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. The
4687 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
4688 is set on the first frame whose timestamp equals or exceeds the detection
4689 duration and it contains the timestamp of the first frame of the silence.
4691 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
4692 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
4693 keys are set on the first frame after the silence. If @option{mono} is
4694 enabled, and each channel is evaluated separately, the @code{.X}
4695 suffixed keys are used, and @code{X} corresponds to the channel number.
4697 The filter accepts the following options:
4701 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
4702 specified value) or amplitude ratio. Default is -60dB, or 0.001.
4705 Set silence duration until notification (default is 2 seconds). See
4706 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4707 for the accepted syntax.
4710 Process each channel separately, instead of combined. By default is disabled.
4713 @subsection Examples
4717 Detect 5 seconds of silence with -50dB noise tolerance:
4719 silencedetect=n=-50dB:d=5
4723 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
4724 tolerance in @file{silence.mp3}:
4726 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
4730 @section silenceremove
4732 Remove silence from the beginning, middle or end of the audio.
4734 The filter accepts the following options:
4738 This value is used to indicate if audio should be trimmed at beginning of
4739 the audio. A value of zero indicates no silence should be trimmed from the
4740 beginning. When specifying a non-zero value, it trims audio up until it
4741 finds non-silence. Normally, when trimming silence from beginning of audio
4742 the @var{start_periods} will be @code{1} but it can be increased to higher
4743 values to trim all audio up to specific count of non-silence periods.
4744 Default value is @code{0}.
4746 @item start_duration
4747 Specify the amount of time that non-silence must be detected before it stops
4748 trimming audio. By increasing the duration, bursts of noises can be treated
4749 as silence and trimmed off. Default value is @code{0}.
4751 @item start_threshold
4752 This indicates what sample value should be treated as silence. For digital
4753 audio, a value of @code{0} may be fine but for audio recorded from analog,
4754 you may wish to increase the value to account for background noise.
4755 Can be specified in dB (in case "dB" is appended to the specified value)
4756 or amplitude ratio. Default value is @code{0}.
4759 Specify max duration of silence at beginning that will be kept after
4760 trimming. Default is 0, which is equal to trimming all samples detected
4764 Specify mode of detection of silence end in start of multi-channel audio.
4765 Can be @var{any} or @var{all}. Default is @var{any}.
4766 With @var{any}, any sample that is detected as non-silence will cause
4767 stopped trimming of silence.
4768 With @var{all}, only if all channels are detected as non-silence will cause
4769 stopped trimming of silence.
4772 Set the count for trimming silence from the end of audio.
4773 To remove silence from the middle of a file, specify a @var{stop_periods}
4774 that is negative. This value is then treated as a positive value and is
4775 used to indicate the effect should restart processing as specified by
4776 @var{start_periods}, making it suitable for removing periods of silence
4777 in the middle of the audio.
4778 Default value is @code{0}.
4781 Specify a duration of silence that must exist before audio is not copied any
4782 more. By specifying a higher duration, silence that is wanted can be left in
4784 Default value is @code{0}.
4786 @item stop_threshold
4787 This is the same as @option{start_threshold} but for trimming silence from
4789 Can be specified in dB (in case "dB" is appended to the specified value)
4790 or amplitude ratio. Default value is @code{0}.
4793 Specify max duration of silence at end that will be kept after
4794 trimming. Default is 0, which is equal to trimming all samples detected
4798 Specify mode of detection of silence start in end of multi-channel audio.
4799 Can be @var{any} or @var{all}. Default is @var{any}.
4800 With @var{any}, any sample that is detected as non-silence will cause
4801 stopped trimming of silence.
4802 With @var{all}, only if all channels are detected as non-silence will cause
4803 stopped trimming of silence.
4806 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
4807 and works better with digital silence which is exactly 0.
4808 Default value is @code{rms}.
4811 Set duration in number of seconds used to calculate size of window in number
4812 of samples for detecting silence.
4813 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
4816 @subsection Examples
4820 The following example shows how this filter can be used to start a recording
4821 that does not contain the delay at the start which usually occurs between
4822 pressing the record button and the start of the performance:
4824 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
4828 Trim all silence encountered from beginning to end where there is more than 1
4829 second of silence in audio:
4831 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
4835 Trim all digital silence samples, using peak detection, from beginning to end
4836 where there is more than 0 samples of digital silence in audio and digital
4837 silence is detected in all channels at same positions in stream:
4839 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
4845 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
4846 loudspeakers around the user for binaural listening via headphones (audio
4847 formats up to 9 channels supported).
4848 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
4849 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
4850 Austrian Academy of Sciences.
4852 To enable compilation of this filter you need to configure FFmpeg with
4853 @code{--enable-libmysofa}.
4855 The filter accepts the following options:
4859 Set the SOFA file used for rendering.
4862 Set gain applied to audio. Value is in dB. Default is 0.
4865 Set rotation of virtual loudspeakers in deg. Default is 0.
4868 Set elevation of virtual speakers in deg. Default is 0.
4871 Set distance in meters between loudspeakers and the listener with near-field
4872 HRTFs. Default is 1.
4875 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4876 processing audio in time domain which is slow.
4877 @var{freq} is processing audio in frequency domain which is fast.
4878 Default is @var{freq}.
4881 Set custom positions of virtual loudspeakers. Syntax for this option is:
4882 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
4883 Each virtual loudspeaker is described with short channel name following with
4884 azimuth and elevation in degrees.
4885 Each virtual loudspeaker description is separated by '|'.
4886 For example to override front left and front right channel positions use:
4887 'speakers=FL 45 15|FR 345 15'.
4888 Descriptions with unrecognised channel names are ignored.
4891 Set custom gain for LFE channels. Value is in dB. Default is 0.
4894 Set custom frame size in number of samples. Default is 1024.
4895 Allowed range is from 1024 to 96000. Only used if option @samp{type}
4896 is set to @var{freq}.
4899 Should all IRs be normalized upon importing SOFA file.
4900 By default is enabled.
4903 Should nearest IRs be interpolated with neighbor IRs if exact position
4904 does not match. By default is disabled.
4907 Minphase all IRs upon loading of SOFA file. By default is disabled.
4910 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
4913 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
4916 @subsection Examples
4920 Using ClubFritz6 sofa file:
4922 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
4926 Using ClubFritz12 sofa file and bigger radius with small rotation:
4928 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
4932 Similar as above but with custom speaker positions for front left, front right, back left and back right
4933 and also with custom gain:
4935 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
4939 @section stereotools
4941 This filter has some handy utilities to manage stereo signals, for converting
4942 M/S stereo recordings to L/R signal while having control over the parameters
4943 or spreading the stereo image of master track.
4945 The filter accepts the following options:
4949 Set input level before filtering for both channels. Defaults is 1.
4950 Allowed range is from 0.015625 to 64.
4953 Set output level after filtering for both channels. Defaults is 1.
4954 Allowed range is from 0.015625 to 64.
4957 Set input balance between both channels. Default is 0.
4958 Allowed range is from -1 to 1.
4961 Set output balance between both channels. Default is 0.
4962 Allowed range is from -1 to 1.
4965 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
4966 clipping. Disabled by default.
4969 Mute the left channel. Disabled by default.
4972 Mute the right channel. Disabled by default.
4975 Change the phase of the left channel. Disabled by default.
4978 Change the phase of the right channel. Disabled by default.
4981 Set stereo mode. Available values are:
4985 Left/Right to Left/Right, this is default.
4988 Left/Right to Mid/Side.
4991 Mid/Side to Left/Right.
4994 Left/Right to Left/Left.
4997 Left/Right to Right/Right.
5000 Left/Right to Left + Right.
5003 Left/Right to Right/Left.
5006 Mid/Side to Left/Left.
5009 Mid/Side to Right/Right.
5013 Set level of side signal. Default is 1.
5014 Allowed range is from 0.015625 to 64.
5017 Set balance of side signal. Default is 0.
5018 Allowed range is from -1 to 1.
5021 Set level of the middle signal. Default is 1.
5022 Allowed range is from 0.015625 to 64.
5025 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5028 Set stereo base between mono and inversed channels. Default is 0.
5029 Allowed range is from -1 to 1.
5032 Set delay in milliseconds how much to delay left from right channel and
5033 vice versa. Default is 0. Allowed range is from -20 to 20.
5036 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5039 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5041 @item bmode_in, bmode_out
5042 Set balance mode for balance_in/balance_out option.
5044 Can be one of the following:
5048 Classic balance mode. Attenuate one channel at time.
5049 Gain is raised up to 1.
5052 Similar as classic mode above but gain is raised up to 2.
5055 Equal power distribution, from -6dB to +6dB range.
5059 @subsection Examples
5063 Apply karaoke like effect:
5065 stereotools=mlev=0.015625
5069 Convert M/S signal to L/R:
5071 "stereotools=mode=ms>lr"
5075 @section stereowiden
5077 This filter enhance the stereo effect by suppressing signal common to both
5078 channels and by delaying the signal of left into right and vice versa,
5079 thereby widening the stereo effect.
5081 The filter accepts the following options:
5085 Time in milliseconds of the delay of left signal into right and vice versa.
5086 Default is 20 milliseconds.
5089 Amount of gain in delayed signal into right and vice versa. Gives a delay
5090 effect of left signal in right output and vice versa which gives widening
5091 effect. Default is 0.3.
5094 Cross feed of left into right with inverted phase. This helps in suppressing
5095 the mono. If the value is 1 it will cancel all the signal common to both
5096 channels. Default is 0.3.
5099 Set level of input signal of original channel. Default is 0.8.
5102 @section superequalizer
5103 Apply 18 band equalizer.
5105 The filter accepts the following options:
5112 Set 131Hz band gain.
5114 Set 185Hz band gain.
5116 Set 262Hz band gain.
5118 Set 370Hz band gain.
5120 Set 523Hz band gain.
5122 Set 740Hz band gain.
5124 Set 1047Hz band gain.
5126 Set 1480Hz band gain.
5128 Set 2093Hz band gain.
5130 Set 2960Hz band gain.
5132 Set 4186Hz band gain.
5134 Set 5920Hz band gain.
5136 Set 8372Hz band gain.
5138 Set 11840Hz band gain.
5140 Set 16744Hz band gain.
5142 Set 20000Hz band gain.
5146 Apply audio surround upmix filter.
5148 This filter allows to produce multichannel output from audio stream.
5150 The filter accepts the following options:
5154 Set output channel layout. By default, this is @var{5.1}.
5156 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5157 for the required syntax.
5160 Set input channel layout. By default, this is @var{stereo}.
5162 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5163 for the required syntax.
5166 Set input volume level. By default, this is @var{1}.
5169 Set output volume level. By default, this is @var{1}.
5172 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5175 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5178 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5181 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5182 In @var{add} mode, LFE channel is created from input audio and added to output.
5183 In @var{sub} mode, LFE channel is created from input audio and added to output but
5184 also all non-LFE output channels are subtracted with output LFE channel.
5187 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5188 Default is @var{90}.
5191 Set front center input volume. By default, this is @var{1}.
5194 Set front center output volume. By default, this is @var{1}.
5197 Set front left input volume. By default, this is @var{1}.
5200 Set front left output volume. By default, this is @var{1}.
5203 Set front right input volume. By default, this is @var{1}.
5206 Set front right output volume. By default, this is @var{1}.
5209 Set side left input volume. By default, this is @var{1}.
5212 Set side left output volume. By default, this is @var{1}.
5215 Set side right input volume. By default, this is @var{1}.
5218 Set side right output volume. By default, this is @var{1}.
5221 Set back left input volume. By default, this is @var{1}.
5224 Set back left output volume. By default, this is @var{1}.
5227 Set back right input volume. By default, this is @var{1}.
5230 Set back right output volume. By default, this is @var{1}.
5233 Set back center input volume. By default, this is @var{1}.
5236 Set back center output volume. By default, this is @var{1}.
5239 Set LFE input volume. By default, this is @var{1}.
5242 Set LFE output volume. By default, this is @var{1}.
5245 Set spread usage of stereo image across X axis for all channels.
5248 Set spread usage of stereo image across Y axis for all channels.
5250 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5251 Set spread usage of stereo image across X axis for each channel.
5253 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5254 Set spread usage of stereo image across Y axis for each channel.
5257 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5260 Set window function.
5262 It accepts the following values:
5285 Default is @code{hann}.
5288 Set window overlap. If set to 1, the recommended overlap for selected
5289 window function will be picked. Default is @code{0.5}.
5292 @section treble, highshelf
5294 Boost or cut treble (upper) frequencies of the audio using a two-pole
5295 shelving filter with a response similar to that of a standard
5296 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
5298 The filter accepts the following options:
5302 Give the gain at whichever is the lower of ~22 kHz and the
5303 Nyquist frequency. Its useful range is about -20 (for a large cut)
5304 to +20 (for a large boost). Beware of clipping when using a positive gain.
5307 Set the filter's central frequency and so can be used
5308 to extend or reduce the frequency range to be boosted or cut.
5309 The default value is @code{3000} Hz.
5312 Set method to specify band-width of filter.
5327 Determine how steep is the filter's shelf transition.
5330 How much to use filtered signal in output. Default is 1.
5331 Range is between 0 and 1.
5334 Specify which channels to filter, by default all available are filtered.
5337 @subsection Commands
5339 This filter supports the following commands:
5342 Change treble frequency.
5343 Syntax for the command is : "@var{frequency}"
5346 Change treble width_type.
5347 Syntax for the command is : "@var{width_type}"
5350 Change treble width.
5351 Syntax for the command is : "@var{width}"
5355 Syntax for the command is : "@var{gain}"
5359 Syntax for the command is : "@var{mix}"
5364 Sinusoidal amplitude modulation.
5366 The filter accepts the following options:
5370 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
5371 (20 Hz or lower) will result in a tremolo effect.
5372 This filter may also be used as a ring modulator by specifying
5373 a modulation frequency higher than 20 Hz.
5374 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5377 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5378 Default value is 0.5.
5383 Sinusoidal phase modulation.
5385 The filter accepts the following options:
5389 Modulation frequency in Hertz.
5390 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5393 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5394 Default value is 0.5.
5399 Adjust the input audio volume.
5401 It accepts the following parameters:
5405 Set audio volume expression.
5407 Output values are clipped to the maximum value.
5409 The output audio volume is given by the relation:
5411 @var{output_volume} = @var{volume} * @var{input_volume}
5414 The default value for @var{volume} is "1.0".
5417 This parameter represents the mathematical precision.
5419 It determines which input sample formats will be allowed, which affects the
5420 precision of the volume scaling.
5424 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
5426 32-bit floating-point; this limits input sample format to FLT. (default)
5428 64-bit floating-point; this limits input sample format to DBL.
5432 Choose the behaviour on encountering ReplayGain side data in input frames.
5436 Remove ReplayGain side data, ignoring its contents (the default).
5439 Ignore ReplayGain side data, but leave it in the frame.
5442 Prefer the track gain, if present.
5445 Prefer the album gain, if present.
5448 @item replaygain_preamp
5449 Pre-amplification gain in dB to apply to the selected replaygain gain.
5451 Default value for @var{replaygain_preamp} is 0.0.
5454 Set when the volume expression is evaluated.
5456 It accepts the following values:
5459 only evaluate expression once during the filter initialization, or
5460 when the @samp{volume} command is sent
5463 evaluate expression for each incoming frame
5466 Default value is @samp{once}.
5469 The volume expression can contain the following parameters.
5473 frame number (starting at zero)
5476 @item nb_consumed_samples
5477 number of samples consumed by the filter
5479 number of samples in the current frame
5481 original frame position in the file
5487 PTS at start of stream
5489 time at start of stream
5495 last set volume value
5498 Note that when @option{eval} is set to @samp{once} only the
5499 @var{sample_rate} and @var{tb} variables are available, all other
5500 variables will evaluate to NAN.
5502 @subsection Commands
5504 This filter supports the following commands:
5507 Modify the volume expression.
5508 The command accepts the same syntax of the corresponding option.
5510 If the specified expression is not valid, it is kept at its current
5512 @item replaygain_noclip
5513 Prevent clipping by limiting the gain applied.
5515 Default value for @var{replaygain_noclip} is 1.
5519 @subsection Examples
5523 Halve the input audio volume:
5527 volume=volume=-6.0206dB
5530 In all the above example the named key for @option{volume} can be
5531 omitted, for example like in:
5537 Increase input audio power by 6 decibels using fixed-point precision:
5539 volume=volume=6dB:precision=fixed
5543 Fade volume after time 10 with an annihilation period of 5 seconds:
5545 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
5549 @section volumedetect
5551 Detect the volume of the input video.
5553 The filter has no parameters. The input is not modified. Statistics about
5554 the volume will be printed in the log when the input stream end is reached.
5556 In particular it will show the mean volume (root mean square), maximum
5557 volume (on a per-sample basis), and the beginning of a histogram of the
5558 registered volume values (from the maximum value to a cumulated 1/1000 of
5561 All volumes are in decibels relative to the maximum PCM value.
5563 @subsection Examples
5565 Here is an excerpt of the output:
5567 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
5568 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
5569 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
5570 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
5571 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
5572 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
5573 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
5574 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
5575 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
5581 The mean square energy is approximately -27 dB, or 10^-2.7.
5583 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
5585 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
5588 In other words, raising the volume by +4 dB does not cause any clipping,
5589 raising it by +5 dB causes clipping for 6 samples, etc.
5591 @c man end AUDIO FILTERS
5593 @chapter Audio Sources
5594 @c man begin AUDIO SOURCES
5596 Below is a description of the currently available audio sources.
5600 Buffer audio frames, and make them available to the filter chain.
5602 This source is mainly intended for a programmatic use, in particular
5603 through the interface defined in @file{libavfilter/asrc_abuffer.h}.
5605 It accepts the following parameters:
5609 The timebase which will be used for timestamps of submitted frames. It must be
5610 either a floating-point number or in @var{numerator}/@var{denominator} form.
5613 The sample rate of the incoming audio buffers.
5616 The sample format of the incoming audio buffers.
5617 Either a sample format name or its corresponding integer representation from
5618 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
5620 @item channel_layout
5621 The channel layout of the incoming audio buffers.
5622 Either a channel layout name from channel_layout_map in
5623 @file{libavutil/channel_layout.c} or its corresponding integer representation
5624 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
5627 The number of channels of the incoming audio buffers.
5628 If both @var{channels} and @var{channel_layout} are specified, then they
5633 @subsection Examples
5636 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
5639 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
5640 Since the sample format with name "s16p" corresponds to the number
5641 6 and the "stereo" channel layout corresponds to the value 0x3, this is
5644 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
5649 Generate an audio signal specified by an expression.
5651 This source accepts in input one or more expressions (one for each
5652 channel), which are evaluated and used to generate a corresponding
5655 This source accepts the following options:
5659 Set the '|'-separated expressions list for each separate channel. In case the
5660 @option{channel_layout} option is not specified, the selected channel layout
5661 depends on the number of provided expressions. Otherwise the last
5662 specified expression is applied to the remaining output channels.
5664 @item channel_layout, c
5665 Set the channel layout. The number of channels in the specified layout
5666 must be equal to the number of specified expressions.
5669 Set the minimum duration of the sourced audio. See
5670 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5671 for the accepted syntax.
5672 Note that the resulting duration may be greater than the specified
5673 duration, as the generated audio is always cut at the end of a
5676 If not specified, or the expressed duration is negative, the audio is
5677 supposed to be generated forever.
5680 Set the number of samples per channel per each output frame,
5683 @item sample_rate, s
5684 Specify the sample rate, default to 44100.
5687 Each expression in @var{exprs} can contain the following constants:
5691 number of the evaluated sample, starting from 0
5694 time of the evaluated sample expressed in seconds, starting from 0
5701 @subsection Examples
5711 Generate a sin signal with frequency of 440 Hz, set sample rate to
5714 aevalsrc="sin(440*2*PI*t):s=8000"
5718 Generate a two channels signal, specify the channel layout (Front
5719 Center + Back Center) explicitly:
5721 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
5725 Generate white noise:
5727 aevalsrc="-2+random(0)"
5731 Generate an amplitude modulated signal:
5733 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
5737 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
5739 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
5746 The null audio source, return unprocessed audio frames. It is mainly useful
5747 as a template and to be employed in analysis / debugging tools, or as
5748 the source for filters which ignore the input data (for example the sox
5751 This source accepts the following options:
5755 @item channel_layout, cl
5757 Specifies the channel layout, and can be either an integer or a string
5758 representing a channel layout. The default value of @var{channel_layout}
5761 Check the channel_layout_map definition in
5762 @file{libavutil/channel_layout.c} for the mapping between strings and
5763 channel layout values.
5765 @item sample_rate, r
5766 Specifies the sample rate, and defaults to 44100.
5769 Set the number of samples per requested frames.
5773 @subsection Examples
5777 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
5779 anullsrc=r=48000:cl=4
5783 Do the same operation with a more obvious syntax:
5785 anullsrc=r=48000:cl=mono
5789 All the parameters need to be explicitly defined.
5793 Synthesize a voice utterance using the libflite library.
5795 To enable compilation of this filter you need to configure FFmpeg with
5796 @code{--enable-libflite}.
5798 Note that versions of the flite library prior to 2.0 are not thread-safe.
5800 The filter accepts the following options:
5805 If set to 1, list the names of the available voices and exit
5806 immediately. Default value is 0.
5809 Set the maximum number of samples per frame. Default value is 512.
5812 Set the filename containing the text to speak.
5815 Set the text to speak.
5818 Set the voice to use for the speech synthesis. Default value is
5819 @code{kal}. See also the @var{list_voices} option.
5822 @subsection Examples
5826 Read from file @file{speech.txt}, and synthesize the text using the
5827 standard flite voice:
5829 flite=textfile=speech.txt
5833 Read the specified text selecting the @code{slt} voice:
5835 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
5839 Input text to ffmpeg:
5841 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
5845 Make @file{ffplay} speak the specified text, using @code{flite} and
5846 the @code{lavfi} device:
5848 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
5852 For more information about libflite, check:
5853 @url{http://www.festvox.org/flite/}
5857 Generate a noise audio signal.
5859 The filter accepts the following options:
5862 @item sample_rate, r
5863 Specify the sample rate. Default value is 48000 Hz.
5866 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
5870 Specify the duration of the generated audio stream. Not specifying this option
5871 results in noise with an infinite length.
5873 @item color, colour, c
5874 Specify the color of noise. Available noise colors are white, pink, brown,
5875 blue and violet. Default color is white.
5878 Specify a value used to seed the PRNG.
5881 Set the number of samples per each output frame, default is 1024.
5884 @subsection Examples
5889 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
5891 anoisesrc=d=60:c=pink:r=44100:a=0.5
5897 Generate odd-tap Hilbert transform FIR coefficients.
5899 The resulting stream can be used with @ref{afir} filter for phase-shifting
5900 the signal by 90 degrees.
5902 This is used in many matrix coding schemes and for analytic signal generation.
5903 The process is often written as a multiplication by i (or j), the imaginary unit.
5905 The filter accepts the following options:
5909 @item sample_rate, s
5910 Set sample rate, default is 44100.
5913 Set length of FIR filter, default is 22051.
5916 Set number of samples per each frame.
5919 Set window function to be used when generating FIR coefficients.
5924 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
5926 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
5928 The filter accepts the following options:
5931 @item sample_rate, r
5932 Set sample rate, default is 44100.
5935 Set number of samples per each frame. Default is 1024.
5938 Set high-pass frequency. Default is 0.
5941 Set low-pass frequency. Default is 0.
5942 If high-pass frequency is lower than low-pass frequency and low-pass frequency
5943 is higher than 0 then filter will create band-pass filter coefficients,
5944 otherwise band-reject filter coefficients.
5947 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
5950 Set Kaiser window beta.
5953 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
5956 Enable rounding, by default is disabled.
5959 Set number of taps for high-pass filter.
5962 Set number of taps for low-pass filter.
5967 Generate an audio signal made of a sine wave with amplitude 1/8.
5969 The audio signal is bit-exact.
5971 The filter accepts the following options:
5976 Set the carrier frequency. Default is 440 Hz.
5978 @item beep_factor, b
5979 Enable a periodic beep every second with frequency @var{beep_factor} times
5980 the carrier frequency. Default is 0, meaning the beep is disabled.
5982 @item sample_rate, r
5983 Specify the sample rate, default is 44100.
5986 Specify the duration of the generated audio stream.
5988 @item samples_per_frame
5989 Set the number of samples per output frame.
5991 The expression can contain the following constants:
5995 The (sequential) number of the output audio frame, starting from 0.
5998 The PTS (Presentation TimeStamp) of the output audio frame,
5999 expressed in @var{TB} units.
6002 The PTS of the output audio frame, expressed in seconds.
6005 The timebase of the output audio frames.
6008 Default is @code{1024}.
6011 @subsection Examples
6016 Generate a simple 440 Hz sine wave:
6022 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6026 sine=frequency=220:beep_factor=4:duration=5
6030 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6033 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6037 @c man end AUDIO SOURCES
6039 @chapter Audio Sinks
6040 @c man begin AUDIO SINKS
6042 Below is a description of the currently available audio sinks.
6044 @section abuffersink
6046 Buffer audio frames, and make them available to the end of filter chain.
6048 This sink is mainly intended for programmatic use, in particular
6049 through the interface defined in @file{libavfilter/buffersink.h}
6050 or the options system.
6052 It accepts a pointer to an AVABufferSinkContext structure, which
6053 defines the incoming buffers' formats, to be passed as the opaque
6054 parameter to @code{avfilter_init_filter} for initialization.
6057 Null audio sink; do absolutely nothing with the input audio. It is
6058 mainly useful as a template and for use in analysis / debugging
6061 @c man end AUDIO SINKS
6063 @chapter Video Filters
6064 @c man begin VIDEO FILTERS
6066 When you configure your FFmpeg build, you can disable any of the
6067 existing filters using @code{--disable-filters}.
6068 The configure output will show the video filters included in your
6071 Below is a description of the currently available video filters.
6075 Mark a region of interest in a video frame.
6077 The frame data is passed through unchanged, but metadata is attached
6078 to the frame indicating regions of interest which can affect the
6079 behaviour of later encoding. Multiple regions can be marked by
6080 applying the filter multiple times.
6084 Region distance in pixels from the left edge of the frame.
6086 Region distance in pixels from the top edge of the frame.
6088 Region width in pixels.
6090 Region height in pixels.
6092 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6093 and may contain the following variables:
6096 Width of the input frame.
6098 Height of the input frame.
6102 Quantisation offset to apply within the region.
6104 This must be a real value in the range -1 to +1. A value of zero
6105 indicates no quality change. A negative value asks for better quality
6106 (less quantisation), while a positive value asks for worse quality
6107 (greater quantisation).
6109 The range is calibrated so that the extreme values indicate the
6110 largest possible offset - if the rest of the frame is encoded with the
6111 worst possible quality, an offset of -1 indicates that this region
6112 should be encoded with the best possible quality anyway. Intermediate
6113 values are then interpolated in some codec-dependent way.
6115 For example, in 10-bit H.264 the quantisation parameter varies between
6116 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6117 this region should be encoded with a QP around one-tenth of the full
6118 range better than the rest of the frame. So, if most of the frame
6119 were to be encoded with a QP of around 30, this region would get a QP
6120 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6121 An extreme value of -1 would indicate that this region should be
6122 encoded with the best possible quality regardless of the treatment of
6123 the rest of the frame - that is, should be encoded at a QP of -12.
6125 If set to true, remove any existing regions of interest marked on the
6126 frame before adding the new one.
6129 @subsection Examples
6133 Mark the centre quarter of the frame as interesting.
6135 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6138 Mark the 100-pixel-wide region on the left edge of the frame as very
6139 uninteresting (to be encoded at much lower quality than the rest of
6142 addroi=0:0:100:ih:+1/5
6146 @section alphaextract
6148 Extract the alpha component from the input as a grayscale video. This
6149 is especially useful with the @var{alphamerge} filter.
6153 Add or replace the alpha component of the primary input with the
6154 grayscale value of a second input. This is intended for use with
6155 @var{alphaextract} to allow the transmission or storage of frame
6156 sequences that have alpha in a format that doesn't support an alpha
6159 For example, to reconstruct full frames from a normal YUV-encoded video
6160 and a separate video created with @var{alphaextract}, you might use:
6162 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6165 Since this filter is designed for reconstruction, it operates on frame
6166 sequences without considering timestamps, and terminates when either
6167 input reaches end of stream. This will cause problems if your encoding
6168 pipeline drops frames. If you're trying to apply an image as an
6169 overlay to a video stream, consider the @var{overlay} filter instead.
6173 Amplify differences between current pixel and pixels of adjacent frames in
6174 same pixel location.
6176 This filter accepts the following options:
6180 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6181 For example radius of 3 will instruct filter to calculate average of 7 frames.
6184 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6187 Set threshold for difference amplification. Any difference greater or equal to
6188 this value will not alter source pixel. Default is 10.
6189 Allowed range is from 0 to 65535.
6192 Set tolerance for difference amplification. Any difference lower to
6193 this value will not alter source pixel. Default is 0.
6194 Allowed range is from 0 to 65535.
6197 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6198 This option controls maximum possible value that will decrease source pixel value.
6201 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6202 This option controls maximum possible value that will increase source pixel value.
6205 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
6208 @subsection Commands
6210 This filter supports the following @ref{commands} that corresponds to option of same name:
6222 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
6223 and libavformat to work. On the other hand, it is limited to ASS (Advanced
6224 Substation Alpha) subtitles files.
6226 This filter accepts the following option in addition to the common options from
6227 the @ref{subtitles} filter:
6231 Set the shaping engine
6233 Available values are:
6236 The default libass shaping engine, which is the best available.
6238 Fast, font-agnostic shaper that can do only substitutions
6240 Slower shaper using OpenType for substitutions and positioning
6243 The default is @code{auto}.
6247 Apply an Adaptive Temporal Averaging Denoiser to the video input.
6249 The filter accepts the following options:
6253 Set threshold A for 1st plane. Default is 0.02.
6254 Valid range is 0 to 0.3.
6257 Set threshold B for 1st plane. Default is 0.04.
6258 Valid range is 0 to 5.
6261 Set threshold A for 2nd plane. Default is 0.02.
6262 Valid range is 0 to 0.3.
6265 Set threshold B for 2nd plane. Default is 0.04.
6266 Valid range is 0 to 5.
6269 Set threshold A for 3rd plane. Default is 0.02.
6270 Valid range is 0 to 0.3.
6273 Set threshold B for 3rd plane. Default is 0.04.
6274 Valid range is 0 to 5.
6276 Threshold A is designed to react on abrupt changes in the input signal and
6277 threshold B is designed to react on continuous changes in the input signal.
6280 Set number of frames filter will use for averaging. Default is 9. Must be odd
6281 number in range [5, 129].
6284 Set what planes of frame filter will use for averaging. Default is all.
6287 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
6288 Alternatively can be set to @code{s} serial.
6290 Parallel can be faster then serial, while other way around is never true.
6291 Parallel will abort early on first change being greater then thresholds, while serial
6292 will continue processing other side of frames if they are equal or bellow thresholds.
6295 @subsection Commands
6296 This filter supports same @ref{commands} as options except option @code{s}.
6297 The command accepts the same syntax of the corresponding option.
6301 Apply average blur filter.
6303 The filter accepts the following options:
6307 Set horizontal radius size.
6310 Set which planes to filter. By default all planes are filtered.
6313 Set vertical radius size, if zero it will be same as @code{sizeX}.
6314 Default is @code{0}.
6317 @subsection Commands
6318 This filter supports same commands as options.
6319 The command accepts the same syntax of the corresponding option.
6321 If the specified expression is not valid, it is kept at its current
6326 Compute the bounding box for the non-black pixels in the input frame
6329 This filter computes the bounding box containing all the pixels with a
6330 luminance value greater than the minimum allowed value.
6331 The parameters describing the bounding box are printed on the filter
6334 The filter accepts the following option:
6338 Set the minimal luminance value. Default is @code{16}.
6342 Apply bilateral filter, spatial smoothing while preserving edges.
6344 The filter accepts the following options:
6347 Set sigma of gaussian function to calculate spatial weight.
6348 Allowed range is 0 to 10. Default is 0.1.
6351 Set sigma of gaussian function to calculate range weight.
6352 Allowed range is 0 to 1. Default is 0.1.
6355 Set planes to filter. Default is first only.
6358 @section bitplanenoise
6360 Show and measure bit plane noise.
6362 The filter accepts the following options:
6366 Set which plane to analyze. Default is @code{1}.
6369 Filter out noisy pixels from @code{bitplane} set above.
6370 Default is disabled.
6373 @section blackdetect
6375 Detect video intervals that are (almost) completely black. Can be
6376 useful to detect chapter transitions, commercials, or invalid
6377 recordings. Output lines contains the time for the start, end and
6378 duration of the detected black interval expressed in seconds.
6380 In order to display the output lines, you need to set the loglevel at
6381 least to the AV_LOG_INFO value.
6383 The filter accepts the following options:
6386 @item black_min_duration, d
6387 Set the minimum detected black duration expressed in seconds. It must
6388 be a non-negative floating point number.
6390 Default value is 2.0.
6392 @item picture_black_ratio_th, pic_th
6393 Set the threshold for considering a picture "black".
6394 Express the minimum value for the ratio:
6396 @var{nb_black_pixels} / @var{nb_pixels}
6399 for which a picture is considered black.
6400 Default value is 0.98.
6402 @item pixel_black_th, pix_th
6403 Set the threshold for considering a pixel "black".
6405 The threshold expresses the maximum pixel luminance value for which a
6406 pixel is considered "black". The provided value is scaled according to
6407 the following equation:
6409 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
6412 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
6413 the input video format, the range is [0-255] for YUV full-range
6414 formats and [16-235] for YUV non full-range formats.
6416 Default value is 0.10.
6419 The following example sets the maximum pixel threshold to the minimum
6420 value, and detects only black intervals of 2 or more seconds:
6422 blackdetect=d=2:pix_th=0.00
6427 Detect frames that are (almost) completely black. Can be useful to
6428 detect chapter transitions or commercials. Output lines consist of
6429 the frame number of the detected frame, the percentage of blackness,
6430 the position in the file if known or -1 and the timestamp in seconds.
6432 In order to display the output lines, you need to set the loglevel at
6433 least to the AV_LOG_INFO value.
6435 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
6436 The value represents the percentage of pixels in the picture that
6437 are below the threshold value.
6439 It accepts the following parameters:
6444 The percentage of the pixels that have to be below the threshold; it defaults to
6447 @item threshold, thresh
6448 The threshold below which a pixel value is considered black; it defaults to
6453 @section blend, tblend
6455 Blend two video frames into each other.
6457 The @code{blend} filter takes two input streams and outputs one
6458 stream, the first input is the "top" layer and second input is
6459 "bottom" layer. By default, the output terminates when the longest input terminates.
6461 The @code{tblend} (time blend) filter takes two consecutive frames
6462 from one single stream, and outputs the result obtained by blending
6463 the new frame on top of the old frame.
6465 A description of the accepted options follows.
6473 Set blend mode for specific pixel component or all pixel components in case
6474 of @var{all_mode}. Default value is @code{normal}.
6476 Available values for component modes are:
6518 Set blend opacity for specific pixel component or all pixel components in case
6519 of @var{all_opacity}. Only used in combination with pixel component blend modes.
6526 Set blend expression for specific pixel component or all pixel components in case
6527 of @var{all_expr}. Note that related mode options will be ignored if those are set.
6529 The expressions can use the following variables:
6533 The sequential number of the filtered frame, starting from @code{0}.
6537 the coordinates of the current sample
6541 the width and height of currently filtered plane
6545 Width and height scale for the plane being filtered. It is the
6546 ratio between the dimensions of the current plane to the luma plane,
6547 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
6548 the luma plane and @code{0.5,0.5} for the chroma planes.
6551 Time of the current frame, expressed in seconds.
6554 Value of pixel component at current location for first video frame (top layer).
6557 Value of pixel component at current location for second video frame (bottom layer).
6561 The @code{blend} filter also supports the @ref{framesync} options.
6563 @subsection Examples
6567 Apply transition from bottom layer to top layer in first 10 seconds:
6569 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
6573 Apply linear horizontal transition from top layer to bottom layer:
6575 blend=all_expr='A*(X/W)+B*(1-X/W)'
6579 Apply 1x1 checkerboard effect:
6581 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
6585 Apply uncover left effect:
6587 blend=all_expr='if(gte(N*SW+X,W),A,B)'
6591 Apply uncover down effect:
6593 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
6597 Apply uncover up-left effect:
6599 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
6603 Split diagonally video and shows top and bottom layer on each side:
6605 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
6609 Display differences between the current and the previous frame:
6611 tblend=all_mode=grainextract
6617 Denoise frames using Block-Matching 3D algorithm.
6619 The filter accepts the following options.
6623 Set denoising strength. Default value is 1.
6624 Allowed range is from 0 to 999.9.
6625 The denoising algorithm is very sensitive to sigma, so adjust it
6626 according to the source.
6629 Set local patch size. This sets dimensions in 2D.
6632 Set sliding step for processing blocks. Default value is 4.
6633 Allowed range is from 1 to 64.
6634 Smaller values allows processing more reference blocks and is slower.
6637 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
6638 When set to 1, no block matching is done. Larger values allows more blocks
6640 Allowed range is from 1 to 256.
6643 Set radius for search block matching. Default is 9.
6644 Allowed range is from 1 to INT32_MAX.
6647 Set step between two search locations for block matching. Default is 1.
6648 Allowed range is from 1 to 64. Smaller is slower.
6651 Set threshold of mean square error for block matching. Valid range is 0 to
6655 Set thresholding parameter for hard thresholding in 3D transformed domain.
6656 Larger values results in stronger hard-thresholding filtering in frequency
6660 Set filtering estimation mode. Can be @code{basic} or @code{final}.
6661 Default is @code{basic}.
6664 If enabled, filter will use 2nd stream for block matching.
6665 Default is disabled for @code{basic} value of @var{estim} option,
6666 and always enabled if value of @var{estim} is @code{final}.
6669 Set planes to filter. Default is all available except alpha.
6672 @subsection Examples
6676 Basic filtering with bm3d:
6678 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
6682 Same as above, but filtering only luma:
6684 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
6688 Same as above, but with both estimation modes:
6690 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
6694 Same as above, but prefilter with @ref{nlmeans} filter instead:
6696 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
6702 Apply a boxblur algorithm to the input video.
6704 It accepts the following parameters:
6708 @item luma_radius, lr
6709 @item luma_power, lp
6710 @item chroma_radius, cr
6711 @item chroma_power, cp
6712 @item alpha_radius, ar
6713 @item alpha_power, ap
6717 A description of the accepted options follows.
6720 @item luma_radius, lr
6721 @item chroma_radius, cr
6722 @item alpha_radius, ar
6723 Set an expression for the box radius in pixels used for blurring the
6724 corresponding input plane.
6726 The radius value must be a non-negative number, and must not be
6727 greater than the value of the expression @code{min(w,h)/2} for the
6728 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
6731 Default value for @option{luma_radius} is "2". If not specified,
6732 @option{chroma_radius} and @option{alpha_radius} default to the
6733 corresponding value set for @option{luma_radius}.
6735 The expressions can contain the following constants:
6739 The input width and height in pixels.
6743 The input chroma image width and height in pixels.
6747 The horizontal and vertical chroma subsample values. For example, for the
6748 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
6751 @item luma_power, lp
6752 @item chroma_power, cp
6753 @item alpha_power, ap
6754 Specify how many times the boxblur filter is applied to the
6755 corresponding plane.
6757 Default value for @option{luma_power} is 2. If not specified,
6758 @option{chroma_power} and @option{alpha_power} default to the
6759 corresponding value set for @option{luma_power}.
6761 A value of 0 will disable the effect.
6764 @subsection Examples
6768 Apply a boxblur filter with the luma, chroma, and alpha radii
6771 boxblur=luma_radius=2:luma_power=1
6776 Set the luma radius to 2, and alpha and chroma radius to 0:
6778 boxblur=2:1:cr=0:ar=0
6782 Set the luma and chroma radii to a fraction of the video dimension:
6784 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
6790 Deinterlace the input video ("bwdif" stands for "Bob Weaver
6791 Deinterlacing Filter").
6793 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
6794 interpolation algorithms.
6795 It accepts the following parameters:
6799 The interlacing mode to adopt. It accepts one of the following values:
6803 Output one frame for each frame.
6805 Output one frame for each field.
6808 The default value is @code{send_field}.
6811 The picture field parity assumed for the input interlaced video. It accepts one
6812 of the following values:
6816 Assume the top field is first.
6818 Assume the bottom field is first.
6820 Enable automatic detection of field parity.
6823 The default value is @code{auto}.
6824 If the interlacing is unknown or the decoder does not export this information,
6825 top field first will be assumed.
