for magenta pixels (pixels where the green component is the minimum)
15923 Adjustments for white pixels (pixels where all components are greater than 128)
15925 Adjustments for all pixels except pure black and pure white
15927 Adjustments for black pixels (pixels where all components are lesser than 128)
15929 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
15932 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
15933 4 space separated floating point adjustment values in the [-1,1] range,
15934 respectively to adjust the amount of cyan, magenta, yellow and black for the
15935 pixels of its range.
15937 @subsection Examples
15941 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
15942 increase magenta by 27% in blue areas:
15944 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
15948 Use a Photoshop selective color preset:
15950 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
15954 @anchor{separatefields}
15955 @section separatefields
15957 The @code{separatefields} takes a frame-based video input and splits
15958 each frame into its components fields, producing a new half height clip
15959 with twice the frame rate and twice the frame count.
15961 This filter use field-dominance information in frame to decide which
15962 of each pair of fields to place first in the output.
15963 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
15965 @section setdar, setsar
15967 The @code{setdar} filter sets the Display Aspect Ratio for the filter
15970 This is done by changing the specified Sample (aka Pixel) Aspect
15971 Ratio, according to the following equation:
15973 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
15976 Keep in mind that the @code{setdar} filter does not modify the pixel
15977 dimensions of the video frame. Also, the display aspect ratio set by
15978 this filter may be changed by later filters in the filterchain,
15979 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
15982 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
15983 the filter output video.
15985 Note that as a consequence of the application of this filter, the
15986 output display aspect ratio will change according to the equation
15989 Keep in mind that the sample aspect ratio set by the @code{setsar}
15990 filter may be changed by later filters in the filterchain, e.g. if
15991 another "setsar" or a "setdar" filter is applied.
15993 It accepts the following parameters:
15996 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
15997 Set the aspect ratio used by the filter.
15999 The parameter can be a floating point number string, an expression, or
16000 a string of the form @var{num}:@var{den}, where @var{num} and
16001 @var{den} are the numerator and denominator of the aspect ratio. If
16002 the parameter is not specified, it is assumed the value "0".
16003 In case the form "@var{num}:@var{den}" is used, the @code{:} character
16007 Set the maximum integer value to use for expressing numerator and
16008 denominator when reducing the expressed aspect ratio to a rational.
16009 Default value is @code{100}.
16013 The parameter @var{sar} is an expression containing
16014 the following constants:
16018 These are approximated values for the mathematical constants e
16019 (Euler's number), pi (Greek pi), and phi (the golden ratio).
16022 The input width and height.
16025 These are the same as @var{w} / @var{h}.
16028 The input sample aspect ratio.
16031 The input display aspect ratio. It is the same as
16032 (@var{w} / @var{h}) * @var{sar}.
16035 Horizontal and vertical chroma subsample values. For example, for the
16036 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16039 @subsection Examples
16044 To change the display aspect ratio to 16:9, specify one of the following:
16051 To change the sample aspect ratio to 10:11, specify:
16057 To set a display aspect ratio of 16:9, and specify a maximum integer value of
16058 1000 in the aspect ratio reduction, use the command:
16060 setdar=ratio=16/9:max=1000
16068 Force field for the output video frame.
16070 The @code{setfield} filter marks the interlace type field for the
16071 output frames. It does not change the input frame, but only sets the
16072 corresponding property, which affects how the frame is treated by
16073 following filters (e.g. @code{fieldorder} or @code{yadif}).
16075 The filter accepts the following options:
16080 Available values are:
16084 Keep the same field property.
16087 Mark the frame as bottom-field-first.
16090 Mark the frame as top-field-first.
16093 Mark the frame as progressive.
16100 Force frame parameter for the output video frame.
16102 The @code{setparams} filter marks interlace and color range for the
16103 output frames. It does not change the input frame, but only sets the
16104 corresponding property, which affects how the frame is treated by
16109 Available values are:
16113 Keep the same field property (default).
16116 Mark the frame as bottom-field-first.
16119 Mark the frame as top-field-first.
16122 Mark the frame as progressive.
16126 Available values are:
16130 Keep the same color range property (default).
16132 @item unspecified, unknown
16133 Mark the frame as unspecified color range.
16135 @item limited, tv, mpeg
16136 Mark the frame as limited range.
16138 @item full, pc, jpeg
16139 Mark the frame as full range.
16142 @item color_primaries
16143 Set the color primaries.
16144 Available values are:
16148 Keep the same color primaries property (default).
16165 Set the color transfer.
16166 Available values are:
16170 Keep the same color trc property (default).
16192 Set the colorspace.
16193 Available values are:
16197 Keep the same colorspace property (default).
16210 @item chroma-derived-nc
16211 @item chroma-derived-c
16218 Show a line containing various information for each input video frame.
16219 The input video is not modified.
16221 This filter supports the following options:
16225 Calculate checksums of each plane. By default enabled.
16228 The shown line contains a sequence of key/value pairs of the form
16229 @var{key}:@var{value}.
16231 The following values are shown in the output:
16235 The (sequential) number of the input frame, starting from 0.
16238 The Presentation TimeStamp of the input frame, expressed as a number of
16239 time base units. The time base unit depends on the filter input pad.
16242 The Presentation TimeStamp of the input frame, expressed as a number of
16246 The position of the frame in the input stream, or -1 if this information is
16247 unavailable and/or meaningless (for example in case of synthetic video).
16250 The pixel format name.
16253 The sample aspect ratio of the input frame, expressed in the form
16254 @var{num}/@var{den}.
16257 The size of the input frame. For the syntax of this option, check the
16258 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16261 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
16262 for bottom field first).
16265 This is 1 if the frame is a key frame, 0 otherwise.
16268 The picture type of the input frame ("I" for an I-frame, "P" for a
16269 P-frame, "B" for a B-frame, or "?" for an unknown type).
16270 Also refer to the documentation of the @code{AVPictureType} enum and of
16271 the @code{av_get_picture_type_char} function defined in
16272 @file{libavutil/avutil.h}.
16275 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
16277 @item plane_checksum
16278 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
16279 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
16282 @section showpalette
16284 Displays the 256 colors palette of each frame. This filter is only relevant for
16285 @var{pal8} pixel format frames.
16287 It accepts the following option:
16291 Set the size of the box used to represent one palette color entry. Default is
16292 @code{30} (for a @code{30x30} pixel box).
16295 @section shuffleframes
16297 Reorder and/or duplicate and/or drop video frames.
16299 It accepts the following parameters:
16303 Set the destination indexes of input frames.
16304 This is space or '|' separated list of indexes that maps input frames to output
16305 frames. Number of indexes also sets maximal value that each index may have.
16306 '-1' index have special meaning and that is to drop frame.
16309 The first frame has the index 0. The default is to keep the input unchanged.
16311 @subsection Examples
16315 Swap second and third frame of every three frames of the input:
16317 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
16321 Swap 10th and 1st frame of every ten frames of the input:
16323 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
16327 @section shuffleplanes
16329 Reorder and/or duplicate video planes.
16331 It accepts the following parameters:
16336 The index of the input plane to be used as the first output plane.
16339 The index of the input plane to be used as the second output plane.
16342 The index of the input plane to be used as the third output plane.
16345 The index of the input plane to be used as the fourth output plane.
16349 The first plane has the index 0. The default is to keep the input unchanged.
16351 @subsection Examples
16355 Swap the second and third planes of the input:
16357 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
16361 @anchor{signalstats}
16362 @section signalstats
16363 Evaluate various visual metrics that assist in determining issues associated
16364 with the digitization of analog video media.
16366 By default the filter will log these metadata values:
16370 Display the minimal Y value contained within the input frame. Expressed in
16374 Display the Y value at the 10% percentile within the input frame. Expressed in
16378 Display the average Y value within the input frame. Expressed in range of
16382 Display the Y value at the 90% percentile within the input frame. Expressed in
16386 Display the maximum Y value contained within the input frame. Expressed in
16390 Display the minimal U value contained within the input frame. Expressed in
16394 Display the U value at the 10% percentile within the input frame. Expressed in
16398 Display the average U value within the input frame. Expressed in range of
16402 Display the U value at the 90% percentile within the input frame. Expressed in
16406 Display the maximum U value contained within the input frame. Expressed in
16410 Display the minimal V value contained within the input frame. Expressed in
16414 Display the V value at the 10% percentile within the input frame. Expressed in
16418 Display the average V value within the input frame. Expressed in range of
16422 Display the V value at the 90% percentile within the input frame. Expressed in
16426 Display the maximum V value contained within the input frame. Expressed in
16430 Display the minimal saturation value contained within the input frame.
16431 Expressed in range of [0-~181.02].
16434 Display the saturation value at the 10% percentile within the input frame.
16435 Expressed in range of [0-~181.02].
16438 Display the average saturation value within the input frame. Expressed in range
16442 Display the saturation value at the 90% percentile within the input frame.
16443 Expressed in range of [0-~181.02].
16446 Display the maximum saturation value contained within the input frame.
16447 Expressed in range of [0-~181.02].
16450 Display the median value for hue within the input frame. Expressed in range of
16454 Display the average value for hue within the input frame. Expressed in range of
16458 Display the average of sample value difference between all values of the Y
16459 plane in the current frame and corresponding values of the previous input frame.
16460 Expressed in range of [0-255].
16463 Display the average of sample value difference between all values of the U
16464 plane in the current frame and corresponding values of the previous input frame.
16465 Expressed in range of [0-255].
16468 Display the average of sample value difference between all values of the V
16469 plane in the current frame and corresponding values of the previous input frame.
16470 Expressed in range of [0-255].
16473 Display bit depth of Y plane in current frame.
16474 Expressed in range of [0-16].
16477 Display bit depth of U plane in current frame.
16478 Expressed in range of [0-16].
16481 Display bit depth of V plane in current frame.
16482 Expressed in range of [0-16].
16485 The filter accepts the following options:
16491 @option{stat} specify an additional form of image analysis.
16492 @option{out} output video with the specified type of pixel highlighted.
16494 Both options accept the following values:
16498 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
16499 unlike the neighboring pixels of the same field. Examples of temporal outliers
16500 include the results of video dropouts, head clogs, or tape tracking issues.
16503 Identify @var{vertical line repetition}. Vertical line repetition includes
16504 similar rows of pixels within a frame. In born-digital video vertical line
16505 repetition is common, but this pattern is uncommon in video digitized from an
16506 analog source. When it occurs in video that results from the digitization of an
16507 analog source it can indicate concealment from a dropout compensator.
16510 Identify pixels that fall outside of legal broadcast range.
16514 Set the highlight color for the @option{out} option. The default color is
16518 @subsection Examples
16522 Output data of various video metrics:
16524 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
16528 Output specific data about the minimum and maximum values of the Y plane per frame:
16530 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
16534 Playback video while highlighting pixels that are outside of broadcast range in red.
16536 ffplay example.mov -vf signalstats="out=brng:color=red"
16540 Playback video with signalstats metadata drawn over the frame.
16542 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
16545 The contents of signalstat_drawtext.txt used in the command are:
16548 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
16549 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
16550 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
16551 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
16559 Calculates the MPEG-7 Video Signature. The filter can handle more than one
16560 input. In this case the matching between the inputs can be calculated additionally.
16561 The filter always passes through the first input. The signature of each stream can
16562 be written into a file.
16564 It accepts the following options:
16568 Enable or disable the matching process.
16570 Available values are:
16574 Disable the calculation of a matching (default).
16576 Calculate the matching for the whole video and output whether the whole video
16577 matches or only parts.
16579 Calculate only until a matching is found or the video ends. Should be faster in
16584 Set the number of inputs. The option value must be a non negative integer.
16585 Default value is 1.
16588 Set the path to which the output is written. If there is more than one input,
16589 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
16590 integer), that will be replaced with the input number. If no filename is
16591 specified, no output will be written. This is the default.
16594 Choose the output format.
16596 Available values are:
16600 Use the specified binary representation (default).
16602 Use the specified xml representation.
16606 Set threshold to detect one word as similar. The option value must be an integer
16607 greater than zero. The default value is 9000.
16610 Set threshold to detect all words as similar. The option value must be an integer
16611 greater than zero. The default value is 60000.
16614 Set threshold to detect frames as similar. The option value must be an integer
16615 greater than zero. The default value is 116.
16618 Set the minimum length of a sequence in frames to recognize it as matching
16619 sequence. The option value must be a non negative integer value.
16620 The default value is 0.
16623 Set the minimum relation, that matching frames to all frames must have.
16624 The option value must be a double value between 0 and 1. The default value is 0.5.
16627 @subsection Examples
16631 To calculate the signature of an input video and store it in signature.bin:
16633 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
16637 To detect whether two videos match and store the signatures in XML format in
16638 signature0.xml and signature1.xml:
16640 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 -
16648 Blur the input video without impacting the outlines.
16650 It accepts the following options:
16653 @item luma_radius, lr
16654 Set the luma radius. The option value must be a float number in
16655 the range [0.1,5.0] that specifies the variance of the gaussian filter
16656 used to blur the image (slower if larger). Default value is 1.0.
16658 @item luma_strength, ls
16659 Set the luma strength. The option value must be a float number
16660 in the range [-1.0,1.0] that configures the blurring. A value included
16661 in [0.0,1.0] will blur the image whereas a value included in
16662 [-1.0,0.0] will sharpen the image. Default value is 1.0.
16664 @item luma_threshold, lt
16665 Set the luma threshold used as a coefficient to determine
16666 whether a pixel should be blurred or not. The option value must be an
16667 integer in the range [-30,30]. A value of 0 will filter all the image,
16668 a value included in [0,30] will filter flat areas and a value included
16669 in [-30,0] will filter edges. Default value is 0.
16671 @item chroma_radius, cr
16672 Set the chroma radius. The option value must be a float number in
16673 the range [0.1,5.0] that specifies the variance of the gaussian filter
16674 used to blur the image (slower if larger). Default value is @option{luma_radius}.
16676 @item chroma_strength, cs
16677 Set the chroma strength. The option value must be a float number
16678 in the range [-1.0,1.0] that configures the blurring. A value included
16679 in [0.0,1.0] will blur the image whereas a value included in
16680 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
16682 @item chroma_threshold, ct
16683 Set the chroma threshold used as a coefficient to determine
16684 whether a pixel should be blurred or not. The option value must be an
16685 integer in the range [-30,30]. A value of 0 will filter all the image,
16686 a value included in [0,30] will filter flat areas and a value included
16687 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
16690 If a chroma option is not explicitly set, the corresponding luma value
16694 Apply sobel operator to input video stream.
16696 The filter accepts the following option:
16700 Set which planes will be processed, unprocessed planes will be copied.
16701 By default value 0xf, all planes will be processed.
16704 Set value which will be multiplied with filtered result.
16707 Set value which will be added to filtered result.
16713 Apply a simple postprocessing filter that compresses and decompresses the image
16714 at several (or - in the case of @option{quality} level @code{6} - all) shifts
16715 and average the results.
16717 The filter accepts the following options:
16721 Set quality. This option defines the number of levels for averaging. It accepts
16722 an integer in the range 0-6. If set to @code{0}, the filter will have no
16723 effect. A value of @code{6} means the higher quality. For each increment of
16724 that value the speed drops by a factor of approximately 2. Default value is
16728 Force a constant quantization parameter. If not set, the filter will use the QP
16729 from the video stream (if available).
16732 Set thresholding mode. Available modes are:
16736 Set hard thresholding (default).
16738 Set soft thresholding (better de-ringing effect, but likely blurrier).
16741 @item use_bframe_qp
16742 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
16743 option may cause flicker since the B-Frames have often larger QP. Default is
16744 @code{0} (not enabled).
16749 Scale the input by applying one of the super-resolution methods based on
16750 convolutional neural networks. Supported models:
16754 Super-Resolution Convolutional Neural Network model (SRCNN).
16755 See @url{https://arxiv.org/abs/1501.00092}.
16758 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
16759 See @url{https://arxiv.org/abs/1609.05158}.
16762 Training scripts as well as scripts for model file (.pb) saving can be found at
16763 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
16764 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
16766 Native model files (.model) can be generated from TensorFlow model
16767 files (.pb) by using tools/python/convert.py
16769 The filter accepts the following options:
16773 Specify which DNN backend to use for model loading and execution. This option accepts
16774 the following values:
16778 Native implementation of DNN loading and execution.
16781 TensorFlow backend. To enable this backend you
16782 need to install the TensorFlow for C library (see
16783 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
16784 @code{--enable-libtensorflow}
16787 Default value is @samp{native}.
16790 Set path to model file specifying network architecture and its parameters.
16791 Note that different backends use different file formats. TensorFlow backend
16792 can load files for both formats, while native backend can load files for only
16796 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
16797 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
16798 input upscaled using bicubic upscaling with proper scale factor.
16803 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
16805 This filter takes in input two input videos, the first input is
16806 considered the "main" source and is passed unchanged to the
16807 output. The second input is used as a "reference" video for computing
16810 Both video inputs must have the same resolution and pixel format for
16811 this filter to work correctly. Also it assumes that both inputs
16812 have the same number of frames, which are compared one by one.
16814 The filter stores the calculated SSIM of each frame.
16816 The description of the accepted parameters follows.
16819 @item stats_file, f
16820 If specified the filter will use the named file to save the SSIM of
16821 each individual frame. When filename equals "-" the data is sent to
16825 The file printed if @var{stats_file} is selected, contains a sequence of
16826 key/value pairs of the form @var{key}:@var{value} for each compared
16829 A description of each shown parameter follows:
16833 sequential number of the input frame, starting from 1
16835 @item Y, U, V, R, G, B
16836 SSIM of the compared frames for the component specified by the suffix.
16839 SSIM of the compared frames for the whole frame.
16842 Same as above but in dB representation.
16845 This filter also supports the @ref{framesync} options.
16849 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
16850 [main][ref] ssim="stats_file=stats.log" [out]
16853 On this example the input file being processed is compared with the
16854 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
16855 is stored in @file{stats.log}.