6828 Specify which frames to deinterlace. Accepts one of the following
6833 Deinterlace all frames.
6835 Only deinterlace frames marked as interlaced.
6838 The default value is @code{all}.
6842 Remove all color information for all colors except for certain one.
6844 The filter accepts the following options:
6848 The color which will not be replaced with neutral chroma.
6851 Similarity percentage with the above color.
6852 0.01 matches only the exact key color, while 1.0 matches everything.
6856 0.0 makes pixels either fully gray, or not gray at all.
6857 Higher values result in more preserved color.
6860 Signals that the color passed is already in YUV instead of RGB.
6862 Literal colors like "green" or "red" don't make sense with this enabled anymore.
6863 This can be used to pass exact YUV values as hexadecimal numbers.
6867 YUV colorspace color/chroma keying.
6869 The filter accepts the following options:
6873 The color which will be replaced with transparency.
6876 Similarity percentage with the key color.
6878 0.01 matches only the exact key color, while 1.0 matches everything.
6883 0.0 makes pixels either fully transparent, or not transparent at all.
6885 Higher values result in semi-transparent pixels, with a higher transparency
6886 the more similar the pixels color is to the key color.
6889 Signals that the color passed is already in YUV instead of RGB.
6891 Literal colors like "green" or "red" don't make sense with this enabled anymore.
6892 This can be used to pass exact YUV values as hexadecimal numbers.
6895 @subsection Examples
6899 Make every green pixel in the input image transparent:
6901 ffmpeg -i input.png -vf chromakey=green out.png
6905 Overlay a greenscreen-video on top of a static black background.
6907 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
6911 @section chromashift
6912 Shift chroma pixels horizontally and/or vertically.
6914 The filter accepts the following options:
6917 Set amount to shift chroma-blue horizontally.
6919 Set amount to shift chroma-blue vertically.
6921 Set amount to shift chroma-red horizontally.
6923 Set amount to shift chroma-red vertically.
6925 Set edge mode, can be @var{smear}, default, or @var{warp}.
6930 Display CIE color diagram with pixels overlaid onto it.
6932 The filter accepts the following options:
6947 @item uhdtv, rec2020
6961 Set what gamuts to draw.
6963 See @code{system} option for available values.
6966 Set ciescope size, by default set to 512.
6969 Set intensity used to map input pixel values to CIE diagram.
6972 Set contrast used to draw tongue colors that are out of active color system gamut.
6975 Correct gamma displayed on scope, by default enabled.
6978 Show white point on CIE diagram, by default disabled.
6981 Set input gamma. Used only with XYZ input color space.
6986 Visualize information exported by some codecs.
6988 Some codecs can export information through frames using side-data or other
6989 means. For example, some MPEG based codecs export motion vectors through the
6990 @var{export_mvs} flag in the codec @option{flags2} option.
6992 The filter accepts the following option:
6996 Set motion vectors to visualize.
6998 Available flags for @var{mv} are:
7002 forward predicted MVs of P-frames
7004 forward predicted MVs of B-frames
7006 backward predicted MVs of B-frames
7010 Display quantization parameters using the chroma planes.
7013 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7015 Available flags for @var{mv_type} are:
7019 forward predicted MVs
7021 backward predicted MVs
7024 @item frame_type, ft
7025 Set frame type to visualize motion vectors of.
7027 Available flags for @var{frame_type} are:
7031 intra-coded frames (I-frames)
7033 predicted frames (P-frames)
7035 bi-directionally predicted frames (B-frames)
7039 @subsection Examples
7043 Visualize forward predicted MVs of all frames using @command{ffplay}:
7045 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7049 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7051 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7055 @section colorbalance
7056 Modify intensity of primary colors (red, green and blue) of input frames.
7058 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7059 regions for the red-cyan, green-magenta or blue-yellow balance.
7061 A positive adjustment value shifts the balance towards the primary color, a negative
7062 value towards the complementary color.
7064 The filter accepts the following options:
7070 Adjust red, green and blue shadows (darkest pixels).
7075 Adjust red, green and blue midtones (medium pixels).
7080 Adjust red, green and blue highlights (brightest pixels).
7082 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7085 @subsection Examples
7089 Add red color cast to shadows:
7095 @section colorchannelmixer
7097 Adjust video input frames by re-mixing color channels.
7099 This filter modifies a color channel by adding the values associated to
7100 the other channels of the same pixels. For example if the value to
7101 modify is red, the output value will be:
7103 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
7106 The filter accepts the following options:
7113 Adjust contribution of input red, green, blue and alpha channels for output red channel.
7114 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
7120 Adjust contribution of input red, green, blue and alpha channels for output green channel.
7121 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
7127 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
7128 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
7134 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
7135 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
7137 Allowed ranges for options are @code{[-2.0, 2.0]}.
7140 @subsection Examples
7144 Convert source to grayscale:
7146 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
7149 Simulate sepia tones:
7151 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
7155 @subsection Commands
7157 This filter supports the all above options as @ref{commands}.
7160 RGB colorspace color keying.
7162 The filter accepts the following options:
7166 The color which will be replaced with transparency.
7169 Similarity percentage with the key color.
7171 0.01 matches only the exact key color, while 1.0 matches everything.
7176 0.0 makes pixels either fully transparent, or not transparent at all.
7178 Higher values result in semi-transparent pixels, with a higher transparency
7179 the more similar the pixels color is to the key color.
7182 @subsection Examples
7186 Make every green pixel in the input image transparent:
7188 ffmpeg -i input.png -vf colorkey=green out.png
7192 Overlay a greenscreen-video on top of a static background image.
7194 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
7199 Remove all color information for all RGB colors except for certain one.
7201 The filter accepts the following options:
7205 The color which will not be replaced with neutral gray.
7208 Similarity percentage with the above color.
7209 0.01 matches only the exact key color, while 1.0 matches everything.
7212 Blend percentage. 0.0 makes pixels fully gray.
7213 Higher values result in more preserved color.
7216 @section colorlevels
7218 Adjust video input frames using levels.
7220 The filter accepts the following options:
7227 Adjust red, green, blue and alpha input black point.
7228 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7234 Adjust red, green, blue and alpha input white point.
7235 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
7237 Input levels are used to lighten highlights (bright tones), darken shadows
7238 (dark tones), change the balance of bright and dark tones.
7244 Adjust red, green, blue and alpha output black point.
7245 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
7251 Adjust red, green, blue and alpha output white point.
7252 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
7254 Output levels allows manual selection of a constrained output level range.
7257 @subsection Examples
7261 Make video output darker:
7263 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
7269 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
7273 Make video output lighter:
7275 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
7279 Increase brightness:
7281 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
7285 @section colormatrix
7287 Convert color matrix.
7289 The filter accepts the following options:
7294 Specify the source and destination color matrix. Both values must be
7297 The accepted values are:
7325 For example to convert from BT.601 to SMPTE-240M, use the command:
7327 colormatrix=bt601:smpte240m
7332 Convert colorspace, transfer characteristics or color primaries.
7333 Input video needs to have an even size.
7335 The filter accepts the following options:
7340 Specify all color properties at once.
7342 The accepted values are:
7372 Specify output colorspace.
7374 The accepted values are:
7383 BT.470BG or BT.601-6 625
7386 SMPTE-170M or BT.601-6 525
7395 BT.2020 with non-constant luminance
7401 Specify output transfer characteristics.
7403 The accepted values are:
7415 Constant gamma of 2.2
7418 Constant gamma of 2.8
7421 SMPTE-170M, BT.601-6 625 or BT.601-6 525
7439 BT.2020 for 10-bits content
7442 BT.2020 for 12-bits content
7448 Specify output color primaries.
7450 The accepted values are:
7459 BT.470BG or BT.601-6 625
7462 SMPTE-170M or BT.601-6 525
7486 Specify output color range.
7488 The accepted values are:
7491 TV (restricted) range
7494 MPEG (restricted) range
7505 Specify output color format.
7507 The accepted values are:
7510 YUV 4:2:0 planar 8-bits
7513 YUV 4:2:0 planar 10-bits
7516 YUV 4:2:0 planar 12-bits
7519 YUV 4:2:2 planar 8-bits
7522 YUV 4:2:2 planar 10-bits
7525 YUV 4:2:2 planar 12-bits
7528 YUV 4:4:4 planar 8-bits
7531 YUV 4:4:4 planar 10-bits
7534 YUV 4:4:4 planar 12-bits
7539 Do a fast conversion, which skips gamma/primary correction. This will take
7540 significantly less CPU, but will be mathematically incorrect. To get output
7541 compatible with that produced by the colormatrix filter, use fast=1.
7544 Specify dithering mode.
7546 The accepted values are:
7552 Floyd-Steinberg dithering
7556 Whitepoint adaptation mode.
7558 The accepted values are:
7561 Bradford whitepoint adaptation
7564 von Kries whitepoint adaptation
7567 identity whitepoint adaptation (i.e. no whitepoint adaptation)
7571 Override all input properties at once. Same accepted values as @ref{all}.
7574 Override input colorspace. Same accepted values as @ref{space}.
7577 Override input color primaries. Same accepted values as @ref{primaries}.
7580 Override input transfer characteristics. Same accepted values as @ref{trc}.
7583 Override input color range. Same accepted values as @ref{range}.
7587 The filter converts the transfer characteristics, color space and color
7588 primaries to the specified user values. The output value, if not specified,
7589 is set to a default value based on the "all" property. If that property is
7590 also not specified, the filter will log an error. The output color range and
7591 format default to the same value as the input color range and format. The
7592 input transfer characteristics, color space, color primaries and color range
7593 should be set on the input data. If any of these are missing, the filter will
7594 log an error and no conversion will take place.
7596 For example to convert the input to SMPTE-240M, use the command:
7598 colorspace=smpte240m
7601 @section convolution
7603 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
7605 The filter accepts the following options:
7612 Set matrix for each plane.
7613 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
7614 and from 1 to 49 odd number of signed integers in @var{row} mode.
7620 Set multiplier for calculated value for each plane.
7621 If unset or 0, it will be sum of all matrix elements.
7627 Set bias for each plane. This value is added to the result of the multiplication.
7628 Useful for making the overall image brighter or darker. Default is 0.0.
7634 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
7635 Default is @var{square}.
7638 @subsection Examples
7644 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"
7650 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"
7656 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"
7662 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"
7666 Apply laplacian edge detector which includes diagonals:
7668 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"
7674 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"
7680 Apply 2D convolution of video stream in frequency domain using second stream
7683 The filter accepts the following options:
7687 Set which planes to process.
7690 Set which impulse video frames will be processed, can be @var{first}
7691 or @var{all}. Default is @var{all}.
7694 The @code{convolve} filter also supports the @ref{framesync} options.
7698 Copy the input video source unchanged to the output. This is mainly useful for
7703 Video filtering on GPU using Apple's CoreImage API on OSX.
7705 Hardware acceleration is based on an OpenGL context. Usually, this means it is
7706 processed by video hardware. However, software-based OpenGL implementations
7707 exist which means there is no guarantee for hardware processing. It depends on
7710 There are many filters and image generators provided by Apple that come with a
7711 large variety of options. The filter has to be referenced by its name along
7714 The coreimage filter accepts the following options:
7717 List all available filters and generators along with all their respective
7718 options as well as possible minimum and maximum values along with the default
7725 Specify all filters by their respective name and options.
7726 Use @var{list_filters} to determine all valid filter names and options.
7727 Numerical options are specified by a float value and are automatically clamped
7728 to their respective value range. Vector and color options have to be specified
7729 by a list of space separated float values. Character escaping has to be done.
7730 A special option name @code{default} is available to use default options for a
7733 It is required to specify either @code{default} or at least one of the filter options.
7734 All omitted options are used with their default values.
7735 The syntax of the filter string is as follows:
7737 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
7741 Specify a rectangle where the output of the filter chain is copied into the
7742 input image. It is given by a list of space separated float values:
7744 output_rect=x\ y\ width\ height
7746 If not given, the output rectangle equals the dimensions of the input image.
7747 The output rectangle is automatically cropped at the borders of the input
7748 image. Negative values are valid for each component.
7750 output_rect=25\ 25\ 100\ 100
7754 Several filters can be chained for successive processing without GPU-HOST
7755 transfers allowing for fast processing of complex filter chains.
7756 Currently, only filters with zero (generators) or exactly one (filters) input
7757 image and one output image are supported. Also, transition filters are not yet
7760 Some filters generate output images with additional padding depending on the
7761 respective filter kernel. The padding is automatically removed to ensure the
7762 filter output has the same size as the input image.
7764 For image generators, the size of the output image is determined by the
7765 previous output image of the filter chain or the input image of the whole
7766 filterchain, respectively. The generators do not use the pixel information of
7767 this image to generate their output. However, the generated output is
7768 blended onto this image, resulting in partial or complete coverage of the
7771 The @ref{coreimagesrc} video source can be used for generating input images
7772 which are directly fed into the filter chain. By using it, providing input
7773 images by another video source or an input video is not required.
7775 @subsection Examples
7780 List all filters available:
7782 coreimage=list_filters=true
7786 Use the CIBoxBlur filter with default options to blur an image:
7788 coreimage=filter=CIBoxBlur@@default
7792 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
7793 its center at 100x100 and a radius of 50 pixels:
7795 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
7799 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
7800 given as complete and escaped command-line for Apple's standard bash shell:
7802 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
7808 Cover a rectangular object
7810 It accepts the following options:
7814 Filepath of the optional cover image, needs to be in yuv420.
7819 It accepts the following values:
7822 cover it by the supplied image
7824 cover it by interpolating the surrounding pixels
7827 Default value is @var{blur}.
7830 @subsection Examples
7834 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
7836 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
7842 Crop the input video to given dimensions.
7844 It accepts the following parameters:
7848 The width of the output video. It defaults to @code{iw}.
7849 This expression is evaluated only once during the filter
7850 configuration, or when the @samp{w} or @samp{out_w} command is sent.
7853 The height of the output video. It defaults to @code{ih}.
7854 This expression is evaluated only once during the filter
7855 configuration, or when the @samp{h} or @samp{out_h} command is sent.
7858 The horizontal position, in the input video, of the left edge of the output
7859 video. It defaults to @code{(in_w-out_w)/2}.
7860 This expression is evaluated per-frame.
7863 The vertical position, in the input video, of the top edge of the output video.
7864 It defaults to @code{(in_h-out_h)/2}.
7865 This expression is evaluated per-frame.
7868 If set to 1 will force the output display aspect ratio
7869 to be the same of the input, by changing the output sample aspect
7870 ratio. It defaults to 0.
7873 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
7874 width/height/x/y as specified and will not be rounded to nearest smaller value.
7878 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
7879 expressions containing the following constants:
7884 The computed values for @var{x} and @var{y}. They are evaluated for
7889 The input width and height.
7893 These are the same as @var{in_w} and @var{in_h}.
7897 The output (cropped) width and height.
7901 These are the same as @var{out_w} and @var{out_h}.
7904 same as @var{iw} / @var{ih}
7907 input sample aspect ratio
7910 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
7914 horizontal and vertical chroma subsample values. For example for the
7915 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
7918 The number of the input frame, starting from 0.
7921 the position in the file of the input frame, NAN if unknown
7924 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
7928 The expression for @var{out_w} may depend on the value of @var{out_h},
7929 and the expression for @var{out_h} may depend on @var{out_w}, but they
7930 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
7931 evaluated after @var{out_w} and @var{out_h}.
7933 The @var{x} and @var{y} parameters specify the expressions for the
7934 position of the top-left corner of the output (non-cropped) area. They
7935 are evaluated for each frame. If the evaluated value is not valid, it
7936 is approximated to the nearest valid value.
7938 The expression for @var{x} may depend on @var{y}, and the expression
7939 for @var{y} may depend on @var{x}.
7941 @subsection Examples
7945 Crop area with size 100x100 at position (12,34).
7950 Using named options, the example above becomes:
7952 crop=w=100:h=100:x=12:y=34
7956 Crop the central input area with size 100x100:
7962 Crop the central input area with size 2/3 of the input video:
7964 crop=2/3*in_w:2/3*in_h
7968 Crop the input video central square:
7975 Delimit the rectangle with the top-left corner placed at position
7976 100:100 and the right-bottom corner corresponding to the right-bottom
7977 corner of the input image.
7979 crop=in_w-100:in_h-100:100:100
7983 Crop 10 pixels from the left and right borders, and 20 pixels from
7984 the top and bottom borders
7986 crop=in_w-2*10:in_h-2*20
7990 Keep only the bottom right quarter of the input image:
7992 crop=in_w/2:in_h/2:in_w/2:in_h/2
7996 Crop height for getting Greek harmony:
7998 crop=in_w:1/PHI*in_w
8002 Apply trembling effect:
8004 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)
8008 Apply erratic camera effect depending on timestamp:
8010 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)"
8014 Set x depending on the value of y:
8016 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8020 @subsection Commands
8022 This filter supports the following commands:
8028 Set width/height of the output video and the horizontal/vertical position
8030 The command accepts the same syntax of the corresponding option.
8032 If the specified expression is not valid, it is kept at its current
8038 Auto-detect the crop size.
8040 It calculates the necessary cropping parameters and prints the
8041 recommended parameters via the logging system. The detected dimensions
8042 correspond to the non-black area of the input video.
8044 It accepts the following parameters:
8049 Set higher black value threshold, which can be optionally specified
8050 from nothing (0) to everything (255 for 8-bit based formats). An intensity
8051 value greater to the set value is considered non-black. It defaults to 24.
8052 You can also specify a value between 0.0 and 1.0 which will be scaled depending
8053 on the bitdepth of the pixel format.
8056 The value which the width/height should be divisible by. It defaults to
8057 16. The offset is automatically adjusted to center the video. Use 2 to
8058 get only even dimensions (needed for 4:2:2 video). 16 is best when
8059 encoding to most video codecs.
8061 @item reset_count, reset
8062 Set the counter that determines after how many frames cropdetect will
8063 reset the previously detected largest video area and start over to
8064 detect the current optimal crop area. Default value is 0.
8066 This can be useful when channel logos distort the video area. 0
8067 indicates 'never reset', and returns the largest area encountered during
8074 Delay video filtering until a given wallclock timestamp. The filter first
8075 passes on @option{preroll} amount of frames, then it buffers at most
8076 @option{buffer} amount of frames and waits for the cue. After reaching the cue
8077 it forwards the buffered frames and also any subsequent frames coming in its
8080 The filter can be used synchronize the output of multiple ffmpeg processes for
8081 realtime output devices like decklink. By putting the delay in the filtering
8082 chain and pre-buffering frames the process can pass on data to output almost
8083 immediately after the target wallclock timestamp is reached.
8085 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
8091 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
8094 The duration of content to pass on as preroll expressed in seconds. Default is 0.
8097 The maximum duration of content to buffer before waiting for the cue expressed
8098 in seconds. Default is 0.
8105 Apply color adjustments using curves.
8107 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
8108 component (red, green and blue) has its values defined by @var{N} key points
8109 tied from each other using a smooth curve. The x-axis represents the pixel
8110 values from the input frame, and the y-axis the new pixel values to be set for
8113 By default, a component curve is defined by the two points @var{(0;0)} and
8114 @var{(1;1)}. This creates a straight line where each original pixel value is
8115 "adjusted" to its own value, which means no change to the image.
8117 The filter allows you to redefine these two points and add some more. A new
8118 curve (using a natural cubic spline interpolation) will be define to pass
8119 smoothly through all these new coordinates. The new defined points needs to be
8120 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
8121 be in the @var{[0;1]} interval. If the computed curves happened to go outside
8122 the vector spaces, the values will be clipped accordingly.
8124 The filter accepts the following options:
8128 Select one of the available color presets. This option can be used in addition
8129 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
8130 options takes priority on the preset values.
8131 Available presets are:
8134 @item color_negative
8137 @item increase_contrast
8139 @item linear_contrast
8140 @item medium_contrast
8142 @item strong_contrast
8145 Default is @code{none}.
8147 Set the master key points. These points will define a second pass mapping. It
8148 is sometimes called a "luminance" or "value" mapping. It can be used with
8149 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
8150 post-processing LUT.
8152 Set the key points for the red component.
8154 Set the key points for the green component.
8156 Set the key points for the blue component.
8158 Set the key points for all components (not including master).
8159 Can be used in addition to the other key points component
8160 options. In this case, the unset component(s) will fallback on this
8161 @option{all} setting.
8163 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
8165 Save Gnuplot script of the curves in specified file.
8168 To avoid some filtergraph syntax conflicts, each key points list need to be
8169 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
8171 @subsection Examples
8175 Increase slightly the middle level of blue:
8177 curves=blue='0/0 0.5/0.58 1/1'
8183 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'
8185 Here we obtain the following coordinates for each components:
8188 @code{(0;0.11) (0.42;0.51) (1;0.95)}
8190 @code{(0;0) (0.50;0.48) (1;1)}
8192 @code{(0;0.22) (0.49;0.44) (1;0.80)}
8196 The previous example can also be achieved with the associated built-in preset:
8198 curves=preset=vintage
8208 Use a Photoshop preset and redefine the points of the green component:
8210 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
8214 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
8215 and @command{gnuplot}:
8217 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
8218 gnuplot -p /tmp/curves.plt
8224 Video data analysis filter.
8226 This filter shows hexadecimal pixel values of part of video.
8228 The filter accepts the following options:
8232 Set output video size.
8235 Set x offset from where to pick pixels.
8238 Set y offset from where to pick pixels.
8241 Set scope mode, can be one of the following:
8244 Draw hexadecimal pixel values with white color on black background.
8247 Draw hexadecimal pixel values with input video pixel color on black
8251 Draw hexadecimal pixel values on color background picked from input video,
8252 the text color is picked in such way so its always visible.
8256 Draw rows and columns numbers on left and top of video.
8259 Set background opacity.
8264 Denoise frames using 2D DCT (frequency domain filtering).
8266 This filter is not designed for real time.
8268 The filter accepts the following options:
8272 Set the noise sigma constant.
8274 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
8275 coefficient (absolute value) below this threshold with be dropped.
8277 If you need a more advanced filtering, see @option{expr}.
8279 Default is @code{0}.
8282 Set number overlapping pixels for each block. Since the filter can be slow, you
8283 may want to reduce this value, at the cost of a less effective filter and the
8284 risk of various artefacts.
8286 If the overlapping value doesn't permit processing the whole input width or
8287 height, a warning will be displayed and according borders won't be denoised.
8289 Default value is @var{blocksize}-1, which is the best possible setting.
8292 Set the coefficient factor expression.
8294 For each coefficient of a DCT block, this expression will be evaluated as a
8295 multiplier value for the coefficient.
8297 If this is option is set, the @option{sigma} option will be ignored.
8299 The absolute value of the coefficient can be accessed through the @var{c}
8303 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
8304 @var{blocksize}, which is the width and height of the processed blocks.
8306 The default value is @var{3} (8x8) and can be raised to @var{4} for a
8307 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
8308 on the speed processing. Also, a larger block size does not necessarily means a
8312 @subsection Examples
8314 Apply a denoise with a @option{sigma} of @code{4.5}:
8319 The same operation can be achieved using the expression system:
8321 dctdnoiz=e='gte(c, 4.5*3)'
8324 Violent denoise using a block size of @code{16x16}:
8331 Remove banding artifacts from input video.
8332 It works by replacing banded pixels with average value of referenced pixels.
8334 The filter accepts the following options:
8341 Set banding detection threshold for each plane. Default is 0.02.
8342 Valid range is 0.00003 to 0.5.
8343 If difference between current pixel and reference pixel is less than threshold,
8344 it will be considered as banded.
8347 Banding detection range in pixels. Default is 16. If positive, random number
8348 in range 0 to set value will be used. If negative, exact absolute value
8350 The range defines square of four pixels around current pixel.
8353 Set direction in radians from which four pixel will be compared. If positive,
8354 random direction from 0 to set direction will be picked. If negative, exact of
8355 absolute value will be picked. For example direction 0, -PI or -2*PI radians
8356 will pick only pixels on same row and -PI/2 will pick only pixels on same
8360 If enabled, current pixel is compared with average value of all four
8361 surrounding pixels. The default is enabled. If disabled current pixel is
8362 compared with all four surrounding pixels. The pixel is considered banded
8363 if only all four differences with surrounding pixels are less than threshold.
8366 If enabled, current pixel is changed if and only if all pixel components are banded,
8367 e.g. banding detection threshold is triggered for all color components.
8368 The default is disabled.
8373 Remove blocking artifacts from input video.
8375 The filter accepts the following options:
8379 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
8380 This controls what kind of deblocking is applied.
8383 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
8389 Set blocking detection thresholds. Allowed range is 0 to 1.
8390 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
8391 Using higher threshold gives more deblocking strength.
8392 Setting @var{alpha} controls threshold detection at exact edge of block.
8393 Remaining options controls threshold detection near the edge. Each one for
8394 below/above or left/right. Setting any of those to @var{0} disables
8398 Set planes to filter. Default is to filter all available planes.
8401 @subsection Examples
8405 Deblock using weak filter and block size of 4 pixels.
8407 deblock=filter=weak:block=4
8411 Deblock using strong filter, block size of 4 pixels and custom thresholds for
8412 deblocking more edges.
8414 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
8418 Similar as above, but filter only first plane.
8420 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
8424 Similar as above, but filter only second and third plane.
8426 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
8433 Drop duplicated frames at regular intervals.
8435 The filter accepts the following options:
8439 Set the number of frames from which one will be dropped. Setting this to
8440 @var{N} means one frame in every batch of @var{N} frames will be dropped.
8441 Default is @code{5}.
8444 Set the threshold for duplicate detection. If the difference metric for a frame
8445 is less than or equal to this value, then it is declared as duplicate. Default
8449 Set scene change threshold. Default is @code{15}.
8453 Set the size of the x and y-axis blocks used during metric calculations.
8454 Larger blocks give better noise suppression, but also give worse detection of
8455 small movements. Must be a power of two. Default is @code{32}.
8458 Mark main input as a pre-processed input and activate clean source input
8459 stream. This allows the input to be pre-processed with various filters to help
8460 the metrics calculation while keeping the frame selection lossless. When set to
8461 @code{1}, the first stream is for the pre-processed input, and the second
8462 stream is the clean source from where the kept frames are chosen. Default is
8466 Set whether or not chroma is considered in the metric calculations. Default is
8472 Apply 2D deconvolution of video stream in frequency domain using second stream
8475 The filter accepts the following options:
8479 Set which planes to process.
8482 Set which impulse video frames will be processed, can be @var{first}
8483 or @var{all}. Default is @var{all}.
8486 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
8487 and height are not same and not power of 2 or if stream prior to convolving
8491 The @code{deconvolve} filter also supports the @ref{framesync} options.
8495 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
8497 It accepts the following options:
8501 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
8502 @var{rainbows} for cross-color reduction.
8505 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
8508 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
8511 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
8514 Set temporal chroma threshold. Lower values increases reduction of cross-color.
8519 Apply deflate effect to the video.
8521 This filter replaces the pixel by the local(3x3) average by taking into account
8522 only values lower than the pixel.
8524 It accepts the following options:
8531 Limit the maximum change for each plane, default is 65535.
8532 If 0, plane will remain unchanged.
8537 Remove temporal frame luminance variations.
8539 It accepts the following options:
8543 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
8546 Set averaging mode to smooth temporal luminance variations.
8548 Available values are:
8573 Do not actually modify frame. Useful when one only wants metadata.
8578 Remove judder produced by partially interlaced telecined content.
8580 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
8581 source was partially telecined content then the output of @code{pullup,dejudder}
8582 will have a variable frame rate. May change the recorded frame rate of the
8583 container. Aside from that change, this filter will not affect constant frame
8586 The option available in this filter is:
8590 Specify the length of the window over which the judder repeats.
8592 Accepts any integer greater than 1. Useful values are:
8596 If the original was telecined from 24 to 30 fps (Film to NTSC).
8599 If the original was telecined from 25 to 30 fps (PAL to NTSC).
8602 If a mixture of the two.
8605 The default is @samp{4}.
8610 Suppress a TV station logo by a simple interpolation of the surrounding
8611 pixels. Just set a rectangle covering the logo and watch it disappear
8612 (and sometimes something even uglier appear - your mileage may vary).
8614 It accepts the following parameters:
8619 Specify the top left corner coordinates of the logo. They must be
8624 Specify the width and height of the logo to clear. They must be
8628 Specify the thickness of the fuzzy edge of the rectangle (added to
8629 @var{w} and @var{h}). The default value is 1. This option is
8630 deprecated, setting higher values should no longer be necessary and
8634 When set to 1, a green rectangle is drawn on the screen to simplify
8635 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
8636 The default value is 0.
8638 The rectangle is drawn on the outermost pixels which will be (partly)
8639 replaced with interpolated values. The values of the next pixels
8640 immediately outside this rectangle in each direction will be used to
8641 compute the interpolated pixel values inside the rectangle.
8645 @subsection Examples
8649 Set a rectangle covering the area with top left corner coordinates 0,0
8650 and size 100x77, and a band of size 10:
8652 delogo=x=0:y=0:w=100:h=77:band=10
8659 Remove the rain in the input image/video by applying the derain methods based on
8660 convolutional neural networks. Supported models:
8664 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
8665 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
8668 Training as well as model generation scripts are provided in
8669 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
8671 Native model files (.model) can be generated from TensorFlow model
8672 files (.pb) by using tools/python/convert.py
8674 The filter accepts the following options:
8678 Specify which filter to use. This option accepts the following values:
8682 Derain filter. To conduct derain filter, you need to use a derain model.
8685 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
8687 Default value is @samp{derain}.
8690 Specify which DNN backend to use for model loading and execution. This option accepts
8691 the following values:
8695 Native implementation of DNN loading and execution.
8698 TensorFlow backend. To enable this backend you
8699 need to install the TensorFlow for C library (see
8700 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
8701 @code{--enable-libtensorflow}
8703 Default value is @samp{native}.
8706 Set path to model file specifying network architecture and its parameters.
8707 Note that different backends use different file formats. TensorFlow and native
8708 backend can load files for only its format.
8713 Attempt to fix small changes in horizontal and/or vertical shift. This
8714 filter helps remove camera shake from hand-holding a camera, bumping a
8715 tripod, moving on a vehicle, etc.
8717 The filter accepts the following options:
8725 Specify a rectangular area where to limit the search for motion
8727 If desired the search for motion vectors can be limited to a
8728 rectangular area of the frame defined by its top left corner, width
8729 and height. These parameters have the same meaning as the drawbox
8730 filter which can be used to visualise the position of the bounding
8733 This is useful when simultaneous movement of subjects within the frame
8734 might be confused for camera motion by the motion vector search.
8736 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
8737 then the full frame is used. This allows later options to be set
8738 without specifying the bounding box for the motion vector search.
8740 Default - search the whole frame.
8744 Specify the maximum extent of movement in x and y directions in the
8745 range 0-64 pixels. Default 16.
8748 Specify how to generate pixels to fill blanks at the edge of the
8749 frame. Available values are:
8752 Fill zeroes at blank locations
8754 Original image at blank locations
8756 Extruded edge value at blank locations
8758 Mirrored edge at blank locations
8760 Default value is @samp{mirror}.
8763 Specify the blocksize to use for motion search. Range 4-128 pixels,
8767 Specify the contrast threshold for blocks. Only blocks with more than
8768 the specified contrast (difference between darkest and lightest
8769 pixels) will be considered. Range 1-255, default 125.
8772 Specify the search strategy. Available values are:
8775 Set exhaustive search
8777 Set less exhaustive search.
8779 Default value is @samp{exhaustive}.
8782 If set then a detailed log of the motion search is written to the
8789 Remove unwanted contamination of foreground colors, caused by reflected color of
8790 greenscreen or bluescreen.
8792 This filter accepts the following options:
8796 Set what type of despill to use.
8799 Set how spillmap will be generated.
8802 Set how much to get rid of still remaining spill.
8805 Controls amount of red in spill area.
8808 Controls amount of green in spill area.
8809 Should be -1 for greenscreen.
8812 Controls amount of blue in spill area.
8813 Should be -1 for bluescreen.
8816 Controls brightness of spill area, preserving colors.
8819 Modify alpha from generated spillmap.
8824 Apply an exact inverse of the telecine operation. It requires a predefined
8825 pattern specified using the pattern option which must be the same as that passed
8826 to the telecine filter.
8828 This filter accepts the following options:
8837 The default value is @code{top}.
8841 A string of numbers representing the pulldown pattern you wish to apply.
8842 The default value is @code{23}.
8845 A number representing position of the first frame with respect to the telecine
8846 pattern. This is to be used if the stream is cut. The default value is @code{0}.
8851 Apply dilation effect to the video.
8853 This filter replaces the pixel by the local(3x3) maximum.
8855 It accepts the following options:
8862 Limit the maximum change for each plane, default is 65535.
8863 If 0, plane will remain unchanged.
8866 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
8869 Flags to local 3x3 coordinates maps like this:
8878 Displace pixels as indicated by second and third input stream.
8880 It takes three input streams and outputs one stream, the first input is the
8881 source, and second and third input are displacement maps.
8883 The second input specifies how much to displace pixels along the
8884 x-axis, while the third input specifies how much to displace pixels
8886 If one of displacement map streams terminates, last frame from that
8887 displacement map will be used.
8889 Note that once generated, displacements maps can be reused over and over again.
8891 A description of the accepted options follows.
8895 Set displace behavior for pixels that are out of range.
8897 Available values are:
8900 Missing pixels are replaced by black pixels.
8903 Adjacent pixels will spread out to replace missing pixels.
8906 Out of range pixels are wrapped so they point to pixels of other side.
8909 Out of range pixels will be replaced with mirrored pixels.
8911 Default is @samp{smear}.
8915 @subsection Examples
8919 Add ripple effect to rgb input of video size hd720:
8921 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
8925 Add wave effect to rgb input of video size hd720:
8927 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
8933 Draw a colored box on the input image.
8935 It accepts the following parameters:
8940 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
8944 The expressions which specify the width and height of the box; if 0 they are interpreted as
8945 the input width and height. It defaults to 0.
8948 Specify the color of the box to write. For the general syntax of this option,
8949 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
8950 value @code{invert} is used, the box edge color is the same as the
8951 video with inverted luma.
8954 The expression which sets the thickness of the box edge.
8955 A value of @code{fill} will create a filled box. Default value is @code{3}.
8957 See below for the list of accepted constants.
8960 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
8961 will overwrite the video's color and alpha pixels.
8962 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
8965 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
8966 following constants:
8970 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
8974 horizontal and vertical chroma subsample values. For example for the
8975 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8979 The input width and height.
8982 The input sample aspect ratio.
8986 The x and y offset coordinates where the box is drawn.
8990 The width and height of the drawn box.
8993 The thickness of the drawn box.
8995 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
8996 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
9000 @subsection Examples
9004 Draw a black box around the edge of the input image:
9010 Draw a box with color red and an opacity of 50%:
9012 drawbox=10:20:200:60:red@@0.5
9015 The previous example can be specified as:
9017 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
9021 Fill the box with pink color:
9023 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
9027 Draw a 2-pixel red 2.40:1 mask:
9029 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
9033 @subsection Commands
9034 This filter supports same commands as options.
9035 The command accepts the same syntax of the corresponding option.
9037 If the specified expression is not valid, it is kept at its current
9042 Draw a grid on the input image.
9044 It accepts the following parameters:
9049 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
9053 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
9054 input width and height, respectively, minus @code{thickness}, so image gets
9055 framed. Default to 0.
9058 Specify the color of the grid. For the general syntax of this option,
9059 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
9060 value @code{invert} is used, the grid color is the same as the
9061 video with inverted luma.
9064 The expression which sets the thickness of the grid line. Default value is @code{1}.
9066 See below for the list of accepted constants.
9069 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
9070 will overwrite the video's color and alpha pixels.
9071 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
9074 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
9075 following constants:
9079 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
9083 horizontal and vertical chroma subsample values. For example for the
9084 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9088 The input grid cell width and height.
9091 The input sample aspect ratio.
9095 The x and y coordinates of some point of grid intersection (meant to configure offset).
9099 The width and height of the drawn cell.
9102 The thickness of the drawn cell.
9104 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
9105 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
9109 @subsection Examples
9113 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
9115 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
9119 Draw a white 3x3 grid with an opacity of 50%:
9121 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
9125 @subsection Commands
9126 This filter supports same commands as options.
9127 The command accepts the same syntax of the corresponding option.
9129 If the specified expression is not valid, it is kept at its current
9135 Draw a text string or text from a specified file on top of a video, using the
9136 libfreetype library.
9138 To enable compilation of this filter, you need to configure FFmpeg with
9139 @code{--enable-libfreetype}.
9140 To enable default font fallback and the @var{font} option you need to
9141 configure FFmpeg with @code{--enable-libfontconfig}.
9142 To enable the @var{text_shaping} option, you need to configure FFmpeg with
9143 @code{--enable-libfribidi}.
9147 It accepts the following parameters:
9152 Used to draw a box around text using the background color.
9153 The value must be either 1 (enable) or 0 (disable).
9154 The default value of @var{box} is 0.
9157 Set the width of the border to be drawn around the box using @var{boxcolor}.
9158 The default value of @var{boxborderw} is 0.
9161 The color to be used for drawing box around text. For the syntax of this
9162 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9164 The default value of @var{boxcolor} is "white".
9167 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
9168 The default value of @var{line_spacing} is 0.
9171 Set the width of the border to be drawn around the text using @var{bordercolor}.
9172 The default value of @var{borderw} is 0.
9175 Set the color to be used for drawing border around text. For the syntax of this
9176 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9178 The default value of @var{bordercolor} is "black".
9181 Select how the @var{text} is expanded. Can be either @code{none},
9182 @code{strftime} (deprecated) or
9183 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
9187 Set a start time for the count. Value is in microseconds. Only applied
9188 in the deprecated strftime expansion mode. To emulate in normal expansion
9189 mode use the @code{pts} function, supplying the start time (in seconds)
9190 as the second argument.
9193 If true, check and fix text coords to avoid clipping.
9196 The color to be used for drawing fonts. For the syntax of this option, check
9197 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9199 The default value of @var{fontcolor} is "black".
9201 @item fontcolor_expr
9202 String which is expanded the same way as @var{text} to obtain dynamic
9203 @var{fontcolor} value. By default this option has empty value and is not
9204 processed. When this option is set, it overrides @var{fontcolor} option.
9207 The font family to be used for drawing text. By default Sans.
9210 The font file to be used for drawing text. The path must be included.
9211 This parameter is mandatory if the fontconfig support is disabled.
9214 Draw the text applying alpha blending. The value can
9215 be a number between 0.0 and 1.0.
9216 The expression accepts the same variables @var{x, y} as well.
9217 The default value is 1.
9218 Please see @var{fontcolor_expr}.
9221 The font size to be used for drawing text.
9222 The default value of @var{fontsize} is 16.
9225 If set to 1, attempt to shape the text (for example, reverse the order of
9226 right-to-left text and join Arabic characters) before drawing it.
9227 Otherwise, just draw the text exactly as given.
9228 By default 1 (if supported).
9231 The flags to be used for loading the fonts.
9233 The flags map the corresponding flags supported by libfreetype, and are
9234 a combination of the following values:
9241 @item vertical_layout
9242 @item force_autohint
9245 @item ignore_global_advance_width
9247 @item ignore_transform
9253 Default value is "default".
9255 For more information consult the documentation for the FT_LOAD_*
9259 The color to be used for drawing a shadow behind the drawn text. For the
9260 syntax of this option, check the @ref{color syntax,,"Color" section in the
9261 ffmpeg-utils manual,ffmpeg-utils}.
9263 The default value of @var{shadowcolor} is "black".
9267 The x and y offsets for the text shadow position with respect to the
9268 position of the text. They can be either positive or negative
9269 values. The default value for both is "0".
9272 The starting frame number for the n/frame_num variable. The default value
9276 The size in number of spaces to use for rendering the tab.
9280 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
9281 format. It can be used with or without text parameter. @var{timecode_rate}
9282 option must be specified.
9284 @item timecode_rate, rate, r
9285 Set the timecode frame rate (timecode only). Value will be rounded to nearest
9286 integer. Minimum value is "1".
9287 Drop-frame timecode is supported for frame rates 30 & 60.
9290 If set to 1, the output of the timecode option will wrap around at 24 hours.
9291 Default is 0 (disabled).
9294 The text string to be drawn. The text must be a sequence of UTF-8
9296 This parameter is mandatory if no file is specified with the parameter
9300 A text file containing text to be drawn. The text must be a sequence
9301 of UTF-8 encoded characters.
9303 This parameter is mandatory if no text string is specified with the
9304 parameter @var{text}.