16857 Another example with both psnr and ssim at same time:
16859 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
16864 Convert between different stereoscopic image formats.
16866 The filters accept the following options:
16870 Set stereoscopic image format of input.
16872 Available values for input image formats are:
16875 side by side parallel (left eye left, right eye right)
16878 side by side crosseye (right eye left, left eye right)
16881 side by side parallel with half width resolution
16882 (left eye left, right eye right)
16885 side by side crosseye with half width resolution
16886 (right eye left, left eye right)
16890 above-below (left eye above, right eye below)
16894 above-below (right eye above, left eye below)
16898 above-below with half height resolution
16899 (left eye above, right eye below)
16903 above-below with half height resolution
16904 (right eye above, left eye below)
16907 alternating frames (left eye first, right eye second)
16910 alternating frames (right eye first, left eye second)
16913 interleaved rows (left eye has top row, right eye starts on next row)
16916 interleaved rows (right eye has top row, left eye starts on next row)
16919 interleaved columns, left eye first
16922 interleaved columns, right eye first
16924 Default value is @samp{sbsl}.
16928 Set stereoscopic image format of output.
16932 side by side parallel (left eye left, right eye right)
16935 side by side crosseye (right eye left, left eye right)
16938 side by side parallel with half width resolution
16939 (left eye left, right eye right)
16942 side by side crosseye with half width resolution
16943 (right eye left, left eye right)
16947 above-below (left eye above, right eye below)
16951 above-below (right eye above, left eye below)
16955 above-below with half height resolution
16956 (left eye above, right eye below)
16960 above-below with half height resolution
16961 (right eye above, left eye below)
16964 alternating frames (left eye first, right eye second)
16967 alternating frames (right eye first, left eye second)
16970 interleaved rows (left eye has top row, right eye starts on next row)
16973 interleaved rows (right eye has top row, left eye starts on next row)
16976 anaglyph red/blue gray
16977 (red filter on left eye, blue filter on right eye)
16980 anaglyph red/green gray
16981 (red filter on left eye, green filter on right eye)
16984 anaglyph red/cyan gray
16985 (red filter on left eye, cyan filter on right eye)
16988 anaglyph red/cyan half colored
16989 (red filter on left eye, cyan filter on right eye)
16992 anaglyph red/cyan color
16993 (red filter on left eye, cyan filter on right eye)
16996 anaglyph red/cyan color optimized with the least squares projection of dubois
16997 (red filter on left eye, cyan filter on right eye)
17000 anaglyph green/magenta gray
17001 (green filter on left eye, magenta filter on right eye)
17004 anaglyph green/magenta half colored
17005 (green filter on left eye, magenta filter on right eye)
17008 anaglyph green/magenta colored
17009 (green filter on left eye, magenta filter on right eye)
17012 anaglyph green/magenta color optimized with the least squares projection of dubois
17013 (green filter on left eye, magenta filter on right eye)
17016 anaglyph yellow/blue gray
17017 (yellow filter on left eye, blue filter on right eye)
17020 anaglyph yellow/blue half colored
17021 (yellow filter on left eye, blue filter on right eye)
17024 anaglyph yellow/blue colored
17025 (yellow filter on left eye, blue filter on right eye)
17028 anaglyph yellow/blue color optimized with the least squares projection of dubois
17029 (yellow filter on left eye, blue filter on right eye)
17032 mono output (left eye only)
17035 mono output (right eye only)
17038 checkerboard, left eye first
17041 checkerboard, right eye first
17044 interleaved columns, left eye first
17047 interleaved columns, right eye first
17053 Default value is @samp{arcd}.
17056 @subsection Examples
17060 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
17066 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
17072 @section streamselect, astreamselect
17073 Select video or audio streams.
17075 The filter accepts the following options:
17079 Set number of inputs. Default is 2.
17082 Set input indexes to remap to outputs.
17085 @subsection Commands
17087 The @code{streamselect} and @code{astreamselect} filter supports the following
17092 Set input indexes to remap to outputs.
17095 @subsection Examples
17099 Select first 5 seconds 1st stream and rest of time 2nd stream:
17101 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
17105 Same as above, but for audio:
17107 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
17114 Draw subtitles on top of input video using the libass library.
17116 To enable compilation of this filter you need to configure FFmpeg with
17117 @code{--enable-libass}. This filter also requires a build with libavcodec and
17118 libavformat to convert the passed subtitles file to ASS (Advanced Substation
17119 Alpha) subtitles format.
17121 The filter accepts the following options:
17125 Set the filename of the subtitle file to read. It must be specified.
17127 @item original_size
17128 Specify the size of the original video, the video for which the ASS file
17129 was composed. For the syntax of this option, check the
17130 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17131 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
17132 correctly scale the fonts if the aspect ratio has been changed.
17135 Set a directory path containing fonts that can be used by the filter.
17136 These fonts will be used in addition to whatever the font provider uses.
17139 Process alpha channel, by default alpha channel is untouched.
17142 Set subtitles input character encoding. @code{subtitles} filter only. Only
17143 useful if not UTF-8.
17145 @item stream_index, si
17146 Set subtitles stream index. @code{subtitles} filter only.
17149 Override default style or script info parameters of the subtitles. It accepts a
17150 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
17153 If the first key is not specified, it is assumed that the first value
17154 specifies the @option{filename}.
17156 For example, to render the file @file{sub.srt} on top of the input
17157 video, use the command:
17162 which is equivalent to:
17164 subtitles=filename=sub.srt
17167 To render the default subtitles stream from file @file{video.mkv}, use:
17169 subtitles=video.mkv
17172 To render the second subtitles stream from that file, use:
17174 subtitles=video.mkv:si=1
17177 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
17178 @code{DejaVu Serif}, use:
17180 subtitles=sub.srt:force_style='FontName=DejaVu Serif,PrimaryColour=&HCCFF0000'
17183 @section super2xsai
17185 Scale the input by 2x and smooth using the Super2xSaI (Scale and
17186 Interpolate) pixel art scaling algorithm.
17188 Useful for enlarging pixel art images without reducing sharpness.
17192 Swap two rectangular objects in video.
17194 This filter accepts the following options:
17204 Set 1st rect x coordinate.
17207 Set 1st rect y coordinate.
17210 Set 2nd rect x coordinate.
17213 Set 2nd rect y coordinate.
17215 All expressions are evaluated once for each frame.
17218 The all options are expressions containing the following constants:
17223 The input width and height.
17226 same as @var{w} / @var{h}
17229 input sample aspect ratio
17232 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
17235 The number of the input frame, starting from 0.
17238 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
17241 the position in the file of the input frame, NAN if unknown
17249 Apply telecine process to the video.
17251 This filter accepts the following options:
17260 The default value is @code{top}.
17264 A string of numbers representing the pulldown pattern you wish to apply.
17265 The default value is @code{23}.
17269 Some typical patterns:
17274 24p: 2332 (preferred)
17281 24p: 222222222223 ("Euro pulldown")
17288 Apply threshold effect to video stream.
17290 This filter needs four video streams to perform thresholding.
17291 First stream is stream we are filtering.
17292 Second stream is holding threshold values, third stream is holding min values,
17293 and last, fourth stream is holding max values.
17295 The filter accepts the following option:
17299 Set which planes will be processed, unprocessed planes will be copied.
17300 By default value 0xf, all planes will be processed.
17303 For example if first stream pixel's component value is less then threshold value
17304 of pixel component from 2nd threshold stream, third stream value will picked,
17305 otherwise fourth stream pixel component value will be picked.
17307 Using color source filter one can perform various types of thresholding:
17309 @subsection Examples
17313 Binary threshold, using gray color as threshold:
17315 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
17319 Inverted binary threshold, using gray color as threshold:
17321 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
17325 Truncate binary threshold, using gray color as threshold:
17327 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
17331 Threshold to zero, using gray color as threshold:
17333 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
17337 Inverted threshold to zero, using gray color as threshold:
17339 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
17344 Select the most representative frame in a given sequence of consecutive frames.
17346 The filter accepts the following options:
17350 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
17351 will pick one of them, and then handle the next batch of @var{n} frames until
17352 the end. Default is @code{100}.
17355 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
17356 value will result in a higher memory usage, so a high value is not recommended.
17358 @subsection Examples
17362 Extract one picture each 50 frames:
17368 Complete example of a thumbnail creation with @command{ffmpeg}:
17370 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
17376 Tile several successive frames together.
17378 The filter accepts the following options:
17383 Set the grid size (i.e. the number of lines and columns). For the syntax of
17384 this option, check the
17385 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17388 Set the maximum number of frames to render in the given area. It must be less
17389 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
17390 the area will be used.
17393 Set the outer border margin in pixels.
17396 Set the inner border thickness (i.e. the number of pixels between frames). For
17397 more advanced padding options (such as having different values for the edges),
17398 refer to the pad video filter.
17401 Specify the color of the unused area. For the syntax of this option, check the
17402 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
17403 The default value of @var{color} is "black".
17406 Set the number of frames to overlap when tiling several successive frames together.
17407 The value must be between @code{0} and @var{nb_frames - 1}.
17410 Set the number of frames to initially be empty before displaying first output frame.
17411 This controls how soon will one get first output frame.
17412 The value must be between @code{0} and @var{nb_frames - 1}.
17415 @subsection Examples
17419 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
17421 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
17423 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
17424 duplicating each output frame to accommodate the originally detected frame
17428 Display @code{5} pictures in an area of @code{3x2} frames,
17429 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
17430 mixed flat and named options:
17432 tile=3x2:nb_frames=5:padding=7:margin=2
17436 @section tinterlace
17438 Perform various types of temporal field interlacing.
17440 Frames are counted starting from 1, so the first input frame is
17443 The filter accepts the following options:
17448 Specify the mode of the interlacing. This option can also be specified
17449 as a value alone. See below for a list of values for this option.
17451 Available values are:
17455 Move odd frames into the upper field, even into the lower field,
17456 generating a double height frame at half frame rate.
17460 Frame 1 Frame 2 Frame 3 Frame 4
17462 11111 22222 33333 44444
17463 11111 22222 33333 44444
17464 11111 22222 33333 44444
17465 11111 22222 33333 44444
17479 Only output odd frames, even frames are dropped, generating a frame with
17480 unchanged height at half frame rate.
17485 Frame 1 Frame 2 Frame 3 Frame 4
17487 11111 22222 33333 44444
17488 11111 22222 33333 44444
17489 11111 22222 33333 44444
17490 11111 22222 33333 44444
17500 Only output even frames, odd frames are dropped, generating a frame with
17501 unchanged height at half frame rate.
17506 Frame 1 Frame 2 Frame 3 Frame 4
17508 11111 22222 33333 44444
17509 11111 22222 33333 44444
17510 11111 22222 33333 44444
17511 11111 22222 33333 44444
17521 Expand each frame to full height, but pad alternate lines with black,
17522 generating a frame with double height at the same input frame rate.
17527 Frame 1 Frame 2 Frame 3 Frame 4
17529 11111 22222 33333 44444
17530 11111 22222 33333 44444
17531 11111 22222 33333 44444
17532 11111 22222 33333 44444
17535 11111 ..... 33333 .....
17536 ..... 22222 ..... 44444
17537 11111 ..... 33333 .....
17538 ..... 22222 ..... 44444
17539 11111 ..... 33333 .....
17540 ..... 22222 ..... 44444
17541 11111 ..... 33333 .....
17542 ..... 22222 ..... 44444
17546 @item interleave_top, 4
17547 Interleave the upper field from odd frames with the lower field from
17548 even frames, generating a frame with unchanged height at half frame rate.
17553 Frame 1 Frame 2 Frame 3 Frame 4
17555 11111<- 22222 33333<- 44444
17556 11111 22222<- 33333 44444<-
17557 11111<- 22222 33333<- 44444
17558 11111 22222<- 33333 44444<-
17568 @item interleave_bottom, 5
17569 Interleave the lower field from odd frames with the upper field from
17570 even frames, generating a frame with unchanged height at half frame rate.
17575 Frame 1 Frame 2 Frame 3 Frame 4
17577 11111 22222<- 33333 44444<-
17578 11111<- 22222 33333<- 44444
17579 11111 22222<- 33333 44444<-
17580 11111<- 22222 33333<- 44444
17590 @item interlacex2, 6
17591 Double frame rate with unchanged height. Frames are inserted each
17592 containing the second temporal field from the previous input frame and
17593 the first temporal field from the next input frame. This mode relies on
17594 the top_field_first flag. Useful for interlaced video displays with no
17595 field synchronisation.
17600 Frame 1 Frame 2 Frame 3 Frame 4
17602 11111 22222 33333 44444
17603 11111 22222 33333 44444
17604 11111 22222 33333 44444
17605 11111 22222 33333 44444
17608 11111 22222 22222 33333 33333 44444 44444
17609 11111 11111 22222 22222 33333 33333 44444
17610 11111 22222 22222 33333 33333 44444 44444
17611 11111 11111 22222 22222 33333 33333 44444
17616 Move odd frames into the upper field, even into the lower field,
17617 generating a double height frame at same frame rate.
17622 Frame 1 Frame 2 Frame 3 Frame 4
17624 11111 22222 33333 44444
17625 11111 22222 33333 44444
17626 11111 22222 33333 44444
17627 11111 22222 33333 44444
17630 11111 33333 33333 55555
17631 22222 22222 44444 44444
17632 11111 33333 33333 55555
17633 22222 22222 44444 44444
17634 11111 33333 33333 55555
17635 22222 22222 44444 44444
17636 11111 33333 33333 55555
17637 22222 22222 44444 44444
17642 Numeric values are deprecated but are accepted for backward
17643 compatibility reasons.
17645 Default mode is @code{merge}.
17648 Specify flags influencing the filter process.
17650 Available value for @var{flags} is:
17653 @item low_pass_filter, vlpf
17654 Enable linear vertical low-pass filtering in the filter.
17655 Vertical low-pass filtering is required when creating an interlaced
17656 destination from a progressive source which contains high-frequency
17657 vertical detail. Filtering will reduce interlace 'twitter' and Moire
17660 @item complex_filter, cvlpf
17661 Enable complex vertical low-pass filtering.
17662 This will slightly less reduce interlace 'twitter' and Moire
17663 patterning but better retain detail and subjective sharpness impression.
17667 Vertical low-pass filtering can only be enabled for @option{mode}
17668 @var{interleave_top} and @var{interleave_bottom}.
17674 Mix successive video frames.
17676 A description of the accepted options follows.
17680 The number of successive frames to mix. If unspecified, it defaults to 3.
17683 Specify weight of each input video frame.
17684 Each weight is separated by space. If number of weights is smaller than
17685 number of @var{frames} last specified weight will be used for all remaining
17689 Specify scale, if it is set it will be multiplied with sum
17690 of each weight multiplied with pixel values to give final destination
17691 pixel value. By default @var{scale} is auto scaled to sum of weights.
17694 @subsection Examples
17698 Average 7 successive frames:
17700 tmix=frames=7:weights="1 1 1 1 1 1 1"
17704 Apply simple temporal convolution:
17706 tmix=frames=3:weights="-1 3 -1"
17710 Similar as above but only showing temporal differences:
17712 tmix=frames=3:weights="-1 2 -1":scale=1
17718 Tone map colors from different dynamic ranges.
17720 This filter expects data in single precision floating point, as it needs to
17721 operate on (and can output) out-of-range values. Another filter, such as
17722 @ref{zscale}, is needed to convert the resulting frame to a usable format.
17724 The tonemapping algorithms implemented only work on linear light, so input
17725 data should be linearized beforehand (and possibly correctly tagged).
17728 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
17731 @subsection Options
17732 The filter accepts the following options.
17736 Set the tone map algorithm to use.
17738 Possible values are:
17741 Do not apply any tone map, only desaturate overbright pixels.
17744 Hard-clip any out-of-range values. Use it for perfect color accuracy for
17745 in-range values, while distorting out-of-range values.
17748 Stretch the entire reference gamut to a linear multiple of the display.
17751 Fit a logarithmic transfer between the tone curves.
17754 Preserve overall image brightness with a simple curve, using nonlinear
17755 contrast, which results in flattening details and degrading color accuracy.
17758 Preserve both dark and bright details better than @var{reinhard}, at the cost
17759 of slightly darkening everything. Use it when detail preservation is more
17760 important than color and brightness accuracy.
17763 Smoothly map out-of-range values, while retaining contrast and colors for
17764 in-range material as much as possible. Use it when color accuracy is more
17765 important than detail preservation.
17771 Tune the tone mapping algorithm.
17773 This affects the following algorithms:
17779 Specifies the scale factor to use while stretching.
17783 Specifies the exponent of the function.
17787 Specify an extra linear coefficient to multiply into the signal before clipping.
17791 Specify the local contrast coefficient at the display peak.
17792 Default to 0.5, which means that in-gamut values will be about half as bright
17799 Specify the transition point from linear to mobius transform. Every value
17800 below this point is guaranteed to be mapped 1:1. The higher the value, the
17801 more accurate the result will be, at the cost of losing bright details.
17802 Default to 0.3, which due to the steep initial slope still preserves in-range
17803 colors fairly accurately.
17807 Apply desaturation for highlights that exceed this level of brightness. The
17808 higher the parameter, the more color information will be preserved. This
17809 setting helps prevent unnaturally blown-out colors for super-highlights, by
17810 (smoothly) turning into white instead. This makes images feel more natural,
17811 at the cost of reducing information about out-of-range colors.
17813 The default of 2.0 is somewhat conservative and will mostly just apply to
17814 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
17816 This option works only if the input frame has a supported color tag.
17819 Override signal/nominal/reference peak with this value. Useful when the
17820 embedded peak information in display metadata is not reliable or when tone
17821 mapping from a lower range to a higher range.
17826 Temporarily pad video frames.
17828 The filter accepts the following options:
17832 Specify number of delay frames before input video stream.
17835 Specify number of padding frames after input video stream.
17836 Set to -1 to pad indefinitely.
17839 Set kind of frames added to beginning of stream.
17840 Can be either @var{add} or @var{clone}.
17841 With @var{add} frames of solid-color are added.
17842 With @var{clone} frames are clones of first frame.
17845 Set kind of frames added to end of stream.
17846 Can be either @var{add} or @var{clone}.
17847 With @var{add} frames of solid-color are added.
17848 With @var{clone} frames are clones of last frame.
17850 @item start_duration, stop_duration
17851 Specify the duration of the start/stop delay. See
17852 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
17853 for the accepted syntax.
17854 These options override @var{start} and @var{stop}.
17857 Specify the color of the padded area. For the syntax of this option,
17858 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
17859 manual,ffmpeg-utils}.
17861 The default value of @var{color} is "black".
17867 Transpose rows with columns in the input video and optionally flip it.
17869 It accepts the following parameters:
17874 Specify the transposition direction.
17876 Can assume the following values:
17878 @item 0, 4, cclock_flip
17879 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
17887 Rotate by 90 degrees clockwise, that is:
17895 Rotate by 90 degrees counterclockwise, that is:
17902 @item 3, 7, clock_flip
17903 Rotate by 90 degrees clockwise and vertically flip, that is:
17911 For values between 4-7, the transposition is only done if the input
17912 video geometry is portrait and not landscape. These values are
17913 deprecated, the @code{passthrough} option should be used instead.
17915 Numerical values are deprecated, and should be dropped in favor of
17916 symbolic constants.
17919 Do not apply the transposition if the input geometry matches the one
17920 specified by the specified value. It accepts the following values:
17923 Always apply transposition.
17925 Preserve portrait geometry (when @var{height} >= @var{width}).
17927 Preserve landscape geometry (when @var{width} >= @var{height}).
17930 Default value is @code{none}.
17933 For example to rotate by 90 degrees clockwise and preserve portrait
17936 transpose=dir=1:passthrough=portrait
17939 The command above can also be specified as:
17941 transpose=1:portrait
17944 @section transpose_npp
17946 Transpose rows with columns in the input video and optionally flip it.
17947 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
17949 It accepts the following parameters:
17954 Specify the transposition direction.
17956 Can assume the following values:
17959 Rotate by 90 degrees counterclockwise and vertically flip. (default)
17962 Rotate by 90 degrees clockwise.
17965 Rotate by 90 degrees counterclockwise.
17968 Rotate by 90 degrees clockwise and vertically flip.
17972 Do not apply the transposition if the input geometry matches the one
17973 specified by the specified value. It accepts the following values:
17976 Always apply transposition. (default)
17978 Preserve portrait geometry (when @var{height} >= @var{width}).
17980 Preserve landscape geometry (when @var{width} >= @var{height}).
17986 Trim the input so that the output contains one continuous subpart of the input.
17988 It accepts the following parameters:
17991 Specify the time of the start of the kept section, i.e. the frame with the
17992 timestamp @var{start} will be the first frame in the output.
17995 Specify the time of the first frame that will be dropped, i.e. the frame
17996 immediately preceding the one with the timestamp @var{end} will be the last
17997 frame in the output.
18000 This is the same as @var{start}, except this option sets the start timestamp
18001 in timebase units instead of seconds.
18004 This is the same as @var{end}, except this option sets the end timestamp
18005 in timebase units instead of seconds.
18008 The maximum duration of the output in seconds.
18011 The number of the first frame that should be passed to the output.
18014 The number of the first frame that should be dropped.
18017 @option{start}, @option{end}, and @option{duration} are expressed as time
18018 duration specifications; see
18019 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
18020 for the accepted syntax.
18022 Note that the first two sets of the start/end options and the @option{duration}
18023 option look at the frame timestamp, while the _frame variants simply count the
18024 frames that pass through the filter. Also note that this filter does not modify
18025 the timestamps. If you wish for the output timestamps to start at zero, insert a
18026 setpts filter after the trim filter.
18028 If multiple start or end options are set, this filter tries to be greedy and
18029 keep all the frames that match at least one of the specified constraints. To keep
18030 only the part that matches all the constraints at once, chain multiple trim
18033 The defaults are such that all the input is kept. So it is possible to set e.g.
18034 just the end values to keep everything before the specified time.
18039 Drop everything except the second minute of input:
18041 ffmpeg -i INPUT -vf trim=60:120
18045 Keep only the first second:
18047 ffmpeg -i INPUT -vf trim=duration=1
18052 @section unpremultiply
18053 Apply alpha unpremultiply effect to input video stream using first plane
18054 of second stream as alpha.
18056 Both streams must have same dimensions and same pixel format.
18058 The filter accepts the following option:
18062 Set which planes will be processed, unprocessed planes will be copied.
18063 By default value 0xf, all planes will be processed.
18065 If the format has 1 or 2 components, then luma is bit 0.
18066 If the format has 3 or 4 components:
18067 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
18068 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
18069 If present, the alpha channel is always the last bit.
18072 Do not require 2nd input for processing, instead use alpha plane from input stream.
18078 Sharpen or blur the input video.
18080 It accepts the following parameters:
18083 @item luma_msize_x, lx
18084 Set the luma matrix horizontal size. It must be an odd integer between
18085 3 and 23. The default value is 5.
18087 @item luma_msize_y, ly
18088 Set the luma matrix vertical size. It must be an odd integer between 3
18089 and 23. The default value is 5.
18091 @item luma_amount, la
18092 Set the luma effect strength. It must be a floating point number, reasonable
18093 values lay between -1.5 and 1.5.
18095 Negative values will blur the input video, while positive values will
18096 sharpen it, a value of zero will disable the effect.
18098 Default value is 1.0.
18100 @item chroma_msize_x, cx
18101 Set the chroma matrix horizontal size. It must be an odd integer
18102 between 3 and 23. The default value is 5.
18104 @item chroma_msize_y, cy
18105 Set the chroma matrix vertical size. It must be an odd integer
18106 between 3 and 23. The default value is 5.
18108 @item chroma_amount, ca
18109 Set the chroma effect strength. It must be a floating point number, reasonable
18110 values lay between -1.5 and 1.5.
18112 Negative values will blur the input video, while positive values will
18113 sharpen it, a value of zero will disable the effect.
18115 Default value is 0.0.
18119 All parameters are optional and default to the equivalent of the
18120 string '5:5:1.0:5:5:0.0'.
18122 @subsection Examples
18126 Apply strong luma sharpen effect:
18128 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
18132 Apply a strong blur of both luma and chroma parameters:
18134 unsharp=7:7:-2:7:7:-2
18140 Apply ultra slow/simple postprocessing filter that compresses and decompresses
18141 the image at several (or - in the case of @option{quality} level @code{8} - all)
18142 shifts and average the results.
18144 The way this differs from the behavior of spp is that uspp actually encodes &
18145 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
18146 DCT similar to MJPEG.
18148 The filter accepts the following options:
18152 Set quality. This option defines the number of levels for averaging. It accepts
18153 an integer in the range 0-8. If set to @code{0}, the filter will have no
18154 effect. A value of @code{8} means the higher quality. For each increment of
18155 that value the speed drops by a factor of approximately 2. Default value is
18159 Force a constant quantization parameter. If not set, the filter will use the QP
18160 from the video stream (if available).
18165 Convert 360 videos between various formats.
18167 The filter accepts the following options:
18173 Set format of the input/output video.
18181 Equirectangular projection.
18186 Cubemap with 3x2/6x1/1x6 layout.
18188 Format specific options:
18193 Set padding proportion for the input/output cubemap. Values in decimals.
18200 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)
18203 Default value is @b{@samp{0}}.
18207 Set fixed padding for the input/output cubemap. Values in pixels.
18209 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
18213 Set order of faces for the input/output cubemap. Choose one direction for each position.
18215 Designation of directions:
18231 Default value is @b{@samp{rludfb}}.
18235 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
18237 Designation of angles:
18240 0 degrees clockwise
18242 90 degrees clockwise
18244 180 degrees clockwise
18246 270 degrees clockwise
18249 Default value is @b{@samp{000000}}.
18253 Equi-Angular Cubemap.
18258 Regular video. @i{(output only)}
18260 Format specific options:
18265 Set horizontal/vertical/diagonal field of view. Values in degrees.
18267 If diagonal field of view is set it overrides horizontal and vertical field of view.
18273 Format specific options:
18277 Set padding proportion. Values in decimals.
18287 Default value is @b{@samp{0}}.
18292 Facebook's 360 format.
18295 Stereographic format.
18297 Format specific options:
18302 Set horizontal/vertical/diagonal field of view. Values in degrees.
18304 If diagonal field of view is set it overrides horizontal and vertical field of view.
18311 Ball format, gives significant distortion toward the back.
18314 Hammer-Aitoff map projection format.
18317 Sinusoidal map projection format.
18322 Set interpolation method.@*
18323 @i{Note: more complex interpolation methods require much more memory to run.}
18333 Bilinear interpolation.
18336 Bicubic interpolation.
18339 Lanczos interpolation.
18342 Default value is @b{@samp{line}}.
18346 Set the output video resolution.
18348 Default resolution depends on formats.
18352 Set the input/output stereo format.
18363 Default value is @b{@samp{2d}} for input and output format.
18368 Set rotation for the output video. Values in degrees.
18371 Set rotation order for the output video. Choose one item for each position.
18382 Default value is @b{@samp{ypr}}.
18387 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
18391 Set if input video is flipped horizontally/vertically. Boolean values.
18394 Set if input video is transposed. Boolean value, by default disabled.
18397 Set if output video needs to be transposed. Boolean value, by default disabled.
18401 @subsection Examples
18405 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
18407 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
18410 Extract back view of Equi-Angular Cubemap:
18412 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
18415 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
18417 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
18421 @section vaguedenoiser
18423 Apply a wavelet based denoiser.
18425 It transforms each frame from the video input into the wavelet domain,
18426 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
18427 the obtained coefficients. It does an inverse wavelet transform after.
18428 Due to wavelet properties, it should give a nice smoothed result, and
18429 reduced noise, without blurring picture features.
18431 This filter accepts the following options:
18435 The filtering strength. The higher, the more filtered the video will be.
18436 Hard thresholding can use a higher threshold than soft thresholding
18437 before the video looks overfiltered. Default value is 2.
18440 The filtering method the filter will use.
18442 It accepts the following values:
18445 All values under the threshold will be zeroed.
18448 All values under the threshold will be zeroed. All values above will be
18449 reduced by the threshold.
18452 Scales or nullifies coefficients - intermediary between (more) soft and
18453 (less) hard thresholding.
18456 Default is garrote.
18459 Number of times, the wavelet will decompose the picture. Picture can't
18460 be decomposed beyond a particular point (typically, 8 for a 640x480
18461 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
18464 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
18467 A list of the planes to process. By default all planes are processed.
18470 @section vectorscope
18472 Display 2 color component values in the two dimensional graph (which is called
18475 This filter accepts the following options:
18479 Set vectorscope mode.
18481 It accepts the following values:
18484 Gray values are displayed on graph, higher brightness means more pixels have
18485 same component color value on location in graph. This is the default mode.
18488 Gray values are displayed on graph. Surrounding pixels values which are not
18489 present in video frame are drawn in gradient of 2 color components which are
18490 set by option @code{x} and @code{y}. The 3rd color component is static.
18493 Actual color components values present in video frame are displayed on graph.
18496 Similar as color2 but higher frequency of same values @code{x} and @code{y}
18497 on graph increases value of another color component, which is luminance by
18498 default values of @code{x} and @code{y}.
18501 Actual colors present in video frame are displayed on graph. If two different
18502 colors map to same position on graph then color with higher value of component
18503 not present in graph is picked.
18506 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
18507 component picked from radial gradient.
18511 Set which color component will be represented on X-axis. Default is @code{1}.
18514 Set which color component will be represented on Y-axis. Default is @code{2}.
18517 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
18518 of color component which represents frequency of (X, Y) location in graph.
18523 No envelope, this is default.
18526 Instant envelope, even darkest single pixel will be clearly highlighted.
18529 Hold maximum and minimum values presented in graph over time. This way you
18530 can still spot out of range values without constantly looking at vectorscope.
18533 Peak and instant envelope combined together.
18537 Set what kind of graticule to draw.
18545 Set graticule opacity.
18548 Set graticule flags.
18552 Draw graticule for white point.
18555 Draw graticule for black point.
18558 Draw color points short names.
18562 Set background opacity.
18564 @item lthreshold, l
18565 Set low threshold for color component not represented on X or Y axis.
18566 Values lower than this value will be ignored. Default is 0.
18567 Note this value is multiplied with actual max possible value one pixel component
18568 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
18571 @item hthreshold, h
18572 Set high threshold for color component not represented on X or Y axis.
18573 Values higher than this value will be ignored. Default is 1.
18574 Note this value is multiplied with actual max possible value one pixel component
18575 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
18576 is 0.9 * 255 = 230.
18578 @item colorspace, c
18579 Set what kind of colorspace to use when drawing graticule.
18588 @anchor{vidstabdetect}
18589 @section vidstabdetect
18591 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
18592 @ref{vidstabtransform} for pass 2.
18594 This filter generates a file with relative translation and rotation
18595 transform information about subsequent frames, which is then used by
18596 the @ref{vidstabtransform} filter.
18598 To enable compilation of this filter you need to configure FFmpeg with
18599 @code{--enable-libvidstab}.
18601 This filter accepts the following options:
18605 Set the path to the file used to write the transforms information.
18606 Default value is @file{transforms.trf}.
18609 Set how shaky the video is and how quick the camera is. It accepts an
18610 integer in the range 1-10, a value of 1 means little shakiness, a
18611 value of 10 means strong shakiness. Default value is 5.
18614 Set the accuracy of the detection process. It must be a value in the
18615 range 1-15. A value of 1 means low accuracy, a value of 15 means high
18616 accuracy. Default value is 15.
18619 Set stepsize of the search process. The region around minimum is
18620 scanned with 1 pixel resolution. Default value is 6.
18623 Set minimum contrast. Below this value a local measurement field is
18624 discarded. Must be a floating point value in the range 0-1. Default
18628 Set reference frame number for tripod mode.
18630 If enabled, the motion of the frames is compared to a reference frame
18631 in the filtered stream, identified by the specified number. The idea
18632 is to compensate all movements in a more-or-less static scene and keep
18633 the camera view absolutely still.
18635 If set to 0, it is disabled. The frames are counted starting from 1.
18638 Show fields and transforms in the resulting frames. It accepts an
18639 integer in the range 0-2. Default value is 0, which disables any
18643 @subsection Examples
18647 Use default values:
18653 Analyze strongly shaky movie and put the results in file
18654 @file{mytransforms.trf}:
18656 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
18660 Visualize the result of internal transformations in the resulting
18663 vidstabdetect=show=1
18667 Analyze a video with medium shakiness using @command{ffmpeg}:
18669 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
18673 @anchor{vidstabtransform}
18674 @section vidstabtransform
18676 Video stabilization/deshaking: pass 2 of 2,
18677 see @ref{vidstabdetect} for pass 1.
18679 Read a file with transform information for each frame and
18680 apply/compensate them. Together with the @ref{vidstabdetect}
18681 filter this can be used to deshake videos. See also
18682 @url{http://public.hronopik.de/vid.stab}. It is important to also use
18683 the @ref{unsharp} filter, see below.
18685 To enable compilation of this filter you need to configure FFmpeg with
18686 @code{--enable-libvidstab}.
18688 @subsection Options
18692 Set path to the file used to read the transforms. Default value is
18693 @file{transforms.trf}.
18696 Set the number of frames (value*2 + 1) used for lowpass filtering the
18697 camera movements. Default value is 10.
18699 For example a number of 10 means that 21 frames are used (10 in the
18700 past and 10 in the future) to smoothen the motion in the video. A
18701 larger value leads to a smoother video, but limits the acceleration of
18702 the camera (pan/tilt movements). 0 is a special case where a static
18703 camera is simulated.
18706 Set the camera path optimization algorithm.
18708 Accepted values are:
18711 gaussian kernel low-pass filter on camera motion (default)
18713 averaging on transformations
18717 Set maximal number of pixels to translate frames. Default value is -1,
18721 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
18722 value is -1, meaning no limit.
18725 Specify how to deal with borders that may be visible due to movement
18728 Available values are:
18731 keep image information from previous frame (default)
18733 fill the border black
18737 Invert transforms if set to 1. Default value is 0.
18740 Consider transforms as relative to previous frame if set to 1,
18741 absolute if set to 0. Default value is 0.
18744 Set percentage to zoom. A positive value will result in a zoom-in
18745 effect, a negative value in a zoom-out effect. Default value is 0 (no
18749 Set optimal zooming to avoid borders.
18751 Accepted values are:
18756 optimal static zoom value is determined (only very strong movements
18757 will lead to visible borders) (default)
18759 optimal adaptive zoom value is determined (no borders will be
18760 visible), see @option{zoomspeed}
18763 Note that the value given at zoom is added to the one calculated here.
18766 Set percent to zoom maximally each frame (enabled when
18767 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
18771 Specify type of interpolation.
18773 Available values are:
18778 linear only horizontal
18780 linear in both directions (default)
18782 cubic in both directions (slow)
18786 Enable virtual tripod mode if set to 1, which is equivalent to
18787 @code{relative=0:smoothing=0}. Default value is 0.
18789 Use also @code{tripod} option of @ref{vidstabdetect}.
18792 Increase log verbosity if set to 1. Also the detected global motions
18793 are written to the temporary file @file{global_motions.trf}. Default
18797 @subsection Examples
18801 Use @command{ffmpeg} for a typical stabilization with default values:
18803 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
18806 Note the use of the @ref{unsharp} filter which is always recommended.
18809 Zoom in a bit more and load transform data from a given file:
18811 vidstabtransform=zoom=5:input="mytransforms.trf"
18815 Smoothen the video even more:
18817 vidstabtransform=smoothing=30
18823 Flip the input video vertically.
18825 For example, to vertically flip a video with @command{ffmpeg}:
18827 ffmpeg -i in.avi -vf "vflip" out.avi
18832 Detect variable frame rate video.
18834 This filter tries to detect if the input is variable or constant frame rate.
18836 At end it will output number of frames detected as having variable delta pts,
18837 and ones with constant delta pts.
18838 If there was frames with variable delta, than it will also show min and max delta
18843 Boost or alter saturation.
18845 The filter accepts the following options:
18848 Set strength of boost if positive value or strength of alter if negative value.
18849 Default is 0. Allowed range is from -2 to 2.
18852 Set the red balance. Default is 1. Allowed range is from -10 to 10.
18855 Set the green balance. Default is 1. Allowed range is from -10 to 10.
18858 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
18861 Set the red luma coefficient.
18864 Set the green luma coefficient.
18867 Set the blue luma coefficient.
18870 If @code{intensity} is negative and this is set to 1, colors will change,
18871 otherwise colors will be less saturated, more towards gray.
18877 Make or reverse a natural vignetting effect.
18879 The filter accepts the following options:
18883 Set lens angle expression as a number of radians.
18885 The value is clipped in the @code{[0,PI/2]} range.
18887 Default value: @code{"PI/5"}
18891 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
18895 Set forward/backward mode.
18897 Available modes are:
18900 The larger the distance from the central point, the darker the image becomes.
18903 The larger the distance from the central point, the brighter the image becomes.
18904 This can be used to reverse a vignette effect, though there is no automatic
18905 detection to extract the lens @option{angle} and other settings (yet). It can
18906 also be used to create a burning effect.