9306 If both @var{text} and @var{textfile} are specified, an error is thrown.
9309 If set to 1, the @var{textfile} will be reloaded before each frame.
9310 Be sure to update it atomically, or it may be read partially, or even fail.
9314 The expressions which specify the offsets where text will be drawn
9315 within the video frame. They are relative to the top/left border of the
9318 The default value of @var{x} and @var{y} is "0".
9320 See below for the list of accepted constants and functions.
9323 The parameters for @var{x} and @var{y} are expressions containing the
9324 following constants and functions:
9328 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
9332 horizontal and vertical chroma subsample values. For example for the
9333 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9336 the height of each text line
9344 @item max_glyph_a, ascent
9345 the maximum distance from the baseline to the highest/upper grid
9346 coordinate used to place a glyph outline point, for all the rendered
9348 It is a positive value, due to the grid's orientation with the Y axis
9351 @item max_glyph_d, descent
9352 the maximum distance from the baseline to the lowest grid coordinate
9353 used to place a glyph outline point, for all the rendered glyphs.
9354 This is a negative value, due to the grid's orientation, with the Y axis
9358 maximum glyph height, that is the maximum height for all the glyphs
9359 contained in the rendered text, it is equivalent to @var{ascent} -
9363 maximum glyph width, that is the maximum width for all the glyphs
9364 contained in the rendered text
9367 the number of input frame, starting from 0
9369 @item rand(min, max)
9370 return a random number included between @var{min} and @var{max}
9373 The input sample aspect ratio.
9376 timestamp expressed in seconds, NAN if the input timestamp is unknown
9379 the height of the rendered text
9382 the width of the rendered text
9386 the x and y offset coordinates where the text is drawn.
9388 These parameters allow the @var{x} and @var{y} expressions to refer
9389 to each other, so you can for example specify @code{y=x/dar}.
9392 A one character description of the current frame's picture type.
9395 The current packet's position in the input file or stream
9396 (in bytes, from the start of the input). A value of -1 indicates
9397 this info is not available.
9400 The current packet's duration, in seconds.
9403 The current packet's size (in bytes).
9406 @anchor{drawtext_expansion}
9407 @subsection Text expansion
9409 If @option{expansion} is set to @code{strftime},
9410 the filter recognizes strftime() sequences in the provided text and
9411 expands them accordingly. Check the documentation of strftime(). This
9412 feature is deprecated.
9414 If @option{expansion} is set to @code{none}, the text is printed verbatim.
9416 If @option{expansion} is set to @code{normal} (which is the default),
9417 the following expansion mechanism is used.
9419 The backslash character @samp{\}, followed by any character, always expands to
9420 the second character.
9422 Sequences of the form @code{%@{...@}} are expanded. The text between the
9423 braces is a function name, possibly followed by arguments separated by ':'.
9424 If the arguments contain special characters or delimiters (':' or '@}'),
9425 they should be escaped.
9427 Note that they probably must also be escaped as the value for the
9428 @option{text} option in the filter argument string and as the filter
9429 argument in the filtergraph description, and possibly also for the shell,
9430 that makes up to four levels of escaping; using a text file avoids these
9433 The following functions are available:
9438 The expression evaluation result.
9440 It must take one argument specifying the expression to be evaluated,
9441 which accepts the same constants and functions as the @var{x} and
9442 @var{y} values. Note that not all constants should be used, for
9443 example the text size is not known when evaluating the expression, so
9444 the constants @var{text_w} and @var{text_h} will have an undefined
9447 @item expr_int_format, eif
9448 Evaluate the expression's value and output as formatted integer.
9450 The first argument is the expression to be evaluated, just as for the @var{expr} function.
9451 The second argument specifies the output format. Allowed values are @samp{x},
9452 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
9453 @code{printf} function.
9454 The third parameter is optional and sets the number of positions taken by the output.
9455 It can be used to add padding with zeros from the left.
9458 The time at which the filter is running, expressed in UTC.
9459 It can accept an argument: a strftime() format string.
9462 The time at which the filter is running, expressed in the local time zone.
9463 It can accept an argument: a strftime() format string.
9466 Frame metadata. Takes one or two arguments.
9468 The first argument is mandatory and specifies the metadata key.
9470 The second argument is optional and specifies a default value, used when the
9471 metadata key is not found or empty.
9473 Available metadata can be identified by inspecting entries
9474 starting with TAG included within each frame section
9475 printed by running @code{ffprobe -show_frames}.
9477 String metadata generated in filters leading to
9478 the drawtext filter are also available.
9481 The frame number, starting from 0.
9484 A one character description of the current picture type.
9487 The timestamp of the current frame.
9488 It can take up to three arguments.
9490 The first argument is the format of the timestamp; it defaults to @code{flt}
9491 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
9492 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
9493 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
9494 @code{localtime} stands for the timestamp of the frame formatted as
9495 local time zone time.
9497 The second argument is an offset added to the timestamp.
9499 If the format is set to @code{hms}, a third argument @code{24HH} may be
9500 supplied to present the hour part of the formatted timestamp in 24h format
9503 If the format is set to @code{localtime} or @code{gmtime},
9504 a third argument may be supplied: a strftime() format string.
9505 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
9508 @subsection Commands
9510 This filter supports altering parameters via commands:
9513 Alter existing filter parameters.
9515 Syntax for the argument is the same as for filter invocation, e.g.
9518 fontsize=56:fontcolor=green:text='Hello World'
9521 Full filter invocation with sendcmd would look like this:
9524 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
9528 If the entire argument can't be parsed or applied as valid values then the filter will
9529 continue with its existing parameters.
9531 @subsection Examples
9535 Draw "Test Text" with font FreeSerif, using the default values for the
9536 optional parameters.
9539 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
9543 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
9544 and y=50 (counting from the top-left corner of the screen), text is
9545 yellow with a red box around it. Both the text and the box have an
9549 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
9550 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
9553 Note that the double quotes are not necessary if spaces are not used
9554 within the parameter list.
9557 Show the text at the center of the video frame:
9559 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
9563 Show the text at a random position, switching to a new position every 30 seconds:
9565 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)"
9569 Show a text line sliding from right to left in the last row of the video
9570 frame. The file @file{LONG_LINE} is assumed to contain a single line
9573 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
9577 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
9579 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
9583 Draw a single green letter "g", at the center of the input video.
9584 The glyph baseline is placed at half screen height.
9586 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
9590 Show text for 1 second every 3 seconds:
9592 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
9596 Use fontconfig to set the font. Note that the colons need to be escaped.
9598 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
9602 Print the date of a real-time encoding (see strftime(3)):
9604 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
9608 Show text fading in and out (appearing/disappearing):
9611 DS=1.0 # display start
9612 DE=10.0 # display end
9613 FID=1.5 # fade in duration
9614 FOD=5 # fade out duration
9615 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 @}"
9619 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
9620 and the @option{fontsize} value are included in the @option{y} offset.
9622 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
9623 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
9628 For more information about libfreetype, check:
9629 @url{http://www.freetype.org/}.
9631 For more information about fontconfig, check:
9632 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
9634 For more information about libfribidi, check:
9635 @url{http://fribidi.org/}.
9639 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
9641 The filter accepts the following options:
9646 Set low and high threshold values used by the Canny thresholding
9649 The high threshold selects the "strong" edge pixels, which are then
9650 connected through 8-connectivity with the "weak" edge pixels selected
9651 by the low threshold.
9653 @var{low} and @var{high} threshold values must be chosen in the range
9654 [0,1], and @var{low} should be lesser or equal to @var{high}.
9656 Default value for @var{low} is @code{20/255}, and default value for @var{high}
9660 Define the drawing mode.
9664 Draw white/gray wires on black background.
9667 Mix the colors to create a paint/cartoon effect.
9670 Apply Canny edge detector on all selected planes.
9672 Default value is @var{wires}.
9675 Select planes for filtering. By default all available planes are filtered.
9678 @subsection Examples
9682 Standard edge detection with custom values for the hysteresis thresholding:
9684 edgedetect=low=0.1:high=0.4
9688 Painting effect without thresholding:
9690 edgedetect=mode=colormix:high=0
9696 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
9698 For each input image, the filter will compute the optimal mapping from
9699 the input to the output given the codebook length, that is the number
9700 of distinct output colors.
9702 This filter accepts the following options.
9705 @item codebook_length, l
9706 Set codebook length. The value must be a positive integer, and
9707 represents the number of distinct output colors. Default value is 256.
9710 Set the maximum number of iterations to apply for computing the optimal
9711 mapping. The higher the value the better the result and the higher the
9712 computation time. Default value is 1.
9715 Set a random seed, must be an integer included between 0 and
9716 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
9717 will try to use a good random seed on a best effort basis.
9720 Set pal8 output pixel format. This option does not work with codebook
9721 length greater than 256.
9726 Measure graylevel entropy in histogram of color channels of video frames.
9728 It accepts the following parameters:
9732 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
9734 @var{diff} mode measures entropy of histogram delta values, absolute differences
9735 between neighbour histogram values.
9739 Set brightness, contrast, saturation and approximate gamma adjustment.
9741 The filter accepts the following options:
9745 Set the contrast expression. The value must be a float value in range
9746 @code{-1000.0} to @code{1000.0}. The default value is "1".
9749 Set the brightness expression. The value must be a float value in
9750 range @code{-1.0} to @code{1.0}. The default value is "0".
9753 Set the saturation expression. The value must be a float in
9754 range @code{0.0} to @code{3.0}. The default value is "1".
9757 Set the gamma expression. The value must be a float in range
9758 @code{0.1} to @code{10.0}. The default value is "1".
9761 Set the gamma expression for red. The value must be a float in
9762 range @code{0.1} to @code{10.0}. The default value is "1".
9765 Set the gamma expression for green. The value must be a float in range
9766 @code{0.1} to @code{10.0}. The default value is "1".
9769 Set the gamma expression for blue. The value must be a float in range
9770 @code{0.1} to @code{10.0}. The default value is "1".
9773 Set the gamma weight expression. It can be used to reduce the effect
9774 of a high gamma value on bright image areas, e.g. keep them from
9775 getting overamplified and just plain white. The value must be a float
9776 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
9777 gamma correction all the way down while @code{1.0} leaves it at its
9778 full strength. Default is "1".
9781 Set when the expressions for brightness, contrast, saturation and
9782 gamma expressions are evaluated.
9784 It accepts the following values:
9787 only evaluate expressions once during the filter initialization or
9788 when a command is processed
9791 evaluate expressions for each incoming frame
9794 Default value is @samp{init}.
9797 The expressions accept the following parameters:
9800 frame count of the input frame starting from 0
9803 byte position of the corresponding packet in the input file, NAN if
9807 frame rate of the input video, NAN if the input frame rate is unknown
9810 timestamp expressed in seconds, NAN if the input timestamp is unknown
9813 @subsection Commands
9814 The filter supports the following commands:
9818 Set the contrast expression.
9821 Set the brightness expression.
9824 Set the saturation expression.
9827 Set the gamma expression.
9830 Set the gamma_r expression.
9833 Set gamma_g expression.
9836 Set gamma_b expression.
9839 Set gamma_weight expression.
9841 The command accepts the same syntax of the corresponding option.
9843 If the specified expression is not valid, it is kept at its current
9850 Apply erosion effect to the video.
9852 This filter replaces the pixel by the local(3x3) minimum.
9854 It accepts the following options:
9861 Limit the maximum change for each plane, default is 65535.
9862 If 0, plane will remain unchanged.
9865 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9868 Flags to local 3x3 coordinates maps like this:
9875 @section extractplanes
9877 Extract color channel components from input video stream into
9878 separate grayscale video streams.
9880 The filter accepts the following option:
9884 Set plane(s) to extract.
9886 Available values for planes are:
9897 Choosing planes not available in the input will result in an error.
9898 That means you cannot select @code{r}, @code{g}, @code{b} planes
9899 with @code{y}, @code{u}, @code{v} planes at same time.
9902 @subsection Examples
9906 Extract luma, u and v color channel component from input video frame
9907 into 3 grayscale outputs:
9909 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
9915 Apply a fade-in/out effect to the input video.
9917 It accepts the following parameters:
9921 The effect type can be either "in" for a fade-in, or "out" for a fade-out
9923 Default is @code{in}.
9925 @item start_frame, s
9926 Specify the number of the frame to start applying the fade
9927 effect at. Default is 0.
9930 The number of frames that the fade effect lasts. At the end of the
9931 fade-in effect, the output video will have the same intensity as the input video.
9932 At the end of the fade-out transition, the output video will be filled with the
9933 selected @option{color}.
9937 If set to 1, fade only alpha channel, if one exists on the input.
9940 @item start_time, st
9941 Specify the timestamp (in seconds) of the frame to start to apply the fade
9942 effect. If both start_frame and start_time are specified, the fade will start at
9943 whichever comes last. Default is 0.
9946 The number of seconds for which the fade effect has to last. At the end of the
9947 fade-in effect the output video will have the same intensity as the input video,
9948 at the end of the fade-out transition the output video will be filled with the
9949 selected @option{color}.
9950 If both duration and nb_frames are specified, duration is used. Default is 0
9951 (nb_frames is used by default).
9954 Specify the color of the fade. Default is "black".
9957 @subsection Examples
9961 Fade in the first 30 frames of video:
9966 The command above is equivalent to:
9972 Fade out the last 45 frames of a 200-frame video:
9975 fade=type=out:start_frame=155:nb_frames=45
9979 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
9981 fade=in:0:25, fade=out:975:25
9985 Make the first 5 frames yellow, then fade in from frame 5-24:
9987 fade=in:5:20:color=yellow
9991 Fade in alpha over first 25 frames of video:
9993 fade=in:0:25:alpha=1
9997 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
9999 fade=t=in:st=5.5:d=0.5
10005 Denoise frames using 3D FFT (frequency domain filtering).
10007 The filter accepts the following options:
10011 Set the noise sigma constant. This sets denoising strength.
10012 Default value is 1. Allowed range is from 0 to 30.
10013 Using very high sigma with low overlap may give blocking artifacts.
10016 Set amount of denoising. By default all detected noise is reduced.
10017 Default value is 1. Allowed range is from 0 to 1.
10020 Set size of block, Default is 4, can be 3, 4, 5 or 6.
10021 Actual size of block in pixels is 2 to power of @var{block}, so by default
10022 block size in pixels is 2^4 which is 16.
10025 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
10028 Set number of previous frames to use for denoising. By default is set to 0.
10031 Set number of next frames to to use for denoising. By default is set to 0.
10034 Set planes which will be filtered, by default are all available filtered
10039 Apply arbitrary expressions to samples in frequency domain
10043 Adjust the dc value (gain) of the luma plane of the image. The filter
10044 accepts an integer value in range @code{0} to @code{1000}. The default
10045 value is set to @code{0}.
10048 Adjust the dc value (gain) of the 1st chroma plane of the image. The
10049 filter accepts an integer value in range @code{0} to @code{1000}. The
10050 default value is set to @code{0}.
10053 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
10054 filter accepts an integer value in range @code{0} to @code{1000}. The
10055 default value is set to @code{0}.
10058 Set the frequency domain weight expression for the luma plane.
10061 Set the frequency domain weight expression for the 1st chroma plane.
10064 Set the frequency domain weight expression for the 2nd chroma plane.
10067 Set when the expressions are evaluated.
10069 It accepts the following values:
10072 Only evaluate expressions once during the filter initialization.
10075 Evaluate expressions for each incoming frame.
10078 Default value is @samp{init}.
10080 The filter accepts the following variables:
10083 The coordinates of the current sample.
10087 The width and height of the image.
10090 The number of input frame, starting from 0.
10093 @subsection Examples
10099 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
10105 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
10111 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
10117 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
10124 Extract a single field from an interlaced image using stride
10125 arithmetic to avoid wasting CPU time. The output frames are marked as
10128 The filter accepts the following options:
10132 Specify whether to extract the top (if the value is @code{0} or
10133 @code{top}) or the bottom field (if the value is @code{1} or
10139 Create new frames by copying the top and bottom fields from surrounding frames
10140 supplied as numbers by the hint file.
10144 Set file containing hints: absolute/relative frame numbers.
10146 There must be one line for each frame in a clip. Each line must contain two
10147 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
10148 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
10149 is current frame number for @code{absolute} mode or out of [-1, 1] range
10150 for @code{relative} mode. First number tells from which frame to pick up top
10151 field and second number tells from which frame to pick up bottom field.
10153 If optionally followed by @code{+} output frame will be marked as interlaced,
10154 else if followed by @code{-} output frame will be marked as progressive, else
10155 it will be marked same as input frame.
10156 If optionally followed by @code{t} output frame will use only top field, or in
10157 case of @code{b} it will use only bottom field.
10158 If line starts with @code{#} or @code{;} that line is skipped.
10161 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
10164 Example of first several lines of @code{hint} file for @code{relative} mode:
10166 0,0 - # first frame
10167 1,0 - # second frame, use third's frame top field and second's frame bottom field
10168 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
10183 @section fieldmatch
10185 Field matching filter for inverse telecine. It is meant to reconstruct the
10186 progressive frames from a telecined stream. The filter does not drop duplicated
10187 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
10188 followed by a decimation filter such as @ref{decimate} in the filtergraph.
10190 The separation of the field matching and the decimation is notably motivated by
10191 the possibility of inserting a de-interlacing filter fallback between the two.
10192 If the source has mixed telecined and real interlaced content,
10193 @code{fieldmatch} will not be able to match fields for the interlaced parts.
10194 But these remaining combed frames will be marked as interlaced, and thus can be
10195 de-interlaced by a later filter such as @ref{yadif} before decimation.
10197 In addition to the various configuration options, @code{fieldmatch} can take an
10198 optional second stream, activated through the @option{ppsrc} option. If
10199 enabled, the frames reconstruction will be based on the fields and frames from
10200 this second stream. This allows the first input to be pre-processed in order to
10201 help the various algorithms of the filter, while keeping the output lossless
10202 (assuming the fields are matched properly). Typically, a field-aware denoiser,
10203 or brightness/contrast adjustments can help.
10205 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
10206 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
10207 which @code{fieldmatch} is based on. While the semantic and usage are very
10208 close, some behaviour and options names can differ.
10210 The @ref{decimate} filter currently only works for constant frame rate input.
10211 If your input has mixed telecined (30fps) and progressive content with a lower
10212 framerate like 24fps use the following filterchain to produce the necessary cfr
10213 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
10215 The filter accepts the following options:
10219 Specify the assumed field order of the input stream. Available values are:
10223 Auto detect parity (use FFmpeg's internal parity value).
10225 Assume bottom field first.
10227 Assume top field first.
10230 Note that it is sometimes recommended not to trust the parity announced by the
10233 Default value is @var{auto}.
10236 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
10237 sense that it won't risk creating jerkiness due to duplicate frames when
10238 possible, but if there are bad edits or blended fields it will end up
10239 outputting combed frames when a good match might actually exist. On the other
10240 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
10241 but will almost always find a good frame if there is one. The other values are
10242 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
10243 jerkiness and creating duplicate frames versus finding good matches in sections
10244 with bad edits, orphaned fields, blended fields, etc.
10246 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
10248 Available values are:
10252 2-way matching (p/c)
10254 2-way matching, and trying 3rd match if still combed (p/c + n)
10256 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
10258 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
10259 still combed (p/c + n + u/b)
10261 3-way matching (p/c/n)
10263 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
10264 detected as combed (p/c/n + u/b)
10267 The parenthesis at the end indicate the matches that would be used for that
10268 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
10271 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
10274 Default value is @var{pc_n}.
10277 Mark the main input stream as a pre-processed input, and enable the secondary
10278 input stream as the clean source to pick the fields from. See the filter
10279 introduction for more details. It is similar to the @option{clip2} feature from
10282 Default value is @code{0} (disabled).
10285 Set the field to match from. It is recommended to set this to the same value as
10286 @option{order} unless you experience matching failures with that setting. In
10287 certain circumstances changing the field that is used to match from can have a
10288 large impact on matching performance. Available values are:
10292 Automatic (same value as @option{order}).
10294 Match from the bottom field.
10296 Match from the top field.
10299 Default value is @var{auto}.
10302 Set whether or not chroma is included during the match comparisons. In most
10303 cases it is recommended to leave this enabled. You should set this to @code{0}
10304 only if your clip has bad chroma problems such as heavy rainbowing or other
10305 artifacts. Setting this to @code{0} could also be used to speed things up at
10306 the cost of some accuracy.
10308 Default value is @code{1}.
10312 These define an exclusion band which excludes the lines between @option{y0} and
10313 @option{y1} from being included in the field matching decision. An exclusion
10314 band can be used to ignore subtitles, a logo, or other things that may
10315 interfere with the matching. @option{y0} sets the starting scan line and
10316 @option{y1} sets the ending line; all lines in between @option{y0} and
10317 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
10318 @option{y0} and @option{y1} to the same value will disable the feature.
10319 @option{y0} and @option{y1} defaults to @code{0}.
10322 Set the scene change detection threshold as a percentage of maximum change on
10323 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
10324 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
10325 @option{scthresh} is @code{[0.0, 100.0]}.
10327 Default value is @code{12.0}.
10330 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
10331 account the combed scores of matches when deciding what match to use as the
10332 final match. Available values are:
10336 No final matching based on combed scores.
10338 Combed scores are only used when a scene change is detected.
10340 Use combed scores all the time.
10343 Default is @var{sc}.
10346 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
10347 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
10348 Available values are:
10352 No forced calculation.
10354 Force p/c/n calculations.
10356 Force p/c/n/u/b calculations.
10359 Default value is @var{none}.
10362 This is the area combing threshold used for combed frame detection. This
10363 essentially controls how "strong" or "visible" combing must be to be detected.
10364 Larger values mean combing must be more visible and smaller values mean combing
10365 can be less visible or strong and still be detected. Valid settings are from
10366 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
10367 be detected as combed). This is basically a pixel difference value. A good
10368 range is @code{[8, 12]}.
10370 Default value is @code{9}.
10373 Sets whether or not chroma is considered in the combed frame decision. Only
10374 disable this if your source has chroma problems (rainbowing, etc.) that are
10375 causing problems for the combed frame detection with chroma enabled. Actually,
10376 using @option{chroma}=@var{0} is usually more reliable, except for the case
10377 where there is chroma only combing in the source.
10379 Default value is @code{0}.
10383 Respectively set the x-axis and y-axis size of the window used during combed
10384 frame detection. This has to do with the size of the area in which
10385 @option{combpel} pixels are required to be detected as combed for a frame to be
10386 declared combed. See the @option{combpel} parameter description for more info.
10387 Possible values are any number that is a power of 2 starting at 4 and going up
10390 Default value is @code{16}.
10393 The number of combed pixels inside any of the @option{blocky} by
10394 @option{blockx} size blocks on the frame for the frame to be detected as
10395 combed. While @option{cthresh} controls how "visible" the combing must be, this
10396 setting controls "how much" combing there must be in any localized area (a
10397 window defined by the @option{blockx} and @option{blocky} settings) on the
10398 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
10399 which point no frames will ever be detected as combed). This setting is known
10400 as @option{MI} in TFM/VFM vocabulary.
10402 Default value is @code{80}.
10405 @anchor{p/c/n/u/b meaning}
10406 @subsection p/c/n/u/b meaning
10408 @subsubsection p/c/n
10410 We assume the following telecined stream:
10413 Top fields: 1 2 2 3 4
10414 Bottom fields: 1 2 3 4 4
10417 The numbers correspond to the progressive frame the fields relate to. Here, the
10418 first two frames are progressive, the 3rd and 4th are combed, and so on.
10420 When @code{fieldmatch} is configured to run a matching from bottom
10421 (@option{field}=@var{bottom}) this is how this input stream get transformed:
10426 B 1 2 3 4 4 <-- matching reference
10435 As a result of the field matching, we can see that some frames get duplicated.
10436 To perform a complete inverse telecine, you need to rely on a decimation filter
10437 after this operation. See for instance the @ref{decimate} filter.
10439 The same operation now matching from top fields (@option{field}=@var{top})
10444 T 1 2 2 3 4 <-- matching reference
10454 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
10455 basically, they refer to the frame and field of the opposite parity:
10458 @item @var{p} matches the field of the opposite parity in the previous frame
10459 @item @var{c} matches the field of the opposite parity in the current frame
10460 @item @var{n} matches the field of the opposite parity in the next frame
10465 The @var{u} and @var{b} matching are a bit special in the sense that they match
10466 from the opposite parity flag. In the following examples, we assume that we are
10467 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
10468 'x' is placed above and below each matched fields.
10470 With bottom matching (@option{field}=@var{bottom}):
10475 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
10476 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
10484 With top matching (@option{field}=@var{top}):
10489 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
10490 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
10498 @subsection Examples
10500 Simple IVTC of a top field first telecined stream:
10502 fieldmatch=order=tff:combmatch=none, decimate
10505 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
10507 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
10510 @section fieldorder
10512 Transform the field order of the input video.
10514 It accepts the following parameters:
10519 The output field order. Valid values are @var{tff} for top field first or @var{bff}
10520 for bottom field first.
10523 The default value is @samp{tff}.
10525 The transformation is done by shifting the picture content up or down
10526 by one line, and filling the remaining line with appropriate picture content.
10527 This method is consistent with most broadcast field order converters.
10529 If the input video is not flagged as being interlaced, or it is already
10530 flagged as being of the required output field order, then this filter does
10531 not alter the incoming video.
10533 It is very useful when converting to or from PAL DV material,
10534 which is bottom field first.
10538 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
10541 @section fifo, afifo
10543 Buffer input images and send them when they are requested.
10545 It is mainly useful when auto-inserted by the libavfilter
10548 It does not take parameters.
10550 @section fillborders
10552 Fill borders of the input video, without changing video stream dimensions.
10553 Sometimes video can have garbage at the four edges and you may not want to
10554 crop video input to keep size multiple of some number.
10556 This filter accepts the following options:
10560 Number of pixels to fill from left border.
10563 Number of pixels to fill from right border.
10566 Number of pixels to fill from top border.
10569 Number of pixels to fill from bottom border.
10574 It accepts the following values:
10577 fill pixels using outermost pixels
10580 fill pixels using mirroring
10583 fill pixels with constant value
10586 Default is @var{smear}.
10589 Set color for pixels in fixed mode. Default is @var{black}.
10594 Find a rectangular object
10596 It accepts the following options:
10600 Filepath of the object image, needs to be in gray8.
10603 Detection threshold, default is 0.5.
10606 Number of mipmaps, default is 3.
10608 @item xmin, ymin, xmax, ymax
10609 Specifies the rectangle in which to search.
10612 @subsection Examples
10616 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
10618 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
10624 Flood area with values of same pixel components with another values.
10626 It accepts the following options:
10629 Set pixel x coordinate.
10632 Set pixel y coordinate.
10635 Set source #0 component value.
10638 Set source #1 component value.
10641 Set source #2 component value.
10644 Set source #3 component value.
10647 Set destination #0 component value.
10650 Set destination #1 component value.
10653 Set destination #2 component value.
10656 Set destination #3 component value.
10662 Convert the input video to one of the specified pixel formats.
10663 Libavfilter will try to pick one that is suitable as input to
10666 It accepts the following parameters:
10670 A '|'-separated list of pixel format names, such as
10671 "pix_fmts=yuv420p|monow|rgb24".
10675 @subsection Examples
10679 Convert the input video to the @var{yuv420p} format
10681 format=pix_fmts=yuv420p
10684 Convert the input video to any of the formats in the list
10686 format=pix_fmts=yuv420p|yuv444p|yuv410p
10693 Convert the video to specified constant frame rate by duplicating or dropping
10694 frames as necessary.
10696 It accepts the following parameters:
10700 The desired output frame rate. The default is @code{25}.
10703 Assume the first PTS should be the given value, in seconds. This allows for
10704 padding/trimming at the start of stream. By default, no assumption is made
10705 about the first frame's expected PTS, so no padding or trimming is done.
10706 For example, this could be set to 0 to pad the beginning with duplicates of
10707 the first frame if a video stream starts after the audio stream or to trim any
10708 frames with a negative PTS.
10711 Timestamp (PTS) rounding method.
10713 Possible values are:
10720 round towards -infinity
10722 round towards +infinity
10726 The default is @code{near}.
10729 Action performed when reading the last frame.
10731 Possible values are:
10734 Use same timestamp rounding method as used for other frames.
10736 Pass through last frame if input duration has not been reached yet.
10738 The default is @code{round}.
10742 Alternatively, the options can be specified as a flat string:
10743 @var{fps}[:@var{start_time}[:@var{round}]].
10745 See also the @ref{setpts} filter.
10747 @subsection Examples
10751 A typical usage in order to set the fps to 25:
10757 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
10759 fps=fps=film:round=near
10765 Pack two different video streams into a stereoscopic video, setting proper
10766 metadata on supported codecs. The two views should have the same size and
10767 framerate and processing will stop when the shorter video ends. Please note
10768 that you may conveniently adjust view properties with the @ref{scale} and
10771 It accepts the following parameters:
10775 The desired packing format. Supported values are:
10780 The views are next to each other (default).
10783 The views are on top of each other.
10786 The views are packed by line.
10789 The views are packed by column.
10792 The views are temporally interleaved.
10801 # Convert left and right views into a frame-sequential video
10802 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
10804 # Convert views into a side-by-side video with the same output resolution as the input
10805 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
10810 Change the frame rate by interpolating new video output frames from the source
10813 This filter is not designed to function correctly with interlaced media. If
10814 you wish to change the frame rate of interlaced media then you are required
10815 to deinterlace before this filter and re-interlace after this filter.
10817 A description of the accepted options follows.
10821 Specify the output frames per second. This option can also be specified
10822 as a value alone. The default is @code{50}.
10825 Specify the start of a range where the output frame will be created as a
10826 linear interpolation of two frames. The range is [@code{0}-@code{255}],
10827 the default is @code{15}.
10830 Specify the end of a range where the output frame will be created as a
10831 linear interpolation of two frames. The range is [@code{0}-@code{255}],
10832 the default is @code{240}.
10835 Specify the level at which a scene change is detected as a value between
10836 0 and 100 to indicate a new scene; a low value reflects a low
10837 probability for the current frame to introduce a new scene, while a higher
10838 value means the current frame is more likely to be one.
10839 The default is @code{8.2}.
10842 Specify flags influencing the filter process.
10844 Available value for @var{flags} is:
10847 @item scene_change_detect, scd
10848 Enable scene change detection using the value of the option @var{scene}.
10849 This flag is enabled by default.
10855 Select one frame every N-th frame.
10857 This filter accepts the following option:
10860 Select frame after every @code{step} frames.
10861 Allowed values are positive integers higher than 0. Default value is @code{1}.
10864 @section freezedetect
10866 Detect frozen video.
10868 This filter logs a message and sets frame metadata when it detects that the
10869 input video has no significant change in content during a specified duration.
10870 Video freeze detection calculates the mean average absolute difference of all
10871 the components of video frames and compares it to a noise floor.
10873 The printed times and duration are expressed in seconds. The
10874 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
10875 whose timestamp equals or exceeds the detection duration and it contains the
10876 timestamp of the first frame of the freeze. The
10877 @code{lavfi.freezedetect.freeze_duration} and
10878 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
10881 The filter accepts the following options:
10885 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
10886 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
10890 Set freeze duration until notification (default is 2 seconds).
10896 Apply a frei0r effect to the input video.
10898 To enable the compilation of this filter, you need to install the frei0r
10899 header and configure FFmpeg with @code{--enable-frei0r}.
10901 It accepts the following parameters:
10906 The name of the frei0r effect to load. If the environment variable
10907 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
10908 directories specified by the colon-separated list in @env{FREI0R_PATH}.
10909 Otherwise, the standard frei0r paths are searched, in this order:
10910 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
10911 @file{/usr/lib/frei0r-1/}.
10913 @item filter_params
10914 A '|'-separated list of parameters to pass to the frei0r effect.
10918 A frei0r effect parameter can be a boolean (its value is either
10919 "y" or "n"), a double, a color (specified as
10920 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
10921 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
10922 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
10923 a position (specified as @var{X}/@var{Y}, where
10924 @var{X} and @var{Y} are floating point numbers) and/or a string.
10926 The number and types of parameters depend on the loaded effect. If an
10927 effect parameter is not specified, the default value is set.
10929 @subsection Examples
10933 Apply the distort0r effect, setting the first two double parameters:
10935 frei0r=filter_name=distort0r:filter_params=0.5|0.01
10939 Apply the colordistance effect, taking a color as the first parameter:
10941 frei0r=colordistance:0.2/0.3/0.4
10942 frei0r=colordistance:violet
10943 frei0r=colordistance:0x112233
10947 Apply the perspective effect, specifying the top left and top right image
10950 frei0r=perspective:0.2/0.2|0.8/0.2
10954 For more information, see
10955 @url{http://frei0r.dyne.org}
10959 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
10961 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
10962 processing filter, one of them is performed once per block, not per pixel.
10963 This allows for much higher speed.
10965 The filter accepts the following options:
10969 Set quality. This option defines the number of levels for averaging. It accepts
10970 an integer in the range 4-5. Default value is @code{4}.
10973 Force a constant quantization parameter. It accepts an integer in range 0-63.
10974 If not set, the filter will use the QP from the video stream (if available).
10977 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
10978 more details but also more artifacts, while higher values make the image smoother
10979 but also blurrier. Default value is @code{0} − PSNR optimal.
10981 @item use_bframe_qp
10982 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
10983 option may cause flicker since the B-Frames have often larger QP. Default is
10984 @code{0} (not enabled).
10990 Apply Gaussian blur filter.
10992 The filter accepts the following options:
10996 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
10999 Set number of steps for Gaussian approximation. Default is @code{1}.
11002 Set which planes to filter. By default all planes are filtered.
11005 Set vertical sigma, if negative it will be same as @code{sigma}.
11006 Default is @code{-1}.
11009 @subsection Commands
11010 This filter supports same commands as options.
11011 The command accepts the same syntax of the corresponding option.
11013 If the specified expression is not valid, it is kept at its current
11018 Apply generic equation to each pixel.
11020 The filter accepts the following options:
11023 @item lum_expr, lum
11024 Set the luminance expression.
11026 Set the chrominance blue expression.
11028 Set the chrominance red expression.
11029 @item alpha_expr, a
11030 Set the alpha expression.
11032 Set the red expression.
11033 @item green_expr, g
11034 Set the green expression.
11036 Set the blue expression.
11039 The colorspace is selected according to the specified options. If one
11040 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
11041 options is specified, the filter will automatically select a YCbCr
11042 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
11043 @option{blue_expr} options is specified, it will select an RGB
11046 If one of the chrominance expression is not defined, it falls back on the other
11047 one. If no alpha expression is specified it will evaluate to opaque value.
11048 If none of chrominance expressions are specified, they will evaluate
11049 to the luminance expression.
11051 The expressions can use the following variables and functions:
11055 The sequential number of the filtered frame, starting from @code{0}.
11059 The coordinates of the current sample.
11063 The width and height of the image.
11067 Width and height scale depending on the currently filtered plane. It is the
11068 ratio between the corresponding luma plane number of pixels and the current
11069 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
11070 @code{0.5,0.5} for chroma planes.
11073 Time of the current frame, expressed in seconds.
11076 Return the value of the pixel at location (@var{x},@var{y}) of the current
11080 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
11084 Return the value of the pixel at location (@var{x},@var{y}) of the
11085 blue-difference chroma plane. Return 0 if there is no such plane.
11088 Return the value of the pixel at location (@var{x},@var{y}) of the
11089 red-difference chroma plane. Return 0 if there is no such plane.
11094 Return the value of the pixel at location (@var{x},@var{y}) of the
11095 red/green/blue component. Return 0 if there is no such component.
11098 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
11099 plane. Return 0 if there is no such plane.
11101 @item interpolation
11102 Set one of interpolation methods:
11107 Default is bilinear.
11110 For functions, if @var{x} and @var{y} are outside the area, the value will be
11111 automatically clipped to the closer edge.
11113 @subsection Examples
11117 Flip the image horizontally:
11123 Generate a bidimensional sine wave, with angle @code{PI/3} and a
11124 wavelength of 100 pixels:
11126 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
11130 Generate a fancy enigmatic moving light:
11132 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
11136 Generate a quick emboss effect:
11138 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
11142 Modify RGB components depending on pixel position:
11144 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
11148 Create a radial gradient that is the same size as the input (also see
11149 the @ref{vignette} filter):
11151 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
11157 Fix the banding artifacts that are sometimes introduced into nearly flat
11158 regions by truncation to 8-bit color depth.
11159 Interpolate the gradients that should go where the bands are, and
11162 It is designed for playback only. Do not use it prior to
11163 lossy compression, because compression tends to lose the dither and
11164 bring back the bands.
11166 It accepts the following parameters:
11171 The maximum amount by which the filter will change any one pixel. This is also
11172 the threshold for detecting nearly flat regions. Acceptable values range from
11173 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
11177 The neighborhood to fit the gradient to. A larger radius makes for smoother
11178 gradients, but also prevents the filter from modifying the pixels near detailed
11179 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
11180 values will be clipped to the valid range.
11184 Alternatively, the options can be specified as a flat string:
11185 @var{strength}[:@var{radius}]
11187 @subsection Examples
11191 Apply the filter with a @code{3.5} strength and radius of @code{8}:
11197 Specify radius, omitting the strength (which will fall-back to the default
11205 @section graphmonitor, agraphmonitor
11206 Show various filtergraph stats.
11208 With this filter one can debug complete filtergraph.
11209 Especially issues with links filling with queued frames.
11211 The filter accepts the following options:
11215 Set video output size. Default is @var{hd720}.
11218 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
11221 Set output mode, can be @var{fulll} or @var{compact}.
11222 In @var{compact} mode only filters with some queued frames have displayed stats.
11225 Set flags which enable which stats are shown in video.
11227 Available values for flags are:
11230 Display number of queued frames in each link.
11232 @item frame_count_in
11233 Display number of frames taken from filter.
11235 @item frame_count_out
11236 Display number of frames given out from filter.
11239 Display current filtered frame pts.
11242 Display current filtered frame time.
11245 Display time base for filter link.
11248 Display used format for filter link.
11251 Display video size or number of audio channels in case of audio used by filter link.
11254 Display video frame rate or sample rate in case of audio used by filter link.
11258 Set upper limit for video rate of output stream, Default value is @var{25}.
11259 This guarantee that output video frame rate will not be higher than this value.
11263 A color constancy variation filter which estimates scene illumination via grey edge algorithm
11264 and corrects the scene colors accordingly.
11266 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
11268 The filter accepts the following options:
11272 The order of differentiation to be applied on the scene. Must be chosen in the range
11273 [0,2] and default value is 1.
11276 The Minkowski parameter to be used for calculating the Minkowski distance. Must
11277 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
11278 max value instead of calculating Minkowski distance.
11281 The standard deviation of Gaussian blur to be applied on the scene. Must be
11282 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
11283 can't be equal to 0 if @var{difford} is greater than 0.
11286 @subsection Examples
11292 greyedge=difford=1:minknorm=5:sigma=2
11298 greyedge=difford=1:minknorm=0:sigma=2
11306 Apply a Hald CLUT to a video stream.
11308 First input is the video stream to process, and second one is the Hald CLUT.
11309 The Hald CLUT input can be a simple picture or a complete video stream.
11311 The filter accepts the following options:
11315 Force termination when the shortest input terminates. Default is @code{0}.
11317 Continue applying the last CLUT after the end of the stream. A value of
11318 @code{0} disable the filter after the last frame of the CLUT is reached.
11319 Default is @code{1}.
11322 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
11323 filters share the same internals).
11325 This filter also supports the @ref{framesync} options.
11327 More information about the Hald CLUT can be found on Eskil Steenberg's website
11328 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
11330 @subsection Workflow examples
11332 @subsubsection Hald CLUT video stream
11334 Generate an identity Hald CLUT stream altered with various effects:
11336 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
11339 Note: make sure you use a lossless codec.
11341 Then use it with @code{haldclut} to apply it on some random stream:
11343 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
11346 The Hald CLUT will be applied to the 10 first seconds (duration of
11347 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
11348 to the remaining frames of the @code{mandelbrot} stream.
11350 @subsubsection Hald CLUT with preview
11352 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
11353 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
11354 biggest possible square starting at the top left of the picture. The remaining
11355 padding pixels (bottom or right) will be ignored. This area can be used to add
11356 a preview of the Hald CLUT.
11358 Typically, the following generated Hald CLUT will be supported by the
11359 @code{haldclut} filter:
11362 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
11363 pad=iw+320 [padded_clut];
11364 smptebars=s=320x256, split [a][b];
11365 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
11366 [main][b] overlay=W-320" -frames:v 1 clut.png
11369 It contains the original and a preview of the effect of the CLUT: SMPTE color
11370 bars are displayed on the right-top, and below the same color bars processed by
11373 Then, the effect of this Hald CLUT can be visualized with:
11375 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
11380 Flip the input video horizontally.
11382 For example, to horizontally flip the input video with @command{ffmpeg}:
11384 ffmpeg -i in.avi -vf "hflip" out.avi
11388 This filter applies a global color histogram equalization on a
11391 It can be used to correct video that has a compressed range of pixel
11392 intensities. The filter redistributes the pixel intensities to
11393 equalize their distribution across the intensity range. It may be
11394 viewed as an "automatically adjusting contrast filter". This filter is
11395 useful only for correcting degraded or poorly captured source
11398 The filter accepts the following options:
11402 Determine the amount of equalization to be applied. As the strength
11403 is reduced, the distribution of pixel intensities more-and-more
11404 approaches that of the input frame. The value must be a float number
11405 in the range [0,1] and defaults to 0.200.