18909 Default value is @samp{forward}.
18912 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
18914 It accepts the following values:
18917 Evaluate expressions only once during the filter initialization.
18920 Evaluate expressions for each incoming frame. This is way slower than the
18921 @samp{init} mode since it requires all the scalers to be re-computed, but it
18922 allows advanced dynamic expressions.
18925 Default value is @samp{init}.
18928 Set dithering to reduce the circular banding effects. Default is @code{1}
18932 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
18933 Setting this value to the SAR of the input will make a rectangular vignetting
18934 following the dimensions of the video.
18936 Default is @code{1/1}.
18939 @subsection Expressions
18941 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
18942 following parameters.
18947 input width and height
18950 the number of input frame, starting from 0
18953 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
18954 @var{TB} units, NAN if undefined
18957 frame rate of the input video, NAN if the input frame rate is unknown
18960 the PTS (Presentation TimeStamp) of the filtered video frame,
18961 expressed in seconds, NAN if undefined
18964 time base of the input video
18968 @subsection Examples
18972 Apply simple strong vignetting effect:
18978 Make a flickering vignetting:
18980 vignette='PI/4+random(1)*PI/50':eval=frame
18985 @section vmafmotion
18987 Obtain the average vmaf motion score of a video.
18988 It is one of the component filters of VMAF.
18990 The obtained average motion score is printed through the logging system.
18992 In the below example the input file @file{ref.mpg} is being processed and score
18996 ffmpeg -i ref.mpg -lavfi vmafmotion -f null -
19000 Stack input videos vertically.
19002 All streams must be of same pixel format and of same width.
19004 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
19005 to create same output.
19007 The filter accepts the following options:
19011 Set number of input streams. Default is 2.
19014 If set to 1, force the output to terminate when the shortest input
19015 terminates. Default value is 0.
19020 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
19021 Deinterlacing Filter").
19023 Based on the process described by Martin Weston for BBC R&D, and
19024 implemented based on the de-interlace algorithm written by Jim
19025 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
19026 uses filter coefficients calculated by BBC R&D.
19028 This filter uses field-dominance information in frame to decide which
19029 of each pair of fields to place first in the output.
19030 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
19032 There are two sets of filter coefficients, so called "simple"
19033 and "complex". Which set of filter coefficients is used can
19034 be set by passing an optional parameter:
19038 Set the interlacing filter coefficients. Accepts one of the following values:
19042 Simple filter coefficient set.
19044 More-complex filter coefficient set.
19046 Default value is @samp{complex}.
19049 Specify which frames to deinterlace. Accepts one of the following values:
19053 Deinterlace all frames,
19055 Only deinterlace frames marked as interlaced.
19058 Default value is @samp{all}.
19062 Video waveform monitor.
19064 The waveform monitor plots color component intensity. By default luminance
19065 only. Each column of the waveform corresponds to a column of pixels in the
19068 It accepts the following options:
19072 Can be either @code{row}, or @code{column}. Default is @code{column}.
19073 In row mode, the graph on the left side represents color component value 0 and
19074 the right side represents value = 255. In column mode, the top side represents
19075 color component value = 0 and bottom side represents value = 255.
19078 Set intensity. Smaller values are useful to find out how many values of the same
19079 luminance are distributed across input rows/columns.
19080 Default value is @code{0.04}. Allowed range is [0, 1].
19083 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
19084 In mirrored mode, higher values will be represented on the left
19085 side for @code{row} mode and at the top for @code{column} mode. Default is
19086 @code{1} (mirrored).
19090 It accepts the following values:
19093 Presents information identical to that in the @code{parade}, except
19094 that the graphs representing color components are superimposed directly
19097 This display mode makes it easier to spot relative differences or similarities
19098 in overlapping areas of the color components that are supposed to be identical,
19099 such as neutral whites, grays, or blacks.
19102 Display separate graph for the color components side by side in
19103 @code{row} mode or one below the other in @code{column} mode.
19106 Display separate graph for the color components side by side in
19107 @code{column} mode or one below the other in @code{row} mode.
19109 Using this display mode makes it easy to spot color casts in the highlights
19110 and shadows of an image, by comparing the contours of the top and the bottom
19111 graphs of each waveform. Since whites, grays, and blacks are characterized
19112 by exactly equal amounts of red, green, and blue, neutral areas of the picture
19113 should display three waveforms of roughly equal width/height. If not, the
19114 correction is easy to perform by making level adjustments the three waveforms.
19116 Default is @code{stack}.
19118 @item components, c
19119 Set which color components to display. Default is 1, which means only luminance
19120 or red color component if input is in RGB colorspace. If is set for example to
19121 7 it will display all 3 (if) available color components.
19126 No envelope, this is default.
19129 Instant envelope, minimum and maximum values presented in graph will be easily
19130 visible even with small @code{step} value.
19133 Hold minimum and maximum values presented in graph across time. This way you
19134 can still spot out of range values without constantly looking at waveforms.
19137 Peak and instant envelope combined together.
19143 No filtering, this is default.
19146 Luma and chroma combined together.
19149 Similar as above, but shows difference between blue and red chroma.
19152 Similar as above, but use different colors.
19155 Similar as above, but again with different colors.
19158 Displays only chroma.
19161 Displays actual color value on waveform.
19164 Similar as above, but with luma showing frequency of chroma values.
19168 Set which graticule to display.
19172 Do not display graticule.
19175 Display green graticule showing legal broadcast ranges.
19178 Display orange graticule showing legal broadcast ranges.
19181 Display invert graticule showing legal broadcast ranges.
19185 Set graticule opacity.
19188 Set graticule flags.
19192 Draw numbers above lines. By default enabled.
19195 Draw dots instead of lines.
19199 Set scale used for displaying graticule.
19206 Default is digital.
19209 Set background opacity.
19212 @section weave, doubleweave
19214 The @code{weave} takes a field-based video input and join
19215 each two sequential fields into single frame, producing a new double
19216 height clip with half the frame rate and half the frame count.
19218 The @code{doubleweave} works same as @code{weave} but without
19219 halving frame rate and frame count.
19221 It accepts the following option:
19225 Set first field. Available values are:
19229 Set the frame as top-field-first.
19232 Set the frame as bottom-field-first.
19236 @subsection Examples
19240 Interlace video using @ref{select} and @ref{separatefields} filter:
19242 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
19247 Apply the xBR high-quality magnification filter which is designed for pixel
19248 art. It follows a set of edge-detection rules, see
19249 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
19251 It accepts the following option:
19255 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
19256 @code{3xBR} and @code{4} for @code{4xBR}.
19257 Default is @code{3}.
19261 Pick median pixels from several input videos.
19263 The filter accepts the following options:
19267 Set number of inputs.
19268 Default is 3. Allowed range is from 3 to 255.
19269 If number of inputs is even number, than result will be mean value between two median values.
19272 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
19276 Stack video inputs into custom layout.
19278 All streams must be of same pixel format.
19280 The filter accepts the following options:
19284 Set number of input streams. Default is 2.
19287 Specify layout of inputs.
19288 This option requires the desired layout configuration to be explicitly set by the user.
19289 This sets position of each video input in output. Each input
19290 is separated by '|'.
19291 The first number represents the column, and the second number represents the row.
19292 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
19293 where X is video input from which to take width or height.
19294 Multiple values can be used when separated by '+'. In such
19295 case values are summed together.
19297 Note that if inputs are of different sizes gaps may appear, as not all of
19298 the output video frame will be filled. Similarly, videos can overlap each
19299 other if their position doesn't leave enough space for the full frame of
19302 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
19303 a layout must be set by the user.
19306 If set to 1, force the output to terminate when the shortest input
19307 terminates. Default value is 0.
19310 @subsection Examples
19314 Display 4 inputs into 2x2 grid.
19318 input1(0, 0) | input3(w0, 0)
19319 input2(0, h0) | input4(w0, h0)
19323 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
19326 Note that if inputs are of different sizes, gaps or overlaps may occur.
19329 Display 4 inputs into 1x4 grid.
19336 input4(0, h0+h1+h2)
19340 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
19343 Note that if inputs are of different widths, unused space will appear.
19346 Display 9 inputs into 3x3 grid.
19350 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
19351 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
19352 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
19356 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
19359 Note that if inputs are of different sizes, gaps or overlaps may occur.
19362 Display 16 inputs into 4x4 grid.
19366 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
19367 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
19368 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
19369 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
19373 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|
19374 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
19377 Note that if inputs are of different sizes, gaps or overlaps may occur.
19384 Deinterlace the input video ("yadif" means "yet another deinterlacing
19387 It accepts the following parameters:
19393 The interlacing mode to adopt. It accepts one of the following values:
19396 @item 0, send_frame
19397 Output one frame for each frame.
19398 @item 1, send_field
19399 Output one frame for each field.
19400 @item 2, send_frame_nospatial
19401 Like @code{send_frame}, but it skips the spatial interlacing check.
19402 @item 3, send_field_nospatial
19403 Like @code{send_field}, but it skips the spatial interlacing check.
19406 The default value is @code{send_frame}.
19409 The picture field parity assumed for the input interlaced video. It accepts one
19410 of the following values:
19414 Assume the top field is first.
19416 Assume the bottom field is first.
19418 Enable automatic detection of field parity.
19421 The default value is @code{auto}.
19422 If the interlacing is unknown or the decoder does not export this information,
19423 top field first will be assumed.
19426 Specify which frames to deinterlace. Accepts one of the following
19431 Deinterlace all frames.
19432 @item 1, interlaced
19433 Only deinterlace frames marked as interlaced.
19436 The default value is @code{all}.
19439 @section yadif_cuda
19441 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
19442 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
19445 It accepts the following parameters:
19451 The interlacing mode to adopt. It accepts one of the following values:
19454 @item 0, send_frame
19455 Output one frame for each frame.
19456 @item 1, send_field
19457 Output one frame for each field.
19458 @item 2, send_frame_nospatial
19459 Like @code{send_frame}, but it skips the spatial interlacing check.
19460 @item 3, send_field_nospatial
19461 Like @code{send_field}, but it skips the spatial interlacing check.
19464 The default value is @code{send_frame}.
19467 The picture field parity assumed for the input interlaced video. It accepts one
19468 of the following values:
19472 Assume the top field is first.
19474 Assume the bottom field is first.
19476 Enable automatic detection of field parity.
19479 The default value is @code{auto}.
19480 If the interlacing is unknown or the decoder does not export this information,
19481 top field first will be assumed.
19484 Specify which frames to deinterlace. Accepts one of the following
19489 Deinterlace all frames.
19490 @item 1, interlaced
19491 Only deinterlace frames marked as interlaced.
19494 The default value is @code{all}.
19499 Apply Zoom & Pan effect.
19501 This filter accepts the following options:
19505 Set the zoom expression. Range is 1-10. Default is 1.
19509 Set the x and y expression. Default is 0.
19512 Set the duration expression in number of frames.
19513 This sets for how many number of frames effect will last for
19514 single input image.
19517 Set the output image size, default is 'hd720'.
19520 Set the output frame rate, default is '25'.
19523 Each expression can contain the following constants:
19542 Output frame count.
19546 Last calculated 'x' and 'y' position from 'x' and 'y' expression
19547 for current input frame.
19551 'x' and 'y' of last output frame of previous input frame or 0 when there was
19552 not yet such frame (first input frame).
19555 Last calculated zoom from 'z' expression for current input frame.
19558 Last calculated zoom of last output frame of previous input frame.
19561 Number of output frames for current input frame. Calculated from 'd' expression
19562 for each input frame.
19565 number of output frames created for previous input frame
19568 Rational number: input width / input height
19571 sample aspect ratio
19574 display aspect ratio
19578 @subsection Examples
19582 Zoom-in up to 1.5 and pan at same time to some spot near center of picture:
19584 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
19588 Zoom-in up to 1.5 and pan always at center of picture:
19590 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
19594 Same as above but without pausing:
19596 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
19602 Scale (resize) the input video, using the z.lib library:
19603 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
19604 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
19606 The zscale filter forces the output display aspect ratio to be the same
19607 as the input, by changing the output sample aspect ratio.
19609 If the input image format is different from the format requested by
19610 the next filter, the zscale filter will convert the input to the
19613 @subsection Options
19614 The filter accepts the following options.
19619 Set the output video dimension expression. Default value is the input
19622 If the @var{width} or @var{w} value is 0, the input width is used for
19623 the output. If the @var{height} or @var{h} value is 0, the input height
19624 is used for the output.
19626 If one and only one of the values is -n with n >= 1, the zscale filter
19627 will use a value that maintains the aspect ratio of the input image,
19628 calculated from the other specified dimension. After that it will,
19629 however, make sure that the calculated dimension is divisible by n and
19630 adjust the value if necessary.
19632 If both values are -n with n >= 1, the behavior will be identical to
19633 both values being set to 0 as previously detailed.
19635 See below for the list of accepted constants for use in the dimension
19639 Set the video size. For the syntax of this option, check the
19640 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19643 Set the dither type.
19645 Possible values are:
19650 @item error_diffusion
19656 Set the resize filter type.
19658 Possible values are:
19668 Default is bilinear.
19671 Set the color range.
19673 Possible values are:
19680 Default is same as input.
19683 Set the color primaries.
19685 Possible values are:
19695 Default is same as input.
19698 Set the transfer characteristics.
19700 Possible values are:
19714 Default is same as input.
19717 Set the colorspace matrix.
19719 Possible value are:
19730 Default is same as input.
19733 Set the input color range.
19735 Possible values are:
19742 Default is same as input.
19744 @item primariesin, pin
19745 Set the input color primaries.
19747 Possible values are:
19757 Default is same as input.
19759 @item transferin, tin
19760 Set the input transfer characteristics.
19762 Possible values are:
19773 Default is same as input.
19775 @item matrixin, min
19776 Set the input colorspace matrix.
19778 Possible value are:
19790 Set the output chroma location.
19792 Possible values are:
19803 @item chromalin, cin
19804 Set the input chroma location.
19806 Possible values are:
19818 Set the nominal peak luminance.
19821 The values of the @option{w} and @option{h} options are expressions
19822 containing the following constants:
19827 The input width and height
19831 These are the same as @var{in_w} and @var{in_h}.
19835 The output (scaled) width and height
19839 These are the same as @var{out_w} and @var{out_h}
19842 The same as @var{iw} / @var{ih}
19845 input sample aspect ratio
19848 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
19852 horizontal and vertical input chroma subsample values. For example for the
19853 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
19857 horizontal and vertical output chroma subsample values. For example for the
19858 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
19864 @c man end VIDEO FILTERS
19866 @chapter OpenCL Video Filters
19867 @c man begin OPENCL VIDEO FILTERS
19869 Below is a description of the currently available OpenCL video filters.
19871 To enable compilation of these filters you need to configure FFmpeg with
19872 @code{--enable-opencl}.
19874 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
19877 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
19878 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
19879 given device parameters.
19881 @item -filter_hw_device @var{name}
19882 Pass the hardware device called @var{name} to all filters in any filter graph.
19886 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
19890 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
19892 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
19896 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.
19898 @section avgblur_opencl
19900 Apply average blur filter.
19902 The filter accepts the following options:
19906 Set horizontal radius size.
19907 Range is @code{[1, 1024]} and default value is @code{1}.
19910 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
19913 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
19916 @subsection Example
19920 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.
19922 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
19926 @section boxblur_opencl
19928 Apply a boxblur algorithm to the input video.
19930 It accepts the following parameters:
19934 @item luma_radius, lr
19935 @item luma_power, lp
19936 @item chroma_radius, cr
19937 @item chroma_power, cp
19938 @item alpha_radius, ar
19939 @item alpha_power, ap
19943 A description of the accepted options follows.
19946 @item luma_radius, lr
19947 @item chroma_radius, cr
19948 @item alpha_radius, ar
19949 Set an expression for the box radius in pixels used for blurring the
19950 corresponding input plane.
19952 The radius value must be a non-negative number, and must not be
19953 greater than the value of the expression @code{min(w,h)/2} for the
19954 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
19957 Default value for @option{luma_radius} is "2". If not specified,
19958 @option{chroma_radius} and @option{alpha_radius} default to the
19959 corresponding value set for @option{luma_radius}.
19961 The expressions can contain the following constants:
19965 The input width and height in pixels.
19969 The input chroma image width and height in pixels.
19973 The horizontal and vertical chroma subsample values. For example, for the
19974 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
19977 @item luma_power, lp
19978 @item chroma_power, cp
19979 @item alpha_power, ap
19980 Specify how many times the boxblur filter is applied to the
19981 corresponding plane.
19983 Default value for @option{luma_power} is 2. If not specified,
19984 @option{chroma_power} and @option{alpha_power} default to the
19985 corresponding value set for @option{luma_power}.
19987 A value of 0 will disable the effect.
19990 @subsection Examples
19992 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.
19996 Apply a boxblur filter with the luma, chroma, and alpha radius
19997 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.
19999 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
20000 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
20004 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.
20006 For the luma plane, a 2x2 box radius will be run once.
20008 For the chroma plane, a 4x4 box radius will be run 5 times.
20010 For the alpha plane, a 3x3 box radius will be run 7 times.
20012 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
20016 @section convolution_opencl
20018 Apply convolution of 3x3, 5x5, 7x7 matrix.
20020 The filter accepts the following options:
20027 Set matrix for each plane.
20028 Matrix is sequence of 9, 25 or 49 signed numbers.
20029 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
20035 Set multiplier for calculated value for each plane.
20036 If unset or 0, it will be sum of all matrix elements.
20037 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
20043 Set bias for each plane. This value is added to the result of the multiplication.
20044 Useful for making the overall image brighter or darker.
20045 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
20049 @subsection Examples
20055 -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
20061 -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
20065 Apply edge enhance:
20067 -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
20073 -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
20077 Apply laplacian edge detector which includes diagonals:
20079 -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
20085 -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
20089 @section dilation_opencl
20091 Apply dilation effect to the video.
20093 This filter replaces the pixel by the local(3x3) maximum.
20095 It accepts the following options:
20102 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
20103 If @code{0}, plane will remain unchanged.
20106 Flag which specifies the pixel to refer to.
20107 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
20109 Flags to local 3x3 coordinates region centered on @code{x}:
20118 @subsection Example
20122 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.
20124 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
20128 @section erosion_opencl
20130 Apply erosion effect to the video.
20132 This filter replaces the pixel by the local(3x3) minimum.
20134 It accepts the following options:
20141 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
20142 If @code{0}, plane will remain unchanged.
20145 Flag which specifies the pixel to refer to.
20146 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
20148 Flags to local 3x3 coordinates region centered on @code{x}:
20157 @subsection Example
20161 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.
20163 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
20167 @section colorkey_opencl
20168 RGB colorspace color keying.
20170 The filter accepts the following options:
20174 The color which will be replaced with transparency.
20177 Similarity percentage with the key color.
20179 0.01 matches only the exact key color, while 1.0 matches everything.
20184 0.0 makes pixels either fully transparent, or not transparent at all.
20186 Higher values result in semi-transparent pixels, with a higher transparency
20187 the more similar the pixels color is to the key color.
20190 @subsection Examples
20194 Make every semi-green pixel in the input transparent with some slight blending:
20196 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
20200 @section deshake_opencl
20201 Feature-point based video stabilization filter.
20203 The filter accepts the following options:
20207 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
20210 Whether or not additional debug info should be displayed, both in the processed output and in the console.
20212 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
20214 Viewing point matches in the output video is only supported for RGB input.
20216 Defaults to @code{0}.
20218 @item adaptive_crop
20219 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
20221 Defaults to @code{1}.
20223 @item refine_features
20224 Whether or not feature points should be refined at a sub-pixel level.
20226 This can be turned off for a slight performance gain at the cost of precision.
20228 Defaults to @code{1}.
20230 @item smooth_strength
20231 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
20233 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
20235 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
20237 Defaults to @code{0.0}.
20239 @item smooth_window_multiplier
20240 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
20242 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
20244 Acceptable values range from @code{0.1} to @code{10.0}.
20246 Larger values increase the amount of motion data available for determining how to smooth the camera path,
20247 potentially improving smoothness, but also increase latency and memory usage.
20249 Defaults to @code{2.0}.
20253 @subsection Examples
20257 Stabilize a video with a fixed, medium smoothing strength:
20259 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
20263 Stabilize a video with debugging (both in console and in rendered video):
20265 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
20269 @section nlmeans_opencl
20271 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
20273 @section overlay_opencl
20275 Overlay one video on top of another.
20277 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
20278 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
20280 The filter accepts the following options:
20285 Set the x coordinate of the overlaid video on the main video.
20286 Default value is @code{0}.
20289 Set the x coordinate of the overlaid video on the main video.
20290 Default value is @code{0}.
20294 @subsection Examples
20298 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
20300 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
20303 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
20305 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
20310 @section prewitt_opencl
20312 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
20314 The filter accepts the following option:
20318 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
20321 Set value which will be multiplied with filtered result.
20322 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
20325 Set value which will be added to filtered result.
20326 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
20329 @subsection Example
20333 Apply the Prewitt operator with scale set to 2 and delta set to 10.
20335 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
20339 @section roberts_opencl
20340 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
20342 The filter accepts the following option:
20346 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
20349 Set value which will be multiplied with filtered result.
20350 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
20353 Set value which will be added to filtered result.
20354 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
20357 @subsection Example
20361 Apply the Roberts cross operator with scale set to 2 and delta set to 10
20363 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
20367 @section sobel_opencl
20369 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
20371 The filter accepts the following option:
20375 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
20378 Set value which will be multiplied with filtered result.
20379 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
20382 Set value which will be added to filtered result.
20383 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
20386 @subsection Example
20390 Apply sobel operator with scale set to 2 and delta set to 10
20392 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
20396 @section tonemap_opencl
20398 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
20400 It accepts the following parameters:
20404 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
20407 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
20410 Apply desaturation for highlights that exceed this level of brightness. The
20411 higher the parameter, the more color information will be preserved. This
20412 setting helps prevent unnaturally blown-out colors for super-highlights, by
20413 (smoothly) turning into white instead. This makes images feel more natural,
20414 at the cost of reducing information about out-of-range colors.
20416 The default value is 0.5, and the algorithm here is a little different from
20417 the cpu version tonemap currently. A setting of 0.0 disables this option.
20420 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
20421 is used to detect whether the scene has changed or not. If the distance between
20422 the current frame average brightness and the current running average exceeds
20423 a threshold value, we would re-calculate scene average and peak brightness.
20424 The default value is 0.2.
20427 Specify the output pixel format.
20429 Currently supported formats are:
20436 Set the output color range.
20438 Possible values are:
20444 Default is same as input.
20447 Set the output color primaries.
20449 Possible values are:
20455 Default is same as input.
20458 Set the output transfer characteristics.
20460 Possible values are:
20469 Set the output colorspace matrix.
20471 Possible value are:
20477 Default is same as input.
20481 @subsection Example
20485 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
20487 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
20491 @section unsharp_opencl
20493 Sharpen or blur the input video.
20495 It accepts the following parameters:
20498 @item luma_msize_x, lx
20499 Set the luma matrix horizontal size.
20500 Range is @code{[1, 23]} and default value is @code{5}.
20502 @item luma_msize_y, ly
20503 Set the luma matrix vertical size.
20504 Range is @code{[1, 23]} and default value is @code{5}.
20506 @item luma_amount, la
20507 Set the luma effect strength.
20508 Range is @code{[-10, 10]} and default value is @code{1.0}.
20510 Negative values will blur the input video, while positive values will
20511 sharpen it, a value of zero will disable the effect.
20513 @item chroma_msize_x, cx
20514 Set the chroma matrix horizontal size.
20515 Range is @code{[1, 23]} and default value is @code{5}.
20517 @item chroma_msize_y, cy
20518 Set the chroma matrix vertical size.
20519 Range is @code{[1, 23]} and default value is @code{5}.
20521 @item chroma_amount, ca
20522 Set the chroma effect strength.
20523 Range is @code{[-10, 10]} and default value is @code{0.0}.
20525 Negative values will blur the input video, while positive values will
20526 sharpen it, a value of zero will disable the effect.
20530 All parameters are optional and default to the equivalent of the
20531 string '5:5:1.0:5:5:0.0'.
20533 @subsection Examples
20537 Apply strong luma sharpen effect:
20539 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
20543 Apply a strong blur of both luma and chroma parameters:
20545 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
20549 @c man end OPENCL VIDEO FILTERS
20551 @chapter Video Sources
20552 @c man begin VIDEO SOURCES
20554 Below is a description of the currently available video sources.
20558 Buffer video frames, and make them available to the filter chain.
20560 This source is mainly intended for a programmatic use, in particular
20561 through the interface defined in @file{libavfilter/vsrc_buffer.h}.
20563 It accepts the following parameters:
20568 Specify the size (width and height) of the buffered video frames. For the
20569 syntax of this option, check the
20570 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20573 The input video width.
20576 The input video height.
20579 A string representing the pixel format of the buffered video frames.
20580 It may be a number corresponding to a pixel format, or a pixel format
20584 Specify the timebase assumed by the timestamps of the buffered frames.
20587 Specify the frame rate expected for the video stream.
20589 @item pixel_aspect, sar
20590 The sample (pixel) aspect ratio of the input video.
20593 Specify the optional parameters to be used for the scale filter which
20594 is automatically inserted when an input change is detected in the
20595 input size or format.
20597 @item hw_frames_ctx
20598 When using a hardware pixel format, this should be a reference to an
20599 AVHWFramesContext describing input frames.
20604 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
20607 will instruct the source to accept video frames with size 320x240 and
20608 with format "yuv410p", assuming 1/24 as the timestamps timebase and
20609 square pixels (1:1 sample aspect ratio).
20610 Since the pixel format with name "yuv410p" corresponds to the number 6
20611 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
20612 this example corresponds to:
20614 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
20617 Alternatively, the options can be specified as a flat string, but this
20618 syntax is deprecated:
20620 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}[:@var{sws_param}]
20624 Create a pattern generated by an elementary cellular automaton.
20626 The initial state of the cellular automaton can be defined through the
20627 @option{filename} and @option{pattern} options. If such options are
20628 not specified an initial state is created randomly.
20630 At each new frame a new row in the video is filled with the result of
20631 the cellular automaton next generation. The behavior when the whole
20632 frame is filled is defined by the @option{scroll} option.
20634 This source accepts the following options:
20638 Read the initial cellular automaton state, i.e. the starting row, from
20639 the specified file.
20640 In the file, each non-whitespace character is considered an alive
20641 cell, a newline will terminate the row, and further characters in the
20642 file will be ignored.
20645 Read the initial cellular automaton state, i.e. the starting row, from
20646 the specified string.
20648 Each non-whitespace character in the string is considered an alive
20649 cell, a newline will terminate the row, and further characters in the
20650 string will be ignored.
20653 Set the video rate, that is the number of frames generated per second.
20656 @item random_fill_ratio, ratio
20657 Set the random fill ratio for the initial cellular automaton row. It
20658 is a floating point number value ranging from 0 to 1, defaults to
20661 This option is ignored when a file or a pattern is specified.
20663 @item random_seed, seed
20664 Set the seed for filling randomly the initial row, must be an integer
20665 included between 0 and UINT32_MAX. If not specified, or if explicitly
20666 set to -1, the filter will try to use a good random seed on a best
20670 Set the cellular automaton rule, it is a number ranging from 0 to 255.
20671 Default value is 110.
20674 Set the size of the output video. For the syntax of this option, check the
20675 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20677 If @option{filename} or @option{pattern} is specified, the size is set
20678 by default to the width of the specified initial state row, and the
20679 height is set to @var{width} * PHI.
20681 If @option{size} is set, it must contain the width of the specified
20682 pattern string, and the specified pattern will be centered in the
20685 If a filename or a pattern string is not specified, the size value
20686 defaults to "320x518" (used for a randomly generated initial state).
20689 If set to 1, scroll the output upward when all the rows in the output
20690 have been already filled. If set to 0, the new generated row will be
20691 written over the top row just after the bottom row is filled.
20694 @item start_full, full
20695 If set to 1, completely fill the output with generated rows before
20696 outputting the first frame.
20697 This is the default behavior, for disabling set the value to 0.
20700 If set to 1, stitch the left and right row edges together.
20701 This is the default behavior, for disabling set the value to 0.
20704 @subsection Examples
20708 Read the initial state from @file{pattern}, and specify an output of
20711 cellauto=f=pattern:s=200x400
20715 Generate a random initial row with a width of 200 cells, with a fill
20718 cellauto=ratio=2/3:s=200x200
20722 Create a pattern generated by rule 18 starting by a single alive cell
20723 centered on an initial row with width 100:
20725 cellauto=p=@@:s=100x400:full=0:rule=18
20729 Specify a more elaborated initial pattern:
20731 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
20736 @anchor{coreimagesrc}
20737 @section coreimagesrc
20738 Video source generated on GPU using Apple's CoreImage API on OSX.
20740 This video source is a specialized version of the @ref{coreimage} video filter.
20741 Use a core image generator at the beginning of the applied filterchain to
20742 generate the content.
20744 The coreimagesrc video source accepts the following options:
20746 @item list_generators
20747 List all available generators along with all their respective options as well as
20748 possible minimum and maximum values along with the default values.
20750 list_generators=true
20754 Specify the size of the sourced video. For the syntax of this option, check the
20755 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20756 The default value is @code{320x240}.
20759 Specify the frame rate of the sourced video, as the number of frames
20760 generated per second. It has to be a string in the format
20761 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
20762 number or a valid video frame rate abbreviation. The default value is
20766 Set the sample aspect ratio of the sourced video.
20769 Set the duration of the sourced video. See
20770 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
20771 for the accepted syntax.
20773 If not specified, or the expressed duration is negative, the video is
20774 supposed to be generated forever.
20777 Additionally, all options of the @ref{coreimage} video filter are accepted.
20778 A complete filterchain can be used for further processing of the
20779 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
20780 and examples for details.
20782 @subsection Examples
20787 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
20788 given as complete and escaped command-line for Apple's standard bash shell:
20790 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
20792 This example is equivalent to the QRCode example of @ref{coreimage} without the
20793 need for a nullsrc video source.
20797 @section mandelbrot
20799 Generate a Mandelbrot set fractal, and progressively zoom towards the
20800 point specified with @var{start_x} and @var{start_y}.
20802 This source accepts the following options:
20807 Set the terminal pts value. Default value is 400.
20810 Set the terminal scale value.
20811 Must be a floating point value. Default value is 0.3.
20814 Set the inner coloring mode, that is the algorithm used to draw the
20815 Mandelbrot fractal internal region.
20817 It shall assume one of the following values:
20822 Show time until convergence.
20824 Set color based on point closest to the origin of the iterations.
20829 Default value is @var{mincol}.
20832 Set the bailout value. Default value is 10.0.
20835 Set the maximum of iterations performed by the rendering
20836 algorithm. Default value is 7189.
20839 Set outer coloring mode.
20840 It shall assume one of following values:
20842 @item iteration_count
20843 Set iteration count mode.
20844 @item normalized_iteration_count
20845 set normalized iteration count mode.
20847 Default value is @var{normalized_iteration_count}.
20850 Set frame rate, expressed as number of frames per second. Default
20854 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
20855 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
20858 Set the initial scale value. Default value is 3.0.
20861 Set the initial x position. Must be a floating point value between
20862 -100 and 100. Default value is -0.743643887037158704752191506114774.
20865 Set the initial y position. Must be a floating point value between
20866 -100 and 100. Default value is -0.131825904205311970493132056385139.
20871 Generate various test patterns, as generated by the MPlayer test filter.
20873 The size of the generated video is fixed, and is 256x256.
20874 This source is useful in particular for testing encoding features.
20876 This source accepts the following options:
20881 Specify the frame rate of the sourced video, as the number of frames
20882 generated per second. It has to be a string in the format
20883 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
20884 number or a valid video frame rate abbreviation. The default value is
20888 Set the duration of the sourced video. See
20889 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
20890 for the accepted syntax.
20892 If not specified, or the expressed duration is negative, the video is
20893 supposed to be generated forever.
20897 Set the number or the name of the test to perform. Supported tests are:
20913 Default value is "all", which will cycle through the list of all tests.
20918 mptestsrc=t=dc_luma
20921 will generate a "dc_luma" test pattern.
20923 @section frei0r_src
20925 Provide a frei0r source.
20927 To enable compilation of this filter you need to install the frei0r
20928 header and configure FFmpeg with @code{--enable-frei0r}.
20930 This source accepts the following parameters:
20935 The size of the video to generate. For the syntax of this option, check the
20936 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20939 The framerate of the generated video. It may be a string of the form
20940 @var{num}/@var{den} or a frame rate abbreviation.
20943 The name to the frei0r source to load. For more information regarding frei0r and
20944 how to set the parameters, read the @ref{frei0r} section in the video filters
20947 @item filter_params
20948 A '|'-separated list of parameters to pass to the frei0r source.
20952 For example, to generate a frei0r partik0l source with size 200x200
20953 and frame rate 10 which is overlaid on the overlay filter main input:
20955 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
20960 Generate a life pattern.
20962 This source is based on a generalization of John Conway's life game.
20964 The sourced input represents a life grid, each pixel represents a cell
20965 which can be in one of two possible states, alive or dead. Every cell
20966 interacts with its eight neighbours, which are the cells that are
20967 horizontally, vertically, or diagonally adjacent.
20969 At each interaction the grid evolves according to the adopted rule,
20970 which specifies the number of neighbor alive cells which will make a
20971 cell stay alive or born. The @option{rule} option allows one to specify
20974 This source accepts the following options:
20978 Set the file from which to read the initial grid state. In the file,
20979 each non-whitespace character is considered an alive cell, and newline
20980 is used to delimit the end of each row.
20982 If this option is not specified, the initial grid is generated
20986 Set the video rate, that is the number of frames generated per second.
20989 @item random_fill_ratio, ratio
20990 Set the random fill ratio for the initial random grid. It is a
20991 floating point number value ranging from 0 to 1, defaults to 1/PHI.
20992 It is ignored when a file is specified.
20994 @item random_seed, seed
20995 Set the seed for filling the initial random grid, must be an integer
20996 included between 0 and UINT32_MAX. If not specified, or if explicitly
20997 set to -1, the filter will try to use a good random seed on a best
21003 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
21004 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
21005 @var{NS} specifies the number of alive neighbor cells which make a
21006 live cell stay alive, and @var{NB} the number of alive neighbor cells
21007 which make a dead cell to become alive (i.e. to "born").
21008 "s" and "b" can be used in place of "S" and "B", respectively.
21010 Alternatively a rule can be specified by an 18-bits integer. The 9
21011 high order bits are used to encode the next cell state if it is alive
21012 for each number of neighbor alive cells, the low order bits specify
21013 the rule for "borning" new cells. Higher order bits encode for an
21014 higher number of neighbor cells.
21015 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
21016 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
21018 Default value is "S23/B3", which is the original Conway's game of life
21019 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
21020 cells, and will born a new cell if there are three alive cells around
21024 Set the size of the output video. For the syntax of this option, check the
21025 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21027 If @option{filename} is specified, the size is set by default to the
21028 same size of the input file. If @option{size} is set, it must contain
21029 the size specified in the input file, and the initial grid defined in
21030 that file is centered in the larger resulting area.
21032 If a filename is not specified, the size value defaults to "320x240"
21033 (used for a randomly generated initial grid).
21036 If set to 1, stitch the left and right grid edges together, and the
21037 top and bottom edges also. Defaults to 1.
21040 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
21041 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
21042 value from 0 to 255.
21045 Set the color of living (or new born) cells.
21048 Set the color of dead cells. If @option{mold} is set, this is the first color
21049 used to represent a dead cell.
21052 Set mold color, for definitely dead and moldy cells.
21054 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
21055 ffmpeg-utils manual,ffmpeg-utils}.
21058 @subsection Examples
21062 Read a grid from @file{pattern}, and center it on a grid of size
21065 life=f=pattern:s=300x300
21069 Generate a random grid of size 200x200, with a fill ratio of 2/3:
21071 life=ratio=2/3:s=200x200
21075 Specify a custom rule for evolving a randomly generated grid:
21081 Full example with slow death effect (mold) using @command{ffplay}:
21083 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
21090 @anchor{haldclutsrc}
21093 @anchor{pal100bars}
21094 @anchor{rgbtestsrc}
21096 @anchor{smptehdbars}
21099 @anchor{yuvtestsrc}
21100 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
21102 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
21104 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
21106 The @code{color} source provides an uniformly colored input.
21108 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
21109 @ref{haldclut} filter.
21111 The @code{nullsrc} source returns unprocessed video frames. It is
21112 mainly useful to be employed in analysis / debugging tools, or as the
21113 source for filters which ignore the input data.
21115 The @code{pal75bars} source generates a color bars pattern, based on
21116 EBU PAL recommendations with 75% color levels.
21118 The @code{pal100bars} source generates a color bars pattern, based on
21119 EBU PAL recommendations with 100% color levels.
21121 The @code{rgbtestsrc} source generates an RGB test pattern useful for
21122 detecting RGB vs BGR issues. You should see a red, green and blue
21123 stripe from top to bottom.