11408 Set the maximum intensity that can generated and scale the output
11409 values appropriately. The strength should be set as desired and then
11410 the intensity can be limited if needed to avoid washing-out. The value
11411 must be a float number in the range [0,1] and defaults to 0.210.
11414 Set the antibanding level. If enabled the filter will randomly vary
11415 the luminance of output pixels by a small amount to avoid banding of
11416 the histogram. Possible values are @code{none}, @code{weak} or
11417 @code{strong}. It defaults to @code{none}.
11422 Compute and draw a color distribution histogram for the input video.
11424 The computed histogram is a representation of the color component
11425 distribution in an image.
11427 Standard histogram displays the color components distribution in an image.
11428 Displays color graph for each color component. Shows distribution of
11429 the Y, U, V, A or R, G, B components, depending on input format, in the
11430 current frame. Below each graph a color component scale meter is shown.
11432 The filter accepts the following options:
11436 Set height of level. Default value is @code{200}.
11437 Allowed range is [50, 2048].
11440 Set height of color scale. Default value is @code{12}.
11441 Allowed range is [0, 40].
11445 It accepts the following values:
11448 Per color component graphs are placed below each other.
11451 Per color component graphs are placed side by side.
11454 Presents information identical to that in the @code{parade}, except
11455 that the graphs representing color components are superimposed directly
11458 Default is @code{stack}.
11461 Set mode. Can be either @code{linear}, or @code{logarithmic}.
11462 Default is @code{linear}.
11465 Set what color components to display.
11466 Default is @code{7}.
11469 Set foreground opacity. Default is @code{0.7}.
11472 Set background opacity. Default is @code{0.5}.
11475 @subsection Examples
11480 Calculate and draw histogram:
11482 ffplay -i input -vf histogram
11490 This is a high precision/quality 3d denoise filter. It aims to reduce
11491 image noise, producing smooth images and making still images really
11492 still. It should enhance compressibility.
11494 It accepts the following optional parameters:
11498 A non-negative floating point number which specifies spatial luma strength.
11499 It defaults to 4.0.
11501 @item chroma_spatial
11502 A non-negative floating point number which specifies spatial chroma strength.
11503 It defaults to 3.0*@var{luma_spatial}/4.0.
11506 A floating point number which specifies luma temporal strength. It defaults to
11507 6.0*@var{luma_spatial}/4.0.
11510 A floating point number which specifies chroma temporal strength. It defaults to
11511 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
11514 @anchor{hwdownload}
11515 @section hwdownload
11517 Download hardware frames to system memory.
11519 The input must be in hardware frames, and the output a non-hardware format.
11520 Not all formats will be supported on the output - it may be necessary to insert
11521 an additional @option{format} filter immediately following in the graph to get
11522 the output in a supported format.
11526 Map hardware frames to system memory or to another device.
11528 This filter has several different modes of operation; which one is used depends
11529 on the input and output formats:
11532 Hardware frame input, normal frame output
11534 Map the input frames to system memory and pass them to the output. If the
11535 original hardware frame is later required (for example, after overlaying
11536 something else on part of it), the @option{hwmap} filter can be used again
11537 in the next mode to retrieve it.
11539 Normal frame input, hardware frame output
11541 If the input is actually a software-mapped hardware frame, then unmap it -
11542 that is, return the original hardware frame.
11544 Otherwise, a device must be provided. Create new hardware surfaces on that
11545 device for the output, then map them back to the software format at the input
11546 and give those frames to the preceding filter. This will then act like the
11547 @option{hwupload} filter, but may be able to avoid an additional copy when
11548 the input is already in a compatible format.
11550 Hardware frame input and output
11552 A device must be supplied for the output, either directly or with the
11553 @option{derive_device} option. The input and output devices must be of
11554 different types and compatible - the exact meaning of this is
11555 system-dependent, but typically it means that they must refer to the same
11556 underlying hardware context (for example, refer to the same graphics card).
11558 If the input frames were originally created on the output device, then unmap
11559 to retrieve the original frames.
11561 Otherwise, map the frames to the output device - create new hardware frames
11562 on the output corresponding to the frames on the input.
11565 The following additional parameters are accepted:
11569 Set the frame mapping mode. Some combination of:
11572 The mapped frame should be readable.
11574 The mapped frame should be writeable.
11576 The mapping will always overwrite the entire frame.
11578 This may improve performance in some cases, as the original contents of the
11579 frame need not be loaded.
11581 The mapping must not involve any copying.
11583 Indirect mappings to copies of frames are created in some cases where either
11584 direct mapping is not possible or it would have unexpected properties.
11585 Setting this flag ensures that the mapping is direct and will fail if that is
11588 Defaults to @var{read+write} if not specified.
11590 @item derive_device @var{type}
11591 Rather than using the device supplied at initialisation, instead derive a new
11592 device of type @var{type} from the device the input frames exist on.
11595 In a hardware to hardware mapping, map in reverse - create frames in the sink
11596 and map them back to the source. This may be necessary in some cases where
11597 a mapping in one direction is required but only the opposite direction is
11598 supported by the devices being used.
11600 This option is dangerous - it may break the preceding filter in undefined
11601 ways if there are any additional constraints on that filter's output.
11602 Do not use it without fully understanding the implications of its use.
11608 Upload system memory frames to hardware surfaces.
11610 The device to upload to must be supplied when the filter is initialised. If
11611 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
11614 @anchor{hwupload_cuda}
11615 @section hwupload_cuda
11617 Upload system memory frames to a CUDA device.
11619 It accepts the following optional parameters:
11623 The number of the CUDA device to use
11628 Apply a high-quality magnification filter designed for pixel art. This filter
11629 was originally created by Maxim Stepin.
11631 It accepts the following option:
11635 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
11636 @code{hq3x} and @code{4} for @code{hq4x}.
11637 Default is @code{3}.
11641 Stack input videos horizontally.
11643 All streams must be of same pixel format and of same height.
11645 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
11646 to create same output.
11648 The filter accepts the following option:
11652 Set number of input streams. Default is 2.
11655 If set to 1, force the output to terminate when the shortest input
11656 terminates. Default value is 0.
11661 Modify the hue and/or the saturation of the input.
11663 It accepts the following parameters:
11667 Specify the hue angle as a number of degrees. It accepts an expression,
11668 and defaults to "0".
11671 Specify the saturation in the [-10,10] range. It accepts an expression and
11675 Specify the hue angle as a number of radians. It accepts an
11676 expression, and defaults to "0".
11679 Specify the brightness in the [-10,10] range. It accepts an expression and
11683 @option{h} and @option{H} are mutually exclusive, and can't be
11684 specified at the same time.
11686 The @option{b}, @option{h}, @option{H} and @option{s} option values are
11687 expressions containing the following constants:
11691 frame count of the input frame starting from 0
11694 presentation timestamp of the input frame expressed in time base units
11697 frame rate of the input video, NAN if the input frame rate is unknown
11700 timestamp expressed in seconds, NAN if the input timestamp is unknown
11703 time base of the input video
11706 @subsection Examples
11710 Set the hue to 90 degrees and the saturation to 1.0:
11716 Same command but expressing the hue in radians:
11722 Rotate hue and make the saturation swing between 0
11723 and 2 over a period of 1 second:
11725 hue="H=2*PI*t: s=sin(2*PI*t)+1"
11729 Apply a 3 seconds saturation fade-in effect starting at 0:
11731 hue="s=min(t/3\,1)"
11734 The general fade-in expression can be written as:
11736 hue="s=min(0\, max((t-START)/DURATION\, 1))"
11740 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
11742 hue="s=max(0\, min(1\, (8-t)/3))"
11745 The general fade-out expression can be written as:
11747 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
11752 @subsection Commands
11754 This filter supports the following commands:
11760 Modify the hue and/or the saturation and/or brightness of the input video.
11761 The command accepts the same syntax of the corresponding option.
11763 If the specified expression is not valid, it is kept at its current
11767 @section hysteresis
11769 Grow first stream into second stream by connecting components.
11770 This makes it possible to build more robust edge masks.
11772 This filter accepts the following options:
11776 Set which planes will be processed as bitmap, unprocessed planes will be
11777 copied from first stream.
11778 By default value 0xf, all planes will be processed.
11781 Set threshold which is used in filtering. If pixel component value is higher than
11782 this value filter algorithm for connecting components is activated.
11783 By default value is 0.
11788 Detect video interlacing type.
11790 This filter tries to detect if the input frames are interlaced, progressive,
11791 top or bottom field first. It will also try to detect fields that are
11792 repeated between adjacent frames (a sign of telecine).
11794 Single frame detection considers only immediately adjacent frames when classifying each frame.
11795 Multiple frame detection incorporates the classification history of previous frames.
11797 The filter will log these metadata values:
11800 @item single.current_frame
11801 Detected type of current frame using single-frame detection. One of:
11802 ``tff'' (top field first), ``bff'' (bottom field first),
11803 ``progressive'', or ``undetermined''
11806 Cumulative number of frames detected as top field first using single-frame detection.
11809 Cumulative number of frames detected as top field first using multiple-frame detection.
11812 Cumulative number of frames detected as bottom field first using single-frame detection.
11814 @item multiple.current_frame
11815 Detected type of current frame using multiple-frame detection. One of:
11816 ``tff'' (top field first), ``bff'' (bottom field first),
11817 ``progressive'', or ``undetermined''
11820 Cumulative number of frames detected as bottom field first using multiple-frame detection.
11822 @item single.progressive
11823 Cumulative number of frames detected as progressive using single-frame detection.
11825 @item multiple.progressive
11826 Cumulative number of frames detected as progressive using multiple-frame detection.
11828 @item single.undetermined
11829 Cumulative number of frames that could not be classified using single-frame detection.
11831 @item multiple.undetermined
11832 Cumulative number of frames that could not be classified using multiple-frame detection.
11834 @item repeated.current_frame
11835 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
11837 @item repeated.neither
11838 Cumulative number of frames with no repeated field.
11841 Cumulative number of frames with the top field repeated from the previous frame's top field.
11843 @item repeated.bottom
11844 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
11847 The filter accepts the following options:
11851 Set interlacing threshold.
11853 Set progressive threshold.
11855 Threshold for repeated field detection.
11857 Number of frames after which a given frame's contribution to the
11858 statistics is halved (i.e., it contributes only 0.5 to its
11859 classification). The default of 0 means that all frames seen are given
11860 full weight of 1.0 forever.
11861 @item analyze_interlaced_flag
11862 When this is not 0 then idet will use the specified number of frames to determine
11863 if the interlaced flag is accurate, it will not count undetermined frames.
11864 If the flag is found to be accurate it will be used without any further
11865 computations, if it is found to be inaccurate it will be cleared without any
11866 further computations. This allows inserting the idet filter as a low computational
11867 method to clean up the interlaced flag
11872 Deinterleave or interleave fields.
11874 This filter allows one to process interlaced images fields without
11875 deinterlacing them. Deinterleaving splits the input frame into 2
11876 fields (so called half pictures). Odd lines are moved to the top
11877 half of the output image, even lines to the bottom half.
11878 You can process (filter) them independently and then re-interleave them.
11880 The filter accepts the following options:
11884 @item chroma_mode, c
11885 @item alpha_mode, a
11886 Available values for @var{luma_mode}, @var{chroma_mode} and
11887 @var{alpha_mode} are:
11893 @item deinterleave, d
11894 Deinterleave fields, placing one above the other.
11896 @item interleave, i
11897 Interleave fields. Reverse the effect of deinterleaving.
11899 Default value is @code{none}.
11901 @item luma_swap, ls
11902 @item chroma_swap, cs
11903 @item alpha_swap, as
11904 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
11909 Apply inflate effect to the video.
11911 This filter replaces the pixel by the local(3x3) average by taking into account
11912 only values higher than the pixel.
11914 It accepts the following options:
11921 Limit the maximum change for each plane, default is 65535.
11922 If 0, plane will remain unchanged.
11927 Simple interlacing filter from progressive contents. This interleaves upper (or
11928 lower) lines from odd frames with lower (or upper) lines from even frames,
11929 halving the frame rate and preserving image height.
11932 Original Original New Frame
11933 Frame 'j' Frame 'j+1' (tff)
11934 ========== =========== ==================
11935 Line 0 --------------------> Frame 'j' Line 0
11936 Line 1 Line 1 ----> Frame 'j+1' Line 1
11937 Line 2 ---------------------> Frame 'j' Line 2
11938 Line 3 Line 3 ----> Frame 'j+1' Line 3
11940 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
11943 It accepts the following optional parameters:
11947 This determines whether the interlaced frame is taken from the even
11948 (tff - default) or odd (bff) lines of the progressive frame.
11951 Vertical lowpass filter to avoid twitter interlacing and
11952 reduce moire patterns.
11956 Disable vertical lowpass filter
11959 Enable linear filter (default)
11962 Enable complex filter. This will slightly less reduce twitter and moire
11963 but better retain detail and subjective sharpness impression.
11970 Deinterlace input video by applying Donald Graft's adaptive kernel
11971 deinterling. Work on interlaced parts of a video to produce
11972 progressive frames.
11974 The description of the accepted parameters follows.
11978 Set the threshold which affects the filter's tolerance when
11979 determining if a pixel line must be processed. It must be an integer
11980 in the range [0,255] and defaults to 10. A value of 0 will result in
11981 applying the process on every pixels.
11984 Paint pixels exceeding the threshold value to white if set to 1.
11988 Set the fields order. Swap fields if set to 1, leave fields alone if
11992 Enable additional sharpening if set to 1. Default is 0.
11995 Enable twoway sharpening if set to 1. Default is 0.
11998 @subsection Examples
12002 Apply default values:
12004 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
12008 Enable additional sharpening:
12014 Paint processed pixels in white:
12022 Slowly update darker pixels.
12024 This filter makes short flashes of light appear longer.
12025 This filter accepts the following options:
12029 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
12032 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
12035 @section lenscorrection
12037 Correct radial lens distortion
12039 This filter can be used to correct for radial distortion as can result from the use
12040 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
12041 one can use tools available for example as part of opencv or simply trial-and-error.
12042 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
12043 and extract the k1 and k2 coefficients from the resulting matrix.
12045 Note that effectively the same filter is available in the open-source tools Krita and
12046 Digikam from the KDE project.
12048 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
12049 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
12050 brightness distribution, so you may want to use both filters together in certain
12051 cases, though you will have to take care of ordering, i.e. whether vignetting should
12052 be applied before or after lens correction.
12054 @subsection Options
12056 The filter accepts the following options:
12060 Relative x-coordinate of the focal point of the image, and thereby the center of the
12061 distortion. This value has a range [0,1] and is expressed as fractions of the image
12062 width. Default is 0.5.
12064 Relative y-coordinate of the focal point of the image, and thereby the center of the
12065 distortion. This value has a range [0,1] and is expressed as fractions of the image
12066 height. Default is 0.5.
12068 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
12069 no correction. Default is 0.
12071 Coefficient of the double quadratic correction term. This value has a range [-1,1].
12072 0 means no correction. Default is 0.
12075 The formula that generates the correction is:
12077 @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)
12079 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
12080 distances from the focal point in the source and target images, respectively.
12084 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
12086 The @code{lensfun} filter requires the camera make, camera model, and lens model
12087 to apply the lens correction. The filter will load the lensfun database and
12088 query it to find the corresponding camera and lens entries in the database. As
12089 long as these entries can be found with the given options, the filter can
12090 perform corrections on frames. Note that incomplete strings will result in the
12091 filter choosing the best match with the given options, and the filter will
12092 output the chosen camera and lens models (logged with level "info"). You must
12093 provide the make, camera model, and lens model as they are required.
12095 The filter accepts the following options:
12099 The make of the camera (for example, "Canon"). This option is required.
12102 The model of the camera (for example, "Canon EOS 100D"). This option is
12106 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
12107 option is required.
12110 The type of correction to apply. The following values are valid options:
12114 Enables fixing lens vignetting.
12117 Enables fixing lens geometry. This is the default.
12120 Enables fixing chromatic aberrations.
12123 Enables fixing lens vignetting and lens geometry.
12126 Enables fixing lens vignetting and chromatic aberrations.
12129 Enables fixing both lens geometry and chromatic aberrations.
12132 Enables all possible corrections.
12136 The focal length of the image/video (zoom; expected constant for video). For
12137 example, a 18--55mm lens has focal length range of [18--55], so a value in that
12138 range should be chosen when using that lens. Default 18.
12141 The aperture of the image/video (expected constant for video). Note that
12142 aperture is only used for vignetting correction. Default 3.5.
12144 @item focus_distance
12145 The focus distance of the image/video (expected constant for video). Note that
12146 focus distance is only used for vignetting and only slightly affects the
12147 vignetting correction process. If unknown, leave it at the default value (which
12151 The scale factor which is applied after transformation. After correction the
12152 video is no longer necessarily rectangular. This parameter controls how much of
12153 the resulting image is visible. The value 0 means that a value will be chosen
12154 automatically such that there is little or no unmapped area in the output
12155 image. 1.0 means that no additional scaling is done. Lower values may result
12156 in more of the corrected image being visible, while higher values may avoid
12157 unmapped areas in the output.
12159 @item target_geometry
12160 The target geometry of the output image/video. The following values are valid
12164 @item rectilinear (default)
12167 @item equirectangular
12168 @item fisheye_orthographic
12169 @item fisheye_stereographic
12170 @item fisheye_equisolid
12171 @item fisheye_thoby
12174 Apply the reverse of image correction (instead of correcting distortion, apply
12177 @item interpolation
12178 The type of interpolation used when correcting distortion. The following values
12183 @item linear (default)
12188 @subsection Examples
12192 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
12193 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
12197 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
12201 Apply the same as before, but only for the first 5 seconds of video.
12204 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
12211 Obtain the VMAF (Video Multi-Method Assessment Fusion)
12212 score between two input videos.
12214 The obtained VMAF score is printed through the logging system.
12216 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
12217 After installing the library it can be enabled using:
12218 @code{./configure --enable-libvmaf --enable-version3}.
12219 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
12221 The filter has following options:
12225 Set the model path which is to be used for SVM.
12226 Default value: @code{"vmaf_v0.6.1.pkl"}
12229 Set the file path to be used to store logs.
12232 Set the format of the log file (xml or json).
12234 @item enable_transform
12235 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
12236 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
12237 Default value: @code{false}
12240 Invokes the phone model which will generate VMAF scores higher than in the
12241 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
12244 Enables computing psnr along with vmaf.
12247 Enables computing ssim along with vmaf.
12250 Enables computing ms_ssim along with vmaf.
12253 Set the pool method (mean, min or harmonic mean) to be used for computing vmaf.
12256 Set number of threads to be used when computing vmaf.
12259 Set interval for frame subsampling used when computing vmaf.
12261 @item enable_conf_interval
12262 Enables confidence interval.
12265 This filter also supports the @ref{framesync} options.
12267 On the below examples the input file @file{main.mpg} being processed is
12268 compared with the reference file @file{ref.mpg}.
12271 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
12274 Example with options:
12276 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
12281 Limits the pixel components values to the specified range [min, max].
12283 The filter accepts the following options:
12287 Lower bound. Defaults to the lowest allowed value for the input.
12290 Upper bound. Defaults to the highest allowed value for the input.
12293 Specify which planes will be processed. Defaults to all available.
12300 The filter accepts the following options:
12304 Set the number of loops. Setting this value to -1 will result in infinite loops.
12308 Set maximal size in number of frames. Default is 0.
12311 Set first frame of loop. Default is 0.
12314 @subsection Examples
12318 Loop single first frame infinitely:
12320 loop=loop=-1:size=1:start=0
12324 Loop single first frame 10 times:
12326 loop=loop=10:size=1:start=0
12330 Loop 10 first frames 5 times:
12332 loop=loop=5:size=10:start=0
12338 Apply a 1D LUT to an input video.
12340 The filter accepts the following options:
12344 Set the 1D LUT file name.
12346 Currently supported formats:
12355 Select interpolation mode.
12357 Available values are:
12361 Use values from the nearest defined point.
12363 Interpolate values using the linear interpolation.
12365 Interpolate values using the cosine interpolation.
12367 Interpolate values using the cubic interpolation.
12369 Interpolate values using the spline interpolation.
12376 Apply a 3D LUT to an input video.
12378 The filter accepts the following options:
12382 Set the 3D LUT file name.
12384 Currently supported formats:
12398 Select interpolation mode.
12400 Available values are:
12404 Use values from the nearest defined point.
12406 Interpolate values using the 8 points defining a cube.
12408 Interpolate values using a tetrahedron.
12414 Turn certain luma values into transparency.
12416 The filter accepts the following options:
12420 Set the luma which will be used as base for transparency.
12421 Default value is @code{0}.
12424 Set the range of luma values to be keyed out.
12425 Default value is @code{0}.
12428 Set the range of softness. Default value is @code{0}.
12429 Use this to control gradual transition from zero to full transparency.
12432 @section lut, lutrgb, lutyuv
12434 Compute a look-up table for binding each pixel component input value
12435 to an output value, and apply it to the input video.
12437 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
12438 to an RGB input video.
12440 These filters accept the following parameters:
12443 set first pixel component expression
12445 set second pixel component expression
12447 set third pixel component expression
12449 set fourth pixel component expression, corresponds to the alpha component
12452 set red component expression
12454 set green component expression
12456 set blue component expression
12458 alpha component expression
12461 set Y/luminance component expression
12463 set U/Cb component expression
12465 set V/Cr component expression
12468 Each of them specifies the expression to use for computing the lookup table for
12469 the corresponding pixel component values.
12471 The exact component associated to each of the @var{c*} options depends on the
12474 The @var{lut} filter requires either YUV or RGB pixel formats in input,
12475 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
12477 The expressions can contain the following constants and functions:
12482 The input width and height.
12485 The input value for the pixel component.
12488 The input value, clipped to the @var{minval}-@var{maxval} range.
12491 The maximum value for the pixel component.
12494 The minimum value for the pixel component.
12497 The negated value for the pixel component value, clipped to the
12498 @var{minval}-@var{maxval} range; it corresponds to the expression
12499 "maxval-clipval+minval".
12502 The computed value in @var{val}, clipped to the
12503 @var{minval}-@var{maxval} range.
12505 @item gammaval(gamma)
12506 The computed gamma correction value of the pixel component value,
12507 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
12509 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
12513 All expressions default to "val".
12515 @subsection Examples
12519 Negate input video:
12521 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
12522 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
12525 The above is the same as:
12527 lutrgb="r=negval:g=negval:b=negval"
12528 lutyuv="y=negval:u=negval:v=negval"
12538 Remove chroma components, turning the video into a graytone image:
12540 lutyuv="u=128:v=128"
12544 Apply a luma burning effect:
12550 Remove green and blue components:
12556 Set a constant alpha channel value on input:
12558 format=rgba,lutrgb=a="maxval-minval/2"
12562 Correct luminance gamma by a factor of 0.5:
12564 lutyuv=y=gammaval(0.5)
12568 Discard least significant bits of luma:
12570 lutyuv=y='bitand(val, 128+64+32)'
12574 Technicolor like effect:
12576 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
12580 @section lut2, tlut2
12582 The @code{lut2} filter takes two input streams and outputs one
12585 The @code{tlut2} (time lut2) filter takes two consecutive frames
12586 from one single stream.
12588 This filter accepts the following parameters:
12591 set first pixel component expression
12593 set second pixel component expression
12595 set third pixel component expression
12597 set fourth pixel component expression, corresponds to the alpha component
12600 set output bit depth, only available for @code{lut2} filter. By default is 0,
12601 which means bit depth is automatically picked from first input format.
12604 Each of them specifies the expression to use for computing the lookup table for
12605 the corresponding pixel component values.
12607 The exact component associated to each of the @var{c*} options depends on the
12610 The expressions can contain the following constants:
12615 The input width and height.
12618 The first input value for the pixel component.
12621 The second input value for the pixel component.
12624 The first input video bit depth.
12627 The second input video bit depth.
12630 All expressions default to "x".
12632 @subsection Examples
12636 Highlight differences between two RGB video streams:
12638 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)'
12642 Highlight differences between two YUV video streams:
12644 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)'
12648 Show max difference between two video streams:
12650 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)))'
12654 @section maskedclamp
12656 Clamp the first input stream with the second input and third input stream.
12658 Returns the value of first stream to be between second input
12659 stream - @code{undershoot} and third input stream + @code{overshoot}.
12661 This filter accepts the following options:
12664 Default value is @code{0}.
12667 Default value is @code{0}.
12670 Set which planes will be processed as bitmap, unprocessed planes will be
12671 copied from first stream.
12672 By default value 0xf, all planes will be processed.
12677 Merge the second and third input stream into output stream using absolute differences
12678 between second input stream and first input stream and absolute difference between
12679 third input stream and first input stream. The picked value will be from second input
12680 stream if second absolute difference is greater than first one or from third input stream
12683 This filter accepts the following options:
12686 Set which planes will be processed as bitmap, unprocessed planes will be
12687 copied from first stream.
12688 By default value 0xf, all planes will be processed.
12691 @section maskedmerge
12693 Merge the first input stream with the second input stream using per pixel
12694 weights in the third input stream.
12696 A value of 0 in the third stream pixel component means that pixel component
12697 from first stream is returned unchanged, while maximum value (eg. 255 for
12698 8-bit videos) means that pixel component from second stream is returned
12699 unchanged. Intermediate values define the amount of merging between both
12700 input stream's pixel components.
12702 This filter accepts the following options:
12705 Set which planes will be processed as bitmap, unprocessed planes will be
12706 copied from first stream.
12707 By default value 0xf, all planes will be processed.
12712 Merge the second and third input stream into output stream using absolute differences
12713 between second input stream and first input stream and absolute difference between
12714 third input stream and first input stream. The picked value will be from second input
12715 stream if second absolute difference is less than first one or from third input stream
12718 This filter accepts the following options:
12721 Set which planes will be processed as bitmap, unprocessed planes will be
12722 copied from first stream.
12723 By default value 0xf, all planes will be processed.
12727 Create mask from input video.
12729 For example it is useful to create motion masks after @code{tblend} filter.
12731 This filter accepts the following options:
12735 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
12738 Set high threshold. Any pixel component higher than this value will be set to max value
12739 allowed for current pixel format.
12742 Set planes to filter, by default all available planes are filtered.
12745 Fill all frame pixels with this value.
12748 Set max average pixel value for frame. If sum of all pixel components is higher that this
12749 average, output frame will be completely filled with value set by @var{fill} option.
12750 Typically useful for scene changes when used in combination with @code{tblend} filter.
12755 Apply motion-compensation deinterlacing.
12757 It needs one field per frame as input and must thus be used together
12758 with yadif=1/3 or equivalent.
12760 This filter accepts the following options:
12763 Set the deinterlacing mode.
12765 It accepts one of the following values:
12770 use iterative motion estimation
12772 like @samp{slow}, but use multiple reference frames.
12774 Default value is @samp{fast}.
12777 Set the picture field parity assumed for the input video. It must be
12778 one of the following values:
12782 assume top field first
12784 assume bottom field first
12787 Default value is @samp{bff}.
12790 Set per-block quantization parameter (QP) used by the internal
12793 Higher values should result in a smoother motion vector field but less
12794 optimal individual vectors. Default value is 1.
12797 @section mergeplanes
12799 Merge color channel components from several video streams.
12801 The filter accepts up to 4 input streams, and merge selected input
12802 planes to the output video.
12804 This filter accepts the following options:
12807 Set input to output plane mapping. Default is @code{0}.
12809 The mappings is specified as a bitmap. It should be specified as a
12810 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
12811 mapping for the first plane of the output stream. 'A' sets the number of
12812 the input stream to use (from 0 to 3), and 'a' the plane number of the
12813 corresponding input to use (from 0 to 3). The rest of the mappings is
12814 similar, 'Bb' describes the mapping for the output stream second
12815 plane, 'Cc' describes the mapping for the output stream third plane and
12816 'Dd' describes the mapping for the output stream fourth plane.
12819 Set output pixel format. Default is @code{yuva444p}.
12822 @subsection Examples
12826 Merge three gray video streams of same width and height into single video stream:
12828 [a0][a1][a2]mergeplanes=0x001020:yuv444p
12832 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
12834 [a0][a1]mergeplanes=0x00010210:yuva444p
12838 Swap Y and A plane in yuva444p stream:
12840 format=yuva444p,mergeplanes=0x03010200:yuva444p
12844 Swap U and V plane in yuv420p stream:
12846 format=yuv420p,mergeplanes=0x000201:yuv420p
12850 Cast a rgb24 clip to yuv444p:
12852 format=rgb24,mergeplanes=0x000102:yuv444p
12858 Estimate and export motion vectors using block matching algorithms.
12859 Motion vectors are stored in frame side data to be used by other filters.
12861 This filter accepts the following options:
12864 Specify the motion estimation method. Accepts one of the following values:
12868 Exhaustive search algorithm.
12870 Three step search algorithm.
12872 Two dimensional logarithmic search algorithm.
12874 New three step search algorithm.
12876 Four step search algorithm.
12878 Diamond search algorithm.
12880 Hexagon-based search algorithm.
12882 Enhanced predictive zonal search algorithm.
12884 Uneven multi-hexagon search algorithm.
12886 Default value is @samp{esa}.
12889 Macroblock size. Default @code{16}.
12892 Search parameter. Default @code{7}.
12895 @section midequalizer
12897 Apply Midway Image Equalization effect using two video streams.
12899 Midway Image Equalization adjusts a pair of images to have the same
12900 histogram, while maintaining their dynamics as much as possible. It's
12901 useful for e.g. matching exposures from a pair of stereo cameras.
12903 This filter has two inputs and one output, which must be of same pixel format, but
12904 may be of different sizes. The output of filter is first input adjusted with
12905 midway histogram of both inputs.
12907 This filter accepts the following option:
12911 Set which planes to process. Default is @code{15}, which is all available planes.
12914 @section minterpolate
12916 Convert the video to specified frame rate using motion interpolation.
12918 This filter accepts the following options:
12921 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}.
12924 Motion interpolation mode. Following values are accepted:
12927 Duplicate previous or next frame for interpolating new ones.
12929 Blend source frames. Interpolated frame is mean of previous and next frames.
12931 Motion compensated interpolation. Following options are effective when this mode is selected:
12935 Motion compensation mode. Following values are accepted:
12938 Overlapped block motion compensation.
12940 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
12942 Default mode is @samp{obmc}.
12945 Motion estimation mode. Following values are accepted:
12948 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
12950 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
12952 Default mode is @samp{bilat}.
12955 The algorithm to be used for motion estimation. Following values are accepted:
12958 Exhaustive search algorithm.
12960 Three step search algorithm.
12962 Two dimensional logarithmic search algorithm.
12964 New three step search algorithm.
12966 Four step search algorithm.
12968 Diamond search algorithm.
12970 Hexagon-based search algorithm.
12972 Enhanced predictive zonal search algorithm.
12974 Uneven multi-hexagon search algorithm.
12976 Default algorithm is @samp{epzs}.
12979 Macroblock size. Default @code{16}.
12982 Motion estimation search parameter. Default @code{32}.
12985 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).
12990 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:
12993 Disable scene change detection.
12995 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
12997 Default method is @samp{fdiff}.
12999 @item scd_threshold
13000 Scene change detection threshold. Default is @code{5.0}.
13005 Mix several video input streams into one video stream.
13007 A description of the accepted options follows.
13011 The number of inputs. If unspecified, it defaults to 2.
13014 Specify weight of each input video stream as sequence.
13015 Each weight is separated by space. If number of weights
13016 is smaller than number of @var{frames} last specified
13017 weight will be used for all remaining unset weights.
13020 Specify scale, if it is set it will be multiplied with sum
13021 of each weight multiplied with pixel values to give final destination
13022 pixel value. By default @var{scale} is auto scaled to sum of weights.
13025 Specify how end of stream is determined.
13028 The duration of the longest input. (default)
13031 The duration of the shortest input.
13034 The duration of the first input.
13038 @section mpdecimate
13040 Drop frames that do not differ greatly from the previous frame in
13041 order to reduce frame rate.
13043 The main use of this filter is for very-low-bitrate encoding
13044 (e.g. streaming over dialup modem), but it could in theory be used for
13045 fixing movies that were inverse-telecined incorrectly.
13047 A description of the accepted options follows.
13051 Set the maximum number of consecutive frames which can be dropped (if
13052 positive), or the minimum interval between dropped frames (if
13053 negative). If the value is 0, the frame is dropped disregarding the
13054 number of previous sequentially dropped frames.
13056 Default value is 0.
13061 Set the dropping threshold values.
13063 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
13064 represent actual pixel value differences, so a threshold of 64
13065 corresponds to 1 unit of difference for each pixel, or the same spread
13066 out differently over the block.
13068 A frame is a candidate for dropping if no 8x8 blocks differ by more
13069 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
13070 meaning the whole image) differ by more than a threshold of @option{lo}.
13072 Default value for @option{hi} is 64*12, default value for @option{lo} is
13073 64*5, and default value for @option{frac} is 0.33.
13079 Negate (invert) the input video.
13081 It accepts the following option:
13086 With value 1, it negates the alpha component, if present. Default value is 0.
13092 Denoise frames using Non-Local Means algorithm.
13094 Each pixel is adjusted by looking for other pixels with similar contexts. This
13095 context similarity is defined by comparing their surrounding patches of size
13096 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
13099 Note that the research area defines centers for patches, which means some
13100 patches will be made of pixels outside that research area.
13102 The filter accepts the following options.
13106 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
13109 Set patch size. Default is 7. Must be odd number in range [0, 99].
13112 Same as @option{p} but for chroma planes.
13114 The default value is @var{0} and means automatic.
13117 Set research size. Default is 15. Must be odd number in range [0, 99].
13120 Same as @option{r} but for chroma planes.
13122 The default value is @var{0} and means automatic.
13127 Deinterlace video using neural network edge directed interpolation.
13129 This filter accepts the following options:
13133 Mandatory option, without binary file filter can not work.
13134 Currently file can be found here:
13135 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
13138 Set which frames to deinterlace, by default it is @code{all}.
13139 Can be @code{all} or @code{interlaced}.
13142 Set mode of operation.
13144 Can be one of the following:
13148 Use frame flags, both fields.
13150 Use frame flags, single field.
13152 Use top field only.
13154 Use bottom field only.
13156 Use both fields, top first.
13158 Use both fields, bottom first.
13162 Set which planes to process, by default filter process all frames.
13165 Set size of local neighborhood around each pixel, used by the predictor neural
13168 Can be one of the following:
13181 Set the number of neurons in predictor neural network.
13182 Can be one of the following:
13193 Controls the number of different neural network predictions that are blended
13194 together to compute the final output value. Can be @code{fast}, default or
13198 Set which set of weights to use in the predictor.
13199 Can be one of the following:
13203 weights trained to minimize absolute error
13205 weights trained to minimize squared error
13209 Controls whether or not the prescreener neural network is used to decide
13210 which pixels should be processed by the predictor neural network and which
13211 can be handled by simple cubic interpolation.
13212 The prescreener is trained to know whether cubic interpolation will be
13213 sufficient for a pixel or whether it should be predicted by the predictor nn.
13214 The computational complexity of the prescreener nn is much less than that of
13215 the predictor nn. Since most pixels can be handled by cubic interpolation,
13216 using the prescreener generally results in much faster processing.
13217 The prescreener is pretty accurate, so the difference between using it and not
13218 using it is almost always unnoticeable.
13220 Can be one of the following:
13228 Default is @code{new}.
13231 Set various debugging flags.
13236 Force libavfilter not to use any of the specified pixel formats for the
13237 input to the next filter.
13239 It accepts the following parameters:
13243 A '|'-separated list of pixel format names, such as
13244 pix_fmts=yuv420p|monow|rgb24".
13248 @subsection Examples
13252 Force libavfilter to use a format different from @var{yuv420p} for the
13253 input to the vflip filter:
13255 noformat=pix_fmts=yuv420p,vflip
13259 Convert the input video to any of the formats not contained in the list:
13261 noformat=yuv420p|yuv444p|yuv410p
13267 Add noise on video input frame.
13269 The filter accepts the following options:
13277 Set noise seed for specific pixel component or all pixel components in case
13278 of @var{all_seed}. Default value is @code{123457}.
13280 @item all_strength, alls
13281 @item c0_strength, c0s
13282 @item c1_strength, c1s
13283 @item c2_strength, c2s
13284 @item c3_strength, c3s
13285 Set noise strength for specific pixel component or all pixel components in case
13286 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
13288 @item all_flags, allf
13289 @item c0_flags, c0f
13290 @item c1_flags, c1f
13291 @item c2_flags, c2f
13292 @item c3_flags, c3f
13293 Set pixel component flags or set flags for all components if @var{all_flags}.
13294 Available values for component flags are:
13297 averaged temporal noise (smoother)
13299 mix random noise with a (semi)regular pattern
13301 temporal noise (noise pattern changes between frames)
13303 uniform noise (gaussian otherwise)
13307 @subsection Examples
13309 Add temporal and uniform noise to input video:
13311 noise=alls=20:allf=t+u
13316 Normalize RGB video (aka histogram stretching, contrast stretching).
13317 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
13319 For each channel of each frame, the filter computes the input range and maps
13320 it linearly to the user-specified output range. The output range defaults
13321 to the full dynamic range from pure black to pure white.
13323 Temporal smoothing can be used on the input range to reduce flickering (rapid
13324 changes in brightness) caused when small dark or bright objects enter or leave
13325 the scene. This is similar to the auto-exposure (automatic gain control) on a
13326 video camera, and, like a video camera, it may cause a period of over- or
13327 under-exposure of the video.
13329 The R,G,B channels can be normalized independently, which may cause some
13330 color shifting, or linked together as a single channel, which prevents
13331 color shifting. Linked normalization preserves hue. Independent normalization
13332 does not, so it can be used to remove some color casts. Independent and linked
13333 normalization can be combined in any ratio.
13335 The normalize filter accepts the following options:
13340 Colors which define the output range. The minimum input value is mapped to
13341 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
13342 The defaults are black and white respectively. Specifying white for
13343 @var{blackpt} and black for @var{whitept} will give color-inverted,
13344 normalized video. Shades of grey can be used to reduce the dynamic range
13345 (contrast). Specifying saturated colors here can create some interesting
13349 The number of previous frames to use for temporal smoothing. The input range
13350 of each channel is smoothed using a rolling average over the current frame
13351 and the @var{smoothing} previous frames. The default is 0 (no temporal
13355 Controls the ratio of independent (color shifting) channel normalization to
13356 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
13357 independent. Defaults to 1.0 (fully independent).
13360 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
13361 expensive no-op. Defaults to 1.0 (full strength).
13365 @subsection Examples
13367 Stretch video contrast to use the full dynamic range, with no temporal
13368 smoothing; may flicker depending on the source content:
13370 normalize=blackpt=black:whitept=white:smoothing=0
13373 As above, but with 50 frames of temporal smoothing; flicker should be
13374 reduced, depending on the source content:
13376 normalize=blackpt=black:whitept=white:smoothing=50
13379 As above, but with hue-preserving linked channel normalization:
13381 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
13384 As above, but with half strength:
13386 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
13389 Map the darkest input color to red, the brightest input color to cyan:
13391 normalize=blackpt=red:whitept=cyan
13396 Pass the video source unchanged to the output.
13399 Optical Character Recognition
13401 This filter uses Tesseract for optical character recognition. To enable
13402 compilation of this filter, you need to configure FFmpeg with
13403 @code{--enable-libtesseract}.
13405 It accepts the following options:
13409 Set datapath to tesseract data. Default is to use whatever was
13410 set at installation.
13413 Set language, default is "eng".
13416 Set character whitelist.
13419 Set character blacklist.
13422 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
13423 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
13427 Apply a video transform using libopencv.
13429 To enable this filter, install the libopencv library and headers and
13430 configure FFmpeg with @code{--enable-libopencv}.
13432 It accepts the following parameters:
13437 The name of the libopencv filter to apply.
13439 @item filter_params
13440 The parameters to pass to the libopencv filter. If not specified, the default
13441 values are assumed.