21125 The @code{smptebars} source generates a color bars pattern, based on
21126 the SMPTE Engineering Guideline EG 1-1990.
21128 The @code{smptehdbars} source generates a color bars pattern, based on
21129 the SMPTE RP 219-2002.
21131 The @code{testsrc} source generates a test video pattern, showing a
21132 color pattern, a scrolling gradient and a timestamp. This is mainly
21133 intended for testing purposes.
21135 The @code{testsrc2} source is similar to testsrc, but supports more
21136 pixel formats instead of just @code{rgb24}. This allows using it as an
21137 input for other tests without requiring a format conversion.
21139 The @code{yuvtestsrc} source generates an YUV test pattern. You should
21140 see a y, cb and cr stripe from top to bottom.
21142 The sources accept the following parameters:
21147 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
21148 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
21149 pixels to be used as identity matrix for 3D lookup tables. Each component is
21150 coded on a @code{1/(N*N)} scale.
21153 Specify the color of the source, only available in the @code{color}
21154 source. For the syntax of this option, check the
21155 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
21158 Specify the size of the sourced video. For the syntax of this option, check the
21159 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21160 The default value is @code{320x240}.
21162 This option is not available with the @code{allrgb}, @code{allyuv}, and
21163 @code{haldclutsrc} filters.
21166 Specify the frame rate of the sourced video, as the number of frames
21167 generated per second. It has to be a string in the format
21168 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
21169 number or a valid video frame rate abbreviation. The default value is
21173 Set the duration of the sourced video. See
21174 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
21175 for the accepted syntax.
21177 If not specified, or the expressed duration is negative, the video is
21178 supposed to be generated forever.
21181 Set the sample aspect ratio of the sourced video.
21184 Specify the alpha (opacity) of the background, only available in the
21185 @code{testsrc2} source. The value must be between 0 (fully transparent) and
21186 255 (fully opaque, the default).
21189 Set the number of decimals to show in the timestamp, only available in the
21190 @code{testsrc} source.
21192 The displayed timestamp value will correspond to the original
21193 timestamp value multiplied by the power of 10 of the specified
21194 value. Default value is 0.
21197 @subsection Examples
21201 Generate a video with a duration of 5.3 seconds, with size
21202 176x144 and a frame rate of 10 frames per second:
21204 testsrc=duration=5.3:size=qcif:rate=10
21208 The following graph description will generate a red source
21209 with an opacity of 0.2, with size "qcif" and a frame rate of 10
21212 color=c=red@@0.2:s=qcif:r=10
21216 If the input content is to be ignored, @code{nullsrc} can be used. The
21217 following command generates noise in the luminance plane by employing
21218 the @code{geq} filter:
21220 nullsrc=s=256x256, geq=random(1)*255:128:128
21224 @subsection Commands
21226 The @code{color} source supports the following commands:
21230 Set the color of the created image. Accepts the same syntax of the
21231 corresponding @option{color} option.
21236 Generate video using an OpenCL program.
21241 OpenCL program source file.
21244 Kernel name in program.
21247 Size of frames to generate. This must be set.
21250 Pixel format to use for the generated frames. This must be set.
21253 Number of frames generated every second. Default value is '25'.
21257 For details of how the program loading works, see the @ref{program_opencl}
21264 Generate a colour ramp by setting pixel values from the position of the pixel
21265 in the output image. (Note that this will work with all pixel formats, but
21266 the generated output will not be the same.)
21268 __kernel void ramp(__write_only image2d_t dst,
21269 unsigned int index)
21271 int2 loc = (int2)(get_global_id(0), get_global_id(1));
21274 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
21276 write_imagef(dst, loc, val);
21281 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
21283 __kernel void sierpinski_carpet(__write_only image2d_t dst,
21284 unsigned int index)
21286 int2 loc = (int2)(get_global_id(0), get_global_id(1));
21288 float4 value = 0.0f;
21289 int x = loc.x + index;
21290 int y = loc.y + index;
21291 while (x > 0 || y > 0) {
21292 if (x % 3 == 1 && y % 3 == 1) {
21300 write_imagef(dst, loc, value);
21306 @section sierpinski
21308 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
21310 This source accepts the following options:
21314 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
21315 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
21318 Set frame rate, expressed as number of frames per second. Default
21322 Set seed which is used for random panning.
21325 Set max jump for single pan destination. Allowed range is from 1 to 10000.
21328 Set fractal type, can be default @code{carpet} or @code{triangle}.
21331 @c man end VIDEO SOURCES
21333 @chapter Video Sinks
21334 @c man begin VIDEO SINKS
21336 Below is a description of the currently available video sinks.
21338 @section buffersink
21340 Buffer video frames, and make them available to the end of the filter
21343 This sink is mainly intended for programmatic use, in particular
21344 through the interface defined in @file{libavfilter/buffersink.h}
21345 or the options system.
21347 It accepts a pointer to an AVBufferSinkContext structure, which
21348 defines the incoming buffers' formats, to be passed as the opaque
21349 parameter to @code{avfilter_init_filter} for initialization.
21353 Null video sink: do absolutely nothing with the input video. It is
21354 mainly useful as a template and for use in analysis / debugging
21357 @c man end VIDEO SINKS
21359 @chapter Multimedia Filters
21360 @c man begin MULTIMEDIA FILTERS
21362 Below is a description of the currently available multimedia filters.
21366 Convert input audio to a video output, displaying the audio bit scope.
21368 The filter accepts the following options:
21372 Set frame rate, expressed as number of frames per second. Default
21376 Specify the video size for the output. For the syntax of this option, check the
21377 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21378 Default value is @code{1024x256}.
21381 Specify list of colors separated by space or by '|' which will be used to
21382 draw channels. Unrecognized or missing colors will be replaced
21386 @section ahistogram
21388 Convert input audio to a video output, displaying the volume histogram.
21390 The filter accepts the following options:
21394 Specify how histogram is calculated.
21396 It accepts the following values:
21399 Use single histogram for all channels.
21401 Use separate histogram for each channel.
21403 Default is @code{single}.
21406 Set frame rate, expressed as number of frames per second. Default
21410 Specify the video size for the output. For the syntax of this option, check the
21411 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21412 Default value is @code{hd720}.
21417 It accepts the following values:
21428 reverse logarithmic
21430 Default is @code{log}.
21433 Set amplitude scale.
21435 It accepts the following values:
21442 Default is @code{log}.
21445 Set how much frames to accumulate in histogram.
21446 Default is 1. Setting this to -1 accumulates all frames.
21449 Set histogram ratio of window height.
21452 Set sonogram sliding.
21454 It accepts the following values:
21457 replace old rows with new ones.
21459 scroll from top to bottom.
21461 Default is @code{replace}.
21464 @section aphasemeter
21466 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
21467 representing mean phase of current audio frame. A video output can also be produced and is
21468 enabled by default. The audio is passed through as first output.
21470 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
21471 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
21472 and @code{1} means channels are in phase.
21474 The filter accepts the following options, all related to its video output:
21478 Set the output frame rate. Default value is @code{25}.
21481 Set the video size for the output. For the syntax of this option, check the
21482 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21483 Default value is @code{800x400}.
21488 Specify the red, green, blue contrast. Default values are @code{2},
21489 @code{7} and @code{1}.
21490 Allowed range is @code{[0, 255]}.
21493 Set color which will be used for drawing median phase. If color is
21494 @code{none} which is default, no median phase value will be drawn.
21497 Enable video output. Default is enabled.
21500 @section avectorscope
21502 Convert input audio to a video output, representing the audio vector
21505 The filter is used to measure the difference between channels of stereo
21506 audio stream. A monaural signal, consisting of identical left and right
21507 signal, results in straight vertical line. Any stereo separation is visible
21508 as a deviation from this line, creating a Lissajous figure.
21509 If the straight (or deviation from it) but horizontal line appears this
21510 indicates that the left and right channels are out of phase.
21512 The filter accepts the following options:
21516 Set the vectorscope mode.
21518 Available values are:
21521 Lissajous rotated by 45 degrees.
21524 Same as above but not rotated.
21527 Shape resembling half of circle.
21530 Default value is @samp{lissajous}.
21533 Set the video size for the output. For the syntax of this option, check the
21534 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21535 Default value is @code{400x400}.
21538 Set the output frame rate. Default value is @code{25}.
21544 Specify the red, green, blue and alpha contrast. Default values are @code{40},
21545 @code{160}, @code{80} and @code{255}.
21546 Allowed range is @code{[0, 255]}.
21552 Specify the red, green, blue and alpha fade. Default values are @code{15},
21553 @code{10}, @code{5} and @code{5}.
21554 Allowed range is @code{[0, 255]}.
21557 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
21558 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
21561 Set the vectorscope drawing mode.
21563 Available values are:
21566 Draw dot for each sample.
21569 Draw line between previous and current sample.
21572 Default value is @samp{dot}.
21575 Specify amplitude scale of audio samples.
21577 Available values are:
21593 Swap left channel axis with right channel axis.
21603 Mirror only x axis.
21606 Mirror only y axis.
21614 @subsection Examples
21618 Complete example using @command{ffplay}:
21620 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
21621 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
21625 @section bench, abench
21627 Benchmark part of a filtergraph.
21629 The filter accepts the following options:
21633 Start or stop a timer.
21635 Available values are:
21638 Get the current time, set it as frame metadata (using the key
21639 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
21642 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
21643 the input frame metadata to get the time difference. Time difference, average,
21644 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
21645 @code{min}) are then printed. The timestamps are expressed in seconds.
21649 @subsection Examples
21653 Benchmark @ref{selectivecolor} filter:
21655 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
21661 Concatenate audio and video streams, joining them together one after the
21664 The filter works on segments of synchronized video and audio streams. All
21665 segments must have the same number of streams of each type, and that will
21666 also be the number of streams at output.
21668 The filter accepts the following options:
21673 Set the number of segments. Default is 2.
21676 Set the number of output video streams, that is also the number of video
21677 streams in each segment. Default is 1.
21680 Set the number of output audio streams, that is also the number of audio
21681 streams in each segment. Default is 0.
21684 Activate unsafe mode: do not fail if segments have a different format.
21688 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
21689 @var{a} audio outputs.
21691 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
21692 segment, in the same order as the outputs, then the inputs for the second
21695 Related streams do not always have exactly the same duration, for various
21696 reasons including codec frame size or sloppy authoring. For that reason,
21697 related synchronized streams (e.g. a video and its audio track) should be
21698 concatenated at once. The concat filter will use the duration of the longest
21699 stream in each segment (except the last one), and if necessary pad shorter
21700 audio streams with silence.
21702 For this filter to work correctly, all segments must start at timestamp 0.
21704 All corresponding streams must have the same parameters in all segments; the
21705 filtering system will automatically select a common pixel format for video
21706 streams, and a common sample format, sample rate and channel layout for
21707 audio streams, but other settings, such as resolution, must be converted
21708 explicitly by the user.
21710 Different frame rates are acceptable but will result in variable frame rate
21711 at output; be sure to configure the output file to handle it.
21713 @subsection Examples
21717 Concatenate an opening, an episode and an ending, all in bilingual version
21718 (video in stream 0, audio in streams 1 and 2):
21720 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
21721 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
21722 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
21723 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
21727 Concatenate two parts, handling audio and video separately, using the
21728 (a)movie sources, and adjusting the resolution:
21730 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
21731 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
21732 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
21734 Note that a desync will happen at the stitch if the audio and video streams
21735 do not have exactly the same duration in the first file.
21739 @subsection Commands
21741 This filter supports the following commands:
21744 Close the current segment and step to the next one
21747 @section drawgraph, adrawgraph
21749 Draw a graph using input video or audio metadata.
21751 It accepts the following parameters:
21755 Set 1st frame metadata key from which metadata values will be used to draw a graph.
21758 Set 1st foreground color expression.
21761 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
21764 Set 2nd foreground color expression.
21767 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
21770 Set 3rd foreground color expression.
21773 Set 4th frame metadata key from which metadata values will be used to draw a graph.
21776 Set 4th foreground color expression.
21779 Set minimal value of metadata value.
21782 Set maximal value of metadata value.
21785 Set graph background color. Default is white.
21790 Available values for mode is:
21797 Default is @code{line}.
21802 Available values for slide is:
21805 Draw new frame when right border is reached.
21808 Replace old columns with new ones.
21811 Scroll from right to left.
21814 Scroll from left to right.
21817 Draw single picture.
21820 Default is @code{frame}.
21823 Set size of graph video. For the syntax of this option, check the
21824 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21825 The default value is @code{900x256}.
21827 The foreground color expressions can use the following variables:
21830 Minimal value of metadata value.
21833 Maximal value of metadata value.
21836 Current metadata key value.
21839 The color is defined as 0xAABBGGRR.
21842 Example using metadata from @ref{signalstats} filter:
21844 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
21847 Example using metadata from @ref{ebur128} filter:
21849 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
21855 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
21856 level. By default, it logs a message at a frequency of 10Hz with the
21857 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
21858 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
21860 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
21861 sample format is double-precision floating point. The input stream will be converted to
21862 this specification, if needed. Users may need to insert aformat and/or aresample filters
21863 after this filter to obtain the original parameters.
21865 The filter also has a video output (see the @var{video} option) with a real
21866 time graph to observe the loudness evolution. The graphic contains the logged
21867 message mentioned above, so it is not printed anymore when this option is set,
21868 unless the verbose logging is set. The main graphing area contains the
21869 short-term loudness (3 seconds of analysis), and the gauge on the right is for
21870 the momentary loudness (400 milliseconds), but can optionally be configured
21871 to instead display short-term loudness (see @var{gauge}).
21873 The green area marks a +/- 1LU target range around the target loudness
21874 (-23LUFS by default, unless modified through @var{target}).
21876 More information about the Loudness Recommendation EBU R128 on
21877 @url{http://tech.ebu.ch/loudness}.
21879 The filter accepts the following options:
21884 Activate the video output. The audio stream is passed unchanged whether this
21885 option is set or no. The video stream will be the first output stream if
21886 activated. Default is @code{0}.
21889 Set the video size. This option is for video only. For the syntax of this
21891 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21892 Default and minimum resolution is @code{640x480}.
21895 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
21896 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
21897 other integer value between this range is allowed.
21900 Set metadata injection. If set to @code{1}, the audio input will be segmented
21901 into 100ms output frames, each of them containing various loudness information
21902 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
21904 Default is @code{0}.
21907 Force the frame logging level.
21909 Available values are:
21912 information logging level
21914 verbose logging level
21917 By default, the logging level is set to @var{info}. If the @option{video} or
21918 the @option{metadata} options are set, it switches to @var{verbose}.
21923 Available modes can be cumulated (the option is a @code{flag} type). Possible
21927 Disable any peak mode (default).
21929 Enable sample-peak mode.
21931 Simple peak mode looking for the higher sample value. It logs a message
21932 for sample-peak (identified by @code{SPK}).
21934 Enable true-peak mode.
21936 If enabled, the peak lookup is done on an over-sampled version of the input
21937 stream for better peak accuracy. It logs a message for true-peak.
21938 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
21939 This mode requires a build with @code{libswresample}.
21943 Treat mono input files as "dual mono". If a mono file is intended for playback
21944 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
21945 If set to @code{true}, this option will compensate for this effect.
21946 Multi-channel input files are not affected by this option.
21949 Set a specific pan law to be used for the measurement of dual mono files.
21950 This parameter is optional, and has a default value of -3.01dB.
21953 Set a specific target level (in LUFS) used as relative zero in the visualization.
21954 This parameter is optional and has a default value of -23LUFS as specified
21955 by EBU R128. However, material published online may prefer a level of -16LUFS
21956 (e.g. for use with podcasts or video platforms).
21959 Set the value displayed by the gauge. Valid values are @code{momentary} and s
21960 @code{shortterm}. By default the momentary value will be used, but in certain
21961 scenarios it may be more useful to observe the short term value instead (e.g.
21965 Sets the display scale for the loudness. Valid parameters are @code{absolute}
21966 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
21967 video output, not the summary or continuous log output.
21970 @subsection Examples
21974 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
21976 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
21980 Run an analysis with @command{ffmpeg}:
21982 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
21986 @section interleave, ainterleave
21988 Temporally interleave frames from several inputs.
21990 @code{interleave} works with video inputs, @code{ainterleave} with audio.
21992 These filters read frames from several inputs and send the oldest
21993 queued frame to the output.
21995 Input streams must have well defined, monotonically increasing frame
21998 In order to submit one frame to output, these filters need to enqueue
21999 at least one frame for each input, so they cannot work in case one
22000 input is not yet terminated and will not receive incoming frames.
22002 For example consider the case when one input is a @code{select} filter
22003 which always drops input frames. The @code{interleave} filter will keep
22004 reading from that input, but it will never be able to send new frames
22005 to output until the input sends an end-of-stream signal.
22007 Also, depending on inputs synchronization, the filters will drop
22008 frames in case one input receives more frames than the other ones, and
22009 the queue is already filled.
22011 These filters accept the following options:
22015 Set the number of different inputs, it is 2 by default.
22018 @subsection Examples
22022 Interleave frames belonging to different streams using @command{ffmpeg}:
22024 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
22028 Add flickering blur effect:
22030 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
22034 @section metadata, ametadata
22036 Manipulate frame metadata.
22038 This filter accepts the following options:
22042 Set mode of operation of the filter.
22044 Can be one of the following:
22048 If both @code{value} and @code{key} is set, select frames
22049 which have such metadata. If only @code{key} is set, select
22050 every frame that has such key in metadata.
22053 Add new metadata @code{key} and @code{value}. If key is already available
22057 Modify value of already present key.
22060 If @code{value} is set, delete only keys that have such value.
22061 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
22065 Print key and its value if metadata was found. If @code{key} is not set print all
22066 metadata values available in frame.
22070 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
22073 Set metadata value which will be used. This option is mandatory for
22074 @code{modify} and @code{add} mode.
22077 Which function to use when comparing metadata value and @code{value}.
22079 Can be one of following:
22083 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
22086 Values are interpreted as strings, returns true if metadata value starts with
22087 the @code{value} option string.
22090 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
22093 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
22096 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
22099 Values are interpreted as floats, returns true if expression from option @code{expr}
22103 Values are interpreted as strings, returns true if metadata value ends with
22104 the @code{value} option string.