13445 Refer to the official libopencv documentation for more precise
13447 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
13449 Several libopencv filters are supported; see the following subsections.
13454 Dilate an image by using a specific structuring element.
13455 It corresponds to the libopencv function @code{cvDilate}.
13457 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
13459 @var{struct_el} represents a structuring element, and has the syntax:
13460 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
13462 @var{cols} and @var{rows} represent the number of columns and rows of
13463 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
13464 point, and @var{shape} the shape for the structuring element. @var{shape}
13465 must be "rect", "cross", "ellipse", or "custom".
13467 If the value for @var{shape} is "custom", it must be followed by a
13468 string of the form "=@var{filename}". The file with name
13469 @var{filename} is assumed to represent a binary image, with each
13470 printable character corresponding to a bright pixel. When a custom
13471 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
13472 or columns and rows of the read file are assumed instead.
13474 The default value for @var{struct_el} is "3x3+0x0/rect".
13476 @var{nb_iterations} specifies the number of times the transform is
13477 applied to the image, and defaults to 1.
13481 # Use the default values
13484 # Dilate using a structuring element with a 5x5 cross, iterating two times
13485 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
13487 # Read the shape from the file diamond.shape, iterating two times.
13488 # The file diamond.shape may contain a pattern of characters like this
13494 # The specified columns and rows are ignored
13495 # but the anchor point coordinates are not
13496 ocv=dilate:0x0+2x2/custom=diamond.shape|2
13501 Erode an image by using a specific structuring element.
13502 It corresponds to the libopencv function @code{cvErode}.
13504 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
13505 with the same syntax and semantics as the @ref{dilate} filter.
13509 Smooth the input video.
13511 The filter takes the following parameters:
13512 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
13514 @var{type} is the type of smooth filter to apply, and must be one of
13515 the following values: "blur", "blur_no_scale", "median", "gaussian",
13516 or "bilateral". The default value is "gaussian".
13518 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
13519 depends on the smooth type. @var{param1} and
13520 @var{param2} accept integer positive values or 0. @var{param3} and
13521 @var{param4} accept floating point values.
13523 The default value for @var{param1} is 3. The default value for the
13524 other parameters is 0.
13526 These parameters correspond to the parameters assigned to the
13527 libopencv function @code{cvSmooth}.
13529 @section oscilloscope
13531 2D Video Oscilloscope.
13533 Useful to measure spatial impulse, step responses, chroma delays, etc.
13535 It accepts the following parameters:
13539 Set scope center x position.
13542 Set scope center y position.
13545 Set scope size, relative to frame diagonal.
13548 Set scope tilt/rotation.
13554 Set trace center x position.
13557 Set trace center y position.
13560 Set trace width, relative to width of frame.
13563 Set trace height, relative to height of frame.
13566 Set which components to trace. By default it traces first three components.
13569 Draw trace grid. By default is enabled.
13572 Draw some statistics. By default is enabled.
13575 Draw scope. By default is enabled.
13578 @subsection Examples
13582 Inspect full first row of video frame.
13584 oscilloscope=x=0.5:y=0:s=1
13588 Inspect full last row of video frame.
13590 oscilloscope=x=0.5:y=1:s=1
13594 Inspect full 5th line of video frame of height 1080.
13596 oscilloscope=x=0.5:y=5/1080:s=1
13600 Inspect full last column of video frame.
13602 oscilloscope=x=1:y=0.5:s=1:t=1
13610 Overlay one video on top of another.
13612 It takes two inputs and has one output. The first input is the "main"
13613 video on which the second input is overlaid.
13615 It accepts the following parameters:
13617 A description of the accepted options follows.
13622 Set the expression for the x and y coordinates of the overlaid video
13623 on the main video. Default value is "0" for both expressions. In case
13624 the expression is invalid, it is set to a huge value (meaning that the
13625 overlay will not be displayed within the output visible area).
13628 See @ref{framesync}.
13631 Set when the expressions for @option{x}, and @option{y} are evaluated.
13633 It accepts the following values:
13636 only evaluate expressions once during the filter initialization or
13637 when a command is processed
13640 evaluate expressions for each incoming frame
13643 Default value is @samp{frame}.
13646 See @ref{framesync}.
13649 Set the format for the output video.
13651 It accepts the following values:
13654 force YUV420 output
13657 force YUV422 output
13660 force YUV444 output
13663 force packed RGB output
13666 force planar RGB output
13669 automatically pick format
13672 Default value is @samp{yuv420}.
13675 See @ref{framesync}.
13678 Set format of alpha of the overlaid video, it can be @var{straight} or
13679 @var{premultiplied}. Default is @var{straight}.
13682 The @option{x}, and @option{y} expressions can contain the following
13688 The main input width and height.
13692 The overlay input width and height.
13696 The computed values for @var{x} and @var{y}. They are evaluated for
13701 horizontal and vertical chroma subsample values of the output
13702 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
13706 the number of input frame, starting from 0
13709 the position in the file of the input frame, NAN if unknown
13712 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
13716 This filter also supports the @ref{framesync} options.
13718 Note that the @var{n}, @var{pos}, @var{t} variables are available only
13719 when evaluation is done @emph{per frame}, and will evaluate to NAN
13720 when @option{eval} is set to @samp{init}.
13722 Be aware that frames are taken from each input video in timestamp
13723 order, hence, if their initial timestamps differ, it is a good idea
13724 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
13725 have them begin in the same zero timestamp, as the example for
13726 the @var{movie} filter does.
13728 You can chain together more overlays but you should test the
13729 efficiency of such approach.
13731 @subsection Commands
13733 This filter supports the following commands:
13737 Modify the x and y of the overlay input.
13738 The command accepts the same syntax of the corresponding option.
13740 If the specified expression is not valid, it is kept at its current
13744 @subsection Examples
13748 Draw the overlay at 10 pixels from the bottom right corner of the main
13751 overlay=main_w-overlay_w-10:main_h-overlay_h-10
13754 Using named options the example above becomes:
13756 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
13760 Insert a transparent PNG logo in the bottom left corner of the input,
13761 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
13763 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
13767 Insert 2 different transparent PNG logos (second logo on bottom
13768 right corner) using the @command{ffmpeg} tool:
13770 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
13774 Add a transparent color layer on top of the main video; @code{WxH}
13775 must specify the size of the main input to the overlay filter:
13777 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
13781 Play an original video and a filtered version (here with the deshake
13782 filter) side by side using the @command{ffplay} tool:
13784 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
13787 The above command is the same as:
13789 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
13793 Make a sliding overlay appearing from the left to the right top part of the
13794 screen starting since time 2:
13796 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
13800 Compose output by putting two input videos side to side:
13802 ffmpeg -i left.avi -i right.avi -filter_complex "
13803 nullsrc=size=200x100 [background];
13804 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
13805 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
13806 [background][left] overlay=shortest=1 [background+left];
13807 [background+left][right] overlay=shortest=1:x=100 [left+right]
13812 Mask 10-20 seconds of a video by applying the delogo filter to a section
13814 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
13815 -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]'
13820 Chain several overlays in cascade:
13822 nullsrc=s=200x200 [bg];
13823 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
13824 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
13825 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
13826 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
13827 [in3] null, [mid2] overlay=100:100 [out0]
13834 Apply Overcomplete Wavelet denoiser.
13836 The filter accepts the following options:
13842 Larger depth values will denoise lower frequency components more, but
13843 slow down filtering.
13845 Must be an int in the range 8-16, default is @code{8}.
13847 @item luma_strength, ls
13850 Must be a double value in the range 0-1000, default is @code{1.0}.
13852 @item chroma_strength, cs
13853 Set chroma strength.
13855 Must be a double value in the range 0-1000, default is @code{1.0}.
13861 Add paddings to the input image, and place the original input at the
13862 provided @var{x}, @var{y} coordinates.
13864 It accepts the following parameters:
13869 Specify an expression for the size of the output image with the
13870 paddings added. If the value for @var{width} or @var{height} is 0, the
13871 corresponding input size is used for the output.
13873 The @var{width} expression can reference the value set by the
13874 @var{height} expression, and vice versa.
13876 The default value of @var{width} and @var{height} is 0.
13880 Specify the offsets to place the input image at within the padded area,
13881 with respect to the top/left border of the output image.
13883 The @var{x} expression can reference the value set by the @var{y}
13884 expression, and vice versa.
13886 The default value of @var{x} and @var{y} is 0.
13888 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
13889 so the input image is centered on the padded area.
13892 Specify the color of the padded area. For the syntax of this option,
13893 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
13894 manual,ffmpeg-utils}.
13896 The default value of @var{color} is "black".
13899 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
13901 It accepts the following values:
13905 Only evaluate expressions once during the filter initialization or when
13906 a command is processed.
13909 Evaluate expressions for each incoming frame.
13913 Default value is @samp{init}.
13916 Pad to aspect instead to a resolution.
13920 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
13921 options are expressions containing the following constants:
13926 The input video width and height.
13930 These are the same as @var{in_w} and @var{in_h}.
13934 The output width and height (the size of the padded area), as
13935 specified by the @var{width} and @var{height} expressions.
13939 These are the same as @var{out_w} and @var{out_h}.
13943 The x and y offsets as specified by the @var{x} and @var{y}
13944 expressions, or NAN if not yet specified.
13947 same as @var{iw} / @var{ih}
13950 input sample aspect ratio
13953 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
13957 The horizontal and vertical chroma subsample values. For example for the
13958 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
13961 @subsection Examples
13965 Add paddings with the color "violet" to the input video. The output video
13966 size is 640x480, and the top-left corner of the input video is placed at
13969 pad=640:480:0:40:violet
13972 The example above is equivalent to the following command:
13974 pad=width=640:height=480:x=0:y=40:color=violet
13978 Pad the input to get an output with dimensions increased by 3/2,
13979 and put the input video at the center of the padded area:
13981 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
13985 Pad the input to get a squared output with size equal to the maximum
13986 value between the input width and height, and put the input video at
13987 the center of the padded area:
13989 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
13993 Pad the input to get a final w/h ratio of 16:9:
13995 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
13999 In case of anamorphic video, in order to set the output display aspect
14000 correctly, it is necessary to use @var{sar} in the expression,
14001 according to the relation:
14003 (ih * X / ih) * sar = output_dar
14004 X = output_dar / sar
14007 Thus the previous example needs to be modified to:
14009 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
14013 Double the output size and put the input video in the bottom-right
14014 corner of the output padded area:
14016 pad="2*iw:2*ih:ow-iw:oh-ih"
14020 @anchor{palettegen}
14021 @section palettegen
14023 Generate one palette for a whole video stream.
14025 It accepts the following options:
14029 Set the maximum number of colors to quantize in the palette.
14030 Note: the palette will still contain 256 colors; the unused palette entries
14033 @item reserve_transparent
14034 Create a palette of 255 colors maximum and reserve the last one for
14035 transparency. Reserving the transparency color is useful for GIF optimization.
14036 If not set, the maximum of colors in the palette will be 256. You probably want
14037 to disable this option for a standalone image.
14040 @item transparency_color
14041 Set the color that will be used as background for transparency.
14044 Set statistics mode.
14046 It accepts the following values:
14049 Compute full frame histograms.
14051 Compute histograms only for the part that differs from previous frame. This
14052 might be relevant to give more importance to the moving part of your input if
14053 the background is static.
14055 Compute new histogram for each frame.
14058 Default value is @var{full}.
14061 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
14062 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
14063 color quantization of the palette. This information is also visible at
14064 @var{info} logging level.
14066 @subsection Examples
14070 Generate a representative palette of a given video using @command{ffmpeg}:
14072 ffmpeg -i input.mkv -vf palettegen palette.png
14076 @section paletteuse
14078 Use a palette to downsample an input video stream.
14080 The filter takes two inputs: one video stream and a palette. The palette must
14081 be a 256 pixels image.
14083 It accepts the following options:
14087 Select dithering mode. Available algorithms are:
14090 Ordered 8x8 bayer dithering (deterministic)
14092 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
14093 Note: this dithering is sometimes considered "wrong" and is included as a
14095 @item floyd_steinberg
14096 Floyd and Steingberg dithering (error diffusion)
14098 Frankie Sierra dithering v2 (error diffusion)
14100 Frankie Sierra dithering v2 "Lite" (error diffusion)
14103 Default is @var{sierra2_4a}.
14106 When @var{bayer} dithering is selected, this option defines the scale of the
14107 pattern (how much the crosshatch pattern is visible). A low value means more
14108 visible pattern for less banding, and higher value means less visible pattern
14109 at the cost of more banding.
14111 The option must be an integer value in the range [0,5]. Default is @var{2}.
14114 If set, define the zone to process
14118 Only the changing rectangle will be reprocessed. This is similar to GIF
14119 cropping/offsetting compression mechanism. This option can be useful for speed
14120 if only a part of the image is changing, and has use cases such as limiting the
14121 scope of the error diffusal @option{dither} to the rectangle that bounds the
14122 moving scene (it leads to more deterministic output if the scene doesn't change
14123 much, and as a result less moving noise and better GIF compression).
14126 Default is @var{none}.
14129 Take new palette for each output frame.
14131 @item alpha_threshold
14132 Sets the alpha threshold for transparency. Alpha values above this threshold
14133 will be treated as completely opaque, and values below this threshold will be
14134 treated as completely transparent.
14136 The option must be an integer value in the range [0,255]. Default is @var{128}.
14139 @subsection Examples
14143 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
14144 using @command{ffmpeg}:
14146 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
14150 @section perspective
14152 Correct perspective of video not recorded perpendicular to the screen.
14154 A description of the accepted parameters follows.
14165 Set coordinates expression for top left, top right, bottom left and bottom right corners.
14166 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
14167 If the @code{sense} option is set to @code{source}, then the specified points will be sent
14168 to the corners of the destination. If the @code{sense} option is set to @code{destination},
14169 then the corners of the source will be sent to the specified coordinates.
14171 The expressions can use the following variables:
14176 the width and height of video frame.
14180 Output frame count.
14183 @item interpolation
14184 Set interpolation for perspective correction.
14186 It accepts the following values:
14192 Default value is @samp{linear}.
14195 Set interpretation of coordinate options.
14197 It accepts the following values:
14201 Send point in the source specified by the given coordinates to
14202 the corners of the destination.
14204 @item 1, destination
14206 Send the corners of the source to the point in the destination specified
14207 by the given coordinates.
14209 Default value is @samp{source}.
14213 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
14215 It accepts the following values:
14218 only evaluate expressions once during the filter initialization or
14219 when a command is processed
14222 evaluate expressions for each incoming frame
14225 Default value is @samp{init}.
14230 Delay interlaced video by one field time so that the field order changes.
14232 The intended use is to fix PAL movies that have been captured with the
14233 opposite field order to the film-to-video transfer.
14235 A description of the accepted parameters follows.
14241 It accepts the following values:
14244 Capture field order top-first, transfer bottom-first.
14245 Filter will delay the bottom field.
14248 Capture field order bottom-first, transfer top-first.
14249 Filter will delay the top field.
14252 Capture and transfer with the same field order. This mode only exists
14253 for the documentation of the other options to refer to, but if you
14254 actually select it, the filter will faithfully do nothing.
14257 Capture field order determined automatically by field flags, transfer
14259 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
14260 basis using field flags. If no field information is available,
14261 then this works just like @samp{u}.
14264 Capture unknown or varying, transfer opposite.
14265 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
14266 analyzing the images and selecting the alternative that produces best
14267 match between the fields.
14270 Capture top-first, transfer unknown or varying.
14271 Filter selects among @samp{t} and @samp{p} using image analysis.
14274 Capture bottom-first, transfer unknown or varying.
14275 Filter selects among @samp{b} and @samp{p} using image analysis.
14278 Capture determined by field flags, transfer unknown or varying.
14279 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
14280 image analysis. If no field information is available, then this works just
14281 like @samp{U}. This is the default mode.
14284 Both capture and transfer unknown or varying.
14285 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
14289 @section photosensitivity
14290 Reduce various flashes in video, so to help users with epilepsy.
14292 It accepts the following options:
14295 Set how many frames to use when filtering. Default is 30.
14298 Set detection threshold factor. Default is 1.
14302 Set how many pixels to skip when sampling frames. Defalt is 1.
14303 Allowed range is from 1 to 1024.
14306 Leave frames unchanged. Default is disabled.
14309 @section pixdesctest
14311 Pixel format descriptor test filter, mainly useful for internal
14312 testing. The output video should be equal to the input video.
14316 format=monow, pixdesctest
14319 can be used to test the monowhite pixel format descriptor definition.
14323 Display sample values of color channels. Mainly useful for checking color
14324 and levels. Minimum supported resolution is 640x480.
14326 The filters accept the following options:
14330 Set scope X position, relative offset on X axis.
14333 Set scope Y position, relative offset on Y axis.
14342 Set window opacity. This window also holds statistics about pixel area.
14345 Set window X position, relative offset on X axis.
14348 Set window Y position, relative offset on Y axis.
14353 Enable the specified chain of postprocessing subfilters using libpostproc. This
14354 library should be automatically selected with a GPL build (@code{--enable-gpl}).
14355 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
14356 Each subfilter and some options have a short and a long name that can be used
14357 interchangeably, i.e. dr/dering are the same.
14359 The filters accept the following options:
14363 Set postprocessing subfilters string.
14366 All subfilters share common options to determine their scope:
14370 Honor the quality commands for this subfilter.
14373 Do chrominance filtering, too (default).
14376 Do luminance filtering only (no chrominance).
14379 Do chrominance filtering only (no luminance).
14382 These options can be appended after the subfilter name, separated by a '|'.
14384 Available subfilters are:
14387 @item hb/hdeblock[|difference[|flatness]]
14388 Horizontal deblocking filter
14391 Difference factor where higher values mean more deblocking (default: @code{32}).
14393 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14396 @item vb/vdeblock[|difference[|flatness]]
14397 Vertical deblocking filter
14400 Difference factor where higher values mean more deblocking (default: @code{32}).
14402 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14405 @item ha/hadeblock[|difference[|flatness]]
14406 Accurate horizontal deblocking filter
14409 Difference factor where higher values mean more deblocking (default: @code{32}).
14411 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14414 @item va/vadeblock[|difference[|flatness]]
14415 Accurate vertical deblocking filter
14418 Difference factor where higher values mean more deblocking (default: @code{32}).
14420 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14424 The horizontal and vertical deblocking filters share the difference and
14425 flatness values so you cannot set different horizontal and vertical
14429 @item h1/x1hdeblock
14430 Experimental horizontal deblocking filter
14432 @item v1/x1vdeblock
14433 Experimental vertical deblocking filter
14438 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
14441 larger -> stronger filtering
14443 larger -> stronger filtering
14445 larger -> stronger filtering
14448 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
14451 Stretch luminance to @code{0-255}.
14454 @item lb/linblenddeint
14455 Linear blend deinterlacing filter that deinterlaces the given block by
14456 filtering all lines with a @code{(1 2 1)} filter.
14458 @item li/linipoldeint
14459 Linear interpolating deinterlacing filter that deinterlaces the given block by
14460 linearly interpolating every second line.
14462 @item ci/cubicipoldeint
14463 Cubic interpolating deinterlacing filter deinterlaces the given block by
14464 cubically interpolating every second line.
14466 @item md/mediandeint
14467 Median deinterlacing filter that deinterlaces the given block by applying a
14468 median filter to every second line.
14470 @item fd/ffmpegdeint
14471 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
14472 second line with a @code{(-1 4 2 4 -1)} filter.
14475 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
14476 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
14478 @item fq/forceQuant[|quantizer]
14479 Overrides the quantizer table from the input with the constant quantizer you
14487 Default pp filter combination (@code{hb|a,vb|a,dr|a})
14490 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
14493 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
14496 @subsection Examples
14500 Apply horizontal and vertical deblocking, deringing and automatic
14501 brightness/contrast:
14507 Apply default filters without brightness/contrast correction:
14513 Apply default filters and temporal denoiser:
14515 pp=default/tmpnoise|1|2|3
14519 Apply deblocking on luminance only, and switch vertical deblocking on or off
14520 automatically depending on available CPU time:
14527 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
14528 similar to spp = 6 with 7 point DCT, where only the center sample is
14531 The filter accepts the following options:
14535 Force a constant quantization parameter. It accepts an integer in range
14536 0 to 63. If not set, the filter will use the QP from the video stream
14540 Set thresholding mode. Available modes are:
14544 Set hard thresholding.
14546 Set soft thresholding (better de-ringing effect, but likely blurrier).
14548 Set medium thresholding (good results, default).
14552 @section premultiply
14553 Apply alpha premultiply effect to input video stream using first plane
14554 of second stream as alpha.
14556 Both streams must have same dimensions and same pixel format.
14558 The filter accepts the following option:
14562 Set which planes will be processed, unprocessed planes will be copied.
14563 By default value 0xf, all planes will be processed.
14566 Do not require 2nd input for processing, instead use alpha plane from input stream.
14570 Apply prewitt operator to input video stream.
14572 The filter accepts the following option:
14576 Set which planes will be processed, unprocessed planes will be copied.
14577 By default value 0xf, all planes will be processed.
14580 Set value which will be multiplied with filtered result.
14583 Set value which will be added to filtered result.
14586 @anchor{program_opencl}
14587 @section program_opencl
14589 Filter video using an OpenCL program.
14594 OpenCL program source file.
14597 Kernel name in program.
14600 Number of inputs to the filter. Defaults to 1.
14603 Size of output frames. Defaults to the same as the first input.
14607 The program source file must contain a kernel function with the given name,
14608 which will be run once for each plane of the output. Each run on a plane
14609 gets enqueued as a separate 2D global NDRange with one work-item for each
14610 pixel to be generated. The global ID offset for each work-item is therefore
14611 the coordinates of a pixel in the destination image.
14613 The kernel function needs to take the following arguments:
14616 Destination image, @var{__write_only image2d_t}.
14618 This image will become the output; the kernel should write all of it.
14620 Frame index, @var{unsigned int}.
14622 This is a counter starting from zero and increasing by one for each frame.
14624 Source images, @var{__read_only image2d_t}.
14626 These are the most recent images on each input. The kernel may read from
14627 them to generate the output, but they can't be written to.
14634 Copy the input to the output (output must be the same size as the input).
14636 __kernel void copy(__write_only image2d_t destination,
14637 unsigned int index,
14638 __read_only image2d_t source)
14640 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
14642 int2 location = (int2)(get_global_id(0), get_global_id(1));
14644 float4 value = read_imagef(source, sampler, location);
14646 write_imagef(destination, location, value);
14651 Apply a simple transformation, rotating the input by an amount increasing
14652 with the index counter. Pixel values are linearly interpolated by the
14653 sampler, and the output need not have the same dimensions as the input.
14655 __kernel void rotate_image(__write_only image2d_t dst,
14656 unsigned int index,
14657 __read_only image2d_t src)
14659 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
14660 CLK_FILTER_LINEAR);
14662 float angle = (float)index / 100.0f;
14664 float2 dst_dim = convert_float2(get_image_dim(dst));
14665 float2 src_dim = convert_float2(get_image_dim(src));
14667 float2 dst_cen = dst_dim / 2.0f;
14668 float2 src_cen = src_dim / 2.0f;
14670 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
14672 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
14674 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
14675 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
14677 src_pos = src_pos * src_dim / dst_dim;
14679 float2 src_loc = src_pos + src_cen;
14681 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
14682 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
14683 write_imagef(dst, dst_loc, 0.5f);
14685 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
14690 Blend two inputs together, with the amount of each input used varying
14691 with the index counter.
14693 __kernel void blend_images(__write_only image2d_t dst,
14694 unsigned int index,
14695 __read_only image2d_t src1,
14696 __read_only image2d_t src2)
14698 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
14699 CLK_FILTER_LINEAR);
14701 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
14703 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
14704 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
14705 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
14707 float4 val1 = read_imagef(src1, sampler, src1_loc);
14708 float4 val2 = read_imagef(src2, sampler, src2_loc);
14710 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
14716 @section pseudocolor
14718 Alter frame colors in video with pseudocolors.
14720 This filter accepts the following options:
14724 set pixel first component expression
14727 set pixel second component expression
14730 set pixel third component expression
14733 set pixel fourth component expression, corresponds to the alpha component
14736 set component to use as base for altering colors
14739 Each of them specifies the expression to use for computing the lookup table for
14740 the corresponding pixel component values.
14742 The expressions can contain the following constants and functions:
14747 The input width and height.
14750 The input value for the pixel component.
14752 @item ymin, umin, vmin, amin
14753 The minimum allowed component value.
14755 @item ymax, umax, vmax, amax
14756 The maximum allowed component value.
14759 All expressions default to "val".
14761 @subsection Examples
14765 Change too high luma values to gradient:
14767 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'"
14773 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
14774 Ratio) between two input videos.
14776 This filter takes in input two input videos, the first input is
14777 considered the "main" source and is passed unchanged to the
14778 output. The second input is used as a "reference" video for computing
14781 Both video inputs must have the same resolution and pixel format for
14782 this filter to work correctly. Also it assumes that both inputs
14783 have the same number of frames, which are compared one by one.
14785 The obtained average PSNR is printed through the logging system.
14787 The filter stores the accumulated MSE (mean squared error) of each
14788 frame, and at the end of the processing it is averaged across all frames
14789 equally, and the following formula is applied to obtain the PSNR:
14792 PSNR = 10*log10(MAX^2/MSE)
14795 Where MAX is the average of the maximum values of each component of the
14798 The description of the accepted parameters follows.
14801 @item stats_file, f
14802 If specified the filter will use the named file to save the PSNR of
14803 each individual frame. When filename equals "-" the data is sent to
14806 @item stats_version
14807 Specifies which version of the stats file format to use. Details of
14808 each format are written below.
14809 Default value is 1.
14811 @item stats_add_max
14812 Determines whether the max value is output to the stats log.
14813 Default value is 0.
14814 Requires stats_version >= 2. If this is set and stats_version < 2,
14815 the filter will return an error.
14818 This filter also supports the @ref{framesync} options.
14820 The file printed if @var{stats_file} is selected, contains a sequence of
14821 key/value pairs of the form @var{key}:@var{value} for each compared
14824 If a @var{stats_version} greater than 1 is specified, a header line precedes
14825 the list of per-frame-pair stats, with key value pairs following the frame
14826 format with the following parameters:
14829 @item psnr_log_version
14830 The version of the log file format. Will match @var{stats_version}.
14833 A comma separated list of the per-frame-pair parameters included in
14837 A description of each shown per-frame-pair parameter follows:
14841 sequential number of the input frame, starting from 1
14844 Mean Square Error pixel-by-pixel average difference of the compared
14845 frames, averaged over all the image components.
14847 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
14848 Mean Square Error pixel-by-pixel average difference of the compared
14849 frames for the component specified by the suffix.
14851 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
14852 Peak Signal to Noise ratio of the compared frames for the component
14853 specified by the suffix.
14855 @item max_avg, max_y, max_u, max_v
14856 Maximum allowed value for each channel, and average over all
14862 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
14863 [main][ref] psnr="stats_file=stats.log" [out]
14866 On this example the input file being processed is compared with the
14867 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
14868 is stored in @file{stats.log}.
14873 Pulldown reversal (inverse telecine) filter, capable of handling mixed
14874 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
14877 The pullup filter is designed to take advantage of future context in making
14878 its decisions. This filter is stateless in the sense that it does not lock
14879 onto a pattern to follow, but it instead looks forward to the following
14880 fields in order to identify matches and rebuild progressive frames.
14882 To produce content with an even framerate, insert the fps filter after
14883 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
14884 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
14886 The filter accepts the following options:
14893 These options set the amount of "junk" to ignore at the left, right, top, and
14894 bottom of the image, respectively. Left and right are in units of 8 pixels,
14895 while top and bottom are in units of 2 lines.
14896 The default is 8 pixels on each side.
14899 Set the strict breaks. Setting this option to 1 will reduce the chances of
14900 filter generating an occasional mismatched frame, but it may also cause an
14901 excessive number of frames to be dropped during high motion sequences.
14902 Conversely, setting it to -1 will make filter match fields more easily.
14903 This may help processing of video where there is slight blurring between
14904 the fields, but may also cause there to be interlaced frames in the output.
14905 Default value is @code{0}.
14908 Set the metric plane to use. It accepts the following values:
14914 Use chroma blue plane.
14917 Use chroma red plane.
14920 This option may be set to use chroma plane instead of the default luma plane
14921 for doing filter's computations. This may improve accuracy on very clean
14922 source material, but more likely will decrease accuracy, especially if there
14923 is chroma noise (rainbow effect) or any grayscale video.
14924 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
14925 load and make pullup usable in realtime on slow machines.
14928 For best results (without duplicated frames in the output file) it is
14929 necessary to change the output frame rate. For example, to inverse
14930 telecine NTSC input:
14932 ffmpeg -i input -vf pullup -r 24000/1001 ...
14937 Change video quantization parameters (QP).
14939 The filter accepts the following option:
14943 Set expression for quantization parameter.
14946 The expression is evaluated through the eval API and can contain, among others,
14947 the following constants:
14951 1 if index is not 129, 0 otherwise.
14954 Sequential index starting from -129 to 128.
14957 @subsection Examples
14961 Some equation like:
14969 Flush video frames from internal cache of frames into a random order.
14970 No frame is discarded.
14971 Inspired by @ref{frei0r} nervous filter.
14975 Set size in number of frames of internal cache, in range from @code{2} to
14976 @code{512}. Default is @code{30}.
14979 Set seed for random number generator, must be an integer included between
14980 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
14981 less than @code{0}, the filter will try to use a good random seed on a
14985 @section readeia608
14987 Read closed captioning (EIA-608) information from the top lines of a video frame.
14989 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
14990 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
14991 with EIA-608 data (starting from 0). A description of each metadata value follows:
14994 @item lavfi.readeia608.X.cc
14995 The two bytes stored as EIA-608 data (printed in hexadecimal).
14997 @item lavfi.readeia608.X.line
14998 The number of the line on which the EIA-608 data was identified and read.
15001 This filter accepts the following options:
15005 Set the line to start scanning for EIA-608 data. Default is @code{0}.
15008 Set the line to end scanning for EIA-608 data. Default is @code{29}.
15011 Set minimal acceptable amplitude change for sync codes detection.
15012 Default is @code{0.2}. Allowed range is @code{[0.001 - 1]}.
15015 Set the ratio of width reserved for sync code detection.
15016 Default is @code{0.27}. Allowed range is @code{[0.01 - 0.7]}.
15019 Set the max peaks height difference for sync code detection.
15020 Default is @code{0.1}. Allowed range is @code{[0.0 - 0.5]}.
15023 Set max peaks period difference for sync code detection.
15024 Default is @code{0.1}. Allowed range is @code{[0.0 - 0.5]}.
15027 Set the first two max start code bits differences.
15028 Default is @code{0.02}. Allowed range is @code{[0.0 - 0.5]}.
15031 Set the minimum ratio of bits height compared to 3rd start code bit.
15032 Default is @code{0.75}. Allowed range is @code{[0.01 - 1]}.
15035 Set the white color threshold. Default is @code{0.35}. Allowed range is @code{[0.1 - 1]}.
15038 Set the black color threshold. Default is @code{0.15}. Allowed range is @code{[0.0 - 0.5]}.
15041 Enable checking the parity bit. In the event of a parity error, the filter will output
15042 @code{0x00} for that character. Default is false.
15045 Lowpass lines prior to further processing. Default is disabled.
15048 @subsection Examples
15052 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
15054 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
15060 Read vertical interval timecode (VITC) information from the top lines of a
15063 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
15064 timecode value, if a valid timecode has been detected. Further metadata key
15065 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
15066 timecode data has been found or not.
15068 This filter accepts the following options:
15072 Set the maximum number of lines to scan for VITC data. If the value is set to
15073 @code{-1} the full video frame is scanned. Default is @code{45}.
15076 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
15077 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
15080 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
15081 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
15084 @subsection Examples
15088 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
15089 draw @code{--:--:--:--} as a placeholder:
15091 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
15097 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
15099 Destination pixel at position (X, Y) will be picked from source (x, y) position
15100 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
15101 value for pixel will be used for destination pixel.
15103 Xmap and Ymap input video streams must be of same dimensions. Output video stream
15104 will have Xmap/Ymap video stream dimensions.
15105 Xmap and Ymap input video streams are 16bit depth, single channel.
15109 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
15110 Default is @code{color}.
15113 @section removegrain
15115 The removegrain filter is a spatial denoiser for progressive video.
15119 Set mode for the first plane.
15122 Set mode for the second plane.
15125 Set mode for the third plane.
15128 Set mode for the fourth plane.
15131 Range of mode is from 0 to 24. Description of each mode follows:
15135 Leave input plane unchanged. Default.
15138 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
15141 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
15144 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
15147 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
15148 This is equivalent to a median filter.
15151 Line-sensitive clipping giving the minimal change.
15154 Line-sensitive clipping, intermediate.
15157 Line-sensitive clipping, intermediate.
15160 Line-sensitive clipping, intermediate.
15163 Line-sensitive clipping on a line where the neighbours pixels are the closest.
15166 Replaces the target pixel with the closest neighbour.
15169 [1 2 1] horizontal and vertical kernel blur.
15175 Bob mode, interpolates top field from the line where the neighbours
15176 pixels are the closest.
15179 Bob mode, interpolates bottom field from the line where the neighbours
15180 pixels are the closest.
15183 Bob mode, interpolates top field. Same as 13 but with a more complicated
15184 interpolation formula.
15187 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
15188 interpolation formula.
15191 Clips the pixel with the minimum and maximum of respectively the maximum and
15192 minimum of each pair of opposite neighbour pixels.
15195 Line-sensitive clipping using opposite neighbours whose greatest distance from
15196 the current pixel is minimal.
15199 Replaces the pixel with the average of its 8 neighbours.
15202 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
15205 Clips pixels using the averages of opposite neighbour.
15208 Same as mode 21 but simpler and faster.
15211 Small edge and halo removal, but reputed useless.
15217 @section removelogo
15219 Suppress a TV station logo, using an image file to determine which
15220 pixels comprise the logo. It works by filling in the pixels that
15221 comprise the logo with neighboring pixels.
15223 The filter accepts the following options:
15227 Set the filter bitmap file, which can be any image format supported by
15228 libavformat. The width and height of the image file must match those of the
15229 video stream being processed.
15232 Pixels in the provided bitmap image with a value of zero are not
15233 considered part of the logo, non-zero pixels are considered part of
15234 the logo. If you use white (255) for the logo and black (0) for the
15235 rest, you will be safe. For making the filter bitmap, it is
15236 recommended to take a screen capture of a black frame with the logo
15237 visible, and then using a threshold filter followed by the erode
15238 filter once or twice.
15240 If needed, little splotches can be fixed manually. Remember that if
15241 logo pixels are not covered, the filter quality will be much
15242 reduced. Marking too many pixels as part of the logo does not hurt as
15243 much, but it will increase the amount of blurring needed to cover over
15244 the image and will destroy more information than necessary, and extra
15245 pixels will slow things down on a large logo.
15247 @section repeatfields
15249 This filter uses the repeat_field flag from the Video ES headers and hard repeats
15250 fields based on its value.
15254 Reverse a video clip.
15256 Warning: This filter requires memory to buffer the entire clip, so trimming
15259 @subsection Examples
15263 Take the first 5 seconds of a clip, and reverse it.
15270 Shift R/G/B/A pixels horizontally and/or vertically.
15272 The filter accepts the following options:
15275 Set amount to shift red horizontally.
15277 Set amount to shift red vertically.
15279 Set amount to shift green horizontally.
15281 Set amount to shift green vertically.
15283 Set amount to shift blue horizontally.
15285 Set amount to shift blue vertically.
15287 Set amount to shift alpha horizontally.
15289 Set amount to shift alpha vertically.
15291 Set edge mode, can be @var{smear}, default, or @var{warp}.
15295 Apply roberts cross operator to input video stream.
15297 The filter accepts the following option:
15301 Set which planes will be processed, unprocessed planes will be copied.
15302 By default value 0xf, all planes will be processed.
15305 Set value which will be multiplied with filtered result.
15308 Set value which will be added to filtered result.
15313 Rotate video by an arbitrary angle expressed in radians.
15315 The filter accepts the following options:
15317 A description of the optional parameters follows.
15320 Set an expression for the angle by which to rotate the input video
15321 clockwise, expressed as a number of radians. A negative value will
15322 result in a counter-clockwise rotation. By default it is set to "0".
15324 This expression is evaluated for each frame.
15327 Set the output width expression, default value is "iw".
15328 This expression is evaluated just once during configuration.
15331 Set the output height expression, default value is "ih".
15332 This expression is evaluated just once during configuration.
15335 Enable bilinear interpolation if set to 1, a value of 0 disables
15336 it. Default value is 1.
15339 Set the color used to fill the output area not covered by the rotated
15340 image. For the general syntax of this option, check the
15341 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
15342 If the special value "none" is selected then no
15343 background is printed (useful for example if the background is never shown).
15345 Default value is "black".
15348 The expressions for the angle and the output size can contain the
15349 following constants and functions:
15353 sequential number of the input frame, starting from 0. It is always NAN
15354 before the first frame is filtered.
15357 time in seconds of the input frame, it is set to 0 when the filter is
15358 configured. It is always NAN before the first frame is filtered.
15362 horizontal and vertical chroma subsample values. For example for the
15363 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15367 the input video width and height
15371 the output width and height, that is the size of the padded area as
15372 specified by the @var{width} and @var{height} expressions
15376 the minimal width/height required for completely containing the input
15377 video rotated by @var{a} radians.
15379 These are only available when computing the @option{out_w} and
15380 @option{out_h} expressions.
15383 @subsection Examples
15387 Rotate the input by PI/6 radians clockwise:
15393 Rotate the input by PI/6 radians counter-clockwise:
15399 Rotate the input by 45 degrees clockwise:
15405 Apply a constant rotation with period T, starting from an angle of PI/3:
15407 rotate=PI/3+2*PI*t/T
15411 Make the input video rotation oscillating with a period of T
15412 seconds and an amplitude of A radians:
15414 rotate=A*sin(2*PI/T*t)
15418 Rotate the video, output size is chosen so that the whole rotating
15419 input video is always completely contained in the output:
15421 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
15425 Rotate the video, reduce the output size so that no background is ever
15428 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
15432 @subsection Commands
15434 The filter supports the following commands:
15438 Set the angle expression.
15439 The command accepts the same syntax of the corresponding option.
15441 If the specified expression is not valid, it is kept at its current
15447 Apply Shape Adaptive Blur.
15449 The filter accepts the following options:
15452 @item luma_radius, lr
15453 Set luma blur filter strength, must be a value in range 0.1-4.0, default
15454 value is 1.0. A greater value will result in a more blurred image, and
15455 in slower processing.
15457 @item luma_pre_filter_radius, lpfr
15458 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
15461 @item luma_strength, ls
15462 Set luma maximum difference between pixels to still be considered, must
15463 be a value in the 0.1-100.0 range, default value is 1.0.
15465 @item chroma_radius, cr
15466 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
15467 greater value will result in a more blurred image, and in slower
15470 @item chroma_pre_filter_radius, cpfr
15471 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
15473 @item chroma_strength, cs
15474 Set chroma maximum difference between pixels to still be considered,
15475 must be a value in the -0.9-100.0 range.
15478 Each chroma option value, if not explicitly specified, is set to the
15479 corresponding luma option value.
15484 Scale (resize) the input video, using the libswscale library.
15486 The scale filter forces the output display aspect ratio to be the same
15487 of the input, by changing the output sample aspect ratio.
15489 If the input image format is different from the format requested by
15490 the next filter, the scale filter will convert the input to the
15493 @subsection Options
15494 The filter accepts the following options, or any of the options
15495 supported by the libswscale scaler.
15497 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
15498 the complete list of scaler options.
15503 Set the output video dimension expression. Default value is the input
15506 If the @var{width} or @var{w} value is 0, the input width is used for
15507 the output. If the @var{height} or @var{h} value is 0, the input height
15508 is used for the output.
15510 If one and only one of the values is -n with n >= 1, the scale filter
15511 will use a value that maintains the aspect ratio of the input image,
15512 calculated from the other specified dimension. After that it will,
15513 however, make sure that the calculated dimension is divisible by n and
15514 adjust the value if necessary.
15516 If both values are -n with n >= 1, the behavior will be identical to
15517 both values being set to 0 as previously detailed.
15519 See below for the list of accepted constants for use in the dimension
15523 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
15527 Only evaluate expressions once during the filter initialization or when a command is processed.
15530 Evaluate expressions for each incoming frame.
15534 Default value is @samp{init}.
15538 Set the interlacing mode. It accepts the following values:
15542 Force interlaced aware scaling.
15545 Do not apply interlaced scaling.
15548 Select interlaced aware scaling depending on whether the source frames
15549 are flagged as interlaced or not.