22108 Set expression which is used when @code{function} is set to @code{expr}.
22109 The expression is evaluated through the eval API and can contain the following
22114 Float representation of @code{value} from metadata key.
22117 Float representation of @code{value} as supplied by user in @code{value} option.
22121 If specified in @code{print} mode, output is written to the named file. Instead of
22122 plain filename any writable url can be specified. Filename ``-'' is a shorthand
22123 for standard output. If @code{file} option is not set, output is written to the log
22124 with AV_LOG_INFO loglevel.
22128 @subsection Examples
22132 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
22135 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
22138 Print silencedetect output to file @file{metadata.txt}.
22140 silencedetect,ametadata=mode=print:file=metadata.txt
22143 Direct all metadata to a pipe with file descriptor 4.
22145 metadata=mode=print:file='pipe\:4'
22149 @section perms, aperms
22151 Set read/write permissions for the output frames.
22153 These filters are mainly aimed at developers to test direct path in the
22154 following filter in the filtergraph.
22156 The filters accept the following options:
22160 Select the permissions mode.
22162 It accepts the following values:
22165 Do nothing. This is the default.
22167 Set all the output frames read-only.
22169 Set all the output frames directly writable.
22171 Make the frame read-only if writable, and writable if read-only.
22173 Set each output frame read-only or writable randomly.
22177 Set the seed for the @var{random} mode, must be an integer included between
22178 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
22179 @code{-1}, the filter will try to use a good random seed on a best effort
22183 Note: in case of auto-inserted filter between the permission filter and the
22184 following one, the permission might not be received as expected in that
22185 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
22186 perms/aperms filter can avoid this problem.
22188 @section realtime, arealtime
22190 Slow down filtering to match real time approximately.
22192 These filters will pause the filtering for a variable amount of time to
22193 match the output rate with the input timestamps.
22194 They are similar to the @option{re} option to @code{ffmpeg}.
22196 They accept the following options:
22200 Time limit for the pauses. Any pause longer than that will be considered
22201 a timestamp discontinuity and reset the timer. Default is 2 seconds.
22203 Speed factor for processing. The value must be a float larger than zero.
22204 Values larger than 1.0 will result in faster than realtime processing,
22205 smaller will slow processing down. The @var{limit} is automatically adapted
22206 accordingly. Default is 1.0.
22208 A processing speed faster than what is possible without these filters cannot
22213 @section select, aselect
22215 Select frames to pass in output.
22217 This filter accepts the following options:
22222 Set expression, which is evaluated for each input frame.
22224 If the expression is evaluated to zero, the frame is discarded.
22226 If the evaluation result is negative or NaN, the frame is sent to the
22227 first output; otherwise it is sent to the output with index
22228 @code{ceil(val)-1}, assuming that the input index starts from 0.
22230 For example a value of @code{1.2} corresponds to the output with index
22231 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
22234 Set the number of outputs. The output to which to send the selected
22235 frame is based on the result of the evaluation. Default value is 1.
22238 The expression can contain the following constants:
22242 The (sequential) number of the filtered frame, starting from 0.
22245 The (sequential) number of the selected frame, starting from 0.
22247 @item prev_selected_n
22248 The sequential number of the last selected frame. It's NAN if undefined.
22251 The timebase of the input timestamps.
22254 The PTS (Presentation TimeStamp) of the filtered video frame,
22255 expressed in @var{TB} units. It's NAN if undefined.
22258 The PTS of the filtered video frame,
22259 expressed in seconds. It's NAN if undefined.
22262 The PTS of the previously filtered video frame. It's NAN if undefined.
22264 @item prev_selected_pts
22265 The PTS of the last previously filtered video frame. It's NAN if undefined.
22267 @item prev_selected_t
22268 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
22271 The PTS of the first video frame in the video. It's NAN if undefined.
22274 The time of the first video frame in the video. It's NAN if undefined.
22276 @item pict_type @emph{(video only)}
22277 The type of the filtered frame. It can assume one of the following
22289 @item interlace_type @emph{(video only)}
22290 The frame interlace type. It can assume one of the following values:
22293 The frame is progressive (not interlaced).
22295 The frame is top-field-first.
22297 The frame is bottom-field-first.
22300 @item consumed_sample_n @emph{(audio only)}
22301 the number of selected samples before the current frame
22303 @item samples_n @emph{(audio only)}
22304 the number of samples in the current frame
22306 @item sample_rate @emph{(audio only)}
22307 the input sample rate
22310 This is 1 if the filtered frame is a key-frame, 0 otherwise.
22313 the position in the file of the filtered frame, -1 if the information
22314 is not available (e.g. for synthetic video)
22316 @item scene @emph{(video only)}
22317 value between 0 and 1 to indicate a new scene; a low value reflects a low
22318 probability for the current frame to introduce a new scene, while a higher
22319 value means the current frame is more likely to be one (see the example below)
22321 @item concatdec_select
22322 The concat demuxer can select only part of a concat input file by setting an
22323 inpoint and an outpoint, but the output packets may not be entirely contained
22324 in the selected interval. By using this variable, it is possible to skip frames
22325 generated by the concat demuxer which are not exactly contained in the selected
22328 This works by comparing the frame pts against the @var{lavf.concat.start_time}
22329 and the @var{lavf.concat.duration} packet metadata values which are also
22330 present in the decoded frames.
22332 The @var{concatdec_select} variable is -1 if the frame pts is at least
22333 start_time and either the duration metadata is missing or the frame pts is less
22334 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
22337 That basically means that an input frame is selected if its pts is within the
22338 interval set by the concat demuxer.
22342 The default value of the select expression is "1".
22344 @subsection Examples
22348 Select all frames in input:
22353 The example above is the same as:
22365 Select only I-frames:
22367 select='eq(pict_type\,I)'
22371 Select one frame every 100:
22373 select='not(mod(n\,100))'
22377 Select only frames contained in the 10-20 time interval:
22379 select=between(t\,10\,20)
22383 Select only I-frames contained in the 10-20 time interval:
22385 select=between(t\,10\,20)*eq(pict_type\,I)
22389 Select frames with a minimum distance of 10 seconds:
22391 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
22395 Use aselect to select only audio frames with samples number > 100:
22397 aselect='gt(samples_n\,100)'
22401 Create a mosaic of the first scenes:
22403 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
22406 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
22410 Send even and odd frames to separate outputs, and compose them:
22412 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
22416 Select useful frames from an ffconcat file which is using inpoints and
22417 outpoints but where the source files are not intra frame only.
22419 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
22423 @section sendcmd, asendcmd
22425 Send commands to filters in the filtergraph.
22427 These filters read commands to be sent to other filters in the
22430 @code{sendcmd} must be inserted between two video filters,
22431 @code{asendcmd} must be inserted between two audio filters, but apart
22432 from that they act the same way.
22434 The specification of commands can be provided in the filter arguments
22435 with the @var{commands} option, or in a file specified by the
22436 @var{filename} option.
22438 These filters accept the following options:
22441 Set the commands to be read and sent to the other filters.
22443 Set the filename of the commands to be read and sent to the other
22447 @subsection Commands syntax
22449 A commands description consists of a sequence of interval
22450 specifications, comprising a list of commands to be executed when a
22451 particular event related to that interval occurs. The occurring event
22452 is typically the current frame time entering or leaving a given time
22455 An interval is specified by the following syntax:
22457 @var{START}[-@var{END}] @var{COMMANDS};
22460 The time interval is specified by the @var{START} and @var{END} times.
22461 @var{END} is optional and defaults to the maximum time.
22463 The current frame time is considered within the specified interval if
22464 it is included in the interval [@var{START}, @var{END}), that is when
22465 the time is greater or equal to @var{START} and is lesser than
22468 @var{COMMANDS} consists of a sequence of one or more command
22469 specifications, separated by ",", relating to that interval. The
22470 syntax of a command specification is given by:
22472 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
22475 @var{FLAGS} is optional and specifies the type of events relating to
22476 the time interval which enable sending the specified command, and must
22477 be a non-null sequence of identifier flags separated by "+" or "|" and
22478 enclosed between "[" and "]".
22480 The following flags are recognized:
22483 The command is sent when the current frame timestamp enters the
22484 specified interval. In other words, the command is sent when the
22485 previous frame timestamp was not in the given interval, and the
22489 The command is sent when the current frame timestamp leaves the
22490 specified interval. In other words, the command is sent when the
22491 previous frame timestamp was in the given interval, and the
22495 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
22498 @var{TARGET} specifies the target of the command, usually the name of
22499 the filter class or a specific filter instance name.
22501 @var{COMMAND} specifies the name of the command for the target filter.
22503 @var{ARG} is optional and specifies the optional list of argument for
22504 the given @var{COMMAND}.
22506 Between one interval specification and another, whitespaces, or
22507 sequences of characters starting with @code{#} until the end of line,
22508 are ignored and can be used to annotate comments.
22510 A simplified BNF description of the commands specification syntax
22513 @var{COMMAND_FLAG} ::= "enter" | "leave"
22514 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
22515 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
22516 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
22517 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
22518 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
22521 @subsection Examples
22525 Specify audio tempo change at second 4:
22527 asendcmd=c='4.0 atempo tempo 1.5',atempo
22531 Target a specific filter instance:
22533 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
22537 Specify a list of drawtext and hue commands in a file.
22539 # show text in the interval 5-10
22540 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
22541 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
22543 # desaturate the image in the interval 15-20
22544 15.0-20.0 [enter] hue s 0,
22545 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
22547 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
22549 # apply an exponential saturation fade-out effect, starting from time 25
22550 25 [enter] hue s exp(25-t)
22553 A filtergraph allowing to read and process the above command list
22554 stored in a file @file{test.cmd}, can be specified with:
22556 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
22561 @section setpts, asetpts
22563 Change the PTS (presentation timestamp) of the input frames.
22565 @code{setpts} works on video frames, @code{asetpts} on audio frames.
22567 This filter accepts the following options:
22572 The expression which is evaluated for each frame to construct its timestamp.
22576 The expression is evaluated through the eval API and can contain the following
22580 @item FRAME_RATE, FR
22581 frame rate, only defined for constant frame-rate video
22584 The presentation timestamp in input
22587 The count of the input frame for video or the number of consumed samples,
22588 not including the current frame for audio, starting from 0.
22590 @item NB_CONSUMED_SAMPLES
22591 The number of consumed samples, not including the current frame (only
22594 @item NB_SAMPLES, S
22595 The number of samples in the current frame (only audio)
22597 @item SAMPLE_RATE, SR
22598 The audio sample rate.
22601 The PTS of the first frame.
22604 the time in seconds of the first frame
22607 State whether the current frame is interlaced.
22610 the time in seconds of the current frame
22613 original position in the file of the frame, or undefined if undefined
22614 for the current frame
22617 The previous input PTS.
22620 previous input time in seconds
22623 The previous output PTS.
22626 previous output time in seconds
22629 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
22633 The wallclock (RTC) time at the start of the movie in microseconds.
22636 The timebase of the input timestamps.
22640 @subsection Examples
22644 Start counting PTS from zero
22646 setpts=PTS-STARTPTS
22650 Apply fast motion effect:
22656 Apply slow motion effect:
22662 Set fixed rate of 25 frames per second:
22668 Set fixed rate 25 fps with some jitter:
22670 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
22674 Apply an offset of 10 seconds to the input PTS:
22680 Generate timestamps from a "live source" and rebase onto the current timebase:
22682 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
22686 Generate timestamps by counting samples:
22695 Force color range for the output video frame.
22697 The @code{setrange} filter marks the color range property for the
22698 output frames. It does not change the input frame, but only sets the
22699 corresponding property, which affects how the frame is treated by
22702 The filter accepts the following options:
22707 Available values are:
22711 Keep the same color range property.
22713 @item unspecified, unknown
22714 Set the color range as unspecified.
22716 @item limited, tv, mpeg
22717 Set the color range as limited.
22719 @item full, pc, jpeg
22720 Set the color range as full.
22724 @section settb, asettb
22726 Set the timebase to use for the output frames timestamps.
22727 It is mainly useful for testing timebase configuration.
22729 It accepts the following parameters:
22734 The expression which is evaluated into the output timebase.
22738 The value for @option{tb} is an arithmetic expression representing a
22739 rational. The expression can contain the constants "AVTB" (the default
22740 timebase), "intb" (the input timebase) and "sr" (the sample rate,
22741 audio only). Default value is "intb".
22743 @subsection Examples
22747 Set the timebase to 1/25:
22753 Set the timebase to 1/10:
22759 Set the timebase to 1001/1000:
22765 Set the timebase to 2*intb:
22771 Set the default timebase value:
22778 Convert input audio to a video output representing frequency spectrum
22779 logarithmically using Brown-Puckette constant Q transform algorithm with
22780 direct frequency domain coefficient calculation (but the transform itself
22781 is not really constant Q, instead the Q factor is actually variable/clamped),
22782 with musical tone scale, from E0 to D#10.
22784 The filter accepts the following options:
22788 Specify the video size for the output. It must be even. For the syntax of this option,
22789 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22790 Default value is @code{1920x1080}.
22793 Set the output frame rate. Default value is @code{25}.
22796 Set the bargraph height. It must be even. Default value is @code{-1} which
22797 computes the bargraph height automatically.
22800 Set the axis height. It must be even. Default value is @code{-1} which computes
22801 the axis height automatically.
22804 Set the sonogram height. It must be even. Default value is @code{-1} which
22805 computes the sonogram height automatically.
22808 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
22809 instead. Default value is @code{1}.
22811 @item sono_v, volume
22812 Specify the sonogram volume expression. It can contain variables:
22815 the @var{bar_v} evaluated expression
22816 @item frequency, freq, f
22817 the frequency where it is evaluated
22818 @item timeclamp, tc
22819 the value of @var{timeclamp} option
22823 @item a_weighting(f)
22824 A-weighting of equal loudness
22825 @item b_weighting(f)
22826 B-weighting of equal loudness
22827 @item c_weighting(f)
22828 C-weighting of equal loudness.
22830 Default value is @code{16}.
22832 @item bar_v, volume2
22833 Specify the bargraph volume expression. It can contain variables:
22836 the @var{sono_v} evaluated expression
22837 @item frequency, freq, f
22838 the frequency where it is evaluated
22839 @item timeclamp, tc
22840 the value of @var{timeclamp} option
22844 @item a_weighting(f)
22845 A-weighting of equal loudness
22846 @item b_weighting(f)
22847 B-weighting of equal loudness
22848 @item c_weighting(f)
22849 C-weighting of equal loudness.
22851 Default value is @code{sono_v}.
22853 @item sono_g, gamma
22854 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
22855 higher gamma makes the spectrum having more range. Default value is @code{3}.
22856 Acceptable range is @code{[1, 7]}.
22858 @item bar_g, gamma2
22859 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
22863 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
22864 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
22866 @item timeclamp, tc
22867 Specify the transform timeclamp. At low frequency, there is trade-off between
22868 accuracy in time domain and frequency domain. If timeclamp is lower,
22869 event in time domain is represented more accurately (such as fast bass drum),
22870 otherwise event in frequency domain is represented more accurately
22871 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
22874 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
22875 limits future samples by applying asymmetric windowing in time domain, useful
22876 when low latency is required. Accepted range is @code{[0, 1]}.
22879 Specify the transform base frequency. Default value is @code{20.01523126408007475},
22880 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
22883 Specify the transform end frequency. Default value is @code{20495.59681441799654},
22884 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
22887 This option is deprecated and ignored.
22890 Specify the transform length in time domain. Use this option to control accuracy
22891 trade-off between time domain and frequency domain at every frequency sample.
22892 It can contain variables:
22894 @item frequency, freq, f
22895 the frequency where it is evaluated
22896 @item timeclamp, tc
22897 the value of @var{timeclamp} option.
22899 Default value is @code{384*tc/(384+tc*f)}.
22902 Specify the transform count for every video frame. Default value is @code{6}.
22903 Acceptable range is @code{[1, 30]}.
22906 Specify the transform count for every single pixel. Default value is @code{0},
22907 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
22910 Specify font file for use with freetype to draw the axis. If not specified,
22911 use embedded font. Note that drawing with font file or embedded font is not
22912 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
22916 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
22917 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
22921 Specify font color expression. This is arithmetic expression that should return
22922 integer value 0xRRGGBB. It can contain variables:
22924 @item frequency, freq, f
22925 the frequency where it is evaluated
22926 @item timeclamp, tc
22927 the value of @var{timeclamp} option
22932 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
22933 @item r(x), g(x), b(x)
22934 red, green, and blue value of intensity x.
22936 Default value is @code{st(0, (midi(f)-59.5)/12);
22937 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
22938 r(1-ld(1)) + b(ld(1))}.
22941 Specify image file to draw the axis. This option override @var{fontfile} and
22942 @var{fontcolor} option.
22945 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
22946 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
22947 Default value is @code{1}.
22950 Set colorspace. The accepted values are:
22953 Unspecified (default)
22962 BT.470BG or BT.601-6 625
22965 SMPTE-170M or BT.601-6 525
22971 BT.2020 with non-constant luminance
22976 Set spectrogram color scheme. This is list of floating point values with format
22977 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
22978 The default is @code{1|0.5|0|0|0.5|1}.
22982 @subsection Examples
22986 Playing audio while showing the spectrum:
22988 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
22992 Same as above, but with frame rate 30 fps:
22994 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
22998 Playing at 1280x720:
23000 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
23004 Disable sonogram display:
23010 A1 and its harmonics: A1, A2, (near)E3, A3:
23012 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),
23013 asplit[a][out1]; [a] showcqt [out0]'
23017 Same as above, but with more accuracy in frequency domain:
23019 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),
23020 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
23026 bar_v=10:sono_v=bar_v*a_weighting(f)
23030 Custom gamma, now spectrum is linear to the amplitude.
23036 Custom tlength equation:
23038 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)))'
23042 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
23044 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
23048 Custom font using fontconfig:
23050 font='Courier New,Monospace,mono|bold'
23054 Custom frequency range with custom axis using image file:
23056 axisfile=myaxis.png:basefreq=40:endfreq=10000
23062 Convert input audio to video output representing the audio power spectrum.
23063 Audio amplitude is on Y-axis while frequency is on X-axis.
23065 The filter accepts the following options:
23069 Specify size of video. For the syntax of this option, check the
23070 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23071 Default is @code{1024x512}.