15552 Default value is @samp{0}.
15555 Set libswscale scaling flags. See
15556 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
15557 complete list of values. If not explicitly specified the filter applies
15561 @item param0, param1
15562 Set libswscale input parameters for scaling algorithms that need them. See
15563 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
15564 complete documentation. If not explicitly specified the filter applies
15570 Set the video size. For the syntax of this option, check the
15571 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
15573 @item in_color_matrix
15574 @item out_color_matrix
15575 Set in/output YCbCr color space type.
15577 This allows the autodetected value to be overridden as well as allows forcing
15578 a specific value used for the output and encoder.
15580 If not specified, the color space type depends on the pixel format.
15586 Choose automatically.
15589 Format conforming to International Telecommunication Union (ITU)
15590 Recommendation BT.709.
15593 Set color space conforming to the United States Federal Communications
15594 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
15599 Set color space conforming to:
15603 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
15606 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
15609 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
15614 Set color space conforming to SMPTE ST 240:1999.
15617 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
15622 Set in/output YCbCr sample range.
15624 This allows the autodetected value to be overridden as well as allows forcing
15625 a specific value used for the output and encoder. If not specified, the
15626 range depends on the pixel format. Possible values:
15630 Choose automatically.
15633 Set full range (0-255 in case of 8-bit luma).
15635 @item mpeg/limited/tv
15636 Set "MPEG" range (16-235 in case of 8-bit luma).
15639 @item force_original_aspect_ratio
15640 Enable decreasing or increasing output video width or height if necessary to
15641 keep the original aspect ratio. Possible values:
15645 Scale the video as specified and disable this feature.
15648 The output video dimensions will automatically be decreased if needed.
15651 The output video dimensions will automatically be increased if needed.
15655 One useful instance of this option is that when you know a specific device's
15656 maximum allowed resolution, you can use this to limit the output video to
15657 that, while retaining the aspect ratio. For example, device A allows
15658 1280x720 playback, and your video is 1920x800. Using this option (set it to
15659 decrease) and specifying 1280x720 to the command line makes the output
15662 Please note that this is a different thing than specifying -1 for @option{w}
15663 or @option{h}, you still need to specify the output resolution for this option
15666 @item force_divisible_by
15667 Ensures that both the output dimensions, width and height, are divisible by the
15668 given integer when used together with @option{force_original_aspect_ratio}. This
15669 works similar to using @code{-n} in the @option{w} and @option{h} options.
15671 This option respects the value set for @option{force_original_aspect_ratio},
15672 increasing or decreasing the resolution accordingly. The video's aspect ratio
15673 may be slightly modified.
15675 This option can be handy if you need to have a video fit within or exceed
15676 a defined resolution using @option{force_original_aspect_ratio} but also have
15677 encoder restrictions on width or height divisibility.
15681 The values of the @option{w} and @option{h} options are expressions
15682 containing the following constants:
15687 The input width and height
15691 These are the same as @var{in_w} and @var{in_h}.
15695 The output (scaled) width and height
15699 These are the same as @var{out_w} and @var{out_h}
15702 The same as @var{iw} / @var{ih}
15705 input sample aspect ratio
15708 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
15712 horizontal and vertical input chroma subsample values. For example for the
15713 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15717 horizontal and vertical output chroma subsample values. For example for the
15718 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15721 @subsection Examples
15725 Scale the input video to a size of 200x100
15730 This is equivalent to:
15741 Specify a size abbreviation for the output size:
15746 which can also be written as:
15752 Scale the input to 2x:
15754 scale=w=2*iw:h=2*ih
15758 The above is the same as:
15760 scale=2*in_w:2*in_h
15764 Scale the input to 2x with forced interlaced scaling:
15766 scale=2*iw:2*ih:interl=1
15770 Scale the input to half size:
15772 scale=w=iw/2:h=ih/2
15776 Increase the width, and set the height to the same size:
15782 Seek Greek harmony:
15789 Increase the height, and set the width to 3/2 of the height:
15791 scale=w=3/2*oh:h=3/5*ih
15795 Increase the size, making the size a multiple of the chroma
15798 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
15802 Increase the width to a maximum of 500 pixels,
15803 keeping the same aspect ratio as the input:
15805 scale=w='min(500\, iw*3/2):h=-1'
15809 Make pixels square by combining scale and setsar:
15811 scale='trunc(ih*dar):ih',setsar=1/1
15815 Make pixels square by combining scale and setsar,
15816 making sure the resulting resolution is even (required by some codecs):
15818 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
15822 @subsection Commands
15824 This filter supports the following commands:
15828 Set the output video dimension expression.
15829 The command accepts the same syntax of the corresponding option.
15831 If the specified expression is not valid, it is kept at its current
15837 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
15838 format conversion on CUDA video frames. Setting the output width and height
15839 works in the same way as for the @var{scale} filter.
15841 The following additional options are accepted:
15844 The pixel format of the output CUDA frames. If set to the string "same" (the
15845 default), the input format will be kept. Note that automatic format negotiation
15846 and conversion is not yet supported for hardware frames
15849 The interpolation algorithm used for resizing. One of the following:
15856 @item cubic2p_bspline
15857 2-parameter cubic (B=1, C=0)
15859 @item cubic2p_catmullrom
15860 2-parameter cubic (B=0, C=1/2)
15862 @item cubic2p_b05c03
15863 2-parameter cubic (B=1/2, C=3/10)
15875 Scale (resize) the input video, based on a reference video.
15877 See the scale filter for available options, scale2ref supports the same but
15878 uses the reference video instead of the main input as basis. scale2ref also
15879 supports the following additional constants for the @option{w} and
15880 @option{h} options:
15885 The main input video's width and height
15888 The same as @var{main_w} / @var{main_h}
15891 The main input video's sample aspect ratio
15893 @item main_dar, mdar
15894 The main input video's display aspect ratio. Calculated from
15895 @code{(main_w / main_h) * main_sar}.
15899 The main input video's horizontal and vertical chroma subsample values.
15900 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
15904 @subsection Examples
15908 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
15910 'scale2ref[b][a];[a][b]overlay'
15914 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
15916 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
15921 Scroll input video horizontally and/or vertically by constant speed.
15923 The filter accepts the following options:
15925 @item horizontal, h
15926 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
15927 Negative values changes scrolling direction.
15930 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
15931 Negative values changes scrolling direction.
15934 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
15937 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
15940 @subsection Commands
15942 This filter supports the following @ref{commands}:
15944 @item horizontal, h
15945 Set the horizontal scrolling speed.
15947 Set the vertical scrolling speed.
15950 @anchor{selectivecolor}
15951 @section selectivecolor
15953 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
15954 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
15955 by the "purity" of the color (that is, how saturated it already is).
15957 This filter is similar to the Adobe Photoshop Selective Color tool.
15959 The filter accepts the following options:
15962 @item correction_method
15963 Select color correction method.
15965 Available values are:
15968 Specified adjustments are applied "as-is" (added/subtracted to original pixel
15971 Specified adjustments are relative to the original component value.
15973 Default is @code{absolute}.
15975 Adjustments for red pixels (pixels where the red component is the maximum)
15977 Adjustments for yellow pixels (pixels where the blue component is the minimum)
15979 Adjustments for green pixels (pixels where the green component is the maximum)
15981 Adjustments for cyan pixels (pixels where the red component is the minimum)
15983 Adjustments for blue pixels (pixels where the blue component is the maximum)
15985 Adjustments for magenta pixels (pixels where the green component is the minimum)
15987 Adjustments for white pixels (pixels where all components are greater than 128)
15989 Adjustments for all pixels except pure black and pure white
15991 Adjustments for black pixels (pixels where all components are lesser than 128)
15993 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
15996 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
15997 4 space separated floating point adjustment values in the [-1,1] range,
15998 respectively to adjust the amount of cyan, magenta, yellow and black for the
15999 pixels of its range.
16001 @subsection Examples
16005 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
16006 increase magenta by 27% in blue areas:
16008 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
16012 Use a Photoshop selective color preset:
16014 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
16018 @anchor{separatefields}
16019 @section separatefields
16021 The @code{separatefields} takes a frame-based video input and splits
16022 each frame into its components fields, producing a new half height clip
16023 with twice the frame rate and twice the frame count.
16025 This filter use field-dominance information in frame to decide which
16026 of each pair of fields to place first in the output.
16027 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
16029 @section setdar, setsar
16031 The @code{setdar} filter sets the Display Aspect Ratio for the filter
16034 This is done by changing the specified Sample (aka Pixel) Aspect
16035 Ratio, according to the following equation:
16037 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
16040 Keep in mind that the @code{setdar} filter does not modify the pixel
16041 dimensions of the video frame. Also, the display aspect ratio set by
16042 this filter may be changed by later filters in the filterchain,
16043 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
16046 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
16047 the filter output video.
16049 Note that as a consequence of the application of this filter, the
16050 output display aspect ratio will change according to the equation
16053 Keep in mind that the sample aspect ratio set by the @code{setsar}
16054 filter may be changed by later filters in the filterchain, e.g. if
16055 another "setsar" or a "setdar" filter is applied.
16057 It accepts the following parameters:
16060 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
16061 Set the aspect ratio used by the filter.
16063 The parameter can be a floating point number string, an expression, or
16064 a string of the form @var{num}:@var{den}, where @var{num} and
16065 @var{den} are the numerator and denominator of the aspect ratio. If
16066 the parameter is not specified, it is assumed the value "0".
16067 In case the form "@var{num}:@var{den}" is used, the @code{:} character
16071 Set the maximum integer value to use for expressing numerator and
16072 denominator when reducing the expressed aspect ratio to a rational.
16073 Default value is @code{100}.
16077 The parameter @var{sar} is an expression containing
16078 the following constants:
16082 These are approximated values for the mathematical constants e
16083 (Euler's number), pi (Greek pi), and phi (the golden ratio).
16086 The input width and height.
16089 These are the same as @var{w} / @var{h}.
16092 The input sample aspect ratio.
16095 The input display aspect ratio. It is the same as
16096 (@var{w} / @var{h}) * @var{sar}.
16099 Horizontal and vertical chroma subsample values. For example, for the
16100 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16103 @subsection Examples
16108 To change the display aspect ratio to 16:9, specify one of the following:
16115 To change the sample aspect ratio to 10:11, specify:
16121 To set a display aspect ratio of 16:9, and specify a maximum integer value of
16122 1000 in the aspect ratio reduction, use the command:
16124 setdar=ratio=16/9:max=1000
16132 Force field for the output video frame.
16134 The @code{setfield} filter marks the interlace type field for the
16135 output frames. It does not change the input frame, but only sets the
16136 corresponding property, which affects how the frame is treated by
16137 following filters (e.g. @code{fieldorder} or @code{yadif}).
16139 The filter accepts the following options:
16144 Available values are:
16148 Keep the same field property.
16151 Mark the frame as bottom-field-first.
16154 Mark the frame as top-field-first.
16157 Mark the frame as progressive.
16164 Force frame parameter for the output video frame.
16166 The @code{setparams} filter marks interlace and color range for the
16167 output frames. It does not change the input frame, but only sets the
16168 corresponding property, which affects how the frame is treated by
16173 Available values are:
16177 Keep the same field property (default).
16180 Mark the frame as bottom-field-first.
16183 Mark the frame as top-field-first.
16186 Mark the frame as progressive.
16190 Available values are:
16194 Keep the same color range property (default).
16196 @item unspecified, unknown
16197 Mark the frame as unspecified color range.
16199 @item limited, tv, mpeg
16200 Mark the frame as limited range.
16202 @item full, pc, jpeg
16203 Mark the frame as full range.
16206 @item color_primaries
16207 Set the color primaries.
16208 Available values are:
16212 Keep the same color primaries property (default).
16229 Set the color transfer.
16230 Available values are:
16234 Keep the same color trc property (default).
16256 Set the colorspace.
16257 Available values are:
16261 Keep the same colorspace property (default).
16274 @item chroma-derived-nc
16275 @item chroma-derived-c
16282 Show a line containing various information for each input video frame.
16283 The input video is not modified.
16285 This filter supports the following options:
16289 Calculate checksums of each plane. By default enabled.
16292 The shown line contains a sequence of key/value pairs of the form
16293 @var{key}:@var{value}.
16295 The following values are shown in the output:
16299 The (sequential) number of the input frame, starting from 0.
16302 The Presentation TimeStamp of the input frame, expressed as a number of
16303 time base units. The time base unit depends on the filter input pad.
16306 The Presentation TimeStamp of the input frame, expressed as a number of
16310 The position of the frame in the input stream, or -1 if this information is
16311 unavailable and/or meaningless (for example in case of synthetic video).
16314 The pixel format name.
16317 The sample aspect ratio of the input frame, expressed in the form
16318 @var{num}/@var{den}.
16321 The size of the input frame. For the syntax of this option, check the
16322 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16325 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
16326 for bottom field first).
16329 This is 1 if the frame is a key frame, 0 otherwise.
16332 The picture type of the input frame ("I" for an I-frame, "P" for a
16333 P-frame, "B" for a B-frame, or "?" for an unknown type).
16334 Also refer to the documentation of the @code{AVPictureType} enum and of
16335 the @code{av_get_picture_type_char} function defined in
16336 @file{libavutil/avutil.h}.
16339 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
16341 @item plane_checksum
16342 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
16343 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
16346 @section showpalette
16348 Displays the 256 colors palette of each frame. This filter is only relevant for
16349 @var{pal8} pixel format frames.
16351 It accepts the following option:
16355 Set the size of the box used to represent one palette color entry. Default is
16356 @code{30} (for a @code{30x30} pixel box).
16359 @section shuffleframes
16361 Reorder and/or duplicate and/or drop video frames.
16363 It accepts the following parameters:
16367 Set the destination indexes of input frames.
16368 This is space or '|' separated list of indexes that maps input frames to output
16369 frames. Number of indexes also sets maximal value that each index may have.
16370 '-1' index have special meaning and that is to drop frame.
16373 The first frame has the index 0. The default is to keep the input unchanged.
16375 @subsection Examples
16379 Swap second and third frame of every three frames of the input:
16381 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
16385 Swap 10th and 1st frame of every ten frames of the input:
16387 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
16391 @section shuffleplanes
16393 Reorder and/or duplicate video planes.
16395 It accepts the following parameters:
16400 The index of the input plane to be used as the first output plane.
16403 The index of the input plane to be used as the second output plane.
16406 The index of the input plane to be used as the third output plane.
16409 The index of the input plane to be used as the fourth output plane.
16413 The first plane has the index 0. The default is to keep the input unchanged.
16415 @subsection Examples
16419 Swap the second and third planes of the input:
16421 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
16425 @anchor{signalstats}
16426 @section signalstats
16427 Evaluate various visual metrics that assist in determining issues associated
16428 with the digitization of analog video media.
16430 By default the filter will log these metadata values:
16434 Display the minimal Y value contained within the input frame. Expressed in
16438 Display the Y value at the 10% percentile within the input frame. Expressed in
16442 Display the average Y value within the input frame. Expressed in range of
16446 Display the Y value at the 90% percentile within the input frame. Expressed in
16450 Display the maximum Y value contained within the input frame. Expressed in
16454 Display the minimal U value contained within the input frame. Expressed in
16458 Display the U value at the 10% percentile within the input frame. Expressed in
16462 Display the average U value within the input frame. Expressed in range of
16466 Display the U value at the 90% percentile within the input frame. Expressed in
16470 Display the maximum U value contained within the input frame. Expressed in
16474 Display the minimal V value contained within the input frame. Expressed in
16478 Display the V value at the 10% percentile within the input frame. Expressed in
16482 Display the average V value within the input frame. Expressed in range of
16486 Display the V value at the 90% percentile within the input frame. Expressed in
16490 Display the maximum V value contained within the input frame. Expressed in
16494 Display the minimal saturation value contained within the input frame.
16495 Expressed in range of [0-~181.02].
16498 Display the saturation value at the 10% percentile within the input frame.
16499 Expressed in range of [0-~181.02].
16502 Display the average saturation value within the input frame. Expressed in range
16506 Display the saturation value at the 90% percentile within the input frame.
16507 Expressed in range of [0-~181.02].
16510 Display the maximum saturation value contained within the input frame.
16511 Expressed in range of [0-~181.02].
16514 Display the median value for hue within the input frame. Expressed in range of
16518 Display the average value for hue within the input frame. Expressed in range of
16522 Display the average of sample value difference between all values of the Y
16523 plane in the current frame and corresponding values of the previous input frame.
16524 Expressed in range of [0-255].
16527 Display the average of sample value difference between all values of the U
16528 plane in the current frame and corresponding values of the previous input frame.
16529 Expressed in range of [0-255].
16532 Display the average of sample value difference between all values of the V
16533 plane in the current frame and corresponding values of the previous input frame.
16534 Expressed in range of [0-255].
16537 Display bit depth of Y plane in current frame.
16538 Expressed in range of [0-16].
16541 Display bit depth of U plane in current frame.
16542 Expressed in range of [0-16].
16545 Display bit depth of V plane in current frame.
16546 Expressed in range of [0-16].
16549 The filter accepts the following options:
16555 @option{stat} specify an additional form of image analysis.
16556 @option{out} output video with the specified type of pixel highlighted.
16558 Both options accept the following values:
16562 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
16563 unlike the neighboring pixels of the same field. Examples of temporal outliers
16564 include the results of video dropouts, head clogs, or tape tracking issues.
16567 Identify @var{vertical line repetition}. Vertical line repetition includes
16568 similar rows of pixels within a frame. In born-digital video vertical line
16569 repetition is common, but this pattern is uncommon in video digitized from an
16570 analog source. When it occurs in video that results from the digitization of an
16571 analog source it can indicate concealment from a dropout compensator.
16574 Identify pixels that fall outside of legal broadcast range.
16578 Set the highlight color for the @option{out} option. The default color is
16582 @subsection Examples
16586 Output data of various video metrics:
16588 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
16592 Output specific data about the minimum and maximum values of the Y plane per frame:
16594 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
16598 Playback video while highlighting pixels that are outside of broadcast range in red.
16600 ffplay example.mov -vf signalstats="out=brng:color=red"
16604 Playback video with signalstats metadata drawn over the frame.
16606 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
16609 The contents of signalstat_drawtext.txt used in the command are:
16612 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
16613 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
16614 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
16615 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
16623 Calculates the MPEG-7 Video Signature. The filter can handle more than one
16624 input. In this case the matching between the inputs can be calculated additionally.
16625 The filter always passes through the first input. The signature of each stream can
16626 be written into a file.
16628 It accepts the following options:
16632 Enable or disable the matching process.
16634 Available values are:
16638 Disable the calculation of a matching (default).
16640 Calculate the matching for the whole video and output whether the whole video
16641 matches or only parts.
16643 Calculate only until a matching is found or the video ends. Should be faster in
16648 Set the number of inputs. The option value must be a non negative integer.
16649 Default value is 1.
16652 Set the path to which the output is written. If there is more than one input,
16653 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
16654 integer), that will be replaced with the input number. If no filename is
16655 specified, no output will be written. This is the default.
16658 Choose the output format.
16660 Available values are:
16664 Use the specified binary representation (default).
16666 Use the specified xml representation.
16670 Set threshold to detect one word as similar. The option value must be an integer
16671 greater than zero. The default value is 9000.
16674 Set threshold to detect all words as similar. The option value must be an integer
16675 greater than zero. The default value is 60000.
16678 Set threshold to detect frames as similar. The option value must be an integer
16679 greater than zero. The default value is 116.
16682 Set the minimum length of a sequence in frames to recognize it as matching
16683 sequence. The option value must be a non negative integer value.
16684 The default value is 0.
16687 Set the minimum relation, that matching frames to all frames must have.
16688 The option value must be a double value between 0 and 1. The default value is 0.5.
16691 @subsection Examples
16695 To calculate the signature of an input video and store it in signature.bin:
16697 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
16701 To detect whether two videos match and store the signatures in XML format in
16702 signature0.xml and signature1.xml:
16704 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 -
16712 Blur the input video without impacting the outlines.
16714 It accepts the following options:
16717 @item luma_radius, lr
16718 Set the luma radius. The option value must be a float number in
16719 the range [0.1,5.0] that specifies the variance of the gaussian filter
16720 used to blur the image (slower if larger). Default value is 1.0.
16722 @item luma_strength, ls
16723 Set the luma strength. The option value must be a float number
16724 in the range [-1.0,1.0] that configures the blurring. A value included
16725 in [0.0,1.0] will blur the image whereas a value included in
16726 [-1.0,0.0] will sharpen the image. Default value is 1.0.
16728 @item luma_threshold, lt
16729 Set the luma threshold used as a coefficient to determine
16730 whether a pixel should be blurred or not. The option value must be an
16731 integer in the range [-30,30]. A value of 0 will filter all the image,
16732 a value included in [0,30] will filter flat areas and a value included
16733 in [-30,0] will filter edges. Default value is 0.
16735 @item chroma_radius, cr
16736 Set the chroma radius. The option value must be a float number in
16737 the range [0.1,5.0] that specifies the variance of the gaussian filter
16738 used to blur the image (slower if larger). Default value is @option{luma_radius}.
16740 @item chroma_strength, cs
16741 Set the chroma strength. The option value must be a float number
16742 in the range [-1.0,1.0] that configures the blurring. A value included
16743 in [0.0,1.0] will blur the image whereas a value included in
16744 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
16746 @item chroma_threshold, ct
16747 Set the chroma threshold used as a coefficient to determine
16748 whether a pixel should be blurred or not. The option value must be an
16749 integer in the range [-30,30]. A value of 0 will filter all the image,
16750 a value included in [0,30] will filter flat areas and a value included
16751 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
16754 If a chroma option is not explicitly set, the corresponding luma value
16758 Apply sobel operator to input video stream.
16760 The filter accepts the following option:
16764 Set which planes will be processed, unprocessed planes will be copied.
16765 By default value 0xf, all planes will be processed.
16768 Set value which will be multiplied with filtered result.
16771 Set value which will be added to filtered result.
16777 Apply a simple postprocessing filter that compresses and decompresses the image
16778 at several (or - in the case of @option{quality} level @code{6} - all) shifts
16779 and average the results.
16781 The filter accepts the following options:
16785 Set quality. This option defines the number of levels for averaging. It accepts
16786 an integer in the range 0-6. If set to @code{0}, the filter will have no
16787 effect. A value of @code{6} means the higher quality. For each increment of
16788 that value the speed drops by a factor of approximately 2. Default value is
16792 Force a constant quantization parameter. If not set, the filter will use the QP
16793 from the video stream (if available).
16796 Set thresholding mode. Available modes are:
16800 Set hard thresholding (default).
16802 Set soft thresholding (better de-ringing effect, but likely blurrier).
16805 @item use_bframe_qp
16806 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
16807 option may cause flicker since the B-Frames have often larger QP. Default is
16808 @code{0} (not enabled).
16813 Scale the input by applying one of the super-resolution methods based on
16814 convolutional neural networks. Supported models:
16818 Super-Resolution Convolutional Neural Network model (SRCNN).
16819 See @url{https://arxiv.org/abs/1501.00092}.
16822 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
16823 See @url{https://arxiv.org/abs/1609.05158}.
16826 Training scripts as well as scripts for model file (.pb) saving can be found at
16827 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
16828 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
16830 Native model files (.model) can be generated from TensorFlow model
16831 files (.pb) by using tools/python/convert.py
16833 The filter accepts the following options:
16837 Specify which DNN backend to use for model loading and execution. This option accepts
16838 the following values:
16842 Native implementation of DNN loading and execution.
16845 TensorFlow backend. To enable this backend you
16846 need to install the TensorFlow for C library (see
16847 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
16848 @code{--enable-libtensorflow}
16851 Default value is @samp{native}.
16854 Set path to model file specifying network architecture and its parameters.
16855 Note that different backends use different file formats. TensorFlow backend
16856 can load files for both formats, while native backend can load files for only
16860 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
16861 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
16862 input upscaled using bicubic upscaling with proper scale factor.
16867 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
16869 This filter takes in input two input videos, the first input is
16870 considered the "main" source and is passed unchanged to the
16871 output. The second input is used as a "reference" video for computing
16874 Both video inputs must have the same resolution and pixel format for
16875 this filter to work correctly. Also it assumes that both inputs
16876 have the same number of frames, which are compared one by one.
16878 The filter stores the calculated SSIM of each frame.
16880 The description of the accepted parameters follows.
16883 @item stats_file, f
16884 If specified the filter will use the named file to save the SSIM of
16885 each individual frame. When filename equals "-" the data is sent to
16889 The file printed if @var{stats_file} is selected, contains a sequence of
16890 key/value pairs of the form @var{key}:@var{value} for each compared
16893 A description of each shown parameter follows:
16897 sequential number of the input frame, starting from 1
16899 @item Y, U, V, R, G, B
16900 SSIM of the compared frames for the component specified by the suffix.
16903 SSIM of the compared frames for the whole frame.
16906 Same as above but in dB representation.
16909 This filter also supports the @ref{framesync} options.
16913 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
16914 [main][ref] ssim="stats_file=stats.log" [out]
16917 On this example the input file being processed is compared with the
16918 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
16919 is stored in @file{stats.log}.
16921 Another example with both psnr and ssim at same time:
16923 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
16928 Convert between different stereoscopic image formats.
16930 The filters accept the following options:
16934 Set stereoscopic image format of input.
16936 Available values for input image formats are:
16939 side by side parallel (left eye left, right eye right)
16942 side by side crosseye (right eye left, left eye right)
16945 side by side parallel with half width resolution
16946 (left eye left, right eye right)
16949 side by side crosseye with half width resolution
16950 (right eye left, left eye right)
16954 above-below (left eye above, right eye below)
16958 above-below (right eye above, left eye below)
16962 above-below with half height resolution
16963 (left eye above, right eye below)
16967 above-below with half height resolution
16968 (right eye above, left eye below)
16971 alternating frames (left eye first, right eye second)
16974 alternating frames (right eye first, left eye second)
16977 interleaved rows (left eye has top row, right eye starts on next row)
16980 interleaved rows (right eye has top row, left eye starts on next row)
16983 interleaved columns, left eye first
16986 interleaved columns, right eye first
16988 Default value is @samp{sbsl}.
16992 Set stereoscopic image format of output.
16996 side by side parallel (left eye left, right eye right)
16999 side by side crosseye (right eye left, left eye right)
17002 side by side parallel with half width resolution
17003 (left eye left, right eye right)
17006 side by side crosseye with half width resolution
17007 (right eye left, left eye right)
17011 above-below (left eye above, right eye below)
17015 above-below (right eye above, left eye below)
17019 above-below with half height resolution
17020 (left eye above, right eye below)
17024 above-below with half height resolution
17025 (right eye above, left eye below)
17028 alternating frames (left eye first, right eye second)
17031 alternating frames (right eye first, left eye second)
17034 interleaved rows (left eye has top row, right eye starts on next row)
17037 interleaved rows (right eye has top row, left eye starts on next row)
17040 anaglyph red/blue gray
17041 (red filter on left eye, blue filter on right eye)
17044 anaglyph red/green gray
17045 (red filter on left eye, green filter on right eye)
17048 anaglyph red/cyan gray
17049 (red filter on left eye, cyan filter on right eye)
17052 anaglyph red/cyan half colored
17053 (red filter on left eye, cyan filter on right eye)
17056 anaglyph red/cyan color
17057 (red filter on left eye, cyan filter on right eye)
17060 anaglyph red/cyan color optimized with the least squares projection of dubois
17061 (red filter on left eye, cyan filter on right eye)
17064 anaglyph green/magenta gray
17065 (green filter on left eye, magenta filter on right eye)
17068 anaglyph green/magenta half colored
17069 (green filter on left eye, magenta filter on right eye)
17072 anaglyph green/magenta colored
17073 (green filter on left eye, magenta filter on right eye)
17076 anaglyph green/magenta color optimized with the least squares projection of dubois
17077 (green filter on left eye, magenta filter on right eye)
17080 anaglyph yellow/blue gray
17081 (yellow filter on left eye, blue filter on right eye)
17084 anaglyph yellow/blue half colored
17085 (yellow filter on left eye, blue filter on right eye)
17088 anaglyph yellow/blue colored
17089 (yellow filter on left eye, blue filter on right eye)
17092 anaglyph yellow/blue color optimized with the least squares projection of dubois
17093 (yellow filter on left eye, blue filter on right eye)
17096 mono output (left eye only)
17099 mono output (right eye only)
17102 checkerboard, left eye first
17105 checkerboard, right eye first
17108 interleaved columns, left eye first
17111 interleaved columns, right eye first
17117 Default value is @samp{arcd}.
17120 @subsection Examples
17124 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
17130 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
17136 @section streamselect, astreamselect
17137 Select video or audio streams.
17139 The filter accepts the following options:
17143 Set number of inputs. Default is 2.
17146 Set input indexes to remap to outputs.
17149 @subsection Commands
17151 The @code{streamselect} and @code{astreamselect} filter supports the following
17156 Set input indexes to remap to outputs.
17159 @subsection Examples
17163 Select first 5 seconds 1st stream and rest of time 2nd stream:
17165 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
17169 Same as above, but for audio:
17171 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
17178 Draw subtitles on top of input video using the libass library.
17180 To enable compilation of this filter you need to configure FFmpeg with
17181 @code{--enable-libass}. This filter also requires a build with libavcodec and
17182 libavformat to convert the passed subtitles file to ASS (Advanced Substation
17183 Alpha) subtitles format.
17185 The filter accepts the following options:
17189 Set the filename of the subtitle file to read. It must be specified.
17191 @item original_size
17192 Specify the size of the original video, the video for which the ASS file
17193 was composed. For the syntax of this option, check the
17194 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17195 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
17196 correctly scale the fonts if the aspect ratio has been changed.
17199 Set a directory path containing fonts that can be used by the filter.
17200 These fonts will be used in addition to whatever the font provider uses.
17203 Process alpha channel, by default alpha channel is untouched.
17206 Set subtitles input character encoding. @code{subtitles} filter only. Only
17207 useful if not UTF-8.
17209 @item stream_index, si
17210 Set subtitles stream index. @code{subtitles} filter only.
17213 Override default style or script info parameters of the subtitles. It accepts a
17214 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
17217 If the first key is not specified, it is assumed that the first value
17218 specifies the @option{filename}.
17220 For example, to render the file @file{sub.srt} on top of the input
17221 video, use the command:
17226 which is equivalent to:
17228 subtitles=filename=sub.srt
17231 To render the default subtitles stream from file @file{video.mkv}, use:
17233 subtitles=video.mkv
17236 To render the second subtitles stream from that file, use:
17238 subtitles=video.mkv:si=1
17241 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
17242 @code{DejaVu Serif}, use:
17244 subtitles=sub.srt:force_style='FontName=DejaVu Serif,PrimaryColour=&HCCFF0000'
17247 @section super2xsai
17249 Scale the input by 2x and smooth using the Super2xSaI (Scale and
17250 Interpolate) pixel art scaling algorithm.
17252 Useful for enlarging pixel art images without reducing sharpness.
17256 Swap two rectangular objects in video.
17258 This filter accepts the following options:
17268 Set 1st rect x coordinate.
17271 Set 1st rect y coordinate.
17274 Set 2nd rect x coordinate.
17277 Set 2nd rect y coordinate.
17279 All expressions are evaluated once for each frame.
17282 The all options are expressions containing the following constants:
17287 The input width and height.
17290 same as @var{w} / @var{h}
17293 input sample aspect ratio
17296 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
17299 The number of the input frame, starting from 0.
17302 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
17305 the position in the file of the input frame, NAN if unknown
17313 Apply telecine process to the video.
17315 This filter accepts the following options:
17324 The default value is @code{top}.
17328 A string of numbers representing the pulldown pattern you wish to apply.
17329 The default value is @code{23}.
17333 Some typical patterns:
17338 24p: 2332 (preferred)
17345 24p: 222222222223 ("Euro pulldown")
17352 Apply threshold effect to video stream.
17354 This filter needs four video streams to perform thresholding.
17355 First stream is stream we are filtering.
17356 Second stream is holding threshold values, third stream is holding min values,
17357 and last, fourth stream is holding max values.
17359 The filter accepts the following option:
17363 Set which planes will be processed, unprocessed planes will be copied.
17364 By default value 0xf, all planes will be processed.
17367 For example if first stream pixel's component value is less then threshold value
17368 of pixel component from 2nd threshold stream, third stream value will picked,
17369 otherwise fourth stream pixel component value will be picked.
17371 Using color source filter one can perform various types of thresholding:
17373 @subsection Examples
17377 Binary threshold, using gray color as threshold:
17379 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
17383 Inverted binary threshold, using gray color as threshold:
17385 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
17389 Truncate binary threshold, using gray color as threshold:
17391 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
17395 Threshold to zero, using gray color as threshold:
17397 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
17401 Inverted threshold to zero, using gray color as threshold:
17403 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
17408 Select the most representative frame in a given sequence of consecutive frames.
17410 The filter accepts the following options:
17414 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
17415 will pick one of them, and then handle the next batch of @var{n} frames until
17416 the end. Default is @code{100}.
17419 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
17420 value will result in a higher memory usage, so a high value is not recommended.
17422 @subsection Examples
17426 Extract one picture each 50 frames:
17432 Complete example of a thumbnail creation with @command{ffmpeg}:
17434 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
17440 Tile several successive frames together.
17442 The filter accepts the following options:
17447 Set the grid size (i.e. the number of lines and columns). For the syntax of
17448 this option, check the
17449 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17452 Set the maximum number of frames to render in the given area. It must be less
17453 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
17454 the area will be used.
17457 Set the outer border margin in pixels.
17460 Set the inner border thickness (i.e. the number of pixels between frames). For
17461 more advanced padding options (such as having different values for the edges),
17462 refer to the pad video filter.
17465 Specify the color of the unused area. For the syntax of this option, check the
17466 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
17467 The default value of @var{color} is "black".
17470 Set the number of frames to overlap when tiling several successive frames together.
17471 The value must be between @code{0} and @var{nb_frames - 1}.
17474 Set the number of frames to initially be empty before displaying first output frame.
17475 This controls how soon will one get first output frame.
17476 The value must be between @code{0} and @var{nb_frames - 1}.
17479 @subsection Examples
17483 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
17485 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
17487 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
17488 duplicating each output frame to accommodate the originally detected frame
17492 Display @code{5} pictures in an area of @code{3x2} frames,
17493 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
17494 mixed flat and named options:
17496 tile=3x2:nb_frames=5:padding=7:margin=2
17500 @section tinterlace
17502 Perform various types of temporal field interlacing.
17504 Frames are counted starting from 1, so the first input frame is
17507 The filter accepts the following options:
17512 Specify the mode of the interlacing. This option can also be specified
17513 as a value alone. See below for a list of values for this option.
17515 Available values are:
17519 Move odd frames into the upper field, even into the lower field,
17520 generating a double height frame at half frame rate.
17524 Frame 1 Frame 2 Frame 3 Frame 4
17526 11111 22222 33333 44444
17527 11111 22222 33333 44444
17528 11111 22222 33333 44444
17529 11111 22222 33333 44444
17543 Only output odd frames, even frames are dropped, generating a frame with
17544 unchanged height at half frame rate.
17549 Frame 1 Frame 2 Frame 3 Frame 4
17551 11111 22222 33333 44444
17552 11111 22222 33333 44444
17553 11111 22222 33333 44444
17554 11111 22222 33333 44444
17564 Only output even frames, odd frames are dropped, generating a frame with
17565 unchanged height at half frame rate.
17570 Frame 1 Frame 2 Frame 3 Frame 4
17572 11111 22222 33333 44444
17573 11111 22222 33333 44444
17574 11111 22222 33333 44444
17575 11111 22222 33333 44444
17585 Expand each frame to full height, but pad alternate lines with black,
17586 generating a frame with double height at the same input frame rate.
17591 Frame 1 Frame 2 Frame 3 Frame 4
17593 11111 22222 33333 44444
17594 11111 22222 33333 44444
17595 11111 22222 33333 44444
17596 11111 22222 33333 44444
17599 11111 ..... 33333 .....
17600 ..... 22222 ..... 44444
17601 11111 ..... 33333 .....
17602 ..... 22222 ..... 44444
17603 11111 ..... 33333 .....
17604 ..... 22222 ..... 44444
17605 11111 ..... 33333 .....
17606 ..... 22222 ..... 44444
17610 @item interleave_top, 4
17611 Interleave the upper field from odd frames with the lower field from
17612 even frames, generating a frame with unchanged height at half frame rate.
17617 Frame 1 Frame 2 Frame 3 Frame 4
17619 11111<- 22222 33333<- 44444
17620 11111 22222<- 33333 44444<-
17621 11111<- 22222 33333<- 44444
17622 11111 22222<- 33333 44444<-
17632 @item interleave_bottom, 5
17633 Interleave the lower field from odd frames with the upper field from
17634 even frames, generating a frame with unchanged height at half frame rate.
17639 Frame 1 Frame 2 Frame 3 Frame 4
17641 11111 22222<- 33333 44444<-
17642 11111<- 22222 33333<- 44444
17643 11111 22222<- 33333 44444<-
17644 11111<- 22222 33333<- 44444
17654 @item interlacex2, 6
17655 Double frame rate with unchanged height. Frames are inserted each
17656 containing the second temporal field from the previous input frame and
17657 the first temporal field from the next input frame. This mode relies on
17658 the top_field_first flag. Useful for interlaced video displays with no
17659 field synchronisation.
17664 Frame 1 Frame 2 Frame 3 Frame 4
17666 11111 22222 33333 44444
17667 11111 22222 33333 44444
17668 11111 22222 33333 44444
17669 11111 22222 33333 44444
17672 11111 22222 22222 33333 33333 44444 44444
17673 11111 11111 22222 22222 33333 33333 44444
17674 11111 22222 22222 33333 33333 44444 44444
17675 11111 11111 22222 22222 33333 33333 44444
17680 Move odd frames into the upper field, even into the lower field,
17681 generating a double height frame at same frame rate.
17686 Frame 1 Frame 2 Frame 3 Frame 4
17688 11111 22222 33333 44444
17689 11111 22222 33333 44444
17690 11111 22222 33333 44444
17691 11111 22222 33333 44444
17694 11111 33333 33333 55555
17695 22222 22222 44444 44444
17696 11111 33333 33333 55555
17697 22222 22222 44444 44444
17698 11111 33333 33333 55555
17699 22222 22222 44444 44444
17700 11111 33333 33333 55555
17701 22222 22222 44444 44444
17706 Numeric values are deprecated but are accepted for backward
17707 compatibility reasons.
17709 Default mode is @code{merge}.
17712 Specify flags influencing the filter process.
17714 Available value for @var{flags} is:
17717 @item low_pass_filter, vlpf
17718 Enable linear vertical low-pass filtering in the filter.
17719 Vertical low-pass filtering is required when creating an interlaced
17720 destination from a progressive source which contains high-frequency
17721 vertical detail. Filtering will reduce interlace 'twitter' and Moire
17724 @item complex_filter, cvlpf
17725 Enable complex vertical low-pass filtering.
17726 This will slightly less reduce interlace 'twitter' and Moire
17727 patterning but better retain detail and subjective sharpness impression.
17731 Vertical low-pass filtering can only be enabled for @option{mode}
17732 @var{interleave_top} and @var{interleave_bottom}.
17738 Mix successive video frames.
17740 A description of the accepted options follows.
17744 The number of successive frames to mix. If unspecified, it defaults to 3.
17747 Specify weight of each input video frame.
17748 Each weight is separated by space. If number of weights is smaller than
17749 number of @var{frames} last specified weight will be used for all remaining
17753 Specify scale, if it is set it will be multiplied with sum
17754 of each weight multiplied with pixel values to give final destination
17755 pixel value. By default @var{scale} is auto scaled to sum of weights.
17758 @subsection Examples
17762 Average 7 successive frames:
17764 tmix=frames=7:weights="1 1 1 1 1 1 1"
17768 Apply simple temporal convolution:
17770 tmix=frames=3:weights="-1 3 -1"
17774 Similar as above but only showing temporal differences:
17776 tmix=frames=3:weights="-1 2 -1":scale=1
17782 Tone map colors from different dynamic ranges.
17784 This filter expects data in single precision floating point, as it needs to
17785 operate on (and can output) out-of-range values. Another filter, such as
17786 @ref{zscale}, is needed to convert the resulting frame to a usable format.
17788 The tonemapping algorithms implemented only work on linear light, so input
17789 data should be linearized beforehand (and possibly correctly tagged).
17792 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
17795 @subsection Options
17796 The filter accepts the following options.
17800 Set the tone map algorithm to use.
17802 Possible values are:
17805 Do not apply any tone map, only desaturate overbright pixels.
17808 Hard-clip any out-of-range values. Use it for perfect color accuracy for
17809 in-range values, while distorting out-of-range values.
17812 Stretch the entire reference gamut to a linear multiple of the display.
17815 Fit a logarithmic transfer between the tone curves.
17818 Preserve overall image brightness with a simple curve, using nonlinear
17819 contrast, which results in flattening details and degrading color accuracy.
17822 Preserve both dark and bright details better than @var{reinhard}, at the cost
17823 of slightly darkening everything. Use it when detail preservation is more
17824 important than color and brightness accuracy.
17827 Smoothly map out-of-range values, while retaining contrast and colors for
17828 in-range material as much as possible. Use it when color accuracy is more
17829 important than detail preservation.
17835 Tune the tone mapping algorithm.
17837 This affects the following algorithms:
17843 Specifies the scale factor to use while stretching.
17847 Specifies the exponent of the function.
17851 Specify an extra linear coefficient to multiply into the signal before clipping.
17855 Specify the local contrast coefficient at the display peak.
17856 Default to 0.5, which means that in-gamut values will be about half as bright
17863 Specify the transition point from linear to mobius transform. Every value
17864 below this point is guaranteed to be mapped 1:1. The higher the value, the
17865 more accurate the result will be, at the cost of losing bright details.
17866 Default to 0.3, which due to the steep initial slope still preserves in-range
17867 colors fairly accurately.
17871 Apply desaturation for highlights that exceed this level of brightness. The
17872 higher the parameter, the more color information will be preserved. This
17873 setting helps prevent unnaturally blown-out colors for super-highlights, by
17874 (smoothly) turning into white instead. This makes images feel more natural,
17875 at the cost of reducing information about out-of-range colors.
17877 The default of 2.0 is somewhat conservative and will mostly just apply to
17878 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
17880 This option works only if the input frame has a supported color tag.
17883 Override signal/nominal/reference peak with this value. Useful when the
17884 embedded peak information in display metadata is not reliable or when tone
17885 mapping from a lower range to a higher range.
17890 Temporarily pad video frames.
17892 The filter accepts the following options:
17896 Specify number of delay frames before input video stream.
17899 Specify number of padding frames after input video stream.
17900 Set to -1 to pad indefinitely.
17903 Set kind of frames added to beginning of stream.
17904 Can be either @var{add} or @var{clone}.
17905 With @var{add} frames of solid-color are added.
17906 With @var{clone} frames are clones of first frame.
17909 Set kind of frames added to end of stream.
17910 Can be either @var{add} or @var{clone}.
17911 With @var{add} frames of solid-color are added.
17912 With @var{clone} frames are clones of last frame.
17914 @item start_duration, stop_duration
17915 Specify the duration of the start/stop delay. See
17916 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
17917 for the accepted syntax.
17918 These options override @var{start} and @var{stop}.
17921 Specify the color of the padded area. For the syntax of this option,
17922 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
17923 manual,ffmpeg-utils}.
17925 The default value of @var{color} is "black".
17931 Transpose rows with columns in the input video and optionally flip it.
17933 It accepts the following parameters:
17938 Specify the transposition direction.
17940 Can assume the following values:
17942 @item 0, 4, cclock_flip
17943 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
17951 Rotate by 90 degrees clockwise, that is:
17959 Rotate by 90 degrees counterclockwise, that is:
17966 @item 3, 7, clock_flip
17967 Rotate by 90 degrees clockwise and vertically flip, that is:
17975 For values between 4-7, the transposition is only done if the input
17976 video geometry is portrait and not landscape. These values are
17977 deprecated, the @code{passthrough} option should be used instead.
17979 Numerical values are deprecated, and should be dropped in favor of
17980 symbolic constants.
17983 Do not apply the transposition if the input geometry matches the one
17984 specified by the specified value. It accepts the following values:
17987 Always apply transposition.
17989 Preserve portrait geometry (when @var{height} >= @var{width}).
17991 Preserve landscape geometry (when @var{width} >= @var{height}).
17994 Default value is @code{none}.
17997 For example to rotate by 90 degrees clockwise and preserve portrait
18000 transpose=dir=1:passthrough=portrait
18003 The command above can also be specified as:
18005 transpose=1:portrait
18008 @section transpose_npp
18010 Transpose rows with columns in the input video and optionally flip it.
18011 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
18013 It accepts the following parameters:
18018 Specify the transposition direction.
18020 Can assume the following values:
18023 Rotate by 90 degrees counterclockwise and vertically flip. (default)
18026 Rotate by 90 degrees clockwise.
18029 Rotate by 90 degrees counterclockwise.
18032 Rotate by 90 degrees clockwise and vertically flip.
18036 Do not apply the transposition if the input geometry matches the one
18037 specified by the specified value. It accepts the following values:
18040 Always apply transposition. (default)
18042 Preserve portrait geometry (when @var{height} >= @var{width}).
18044 Preserve landscape geometry (when @var{width} >= @var{height}).
18050 Trim the input so that the output contains one continuous subpart of the input.
18052 It accepts the following parameters:
18055 Specify the time of the start of the kept section, i.e. the frame with the
18056 timestamp @var{start} will be the first frame in the output.
18059 Specify the time of the first frame that will be dropped, i.e. the frame
18060 immediately preceding the one with the timestamp @var{end} will be the last
18061 frame in the output.
18064 This is the same as @var{start}, except this option sets the start timestamp
18065 in timebase units instead of seconds.
18068 This is the same as @var{end}, except this option sets the end timestamp
18069 in timebase units instead of seconds.
18072 The maximum duration of the output in seconds.
18075 The number of the first frame that should be passed to the output.
18078 The number of the first frame that should be dropped.
18081 @option{start}, @option{end}, and @option{duration} are expressed as time
18082 duration specifications; see
18083 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
18084 for the accepted syntax.
18086 Note that the first two sets of the start/end options and the @option{duration}
18087 option look at the frame timestamp, while the _frame variants simply count the
18088 frames that pass through the filter. Also note that this filter does not modify
18089 the timestamps. If you wish for the output timestamps to start at zero, insert a
18090 setpts filter after the trim filter.
18092 If multiple start or end options are set, this filter tries to be greedy and
18093 keep all the frames that match at least one of the specified constraints. To keep
18094 only the part that matches all the constraints at once, chain multiple trim
18097 The defaults are such that all the input is kept. So it is possible to set e.g.
18098 just the end values to keep everything before the specified time.
18103 Drop everything except the second minute of input:
18105 ffmpeg -i INPUT -vf trim=60:120
18109 Keep only the first second:
18111 ffmpeg -i INPUT -vf trim=duration=1
18116 @section unpremultiply
18117 Apply alpha unpremultiply effect to input video stream using first plane
18118 of second stream as alpha.
18120 Both streams must have same dimensions and same pixel format.
18122 The filter accepts the following option:
18126 Set which planes will be processed, unprocessed planes will be copied.
18127 By default value 0xf, all planes will be processed.
18129 If the format has 1 or 2 components, then luma is bit 0.
18130 If the format has 3 or 4 components:
18131 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
18132 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
18133 If present, the alpha channel is always the last bit.
18136 Do not require 2nd input for processing, instead use alpha plane from input stream.
18142 Sharpen or blur the input video.
18144 It accepts the following parameters:
18147 @item luma_msize_x, lx
18148 Set the luma matrix horizontal size. It must be an odd integer between
18149 3 and 23. The default value is 5.
18151 @item luma_msize_y, ly
18152 Set the luma matrix vertical size. It must be an odd integer between 3
18153 and 23. The default value is 5.
18155 @item luma_amount, la
18156 Set the luma effect strength. It must be a floating point number, reasonable
18157 values lay between -1.5 and 1.5.
18159 Negative values will blur the input video, while positive values will
18160 sharpen it, a value of zero will disable the effect.
18162 Default value is 1.0.
18164 @item chroma_msize_x, cx
18165 Set the chroma matrix horizontal size. It must be an odd integer
18166 between 3 and 23. The default value is 5.
18168 @item chroma_msize_y, cy
18169 Set the chroma matrix vertical size. It must be an odd integer
18170 between 3 and 23. The default value is 5.
18172 @item chroma_amount, ca
18173 Set the chroma effect strength. It must be a floating point number, reasonable
18174 values lay between -1.5 and 1.5.
18176 Negative values will blur the input video, while positive values will
18177 sharpen it, a value of zero will disable the effect.
18179 Default value is 0.0.
18183 All parameters are optional and default to the equivalent of the
18184 string '5:5:1.0:5:5:0.0'.
18186 @subsection Examples
18190 Apply strong luma sharpen effect:
18192 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
18196 Apply a strong blur of both luma and chroma parameters:
18198 unsharp=7:7:-2:7:7:-2
18204 Apply ultra slow/simple postprocessing filter that compresses and decompresses
18205 the image at several (or - in the case of @option{quality} level @code{8} - all)
18206 shifts and average the results.
18208 The way this differs from the behavior of spp is that uspp actually encodes &
18209 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
18210 DCT similar to MJPEG.
18212 The filter accepts the following options:
18216 Set quality. This option defines the number of levels for averaging. It accepts
18217 an integer in the range 0-8. If set to @code{0}, the filter will have no
18218 effect. A value of @code{8} means the higher quality. For each increment of
18219 that value the speed drops by a factor of approximately 2. Default value is
18223 Force a constant quantization parameter. If not set, the filter will use the QP
18224 from the video stream (if available).
18229 Convert 360 videos between various formats.
18231 The filter accepts the following options:
18237 Set format of the input/output video.
18245 Equirectangular projection.
18250 Cubemap with 3x2/6x1/1x6 layout.
18252 Format specific options:
18257 Set padding proportion for the input/output cubemap. Values in decimals.
18264 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)
18267 Default value is @b{@samp{0}}.
18271 Set fixed padding for the input/output cubemap. Values in pixels.
18273 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
18277 Set order of faces for the input/output cubemap. Choose one direction for each position.
18279 Designation of directions:
18295 Default value is @b{@samp{rludfb}}.
18299 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
18301 Designation of angles:
18304 0 degrees clockwise
18306 90 degrees clockwise
18308 180 degrees clockwise
18310 270 degrees clockwise
18313 Default value is @b{@samp{000000}}.
18317 Equi-Angular Cubemap.
18322 Regular video. @i{(output only)}
18324 Format specific options:
18329 Set horizontal/vertical/diagonal field of view. Values in degrees.
18331 If diagonal field of view is set it overrides horizontal and vertical field of view.
18337 Format specific options:
18341 Set padding proportion. Values in decimals.
18351 Default value is @b{@samp{0}}.
18356 Facebook's 360 format.
18359 Stereographic format.
18361 Format specific options:
18366 Set horizontal/vertical/diagonal field of view. Values in degrees.
18368 If diagonal field of view is set it overrides horizontal and vertical field of view.
18375 Ball format, gives significant distortion toward the back.
18378 Hammer-Aitoff map projection format.
18381 Sinusoidal map projection format.
18386 Set interpolation method.@*
18387 @i{Note: more complex interpolation methods require much more memory to run.}
18397 Bilinear interpolation.
18400 Bicubic interpolation.
18403 Lanczos interpolation.
18406 Default value is @b{@samp{line}}.
18410 Set the output video resolution.
18412 Default resolution depends on formats.
18416 Set the input/output stereo format.
18427 Default value is @b{@samp{2d}} for input and output format.
18432 Set rotation for the output video. Values in degrees.
18435 Set rotation order for the output video. Choose one item for each position.
18446 Default value is @b{@samp{ypr}}.
18451 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
18455 Set if input video is flipped horizontally/vertically. Boolean values.
18458 Set if input video is transposed. Boolean value, by default disabled.
18461 Set if output video needs to be transposed. Boolean value, by default disabled.
18465 @subsection Examples
18469 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
18471 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
18474 Extract back view of Equi-Angular Cubemap:
18476 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
18479 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
18481 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
18485 @section vaguedenoiser
18487 Apply a wavelet based denoiser.
18489 It transforms each frame from the video input into the wavelet domain,
18490 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
18491 the obtained coefficients. It does an inverse wavelet transform after.
18492 Due to wavelet properties, it should give a nice smoothed result, and
18493 reduced noise, without blurring picture features.
18495 This filter accepts the following options:
18499 The filtering strength. The higher, the more filtered the video will be.
18500 Hard thresholding can use a higher threshold than soft thresholding
18501 before the video looks overfiltered. Default value is 2.
18504 The filtering method the filter will use.
18506 It accepts the following values:
18509 All values under the threshold will be zeroed.
18512 All values under the threshold will be zeroed. All values above will be
18513 reduced by the threshold.
18516 Scales or nullifies coefficients - intermediary between (more) soft and
18517 (less) hard thresholding.
18520 Default is garrote.
18523 Number of times, the wavelet will decompose the picture. Picture can't
18524 be decomposed beyond a particular point (typically, 8 for a 640x480
18525 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
18528 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
18531 A list of the planes to process. By default all planes are processed.
18534 @section vectorscope
18536 Display 2 color component values in the two dimensional graph (which is called
18539 This filter accepts the following options:
18543 Set vectorscope mode.
18545 It accepts the following values:
18548 Gray values are displayed on graph, higher brightness means more pixels have
18549 same component color value on location in graph. This is the default mode.
18552 Gray values are displayed on graph. Surrounding pixels values which are not
18553 present in video frame are drawn in gradient of 2 color components which are
18554 set by option @code{x} and @code{y}. The 3rd color component is static.
18557 Actual color components values present in video frame are displayed on graph.
18560 Similar as color2 but higher frequency of same values @code{x} and @code{y}
18561 on graph increases value of another color component, which is luminance by
18562 default values of @code{x} and @code{y}.
18565 Actual colors present in video frame are displayed on graph. If two different
18566 colors map to same position on graph then color with higher value of component
18567 not present in graph is picked.
18570 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
18571 component picked from radial gradient.
18575 Set which color component will be represented on X-axis. Default is @code{1}.
18578 Set which color component will be represented on Y-axis. Default is @code{2}.
18581 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
18582 of color component which represents frequency of (X, Y) location in graph.
18587 No envelope, this is default.
18590 Instant envelope, even darkest single pixel will be clearly highlighted.
18593 Hold maximum and minimum values presented in graph over time. This way you
18594 can still spot out of range values without constantly looking at vectorscope.
18597 Peak and instant envelope combined together.
18601 Set what kind of graticule to draw.
18609 Set graticule opacity.
18612 Set graticule flags.
18616 Draw graticule for white point.
18619 Draw graticule for black point.
18622 Draw color points short names.
18626 Set background opacity.
18628 @item lthreshold, l
18629 Set low threshold for color component not represented on X or Y axis.
18630 Values lower than this value will be ignored. Default is 0.
18631 Note this value is multiplied with actual max possible value one pixel component
18632 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
18635 @item hthreshold, h
18636 Set high threshold for color component not represented on X or Y axis.
18637 Values higher than this value will be ignored. Default is 1.
18638 Note this value is multiplied with actual max possible value one pixel component
18639 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
18640 is 0.9 * 255 = 230.
18642 @item colorspace, c
18643 Set what kind of colorspace to use when drawing graticule.
18652 @anchor{vidstabdetect}
18653 @section vidstabdetect
18655 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
18656 @ref{vidstabtransform} for pass 2.
18658 This filter generates a file with relative translation and rotation
18659 transform information about subsequent frames, which is then used by
18660 the @ref{vidstabtransform} filter.
18662 To enable compilation of this filter you need to configure FFmpeg with
18663 @code{--enable-libvidstab}.
18665 This filter accepts the following options:
18669 Set the path to the file used to write the transforms information.
18670 Default value is @file{transforms.trf}.
18673 Set how shaky the video is and how quick the camera is. It accepts an
18674 integer in the range 1-10, a value of 1 means little shakiness, a
18675 value of 10 means strong shakiness. Default value is 5.
18678 Set the accuracy of the detection process. It must be a value in the
18679 range 1-15. A value of 1 means low accuracy, a value of 15 means high
18680 accuracy. Default value is 15.
18683 Set stepsize of the search process. The region around minimum is
18684 scanned with 1 pixel resolution. Default value is 6.
18687 Set minimum contrast. Below this value a local measurement field is
18688 discarded. Must be a floating point value in the range 0-1. Default
18692 Set reference frame number for tripod mode.
18694 If enabled, the motion of the frames is compared to a reference frame
18695 in the filtered stream, identified by the specified number. The idea
18696 is to compensate all movements in a more-or-less static scene and keep
18697 the camera view absolutely still.
18699 If set to 0, it is disabled. The frames are counted starting from 1.
18702 Show fields and transforms in the resulting frames. It accepts an
18703 integer in the range 0-2. Default value is 0, which disables any
18707 @subsection Examples
18711 Use default values:
18717 Analyze strongly shaky movie and put the results in file
18718 @file{mytransforms.trf}:
18720 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
18724 Visualize the result of internal transformations in the resulting
18727 vidstabdetect=show=1
18731 Analyze a video with medium shakiness using @command{ffmpeg}:
18733 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
18737 @anchor{vidstabtransform}
18738 @section vidstabtransform
18740 Video stabilization/deshaking: pass 2 of 2,
18741 see @ref{vidstabdetect} for pass 1.
18743 Read a file with transform information for each frame and
18744 apply/compensate them. Together with the @ref{vidstabdetect}
18745 filter this can be used to deshake videos. See also
18746 @url{http://public.hronopik.de/vid.stab}. It is important to also use
18747 the @ref{unsharp} filter, see below.
18749 To enable compilation of this filter you need to configure FFmpeg with
18750 @code{--enable-libvidstab}.
18752 @subsection Options
18756 Set path to the file used to read the transforms. Default value is
18757 @file{transforms.trf}.
18760 Set the number of frames (value*2 + 1) used for lowpass filtering the
18761 camera movements. Default value is 10.
18763 For example a number of 10 means that 21 frames are used (10 in the
18764 past and 10 in the future) to smoothen the motion in the video. A
18765 larger value leads to a smoother video, but limits the acceleration of
18766 the camera (pan/tilt movements). 0 is a special case where a static
18767 camera is simulated.
18770 Set the camera path optimization algorithm.
18772 Accepted values are:
18775 gaussian kernel low-pass filter on camera motion (default)
18777 averaging on transformations
18781 Set maximal number of pixels to translate frames. Default value is -1,
18785 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
18786 value is -1, meaning no limit.
18789 Specify how to deal with borders that may be visible due to movement
18792 Available values are:
18795 keep image information from previous frame (default)
18797 fill the border black
18801 Invert transforms if set to 1. Default value is 0.
18804 Consider transforms as relative to previous frame if set to 1,
18805 absolute if set to 0. Default value is 0.
18808 Set percentage to zoom. A positive value will result in a zoom-in
18809 effect, a negative value in a zoom-out effect. Default value is 0 (no
18813 Set optimal zooming to avoid borders.
18815 Accepted values are:
18820 optimal static zoom value is determined (only very strong movements
18821 will lead to visible borders) (default)
18823 optimal adaptive zoom value is determined (no borders will be
18824 visible), see @option{zoomspeed}
18827 Note that the value given at zoom is added to the one calculated here.
18830 Set percent to zoom maximally each frame (enabled when
18831 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
18835 Specify type of interpolation.
18837 Available values are:
18842 linear only horizontal
18844 linear in both directions (default)
18846 cubic in both directions (slow)
18850 Enable virtual tripod mode if set to 1, which is equivalent to
18851 @code{relative=0:smoothing=0}. Default value is 0.
18853 Use also @code{tripod} option of @ref{vidstabdetect}.
18856 Increase log verbosity if set to 1. Also the detected global motions
18857 are written to the temporary file @file{global_motions.trf}. Default
18861 @subsection Examples
18865 Use @command{ffmpeg} for a typical stabilization with default values:
18867 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
18870 Note the use of the @ref{unsharp} filter which is always recommended.
18873 Zoom in a bit more and load transform data from a given file:
18875 vidstabtransform=zoom=5:input="mytransforms.trf"
18879 Smoothen the video even more:
18881 vidstabtransform=smoothing=30
18887 Flip the input video vertically.
18889 For example, to vertically flip a video with @command{ffmpeg}:
18891 ffmpeg -i in.avi -vf "vflip" out.avi
18896 Detect variable frame rate video.
18898 This filter tries to detect if the input is variable or constant frame rate.
18900 At end it will output number of frames detected as having variable delta pts,
18901 and ones with constant delta pts.
18902 If there was frames with variable delta, than it will also show min and max delta
18907 Boost or alter saturation.
18909 The filter accepts the following options:
18912 Set strength of boost if positive value or strength of alter if negative value.
18913 Default is 0. Allowed range is from -2 to 2.
18916 Set the red balance. Default is 1. Allowed range is from -10 to 10.
18919 Set the green balance. Default is 1. Allowed range is from -10 to 10.
18922 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
18925 Set the red luma coefficient.
18928 Set the green luma coefficient.
18931 Set the blue luma coefficient.
18934 If @code{intensity} is negative and this is set to 1, colors will change,
18935 otherwise colors will be less saturated, more towards gray.
18941 Make or reverse a natural vignetting effect.
18943 The filter accepts the following options:
18947 Set lens angle expression as a number of radians.
18949 The value is clipped in the @code{[0,PI/2]} range.
18951 Default value: @code{"PI/5"}
18955 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
18959 Set forward/backward mode.
18961 Available modes are:
18964 The larger the distance from the central point, the darker the image becomes.
18967 The larger the distance from the central point, the brighter the image becomes.
18968 This can be used to reverse a vignette effect, though there is no automatic
18969 detection to extract the lens @option{angle} and other settings (yet). It can
18970 also be used to create a burning effect.
18973 Default value is @samp{forward}.
18976 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
18978 It accepts the following values:
18981 Evaluate expressions only once during the filter initialization.
18984 Evaluate expressions for each incoming frame. This is way slower than the
18985 @samp{init} mode since it requires all the scalers to be re-computed, but it
18986 allows advanced dynamic expressions.
18989 Default value is @samp{init}.
18992 Set dithering to reduce the circular banding effects. Default is @code{1}
18996 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
18997 Setting this value to the SAR of the input will make a rectangular vignetting
18998 following the dimensions of the video.
19000 Default is @code{1/1}.
19003 @subsection Expressions
19005 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
19006 following parameters.
19011 input width and height
19014 the number of input frame, starting from 0
19017 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
19018 @var{TB} units, NAN if undefined
19021 frame rate of the input video, NAN if the input frame rate is unknown
19024 the PTS (Presentation TimeStamp) of the filtered video frame,
19025 expressed in seconds, NAN if undefined
19028 time base of the input video
19032 @subsection Examples
19036 Apply simple strong vignetting effect:
19042 Make a flickering vignetting:
19044 vignette='PI/4+random(1)*PI/50':eval=frame
19049 @section vmafmotion
19051 Obtain the average vmaf motion score of a video.
19052 It is one of the component filters of VMAF.
19054 The obtained average motion score is printed through the logging system.
19056 In the below example the input file @file{ref.mpg} is being processed and score
19060 ffmpeg -i ref.mpg -lavfi vmafmotion -f null -
19064 Stack input videos vertically.
19066 All streams must be of same pixel format and of same width.
19068 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
19069 to create same output.
19071 The filter accepts the following options:
19075 Set number of input streams. Default is 2.
19078 If set to 1, force the output to terminate when the shortest input
19079 terminates. Default value is 0.
19084 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
19085 Deinterlacing Filter").
19087 Based on the process described by Martin Weston for BBC R&D, and
19088 implemented based on the de-interlace algorithm written by Jim
19089 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
19090 uses filter coefficients calculated by BBC R&D.
19092 This filter uses field-dominance information in frame to decide which
19093 of each pair of fields to place first in the output.
19094 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
19096 There are two sets of filter coefficients, so called "simple"
19097 and "complex". Which set of filter coefficients is used can
19098 be set by passing an optional parameter:
19102 Set the interlacing filter coefficients. Accepts one of the following values:
19106 Simple filter coefficient set.
19108 More-complex filter coefficient set.
19110 Default value is @samp{complex}.
19113 Specify which frames to deinterlace. Accepts one of the following values:
19117 Deinterlace all frames,
19119 Only deinterlace frames marked as interlaced.
19122 Default value is @samp{all}.
19126 Video waveform monitor.
19128 The waveform monitor plots color component intensity. By default luminance
19129 only. Each column of the waveform corresponds to a column of pixels in the
19132 It accepts the following options:
19136 Can be either @code{row}, or @code{column}. Default is @code{column}.
19137 In row mode, the graph on the left side represents color component value 0 and
19138 the right side represents value = 255. In column mode, the top side represents
19139 color component value = 0 and bottom side represents value = 255.
19142 Set intensity. Smaller values are useful to find out how many values of the same
19143 luminance are distributed across input rows/columns.
19144 Default value is @code{0.04}. Allowed range is [0, 1].
19147 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
19148 In mirrored mode, higher values will be represented on the left
19149 side for @code{row} mode and at the top for @code{column} mode. Default is
19150 @code{1} (mirrored).
19154 It accepts the following values:
19157 Presents information identical to that in the @code{parade}, except
19158 that the graphs representing color components are superimposed directly
19161 This display mode makes it easier to spot relative differences or similarities
19162 in overlapping areas of the color components that are supposed to be identical,
19163 such as neutral whites, grays, or blacks.
19166 Display separate graph for the color components side by side in
19167 @code{row} mode or one below the other in @code{column} mode.
19170 Display separate graph for the color components side by side in
19171 @code{column} mode or one below the other in @code{row} mode.
19173 Using this display mode makes it easy to spot color casts in the highlights
19174 and shadows of an image, by comparing the contours of the top and the bottom
19175 graphs of each waveform. Since whites, grays, and blacks are characterized
19176 by exactly equal amounts of red, green, and blue, neutral areas of the picture
19177 should display three waveforms of roughly equal width/height. If not, the
19178 correction is easy to perform by making level adjustments the three waveforms.
19180 Default is @code{stack}.
19182 @item components, c
19183 Set which color components to display. Default is 1, which means only luminance
19184 or red color component if input is in RGB colorspace. If is set for example to
19185 7 it will display all 3 (if) available color components.
19190 No envelope, this is default.
19193 Instant envelope, minimum and maximum values presented in graph will be easily
19194 visible even with small @code{step} value.
19197 Hold minimum and maximum values presented in graph across time. This way you
19198 can still spot out of range values without constantly looking at waveforms.
19201 Peak and instant envelope combined together.
19207 No filtering, this is default.
19210 Luma and chroma combined together.
19213 Similar as above, but shows difference between blue and red chroma.
19216 Similar as above, but use different colors.
19219 Similar as above, but again with different colors.
19222 Displays only chroma.
19225 Displays actual color value on waveform.
19228 Similar as above, but with luma showing frequency of chroma values.
19232 Set which graticule to display.
19236 Do not display graticule.
19239 Display green graticule showing legal broadcast ranges.
19242 Display orange graticule showing legal broadcast ranges.
19245 Display invert graticule showing legal broadcast ranges.
19249 Set graticule opacity.
19252 Set graticule flags.
19256 Draw numbers above lines. By default enabled.
19259 Draw dots instead of lines.
19263 Set scale used for displaying graticule.
19270 Default is digital.
19273 Set background opacity.
19276 @section weave, doubleweave
19278 The @code{weave} takes a field-based video input and join
19279 each two sequential fields into single frame, producing a new double
19280 height clip with half the frame rate and half the frame count.
19282 The @code{doubleweave} works same as @code{weave} but without
19283 halving frame rate and frame count.
19285 It accepts the following option:
19289 Set first field. Available values are:
19293 Set the frame as top-field-first.
19296 Set the frame as bottom-field-first.
19300 @subsection Examples
19304 Interlace video using @ref{select} and @ref{separatefields} filter:
19306 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
19311 Apply the xBR high-quality magnification filter which is designed for pixel
19312 art. It follows a set of edge-detection rules, see
19313 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
19315 It accepts the following option:
19319 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
19320 @code{3xBR} and @code{4} for @code{4xBR}.
19321 Default is @code{3}.
19325 Pick median pixels from several input videos.
19327 The filter accepts the following options:
19331 Set number of inputs.
19332 Default is 3. Allowed range is from 3 to 255.
19333 If number of inputs is even number, than result will be mean value between two median values.
19336 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
19340 Stack video inputs into custom layout.
19342 All streams must be of same pixel format.
19344 The filter accepts the following options:
19348 Set number of input streams. Default is 2.
19351 Specify layout of inputs.
19352 This option requires the desired layout configuration to be explicitly set by the user.
19353 This sets position of each video input in output. Each input
19354 is separated by '|'.
19355 The first number represents the column, and the second number represents the row.
19356 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
19357 where X is video input from which to take width or height.
19358 Multiple values can be used when separated by '+'. In such
19359 case values are summed together.
19361 Note that if inputs are of different sizes gaps may appear, as not all of
19362 the output video frame will be filled. Similarly, videos can overlap each
19363 other if their position doesn't leave enough space for the full frame of
19366 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
19367 a layout must be set by the user.
19370 If set to 1, force the output to terminate when the shortest input
19371 terminates. Default value is 0.
19374 @subsection Examples
19378 Display 4 inputs into 2x2 grid.
19382 input1(0, 0) | input3(w0, 0)
19383 input2(0, h0) | input4(w0, h0)
19387 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
19390 Note that if inputs are of different sizes, gaps or overlaps may occur.
19393 Display 4 inputs into 1x4 grid.
19400 input4(0, h0+h1+h2)
19404 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
19407 Note that if inputs are of different widths, unused space will appear.
19410 Display 9 inputs into 3x3 grid.
19414 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
19415 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
19416 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
19420 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
19423 Note that if inputs are of different sizes, gaps or overlaps may occur.
19426 Display 16 inputs into 4x4 grid.
19430 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
19431 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
19432 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
19433 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
19437 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|
19438 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
19441 Note that if inputs are of different sizes, gaps or overlaps may occur.
19448 Deinterlace the input video ("yadif" means "yet another deinterlacing
19451 It accepts the following parameters:
19457 The interlacing mode to adopt. It accepts one of the following values:
19460 @item 0, send_frame
19461 Output one frame for each frame.
19462 @item 1, send_field
19463 Output one frame for each field.
19464 @item 2, send_frame_nospatial
19465 Like @code{send_frame}, but it skips the spatial interlacing check.
19466 @item 3, send_field_nospatial
19467 Like @code{send_field}, but it skips the spatial interlacing check.
19470 The default value is @code{send_frame}.
19473 The picture field parity assumed for the input interlaced video. It accepts one
19474 of the following values:
19478 Assume the top field is first.
19480 Assume the bottom field is first.
19482 Enable automatic detection of field parity.
19485 The default value is @code{auto}.
19486 If the interlacing is unknown or the decoder does not export this information,
19487 top field first will be assumed.
19490 Specify which frames to deinterlace. Accepts one of the following
19495 Deinterlace all frames.
19496 @item 1, interlaced
19497 Only deinterlace frames marked as interlaced.
19500 The default value is @code{all}.
19503 @section yadif_cuda
19505 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
19506 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
19509 It accepts the following parameters:
19515 The interlacing mode to adopt. It accepts one of the following values:
19518 @item 0, send_frame
19519 Output one frame for each frame.
19520 @item 1, send_field
19521 Output one frame for each field.
19522 @item 2, send_frame_nospatial
19523 Like @code{send_frame}, but it skips the spatial interlacing check.
19524 @item 3, send_field_nospatial
19525 Like @code{send_field}, but it skips the spatial interlacing check.
19528 The default value is @code{send_frame}.
19531 The picture field parity assumed for the input interlaced video. It accepts one
19532 of the following values:
19536 Assume the top field is first.
19538 Assume the bottom field is first.
19540 Enable automatic detection of field parity.
19543 The default value is @code{auto}.
19544 If the interlacing is unknown or the decoder does not export this information,
19545 top field first will be assumed.
19548 Specify which frames to deinterlace. Accepts one of the following
19553 Deinterlace all frames.
19554 @item 1, interlaced
19555 Only deinterlace frames marked as interlaced.
19558 The default value is @code{all}.
19563 Apply Zoom & Pan effect.
19565 This filter accepts the following options:
19569 Set the zoom expression. Range is 1-10. Default is 1.
19573 Set the x and y expression. Default is 0.
19576 Set the duration expression in number of frames.
19577 This sets for how many number of frames effect will last for
19578 single input image.
19581 Set the output image size, default is 'hd720'.
19584 Set the output frame rate, default is '25'.
19587 Each expression can contain the following constants:
19606 Output frame count.
19610 Last calculated 'x' and 'y' position from 'x' and 'y' expression
19611 for current input frame.
19615 'x' and 'y' of last output frame of previous input frame or 0 when there was
19616 not yet such frame (first input frame).
19619 Last calculated zoom from 'z' expression for current input frame.
19622 Last calculated zoom of last output frame of previous input frame.
19625 Number of output frames for current input frame. Calculated from 'd' expression
19626 for each input frame.
19629 number of output frames created for previous input frame
19632 Rational number: input width / input height
19635 sample aspect ratio
19638 display aspect ratio
19642 @subsection Examples
19646 Zoom-in up to 1.5 and pan at same time to some spot near center of picture:
19648 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
19652 Zoom-in up to 1.5 and pan always at center of picture:
19654 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
19658 Same as above but without pausing:
19660 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
19666 Scale (resize) the input video, using the z.lib library:
19667 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
19668 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
19670 The zscale filter forces the output display aspect ratio to be the same
19671 as the input, by changing the output sample aspect ratio.
19673 If the input image format is different from the format requested by
19674 the next filter, the zscale filter will convert the input to the
19677 @subsection Options
19678 The filter accepts the following options.
19683 Set the output video dimension expression. Default value is the input
19686 If the @var{width} or @var{w} value is 0, the input width is used for
19687 the output. If the @var{height} or @var{h} value is 0, the input height
19688 is used for the output.
19690 If one and only one of the values is -n with n >= 1, the zscale filter
19691 will use a value that maintains the aspect ratio of the input image,
19692 calculated from the other specified dimension. After that it will,
19693 however, make sure that the calculated dimension is divisible by n and
19694 adjust the value if necessary.
19696 If both values are -n with n >= 1, the behavior will be identical to
19697 both values being set to 0 as previously detailed.
19699 See below for the list of accepted constants for use in the dimension
19703 Set the video size. For the syntax of this option, check the
19704 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19707 Set the dither type.
19709 Possible values are:
19714 @item error_diffusion
19720 Set the resize filter type.
19722 Possible values are:
19732 Default is bilinear.
19735 Set the color range.
19737 Possible values are:
19744 Default is same as input.
19747 Set the color primaries.
19749 Possible values are:
19759 Default is same as input.
19762 Set the transfer characteristics.
19764 Possible values are:
19778 Default is same as input.
19781 Set the colorspace matrix.
19783 Possible value are:
19794 Default is same as input.
19797 Set the input color range.
19799 Possible values are:
19806 Default is same as input.
19808 @item primariesin, pin
19809 Set the input color primaries.
19811 Possible values are:
19821 Default is same as input.
19823 @item transferin, tin
19824 Set the input transfer characteristics.
19826 Possible values are:
19837 Default is same as input.
19839 @item matrixin, min
19840 Set the input colorspace matrix.
19842 Possible value are:
19854 Set the output chroma location.
19856 Possible values are:
19867 @item chromalin, cin
19868 Set the input chroma location.
19870 Possible values are:
19882 Set the nominal peak luminance.
19885 The values of the @option{w} and @option{h} options are expressions
19886 containing the following constants:
19891 The input width and height
19895 These are the same as @var{in_w} and @var{in_h}.
19899 The output (scaled) width and height
19903 These are the same as @var{out_w} and @var{out_h}
19906 The same as @var{iw} / @var{ih}
19909 input sample aspect ratio
19912 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
19916 horizontal and vertical input chroma subsample values. For example for the
19917 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
19921 horizontal and vertical output chroma subsample values. For example for the
19922 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
19928 @c man end VIDEO FILTERS
19930 @chapter OpenCL Video Filters
19931 @c man begin OPENCL VIDEO FILTERS
19933 Below is a description of the currently available OpenCL video filters.
19935 To enable compilation of these filters you need to configure FFmpeg with
19936 @code{--enable-opencl}.
19938 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
19941 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
19942 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
19943 given device parameters.
19945 @item -filter_hw_device @var{name}
19946 Pass the hardware device called @var{name} to all filters in any filter graph.
19950 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
19954 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
19956 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
19960 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.
19962 @section avgblur_opencl
19964 Apply average blur filter.
19966 The filter accepts the following options:
19970 Set horizontal radius size.
19971 Range is @code{[1, 1024]} and default value is @code{1}.
19974 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
19977 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
19980 @subsection Example
19984 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.
19986 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
19990 @section boxblur_opencl
19992 Apply a boxblur algorithm to the input video.
19994 It accepts the following parameters:
19998 @item luma_radius, lr
19999 @item luma_power, lp
20000 @item chroma_radius, cr
20001 @item chroma_power, cp
20002 @item alpha_radius, ar
20003 @item alpha_power, ap
20007 A description of the accepted options follows.
20010 @item luma_radius, lr
20011 @item chroma_radius, cr
20012 @item alpha_radius, ar
20013 Set an expression for the box radius in pixels used for blurring the
20014 corresponding input plane.
20016 The radius value must be a non-negative number, and must not be
20017 greater than the value of the expression @code{min(w,h)/2} for the
20018 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
20021 Default value for @option{luma_radius} is "2". If not specified,
20022 @option{chroma_radius} and @option{alpha_radius} default to the
20023 corresponding value set for @option{luma_radius}.
20025 The expressions can contain the following constants:
20029 The input width and height in pixels.
20033 The input chroma image width and height in pixels.
20037 The horizontal and vertical chroma subsample values. For example, for the
20038 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
20041 @item luma_power, lp
20042 @item chroma_power, cp
20043 @item alpha_power, ap
20044 Specify how many times the boxblur filter is applied to the
20045 corresponding plane.
20047 Default value for @option{luma_power} is 2. If not specified,
20048 @option{chroma_power} and @option{alpha_power} default to the
20049 corresponding value set for @option{luma_power}.
20051 A value of 0 will disable the effect.
20054 @subsection Examples
20056 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.
20060 Apply a boxblur filter with the luma, chroma, and alpha radius
20061 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.
20063 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
20064 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
20068 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.
20070 For the luma plane, a 2x2 box radius will be run once.
20072 For the chroma plane, a 4x4 box radius will be run 5 times.
20074 For the alpha plane, a 3x3 box radius will be run 7 times.
20076 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
20080 @section convolution_opencl
20082 Apply convolution of 3x3, 5x5, 7x7 matrix.
20084 The filter accepts the following options:
20091 Set matrix for each plane.
20092 Matrix is sequence of 9, 25 or 49 signed numbers.
20093 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
20099 Set multiplier for calculated value for each plane.
20100 If unset or 0, it will be sum of all matrix elements.
20101 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
20107 Set bias for each plane. This value is added to the result of the multiplication.
20108 Useful for making the overall image brighter or darker.
20109 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
20113 @subsection Examples
20119 -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
20125 -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
20129 Apply edge enhance:
20131 -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
20137 -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
20141 Apply laplacian edge detector which includes diagonals:
20143 -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
20149 -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
20153 @section dilation_opencl
20155 Apply dilation effect to the video.
20157 This filter replaces the pixel by the local(3x3) maximum.
20159 It accepts the following options:
20166 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
20167 If @code{0}, plane will remain unchanged.
20170 Flag which specifies the pixel to refer to.
20171 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
20173 Flags to local 3x3 coordinates region centered on @code{x}:
20182 @subsection Example
20186 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.
20188 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
20192 @section erosion_opencl
20194 Apply erosion effect to the video.
20196 This filter replaces the pixel by the local(3x3) minimum.
20198 It accepts the following options:
20205 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
20206 If @code{0}, plane will remain unchanged.
20209 Flag which specifies the pixel to refer to.
20210 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
20212 Flags to local 3x3 coordinates region centered on @code{x}:
20221 @subsection Example
20225 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.
20227 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
20231 @section colorkey_opencl
20232 RGB colorspace color keying.
20234 The filter accepts the following options:
20238 The color which will be replaced with transparency.
20241 Similarity percentage with the key color.
20243 0.01 matches only the exact key color, while 1.0 matches everything.
20248 0.0 makes pixels either fully transparent, or not transparent at all.
20250 Higher values result in semi-transparent pixels, with a higher transparency
20251 the more similar the pixels color is to the key color.
20254 @subsection Examples
20258 Make every semi-green pixel in the input transparent with some slight blending:
20260 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
20264 @section deshake_opencl
20265 Feature-point based video stabilization filter.
20267 The filter accepts the following options:
20271 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
20274 Whether or not additional debug info should be displayed, both in the processed output and in the console.
20276 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
20278 Viewing point matches in the output video is only supported for RGB input.
20280 Defaults to @code{0}.
20282 @item adaptive_crop
20283 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
20285 Defaults to @code{1}.
20287 @item refine_features
20288 Whether or not feature points should be refined at a sub-pixel level.
20290 This can be turned off for a slight performance gain at the cost of precision.
20292 Defaults to @code{1}.
20294 @item smooth_strength
20295 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
20297 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
20299 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
20301 Defaults to @code{0.0}.
20303 @item smooth_window_multiplier
20304 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
20306 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
20308 Acceptable values range from @code{0.1} to @code{10.0}.
20310 Larger values increase the amount of motion data available for determining how to smooth the camera path,
20311 potentially improving smoothness, but also increase latency and memory usage.