23075 This set how each frequency bin will be represented.
23077 It accepts the following values:
23083 Default is @code{bar}.
23086 Set amplitude scale.
23088 It accepts the following values:
23102 Default is @code{log}.
23105 Set frequency scale.
23107 It accepts the following values:
23116 Reverse logarithmic scale.
23118 Default is @code{lin}.
23121 Set window size. Allowed range is from 16 to 65536.
23123 Default is @code{2048}
23126 Set windowing function.
23128 It accepts the following values:
23151 Default is @code{hanning}.
23154 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
23155 which means optimal overlap for selected window function will be picked.
23158 Set time averaging. Setting this to 0 will display current maximal peaks.
23159 Default is @code{1}, which means time averaging is disabled.
23162 Specify list of colors separated by space or by '|' which will be used to
23163 draw channel frequencies. Unrecognized or missing colors will be replaced
23167 Set channel display mode.
23169 It accepts the following values:
23174 Default is @code{combined}.
23177 Set minimum amplitude used in @code{log} amplitude scaler.
23181 @section showspatial
23183 Convert stereo input audio to a video output, representing the spatial relationship
23184 between two channels.
23186 The filter accepts the following options:
23190 Specify the video size for the output. For the syntax of this option, check the
23191 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23192 Default value is @code{512x512}.
23195 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
23198 Set window function.
23200 It accepts the following values:
23225 Default value is @code{hann}.
23228 Set ratio of overlap window. Default value is @code{0.5}.
23229 When value is @code{1} overlap is set to recommended size for specific
23230 window function currently used.
23233 @anchor{showspectrum}
23234 @section showspectrum
23236 Convert input audio to a video output, representing the audio frequency
23239 The filter accepts the following options:
23243 Specify the video size for the output. For the syntax of this option, check the
23244 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23245 Default value is @code{640x512}.
23248 Specify how the spectrum should slide along the window.
23250 It accepts the following values:
23253 the samples start again on the left when they reach the right
23255 the samples scroll from right to left
23257 frames are only produced when the samples reach the right
23259 the samples scroll from left to right
23262 Default value is @code{replace}.
23265 Specify display mode.
23267 It accepts the following values:
23270 all channels are displayed in the same row
23272 all channels are displayed in separate rows
23275 Default value is @samp{combined}.
23278 Specify display color mode.
23280 It accepts the following values:
23283 each channel is displayed in a separate color
23285 each channel is displayed using the same color scheme
23287 each channel is displayed using the rainbow color scheme
23289 each channel is displayed using the moreland color scheme
23291 each channel is displayed using the nebulae color scheme
23293 each channel is displayed using the fire color scheme
23295 each channel is displayed using the fiery color scheme
23297 each channel is displayed using the fruit color scheme
23299 each channel is displayed using the cool color scheme
23301 each channel is displayed using the magma color scheme
23303 each channel is displayed using the green color scheme
23305 each channel is displayed using the viridis color scheme
23307 each channel is displayed using the plasma color scheme
23309 each channel is displayed using the cividis color scheme
23311 each channel is displayed using the terrain color scheme
23314 Default value is @samp{channel}.
23317 Specify scale used for calculating intensity color values.
23319 It accepts the following values:
23324 square root, default
23335 Default value is @samp{sqrt}.
23338 Specify frequency scale.
23340 It accepts the following values:
23348 Default value is @samp{lin}.
23351 Set saturation modifier for displayed colors. Negative values provide
23352 alternative color scheme. @code{0} is no saturation at all.
23353 Saturation must be in [-10.0, 10.0] range.
23354 Default value is @code{1}.
23357 Set window function.
23359 It accepts the following values:
23384 Default value is @code{hann}.
23387 Set orientation of time vs frequency axis. Can be @code{vertical} or
23388 @code{horizontal}. Default is @code{vertical}.
23391 Set ratio of overlap window. Default value is @code{0}.
23392 When value is @code{1} overlap is set to recommended size for specific
23393 window function currently used.
23396 Set scale gain for calculating intensity color values.
23397 Default value is @code{1}.
23400 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
23403 Set color rotation, must be in [-1.0, 1.0] range.
23404 Default value is @code{0}.
23407 Set start frequency from which to display spectrogram. Default is @code{0}.
23410 Set stop frequency to which to display spectrogram. Default is @code{0}.
23413 Set upper frame rate limit. Default is @code{auto}, unlimited.
23416 Draw time and frequency axes and legends. Default is disabled.
23419 The usage is very similar to the showwaves filter; see the examples in that
23422 @subsection Examples
23426 Large window with logarithmic color scaling:
23428 showspectrum=s=1280x480:scale=log
23432 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
23434 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
23435 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
23439 @section showspectrumpic
23441 Convert input audio to a single video frame, representing the audio frequency
23444 The filter accepts the following options:
23448 Specify the video size for the output. For the syntax of this option, check the
23449 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23450 Default value is @code{4096x2048}.
23453 Specify display mode.
23455 It accepts the following values:
23458 all channels are displayed in the same row
23460 all channels are displayed in separate rows
23462 Default value is @samp{combined}.
23465 Specify display color mode.
23467 It accepts the following values:
23470 each channel is displayed in a separate color
23472 each channel is displayed using the same color scheme
23474 each channel is displayed using the rainbow color scheme
23476 each channel is displayed using the moreland color scheme
23478 each channel is displayed using the nebulae color scheme
23480 each channel is displayed using the fire color scheme
23482 each channel is displayed using the fiery color scheme
23484 each channel is displayed using the fruit color scheme
23486 each channel is displayed using the cool color scheme
23488 each channel is displayed using the magma color scheme
23490 each channel is displayed using the green color scheme
23492 each channel is displayed using the viridis color scheme
23494 each channel is displayed using the plasma color scheme
23496 each channel is displayed using the cividis color scheme
23498 each channel is displayed using the terrain color scheme
23500 Default value is @samp{intensity}.
23503 Specify scale used for calculating intensity color values.
23505 It accepts the following values:
23510 square root, default
23520 Default value is @samp{log}.
23523 Specify frequency scale.
23525 It accepts the following values:
23533 Default value is @samp{lin}.
23536 Set saturation modifier for displayed colors. Negative values provide
23537 alternative color scheme. @code{0} is no saturation at all.
23538 Saturation must be in [-10.0, 10.0] range.
23539 Default value is @code{1}.
23542 Set window function.
23544 It accepts the following values:
23568 Default value is @code{hann}.
23571 Set orientation of time vs frequency axis. Can be @code{vertical} or
23572 @code{horizontal}. Default is @code{vertical}.
23575 Set scale gain for calculating intensity color values.
23576 Default value is @code{1}.
23579 Draw time and frequency axes and legends. Default is enabled.
23582 Set color rotation, must be in [-1.0, 1.0] range.
23583 Default value is @code{0}.
23586 Set start frequency from which to display spectrogram. Default is @code{0}.
23589 Set stop frequency to which to display spectrogram. Default is @code{0}.
23592 @subsection Examples
23596 Extract an audio spectrogram of a whole audio track
23597 in a 1024x1024 picture using @command{ffmpeg}:
23599 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
23603 @section showvolume
23605 Convert input audio volume to a video output.
23607 The filter accepts the following options:
23614 Set border width, allowed range is [0, 5]. Default is 1.
23617 Set channel width, allowed range is [80, 8192]. Default is 400.
23620 Set channel height, allowed range is [1, 900]. Default is 20.
23623 Set fade, allowed range is [0, 1]. Default is 0.95.
23626 Set volume color expression.
23628 The expression can use the following variables:
23632 Current max volume of channel in dB.
23638 Current channel number, starting from 0.
23642 If set, displays channel names. Default is enabled.
23645 If set, displays volume values. Default is enabled.
23648 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
23649 default is @code{h}.
23652 Set step size, allowed range is [0, 5]. Default is 0, which means
23656 Set background opacity, allowed range is [0, 1]. Default is 0.
23659 Set metering mode, can be peak: @code{p} or rms: @code{r},
23660 default is @code{p}.
23663 Set display scale, can be linear: @code{lin} or log: @code{log},
23664 default is @code{lin}.
23668 If set to > 0., display a line for the max level
23669 in the previous seconds.
23670 default is disabled: @code{0.}
23673 The color of the max line. Use when @code{dm} option is set to > 0.
23674 default is: @code{orange}
23679 Convert input audio to a video output, representing the samples waves.
23681 The filter accepts the following options:
23685 Specify the video size for the output. For the syntax of this option, check the
23686 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23687 Default value is @code{600x240}.
23692 Available values are:
23695 Draw a point for each sample.
23698 Draw a vertical line for each sample.
23701 Draw a point for each sample and a line between them.
23704 Draw a centered vertical line for each sample.
23707 Default value is @code{point}.
23710 Set the number of samples which are printed on the same column. A
23711 larger value will decrease the frame rate. Must be a positive
23712 integer. This option can be set only if the value for @var{rate}
23713 is not explicitly specified.
23716 Set the (approximate) output frame rate. This is done by setting the
23717 option @var{n}. Default value is "25".
23719 @item split_channels
23720 Set if channels should be drawn separately or overlap. Default value is 0.
23723 Set colors separated by '|' which are going to be used for drawing of each channel.
23726 Set amplitude scale.
23728 Available values are:
23746 Set the draw mode. This is mostly useful to set for high @var{n}.
23748 Available values are:
23751 Scale pixel values for each drawn sample.
23754 Draw every sample directly.
23757 Default value is @code{scale}.
23760 @subsection Examples
23764 Output the input file audio and the corresponding video representation
23767 amovie=a.mp3,asplit[out0],showwaves[out1]
23771 Create a synthetic signal and show it with showwaves, forcing a
23772 frame rate of 30 frames per second:
23774 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
23778 @section showwavespic
23780 Convert input audio to a single video frame, representing the samples waves.
23782 The filter accepts the following options:
23786 Specify the video size for the output. For the syntax of this option, check the
23787 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23788 Default value is @code{600x240}.
23790 @item split_channels
23791 Set if channels should be drawn separately or overlap. Default value is 0.
23794 Set colors separated by '|' which are going to be used for drawing of each channel.
23797 Set amplitude scale.
23799 Available values are:
23819 Available values are:
23822 Scale pixel values for each drawn sample.
23825 Draw every sample directly.
23828 Default value is @code{scale}.
23831 @subsection Examples
23835 Extract a channel split representation of the wave form of a whole audio track
23836 in a 1024x800 picture using @command{ffmpeg}:
23838 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
23842 @section sidedata, asidedata
23844 Delete frame side data, or select frames based on it.
23846 This filter accepts the following options:
23850 Set mode of operation of the filter.
23852 Can be one of the following:
23856 Select every frame with side data of @code{type}.
23859 Delete side data of @code{type}. If @code{type} is not set, delete all side
23865 Set side data type used with all modes. Must be set for @code{select} mode. For
23866 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
23867 in @file{libavutil/frame.h}. For example, to choose
23868 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
23872 @section spectrumsynth
23874 Synthesize audio from 2 input video spectrums, first input stream represents
23875 magnitude across time and second represents phase across time.
23876 The filter will transform from frequency domain as displayed in videos back
23877 to time domain as presented in audio output.
23879 This filter is primarily created for reversing processed @ref{showspectrum}
23880 filter outputs, but can synthesize sound from other spectrograms too.
23881 But in such case results are going to be poor if the phase data is not
23882 available, because in such cases phase data need to be recreated, usually
23883 it's just recreated from random noise.
23884 For best results use gray only output (@code{channel} color mode in
23885 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
23886 @code{lin} scale for phase video. To produce phase, for 2nd video, use
23887 @code{data} option. Inputs videos should generally use @code{fullframe}
23888 slide mode as that saves resources needed for decoding video.
23890 The filter accepts the following options:
23894 Specify sample rate of output audio, the sample rate of audio from which
23895 spectrum was generated may differ.
23898 Set number of channels represented in input video spectrums.
23901 Set scale which was used when generating magnitude input spectrum.
23902 Can be @code{lin} or @code{log}. Default is @code{log}.
23905 Set slide which was used when generating inputs spectrums.
23906 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
23907 Default is @code{fullframe}.
23910 Set window function used for resynthesis.
23913 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
23914 which means optimal overlap for selected window function will be picked.
23917 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
23918 Default is @code{vertical}.
23921 @subsection Examples
23925 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
23926 then resynthesize videos back to audio with spectrumsynth:
23928 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
23929 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
23930 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
23934 @section split, asplit
23936 Split input into several identical outputs.
23938 @code{asplit} works with audio input, @code{split} with video.
23940 The filter accepts a single parameter which specifies the number of outputs. If
23941 unspecified, it defaults to 2.
23943 @subsection Examples
23947 Create two separate outputs from the same input:
23949 [in] split [out0][out1]
23953 To create 3 or more outputs, you need to specify the number of
23956 [in] asplit=3 [out0][out1][out2]
23960 Create two separate outputs from the same input, one cropped and
23963 [in] split [splitout1][splitout2];
23964 [splitout1] crop=100:100:0:0 [cropout];
23965 [splitout2] pad=200:200:100:100 [padout];
23969 Create 5 copies of the input audio with @command{ffmpeg}:
23971 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
23977 Receive commands sent through a libzmq client, and forward them to
23978 filters in the filtergraph.
23980 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
23981 must be inserted between two video filters, @code{azmq} between two
23982 audio filters. Both are capable to send messages to any filter type.
23984 To enable these filters you need to install the libzmq library and
23985 headers and configure FFmpeg with @code{--enable-libzmq}.
23987 For more information about libzmq see:
23988 @url{http://www.zeromq.org/}
23990 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
23991 receives messages sent through a network interface defined by the
23992 @option{bind_address} (or the abbreviation "@option{b}") option.
23993 Default value of this option is @file{tcp://localhost:5555}. You may
23994 want to alter this value to your needs, but do not forget to escape any
23995 ':' signs (see @ref{filtergraph escaping}).
23997 The received message must be in the form:
23999 @var{TARGET} @var{COMMAND} [@var{ARG}]
24002 @var{TARGET} specifies the target of the command, usually the name of
24003 the filter class or a specific filter instance name. The default
24004 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
24005 but you can override this by using the @samp{filter_name@@id} syntax
24006 (see @ref{Filtergraph syntax}).
24008 @var{COMMAND} specifies the name of the command for the target filter.
24010 @var{ARG} is optional and specifies the optional argument list for the
24011 given @var{COMMAND}.
24013 Upon reception, the message is processed and the corresponding command
24014 is injected into the filtergraph. Depending on the result, the filter
24015 will send a reply to the client, adopting the format:
24017 @var{ERROR_CODE} @var{ERROR_REASON}
24021 @var{MESSAGE} is optional.
24023 @subsection Examples
24025 Look at @file{tools/zmqsend} for an example of a zmq client which can
24026 be used to send commands processed by these filters.
24028 Consider the following filtergraph generated by @command{ffplay}.
24029 In this example the last overlay filter has an instance name. All other
24030 filters will have default instance names.
24033 ffplay -dumpgraph 1 -f lavfi "
24034 color=s=100x100:c=red [l];
24035 color=s=100x100:c=blue [r];
24036 nullsrc=s=200x100, zmq [bg];
24037 [bg][l] overlay [bg+l];
24038 [bg+l][r] overlay@@my=x=100 "
24041 To change the color of the left side of the video, the following
24042 command can be used:
24044 echo Parsed_color_0 c yellow | tools/zmqsend
24047 To change the right side:
24049 echo Parsed_color_1 c pink | tools/zmqsend
24052 To change the position of the right side:
24054 echo overlay@@my x 150 | tools/zmqsend
24058 @c man end MULTIMEDIA FILTERS
24060 @chapter Multimedia Sources
24061 @c man begin MULTIMEDIA SOURCES
24063 Below is a description of the currently available multimedia sources.
24067 This is the same as @ref{movie} source, except it selects an audio
24073 Read audio and/or video stream(s) from a movie container.
24075 It accepts the following parameters:
24079 The name of the resource to read (not necessarily a file; it can also be a
24080 device or a stream accessed through some protocol).
24082 @item format_name, f
24083 Specifies the format assumed for the movie to read, and can be either
24084 the name of a container or an input device. If not specified, the
24085 format is guessed from @var{movie_name} or by probing.
24087 @item seek_point, sp
24088 Specifies the seek point in seconds. The frames will be output
24089 starting from this seek point. The parameter is evaluated with
24090 @code{av_strtod}, so the numerical value may be suffixed by an IS
24091 postfix. The default value is "0".
24094 Specifies the streams to read. Several streams can be specified,
24095 separated by "+". The source will then have as many outputs, in the
24096 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
24097 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
24098 respectively the default (best suited) video and audio stream. Default
24099 is "dv", or "da" if the filter is called as "amovie".
24101 @item stream_index, si
24102 Specifies the index of the video stream to read. If the value is -1,
24103 the most suitable video stream will be automatically selected. The default
24104 value is "-1". Deprecated. If the filter is called "amovie", it will select
24105 audio instead of video.
24108 Specifies how many times to read the stream in sequence.
24109 If the value is 0, the stream will be looped infinitely.
24110 Default value is "1".
24112 Note that when the movie is looped the source timestamps are not
24113 changed, so it will generate non monotonically increasing timestamps.
24115 @item discontinuity
24116 Specifies the time difference between frames above which the point is
24117 considered a timestamp discontinuity which is removed by adjusting the later
24121 It allows overlaying a second video on top of the main input of
24122 a filtergraph, as shown in this graph:
24124 input -----------> deltapts0 --> overlay --> output
24127 movie --> scale--> deltapts1 -------+
24129 @subsection Examples
24133 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
24134 on top of the input labelled "in":
24136 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
24137 [in] setpts=PTS-STARTPTS [main];
24138 [main][over] overlay=16:16 [out]
24142 Read from a video4linux2 device, and overlay it on top of the input
24145 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
24146 [in] setpts=PTS-STARTPTS [main];
24147 [main][over] overlay=16:16 [out]
24151 Read the first video stream and the audio stream with id 0x81 from
24152 dvd.vob; the video is connected to the pad named "video" and the audio is
24153 connected to the pad named "audio":
24155 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
24159 @subsection Commands
24161 Both movie and amovie support the following commands:
24164 Perform seek using "av_seek_frame".
24165 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
24168 @var{stream_index}: If stream_index is -1, a default
24169 stream is selected, and @var{timestamp} is automatically converted
24170 from AV_TIME_BASE units to the stream specific time_base.
24172 @var{timestamp}: Timestamp in AVStream.time_base units
24173 or, if no stream is specified, in AV_TIME_BASE units.
24175 @var{flags}: Flags which select direction and seeking mode.
24179 Get movie duration in AV_TIME_BASE units.
24183 @c man end MULTIMEDIA SOURCES