20313 Defaults to @code{2.0}.
20317 @subsection Examples
20321 Stabilize a video with a fixed, medium smoothing strength:
20323 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
20327 Stabilize a video with debugging (both in console and in rendered video):
20329 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
20333 @section nlmeans_opencl
20335 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
20337 @section overlay_opencl
20339 Overlay one video on top of another.
20341 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
20342 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
20344 The filter accepts the following options:
20349 Set the x coordinate of the overlaid video on the main video.
20350 Default value is @code{0}.
20353 Set the x coordinate of the overlaid video on the main video.
20354 Default value is @code{0}.
20358 @subsection Examples
20362 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
20364 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
20367 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
20369 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
20374 @section prewitt_opencl
20376 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
20378 The filter accepts the following option:
20382 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
20385 Set value which will be multiplied with filtered result.
20386 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
20389 Set value which will be added to filtered result.
20390 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
20393 @subsection Example
20397 Apply the Prewitt operator with scale set to 2 and delta set to 10.
20399 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
20403 @section roberts_opencl
20404 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
20406 The filter accepts the following option:
20410 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
20413 Set value which will be multiplied with filtered result.
20414 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
20417 Set value which will be added to filtered result.
20418 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
20421 @subsection Example
20425 Apply the Roberts cross operator with scale set to 2 and delta set to 10
20427 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
20431 @section sobel_opencl
20433 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
20435 The filter accepts the following option:
20439 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
20442 Set value which will be multiplied with filtered result.
20443 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
20446 Set value which will be added to filtered result.
20447 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
20450 @subsection Example
20454 Apply sobel operator with scale set to 2 and delta set to 10
20456 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
20460 @section tonemap_opencl
20462 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
20464 It accepts the following parameters:
20468 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
20471 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
20474 Apply desaturation for highlights that exceed this level of brightness. The
20475 higher the parameter, the more color information will be preserved. This
20476 setting helps prevent unnaturally blown-out colors for super-highlights, by
20477 (smoothly) turning into white instead. This makes images feel more natural,
20478 at the cost of reducing information about out-of-range colors.
20480 The default value is 0.5, and the algorithm here is a little different from
20481 the cpu version tonemap currently. A setting of 0.0 disables this option.
20484 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
20485 is used to detect whether the scene has changed or not. If the distance between
20486 the current frame average brightness and the current running average exceeds
20487 a threshold value, we would re-calculate scene average and peak brightness.
20488 The default value is 0.2.
20491 Specify the output pixel format.
20493 Currently supported formats are:
20500 Set the output color range.
20502 Possible values are:
20508 Default is same as input.
20511 Set the output color primaries.
20513 Possible values are:
20519 Default is same as input.
20522 Set the output transfer characteristics.
20524 Possible values are:
20533 Set the output colorspace matrix.
20535 Possible value are:
20541 Default is same as input.
20545 @subsection Example
20549 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
20551 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
20555 @section unsharp_opencl
20557 Sharpen or blur the input video.
20559 It accepts the following parameters:
20562 @item luma_msize_x, lx
20563 Set the luma matrix horizontal size.
20564 Range is @code{[1, 23]} and default value is @code{5}.
20566 @item luma_msize_y, ly
20567 Set the luma matrix vertical size.
20568 Range is @code{[1, 23]} and default value is @code{5}.
20570 @item luma_amount, la
20571 Set the luma effect strength.
20572 Range is @code{[-10, 10]} and default value is @code{1.0}.
20574 Negative values will blur the input video, while positive values will
20575 sharpen it, a value of zero will disable the effect.
20577 @item chroma_msize_x, cx
20578 Set the chroma matrix horizontal size.
20579 Range is @code{[1, 23]} and default value is @code{5}.
20581 @item chroma_msize_y, cy
20582 Set the chroma matrix vertical size.
20583 Range is @code{[1, 23]} and default value is @code{5}.
20585 @item chroma_amount, ca
20586 Set the chroma effect strength.
20587 Range is @code{[-10, 10]} and default value is @code{0.0}.
20589 Negative values will blur the input video, while positive values will
20590 sharpen it, a value of zero will disable the effect.
20594 All parameters are optional and default to the equivalent of the
20595 string '5:5:1.0:5:5:0.0'.
20597 @subsection Examples
20601 Apply strong luma sharpen effect:
20603 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
20607 Apply a strong blur of both luma and chroma parameters:
20609 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
20613 @c man end OPENCL VIDEO FILTERS
20615 @chapter Video Sources
20616 @c man begin VIDEO SOURCES
20618 Below is a description of the currently available video sources.
20622 Buffer video frames, and make them available to the filter chain.
20624 This source is mainly intended for a programmatic use, in particular
20625 through the interface defined in @file{libavfilter/vsrc_buffer.h}.
20627 It accepts the following parameters:
20632 Specify the size (width and height) of the buffered video frames. For the
20633 syntax of this option, check the
20634 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20637 The input video width.
20640 The input video height.
20643 A string representing the pixel format of the buffered video frames.
20644 It may be a number corresponding to a pixel format, or a pixel format
20648 Specify the timebase assumed by the timestamps of the buffered frames.
20651 Specify the frame rate expected for the video stream.
20653 @item pixel_aspect, sar
20654 The sample (pixel) aspect ratio of the input video.
20657 Specify the optional parameters to be used for the scale filter which
20658 is automatically inserted when an input change is detected in the
20659 input size or format.
20661 @item hw_frames_ctx
20662 When using a hardware pixel format, this should be a reference to an
20663 AVHWFramesContext describing input frames.
20668 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
20671 will instruct the source to accept video frames with size 320x240 and
20672 with format "yuv410p", assuming 1/24 as the timestamps timebase and
20673 square pixels (1:1 sample aspect ratio).
20674 Since the pixel format with name "yuv410p" corresponds to the number 6
20675 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
20676 this example corresponds to:
20678 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
20681 Alternatively, the options can be specified as a flat string, but this
20682 syntax is deprecated:
20684 @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}]
20688 Create a pattern generated by an elementary cellular automaton.
20690 The initial state of the cellular automaton can be defined through the
20691 @option{filename} and @option{pattern} options. If such options are
20692 not specified an initial state is created randomly.
20694 At each new frame a new row in the video is filled with the result of
20695 the cellular automaton next generation. The behavior when the whole
20696 frame is filled is defined by the @option{scroll} option.
20698 This source accepts the following options:
20702 Read the initial cellular automaton state, i.e. the starting row, from
20703 the specified file.
20704 In the file, each non-whitespace character is considered an alive
20705 cell, a newline will terminate the row, and further characters in the
20706 file will be ignored.
20709 Read the initial cellular automaton state, i.e. the starting row, from
20710 the specified string.
20712 Each non-whitespace character in the string is considered an alive
20713 cell, a newline will terminate the row, and further characters in the
20714 string will be ignored.
20717 Set the video rate, that is the number of frames generated per second.
20720 @item random_fill_ratio, ratio
20721 Set the random fill ratio for the initial cellular automaton row. It
20722 is a floating point number value ranging from 0 to 1, defaults to
20725 This option is ignored when a file or a pattern is specified.
20727 @item random_seed, seed
20728 Set the seed for filling randomly the initial row, must be an integer
20729 included between 0 and UINT32_MAX. If not specified, or if explicitly
20730 set to -1, the filter will try to use a good random seed on a best
20734 Set the cellular automaton rule, it is a number ranging from 0 to 255.
20735 Default value is 110.
20738 Set the size of the output video. For the syntax of this option, check the
20739 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20741 If @option{filename} or @option{pattern} is specified, the size is set
20742 by default to the width of the specified initial state row, and the
20743 height is set to @var{width} * PHI.
20745 If @option{size} is set, it must contain the width of the specified
20746 pattern string, and the specified pattern will be centered in the
20749 If a filename or a pattern string is not specified, the size value
20750 defaults to "320x518" (used for a randomly generated initial state).
20753 If set to 1, scroll the output upward when all the rows in the output
20754 have been already filled. If set to 0, the new generated row will be
20755 written over the top row just after the bottom row is filled.
20758 @item start_full, full
20759 If set to 1, completely fill the output with generated rows before
20760 outputting the first frame.
20761 This is the default behavior, for disabling set the value to 0.
20764 If set to 1, stitch the left and right row edges together.
20765 This is the default behavior, for disabling set the value to 0.
20768 @subsection Examples
20772 Read the initial state from @file{pattern}, and specify an output of
20775 cellauto=f=pattern:s=200x400
20779 Generate a random initial row with a width of 200 cells, with a fill
20782 cellauto=ratio=2/3:s=200x200
20786 Create a pattern generated by rule 18 starting by a single alive cell
20787 centered on an initial row with width 100:
20789 cellauto=p=@@:s=100x400:full=0:rule=18
20793 Specify a more elaborated initial pattern:
20795 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
20800 @anchor{coreimagesrc}
20801 @section coreimagesrc
20802 Video source generated on GPU using Apple's CoreImage API on OSX.
20804 This video source is a specialized version of the @ref{coreimage} video filter.
20805 Use a core image generator at the beginning of the applied filterchain to
20806 generate the content.
20808 The coreimagesrc video source accepts the following options:
20810 @item list_generators
20811 List all available generators along with all their respective options as well as
20812 possible minimum and maximum values along with the default values.
20814 list_generators=true
20818 Specify the size of the sourced video. For the syntax of this option, check the
20819 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20820 The default value is @code{320x240}.
20823 Specify the frame rate of the sourced video, as the number of frames
20824 generated per second. It has to be a string in the format
20825 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
20826 number or a valid video frame rate abbreviation. The default value is
20830 Set the sample aspect ratio of the sourced video.
20833 Set the duration of the sourced video. See
20834 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
20835 for the accepted syntax.
20837 If not specified, or the expressed duration is negative, the video is
20838 supposed to be generated forever.
20841 Additionally, all options of the @ref{coreimage} video filter are accepted.
20842 A complete filterchain can be used for further processing of the
20843 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
20844 and examples for details.
20846 @subsection Examples
20851 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
20852 given as complete and escaped command-line for Apple's standard bash shell:
20854 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
20856 This example is equivalent to the QRCode example of @ref{coreimage} without the
20857 need for a nullsrc video source.
20861 @section mandelbrot
20863 Generate a Mandelbrot set fractal, and progressively zoom towards the
20864 point specified with @var{start_x} and @var{start_y}.
20866 This source accepts the following options:
20871 Set the terminal pts value. Default value is 400.
20874 Set the terminal scale value.
20875 Must be a floating point value. Default value is 0.3.
20878 Set the inner coloring mode, that is the algorithm used to draw the
20879 Mandelbrot fractal internal region.
20881 It shall assume one of the following values:
20886 Show time until convergence.
20888 Set color based on point closest to the origin of the iterations.
20893 Default value is @var{mincol}.
20896 Set the bailout value. Default value is 10.0.
20899 Set the maximum of iterations performed by the rendering
20900 algorithm. Default value is 7189.
20903 Set outer coloring mode.
20904 It shall assume one of following values:
20906 @item iteration_count
20907 Set iteration count mode.
20908 @item normalized_iteration_count
20909 set normalized iteration count mode.
20911 Default value is @var{normalized_iteration_count}.
20914 Set frame rate, expressed as number of frames per second. Default
20918 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
20919 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
20922 Set the initial scale value. Default value is 3.0.
20925 Set the initial x position. Must be a floating point value between
20926 -100 and 100. Default value is -0.743643887037158704752191506114774.
20929 Set the initial y position. Must be a floating point value between
20930 -100 and 100. Default value is -0.131825904205311970493132056385139.
20935 Generate various test patterns, as generated by the MPlayer test filter.
20937 The size of the generated video is fixed, and is 256x256.
20938 This source is useful in particular for testing encoding features.
20940 This source accepts the following options:
20945 Specify the frame rate of the sourced video, as the number of frames
20946 generated per second. It has to be a string in the format
20947 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
20948 number or a valid video frame rate abbreviation. The default value is
20952 Set the duration of the sourced video. See
20953 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
20954 for the accepted syntax.
20956 If not specified, or the expressed duration is negative, the video is
20957 supposed to be generated forever.
20961 Set the number or the name of the test to perform. Supported tests are:
20975 @item max_frames, m
20976 Set the maximum number of frames generated for each test, default value is 30.
20980 Default value is "all", which will cycle through the list of all tests.
20985 mptestsrc=t=dc_luma
20988 will generate a "dc_luma" test pattern.
20990 @section frei0r_src
20992 Provide a frei0r source.
20994 To enable compilation of this filter you need to install the frei0r
20995 header and configure FFmpeg with @code{--enable-frei0r}.
20997 This source accepts the following parameters:
21002 The size of the video to generate. For the syntax of this option, check the
21003 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21006 The framerate of the generated video. It may be a string of the form
21007 @var{num}/@var{den} or a frame rate abbreviation.
21010 The name to the frei0r source to load. For more information regarding frei0r and
21011 how to set the parameters, read the @ref{frei0r} section in the video filters
21014 @item filter_params
21015 A '|'-separated list of parameters to pass to the frei0r source.
21019 For example, to generate a frei0r partik0l source with size 200x200
21020 and frame rate 10 which is overlaid on the overlay filter main input:
21022 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
21027 Generate a life pattern.
21029 This source is based on a generalization of John Conway's life game.
21031 The sourced input represents a life grid, each pixel represents a cell
21032 which can be in one of two possible states, alive or dead. Every cell
21033 interacts with its eight neighbours, which are the cells that are
21034 horizontally, vertically, or diagonally adjacent.
21036 At each interaction the grid evolves according to the adopted rule,
21037 which specifies the number of neighbor alive cells which will make a
21038 cell stay alive or born. The @option{rule} option allows one to specify
21041 This source accepts the following options:
21045 Set the file from which to read the initial grid state. In the file,
21046 each non-whitespace character is considered an alive cell, and newline
21047 is used to delimit the end of each row.
21049 If this option is not specified, the initial grid is generated
21053 Set the video rate, that is the number of frames generated per second.
21056 @item random_fill_ratio, ratio
21057 Set the random fill ratio for the initial random grid. It is a
21058 floating point number value ranging from 0 to 1, defaults to 1/PHI.
21059 It is ignored when a file is specified.
21061 @item random_seed, seed
21062 Set the seed for filling the initial random grid, must be an integer
21063 included between 0 and UINT32_MAX. If not specified, or if explicitly
21064 set to -1, the filter will try to use a good random seed on a best
21070 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
21071 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
21072 @var{NS} specifies the number of alive neighbor cells which make a
21073 live cell stay alive, and @var{NB} the number of alive neighbor cells
21074 which make a dead cell to become alive (i.e. to "born").
21075 "s" and "b" can be used in place of "S" and "B", respectively.
21077 Alternatively a rule can be specified by an 18-bits integer. The 9
21078 high order bits are used to encode the next cell state if it is alive
21079 for each number of neighbor alive cells, the low order bits specify
21080 the rule for "borning" new cells. Higher order bits encode for an
21081 higher number of neighbor cells.
21082 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
21083 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
21085 Default value is "S23/B3", which is the original Conway's game of life
21086 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
21087 cells, and will born a new cell if there are three alive cells around
21091 Set the size of the output video. For the syntax of this option, check the
21092 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21094 If @option{filename} is specified, the size is set by default to the
21095 same size of the input file. If @option{size} is set, it must contain
21096 the size specified in the input file, and the initial grid defined in
21097 that file is centered in the larger resulting area.
21099 If a filename is not specified, the size value defaults to "320x240"
21100 (used for a randomly generated initial grid).
21103 If set to 1, stitch the left and right grid edges together, and the
21104 top and bottom edges also. Defaults to 1.
21107 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
21108 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
21109 value from 0 to 255.
21112 Set the color of living (or new born) cells.
21115 Set the color of dead cells. If @option{mold} is set, this is the first color
21116 used to represent a dead cell.
21119 Set mold color, for definitely dead and moldy cells.
21121 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
21122 ffmpeg-utils manual,ffmpeg-utils}.
21125 @subsection Examples
21129 Read a grid from @file{pattern}, and center it on a grid of size
21132 life=f=pattern:s=300x300
21136 Generate a random grid of size 200x200, with a fill ratio of 2/3:
21138 life=ratio=2/3:s=200x200
21142 Specify a custom rule for evolving a randomly generated grid:
21148 Full example with slow death effect (mold) using @command{ffplay}:
21150 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
21157 @anchor{haldclutsrc}
21160 @anchor{pal100bars}
21161 @anchor{rgbtestsrc}
21163 @anchor{smptehdbars}
21166 @anchor{yuvtestsrc}
21167 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
21169 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
21171 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
21173 The @code{color} source provides an uniformly colored input.
21175 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
21176 @ref{haldclut} filter.
21178 The @code{nullsrc} source returns unprocessed video frames. It is
21179 mainly useful to be employed in analysis / debugging tools, or as the
21180 source for filters which ignore the input data.
21182 The @code{pal75bars} source generates a color bars pattern, based on
21183 EBU PAL recommendations with 75% color levels.
21185 The @code{pal100bars} source generates a color bars pattern, based on
21186 EBU PAL recommendations with 100% color levels.
21188 The @code{rgbtestsrc} source generates an RGB test pattern useful for
21189 detecting RGB vs BGR issues. You should see a red, green and blue
21190 stripe from top to bottom.
21192 The @code{smptebars} source generates a color bars pattern, based on
21193 the SMPTE Engineering Guideline EG 1-1990.
21195 The @code{smptehdbars} source generates a color bars pattern, based on
21196 the SMPTE RP 219-2002.
21198 The @code{testsrc} source generates a test video pattern, showing a
21199 color pattern, a scrolling gradient and a timestamp. This is mainly
21200 intended for testing purposes.
21202 The @code{testsrc2} source is similar to testsrc, but supports more
21203 pixel formats instead of just @code{rgb24}. This allows using it as an
21204 input for other tests without requiring a format conversion.
21206 The @code{yuvtestsrc} source generates an YUV test pattern. You should
21207 see a y, cb and cr stripe from top to bottom.
21209 The sources accept the following parameters:
21214 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
21215 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
21216 pixels to be used as identity matrix for 3D lookup tables. Each component is
21217 coded on a @code{1/(N*N)} scale.
21220 Specify the color of the source, only available in the @code{color}
21221 source. For the syntax of this option, check the
21222 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
21225 Specify the size of the sourced video. For the syntax of this option, check the
21226 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21227 The default value is @code{320x240}.
21229 This option is not available with the @code{allrgb}, @code{allyuv}, and
21230 @code{haldclutsrc} filters.
21233 Specify the frame rate of the sourced video, as the number of frames
21234 generated per second. It has to be a string in the format
21235 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
21236 number or a valid video frame rate abbreviation. The default value is
21240 Set the duration of the sourced video. See
21241 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
21242 for the accepted syntax.
21244 If not specified, or the expressed duration is negative, the video is
21245 supposed to be generated forever.
21248 Set the sample aspect ratio of the sourced video.
21251 Specify the alpha (opacity) of the background, only available in the
21252 @code{testsrc2} source. The value must be between 0 (fully transparent) and
21253 255 (fully opaque, the default).
21256 Set the number of decimals to show in the timestamp, only available in the
21257 @code{testsrc} source.
21259 The displayed timestamp value will correspond to the original
21260 timestamp value multiplied by the power of 10 of the specified
21261 value. Default value is 0.
21264 @subsection Examples
21268 Generate a video with a duration of 5.3 seconds, with size
21269 176x144 and a frame rate of 10 frames per second:
21271 testsrc=duration=5.3:size=qcif:rate=10
21275 The following graph description will generate a red source
21276 with an opacity of 0.2, with size "qcif" and a frame rate of 10
21279 color=c=red@@0.2:s=qcif:r=10
21283 If the input content is to be ignored, @code{nullsrc} can be used. The
21284 following command generates noise in the luminance plane by employing
21285 the @code{geq} filter:
21287 nullsrc=s=256x256, geq=random(1)*255:128:128
21291 @subsection Commands
21293 The @code{color} source supports the following commands:
21297 Set the color of the created image. Accepts the same syntax of the
21298 corresponding @option{color} option.
21303 Generate video using an OpenCL program.
21308 OpenCL program source file.
21311 Kernel name in program.
21314 Size of frames to generate. This must be set.
21317 Pixel format to use for the generated frames. This must be set.
21320 Number of frames generated every second. Default value is '25'.
21324 For details of how the program loading works, see the @ref{program_opencl}
21331 Generate a colour ramp by setting pixel values from the position of the pixel
21332 in the output image. (Note that this will work with all pixel formats, but
21333 the generated output will not be the same.)
21335 __kernel void ramp(__write_only image2d_t dst,
21336 unsigned int index)
21338 int2 loc = (int2)(get_global_id(0), get_global_id(1));
21341 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
21343 write_imagef(dst, loc, val);
21348 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
21350 __kernel void sierpinski_carpet(__write_only image2d_t dst,
21351 unsigned int index)
21353 int2 loc = (int2)(get_global_id(0), get_global_id(1));
21355 float4 value = 0.0f;
21356 int x = loc.x + index;
21357 int y = loc.y + index;
21358 while (x > 0 || y > 0) {
21359 if (x % 3 == 1 && y % 3 == 1) {
21367 write_imagef(dst, loc, value);
21373 @section sierpinski
21375 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
21377 This source accepts the following options:
21381 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
21382 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
21385 Set frame rate, expressed as number of frames per second. Default
21389 Set seed which is used for random panning.
21392 Set max jump for single pan destination. Allowed range is from 1 to 10000.
21395 Set fractal type, can be default @code{carpet} or @code{triangle}.
21398 @c man end VIDEO SOURCES
21400 @chapter Video Sinks
21401 @c man begin VIDEO SINKS
21403 Below is a description of the currently available video sinks.
21405 @section buffersink
21407 Buffer video frames, and make them available to the end of the filter
21410 This sink is mainly intended for programmatic use, in particular
21411 through the interface defined in @file{libavfilter/buffersink.h}
21412 or the options system.
21414 It accepts a pointer to an AVBufferSinkContext structure, which
21415 defines the incoming buffers' formats, to be passed as the opaque
21416 parameter to @code{avfilter_init_filter} for initialization.
21420 Null video sink: do absolutely nothing with the input video. It is
21421 mainly useful as a template and for use in analysis / debugging
21424 @c man end VIDEO SINKS
21426 @chapter Multimedia Filters
21427 @c man begin MULTIMEDIA FILTERS
21429 Below is a description of the currently available multimedia filters.
21433 Convert input audio to a video output, displaying the audio bit scope.
21435 The filter accepts the following options:
21439 Set frame rate, expressed as number of frames per second. Default
21443 Specify the video size for the output. For the syntax of this option, check the
21444 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21445 Default value is @code{1024x256}.
21448 Specify list of colors separated by space or by '|' which will be used to
21449 draw channels. Unrecognized or missing colors will be replaced
21453 @section ahistogram
21455 Convert input audio to a video output, displaying the volume histogram.
21457 The filter accepts the following options:
21461 Specify how histogram is calculated.
21463 It accepts the following values:
21466 Use single histogram for all channels.
21468 Use separate histogram for each channel.
21470 Default is @code{single}.
21473 Set frame rate, expressed as number of frames per second. Default
21477 Specify the video size for the output. For the syntax of this option, check the
21478 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21479 Default value is @code{hd720}.
21484 It accepts the following values:
21495 reverse logarithmic
21497 Default is @code{log}.
21500 Set amplitude scale.
21502 It accepts the following values:
21509 Default is @code{log}.
21512 Set how much frames to accumulate in histogram.
21513 Default is 1. Setting this to -1 accumulates all frames.
21516 Set histogram ratio of window height.
21519 Set sonogram sliding.
21521 It accepts the following values:
21524 replace old rows with new ones.
21526 scroll from top to bottom.
21528 Default is @code{replace}.
21531 @section aphasemeter
21533 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
21534 representing mean phase of current audio frame. A video output can also be produced and is
21535 enabled by default. The audio is passed through as first output.
21537 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
21538 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
21539 and @code{1} means channels are in phase.
21541 The filter accepts the following options, all related to its video output:
21545 Set the output frame rate. Default value is @code{25}.
21548 Set the video size for the output. For the syntax of this option, check the
21549 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21550 Default value is @code{800x400}.
21555 Specify the red, green, blue contrast. Default values are @code{2},
21556 @code{7} and @code{1}.
21557 Allowed range is @code{[0, 255]}.
21560 Set color which will be used for drawing median phase. If color is
21561 @code{none} which is default, no median phase value will be drawn.
21564 Enable video output. Default is enabled.
21567 @section avectorscope
21569 Convert input audio to a video output, representing the audio vector
21572 The filter is used to measure the difference between channels of stereo
21573 audio stream. A monaural signal, consisting of identical left and right
21574 signal, results in straight vertical line. Any stereo separation is visible
21575 as a deviation from this line, creating a Lissajous figure.
21576 If the straight (or deviation from it) but horizontal line appears this
21577 indicates that the left and right channels are out of phase.
21579 The filter accepts the following options:
21583 Set the vectorscope mode.
21585 Available values are:
21588 Lissajous rotated by 45 degrees.
21591 Same as above but not rotated.
21594 Shape resembling half of circle.
21597 Default value is @samp{lissajous}.
21600 Set the video size for the output. For the syntax of this option, check the
21601 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21602 Default value is @code{400x400}.
21605 Set the output frame rate. Default value is @code{25}.
21611 Specify the red, green, blue and alpha contrast. Default values are @code{40},
21612 @code{160}, @code{80} and @code{255}.
21613 Allowed range is @code{[0, 255]}.
21619 Specify the red, green, blue and alpha fade. Default values are @code{15},
21620 @code{10}, @code{5} and @code{5}.
21621 Allowed range is @code{[0, 255]}.
21624 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
21625 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
21628 Set the vectorscope drawing mode.
21630 Available values are:
21633 Draw dot for each sample.
21636 Draw line between previous and current sample.
21639 Default value is @samp{dot}.
21642 Specify amplitude scale of audio samples.
21644 Available values are:
21660 Swap left channel axis with right channel axis.
21670 Mirror only x axis.
21673 Mirror only y axis.
21681 @subsection Examples
21685 Complete example using @command{ffplay}:
21687 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
21688 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
21692 @section bench, abench
21694 Benchmark part of a filtergraph.
21696 The filter accepts the following options:
21700 Start or stop a timer.
21702 Available values are:
21705 Get the current time, set it as frame metadata (using the key
21706 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
21709 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
21710 the input frame metadata to get the time difference. Time difference, average,
21711 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
21712 @code{min}) are then printed. The timestamps are expressed in seconds.
21716 @subsection Examples
21720 Benchmark @ref{selectivecolor} filter:
21722 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
21728 Concatenate audio and video streams, joining them together one after the
21731 The filter works on segments of synchronized video and audio streams. All
21732 segments must have the same number of streams of each type, and that will
21733 also be the number of streams at output.
21735 The filter accepts the following options:
21740 Set the number of segments. Default is 2.
21743 Set the number of output video streams, that is also the number of video
21744 streams in each segment. Default is 1.
21747 Set the number of output audio streams, that is also the number of audio
21748 streams in each segment. Default is 0.
21751 Activate unsafe mode: do not fail if segments have a different format.
21755 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
21756 @var{a} audio outputs.
21758 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
21759 segment, in the same order as the outputs, then the inputs for the second
21762 Related streams do not always have exactly the same duration, for various
21763 reasons including codec frame size or sloppy authoring. For that reason,
21764 related synchronized streams (e.g. a video and its audio track) should be
21765 concatenated at once. The concat filter will use the duration of the longest
21766 stream in each segment (except the last one), and if necessary pad shorter
21767 audio streams with silence.
21769 For this filter to work correctly, all segments must start at timestamp 0.
21771 All corresponding streams must have the same parameters in all segments; the
21772 filtering system will automatically select a common pixel format for video
21773 streams, and a common sample format, sample rate and channel layout for
21774 audio streams, but other settings, such as resolution, must be converted
21775 explicitly by the user.
21777 Different frame rates are acceptable but will result in variable frame rate
21778 at output; be sure to configure the output file to handle it.
21780 @subsection Examples
21784 Concatenate an opening, an episode and an ending, all in bilingual version
21785 (video in stream 0, audio in streams 1 and 2):
21787 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
21788 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
21789 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
21790 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
21794 Concatenate two parts, handling audio and video separately, using the
21795 (a)movie sources, and adjusting the resolution:
21797 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
21798 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
21799 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
21801 Note that a desync will happen at the stitch if the audio and video streams
21802 do not have exactly the same duration in the first file.
21806 @subsection Commands
21808 This filter supports the following commands:
21811 Close the current segment and step to the next one
21814 @section drawgraph, adrawgraph
21816 Draw a graph using input video or audio metadata.
21818 It accepts the following parameters:
21822 Set 1st frame metadata key from which metadata values will be used to draw a graph.
21825 Set 1st foreground color expression.
21828 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
21831 Set 2nd foreground color expression.
21834 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
21837 Set 3rd foreground color expression.
21840 Set 4th frame metadata key from which metadata values will be used to draw a graph.
21843 Set 4th foreground color expression.
21846 Set minimal value of metadata value.
21849 Set maximal value of metadata value.
21852 Set graph background color. Default is white.
21857 Available values for mode is:
21864 Default is @code{line}.
21869 Available values for slide is:
21872 Draw new frame when right border is reached.
21875 Replace old columns with new ones.
21878 Scroll from right to left.
21881 Scroll from left to right.
21884 Draw single picture.
21887 Default is @code{frame}.
21890 Set size of graph video. For the syntax of this option, check the
21891 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21892 The default value is @code{900x256}.
21894 The foreground color expressions can use the following variables:
21897 Minimal value of metadata value.
21900 Maximal value of metadata value.
21903 Current metadata key value.
21906 The color is defined as 0xAABBGGRR.
21909 Example using metadata from @ref{signalstats} filter:
21911 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
21914 Example using metadata from @ref{ebur128} filter:
21916 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
21922 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
21923 level. By default, it logs a message at a frequency of 10Hz with the
21924 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
21925 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
21927 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
21928 sample format is double-precision floating point. The input stream will be converted to
21929 this specification, if needed. Users may need to insert aformat and/or aresample filters
21930 after this filter to obtain the original parameters.
21932 The filter also has a video output (see the @var{video} option) with a real
21933 time graph to observe the loudness evolution. The graphic contains the logged
21934 message mentioned above, so it is not printed anymore when this option is set,
21935 unless the verbose logging is set. The main graphing area contains the
21936 short-term loudness (3 seconds of analysis), and the gauge on the right is for
21937 the momentary loudness (400 milliseconds), but can optionally be configured
21938 to instead display short-term loudness (see @var{gauge}).
21940 The green area marks a +/- 1LU target range around the target loudness
21941 (-23LUFS by default, unless modified through @var{target}).
21943 More information about the Loudness Recommendation EBU R128 on
21944 @url{http://tech.ebu.ch/loudness}.
21946 The filter accepts the following options:
21951 Activate the video output. The audio stream is passed unchanged whether this
21952 option is set or no. The video stream will be the first output stream if
21953 activated. Default is @code{0}.
21956 Set the video size. This option is for video only. For the syntax of this
21958 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21959 Default and minimum resolution is @code{640x480}.
21962 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
21963 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
21964 other integer value between this range is allowed.
21967 Set metadata injection. If set to @code{1}, the audio input will be segmented
21968 into 100ms output frames, each of them containing various loudness information
21969 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
21971 Default is @code{0}.
21974 Force the frame logging level.
21976 Available values are:
21979 information logging level
21981 verbose logging level
21984 By default, the logging level is set to @var{info}. If the @option{video} or
21985 the @option{metadata} options are set, it switches to @var{verbose}.
21990 Available modes can be cumulated (the option is a @code{flag} type). Possible
21994 Disable any peak mode (default).
21996 Enable sample-peak mode.
21998 Simple peak mode looking for the higher sample value. It logs a message
21999 for sample-peak (identified by @code{SPK}).
22001 Enable true-peak mode.
22003 If enabled, the peak lookup is done on an over-sampled version of the input
22004 stream for better peak accuracy. It logs a message for true-peak.
22005 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
22006 This mode requires a build with @code{libswresample}.
22010 Treat mono input files as "dual mono". If a mono file is intended for playback
22011 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
22012 If set to @code{true}, this option will compensate for this effect.
22013 Multi-channel input files are not affected by this option.
22016 Set a specific pan law to be used for the measurement of dual mono files.
22017 This parameter is optional, and has a default value of -3.01dB.
22020 Set a specific target level (in LUFS) used as relative zero in the visualization.
22021 This parameter is optional and has a default value of -23LUFS as specified
22022 by EBU R128. However, material published online may prefer a level of -16LUFS
22023 (e.g. for use with podcasts or video platforms).
22026 Set the value displayed by the gauge. Valid values are @code{momentary} and s
22027 @code{shortterm}. By default the momentary value will be used, but in certain
22028 scenarios it may be more useful to observe the short term value instead (e.g.
22032 Sets the display scale for the loudness. Valid parameters are @code{absolute}
22033 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
22034 video output, not the summary or continuous log output.
22037 @subsection Examples
22041 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
22043 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
22047 Run an analysis with @command{ffmpeg}:
22049 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
22053 @section interleave, ainterleave
22055 Temporally interleave frames from several inputs.
22057 @code{interleave} works with video inputs, @code{ainterleave} with audio.
22059 These filters read frames from several inputs and send the oldest
22060 queued frame to the output.
22062 Input streams must have well defined, monotonically increasing frame
22065 In order to submit one frame to output, these filters need to enqueue
22066 at least one frame for each input, so they cannot work in case one
22067 input is not yet terminated and will not receive incoming frames.
22069 For example consider the case when one input is a @code{select} filter
22070 which always drops input frames. The @code{interleave} filter will keep
22071 reading from that input, but it will never be able to send new frames
22072 to output until the input sends an end-of-stream signal.
22074 Also, depending on inputs synchronization, the filters will drop
22075 frames in case one input receives more frames than the other ones, and
22076 the queue is already filled.
22078 These filters accept the following options:
22082 Set the number of different inputs, it is 2 by default.
22085 @subsection Examples
22089 Interleave frames belonging to different streams using @command{ffmpeg}:
22091 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
22095 Add flickering blur effect:
22097 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
22101 @section metadata, ametadata
22103 Manipulate frame metadata.
22105 This filter accepts the following options:
22109 Set mode of operation of the filter.
22111 Can be one of the following:
22115 If both @code{value} and @code{key} is set, select frames
22116 which have such metadata. If only @code{key} is set, select
22117 every frame that has such key in metadata.
22120 Add new metadata @code{key} and @code{value}. If key is already available
22124 Modify value of already present key.
22127 If @code{value} is set, delete only keys that have such value.
22128 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
22132 Print key and its value if metadata was found. If @code{key} is not set print all
22133 metadata values available in frame.
22137 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
22140 Set metadata value which will be used. This option is mandatory for
22141 @code{modify} and @code{add} mode.
22144 Which function to use when comparing metadata value and @code{value}.
22146 Can be one of following:
22150 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
22153 Values are interpreted as strings, returns true if metadata value starts with
22154 the @code{value} option string.
22157 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
22160 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
22163 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
22166 Values are interpreted as floats, returns true if expression from option @code{expr}
22170 Values are interpreted as strings, returns true if metadata value ends with
22171 the @code{value} option string.
22175 Set expression which is used when @code{function} is set to @code{expr}.
22176 The expression is evaluated through the eval API and can contain the following
22181 Float representation of @code{value} from metadata key.
22184 Float representation of @code{value} as supplied by user in @code{value} option.
22188 If specified in @code{print} mode, output is written to the named file. Instead of
22189 plain filename any writable url can be specified. Filename ``-'' is a shorthand
22190 for standard output. If @code{file} option is not set, output is written to the log
22191 with AV_LOG_INFO loglevel.
22195 @subsection Examples
22199 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
22202 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
22205 Print silencedetect output to file @file{metadata.txt}.
22207 silencedetect,ametadata=mode=print:file=metadata.txt
22210 Direct all metadata to a pipe with file descriptor 4.
22212 metadata=mode=print:file='pipe\:4'
22216 @section perms, aperms
22218 Set read/write permissions for the output frames.
22220 These filters are mainly aimed at developers to test direct path in the
22221 following filter in the filtergraph.
22223 The filters accept the following options:
22227 Select the permissions mode.
22229 It accepts the following values:
22232 Do nothing. This is the default.
22234 Set all the output frames read-only.
22236 Set all the output frames directly writable.
22238 Make the frame read-only if writable, and writable if read-only.
22240 Set each output frame read-only or writable randomly.
22244 Set the seed for the @var{random} mode, must be an integer included between
22245 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
22246 @code{-1}, the filter will try to use a good random seed on a best effort
22250 Note: in case of auto-inserted filter between the permission filter and the
22251 following one, the permission might not be received as expected in that
22252 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
22253 perms/aperms filter can avoid this problem.
22255 @section realtime, arealtime
22257 Slow down filtering to match real time approximately.
22259 These filters will pause the filtering for a variable amount of time to
22260 match the output rate with the input timestamps.
22261 They are similar to the @option{re} option to @code{ffmpeg}.
22263 They accept the following options:
22267 Time limit for the pauses. Any pause longer than that will be considered
22268 a timestamp discontinuity and reset the timer. Default is 2 seconds.
22270 Speed factor for processing. The value must be a float larger than zero.
22271 Values larger than 1.0 will result in faster than realtime processing,
22272 smaller will slow processing down. The @var{limit} is automatically adapted
22273 accordingly. Default is 1.0.
22275 A processing speed faster than what is possible without these filters cannot
22280 @section select, aselect
22282 Select frames to pass in output.
22284 This filter accepts the following options:
22289 Set expression, which is evaluated for each input frame.
22291 If the expression is evaluated to zero, the frame is discarded.
22293 If the evaluation result is negative or NaN, the frame is sent to the
22294 first output; otherwise it is sent to the output with index
22295 @code{ceil(val)-1}, assuming that the input index starts from 0.
22297 For example a value of @code{1.2} corresponds to the output with index
22298 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
22301 Set the number of outputs. The output to which to send the selected
22302 frame is based on the result of the evaluation. Default value is 1.
22305 The expression can contain the following constants:
22309 The (sequential) number of the filtered frame, starting from 0.
22312 The (sequential) number of the selected frame, starting from 0.
22314 @item prev_selected_n
22315 The sequential number of the last selected frame. It's NAN if undefined.
22318 The timebase of the input timestamps.
22321 The PTS (Presentation TimeStamp) of the filtered video frame,
22322 expressed in @var{TB} units. It's NAN if undefined.
22325 The PTS of the filtered video frame,
22326 expressed in seconds. It's NAN if undefined.
22329 The PTS of the previously filtered video frame. It's NAN if undefined.
22331 @item prev_selected_pts
22332 The PTS of the last previously filtered video frame. It's NAN if undefined.
22334 @item prev_selected_t
22335 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
22338 The PTS of the first video frame in the video. It's NAN if undefined.
22341 The time of the first video frame in the video. It's NAN if undefined.
22343 @item pict_type @emph{(video only)}
22344 The type of the filtered frame. It can assume one of the following
22356 @item interlace_type @emph{(video only)}
22357 The frame interlace type. It can assume one of the following values:
22360 The frame is progressive (not interlaced).
22362 The frame is top-field-first.
22364 The frame is bottom-field-first.
22367 @item consumed_sample_n @emph{(audio only)}
22368 the number of selected samples before the current frame
22370 @item samples_n @emph{(audio only)}
22371 the number of samples in the current frame
22373 @item sample_rate @emph{(audio only)}
22374 the input sample rate
22377 This is 1 if the filtered frame is a key-frame, 0 otherwise.
22380 the position in the file of the filtered frame, -1 if the information
22381 is not available (e.g. for synthetic video)
22383 @item scene @emph{(video only)}
22384 value between 0 and 1 to indicate a new scene; a low value reflects a low
22385 probability for the current frame to introduce a new scene, while a higher
22386 value means the current frame is more likely to be one (see the example below)
22388 @item concatdec_select
22389 The concat demuxer can select only part of a concat input file by setting an
22390 inpoint and an outpoint, but the output packets may not be entirely contained
22391 in the selected interval. By using this variable, it is possible to skip frames
22392 generated by the concat demuxer which are not exactly contained in the selected
22395 This works by comparing the frame pts against the @var{lavf.concat.start_time}
22396 and the @var{lavf.concat.duration} packet metadata values which are also
22397 present in the decoded frames.
22399 The @var{concatdec_select} variable is -1 if the frame pts is at least
22400 start_time and either the duration metadata is missing or the frame pts is less
22401 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
22404 That basically means that an input frame is selected if its pts is within the
22405 interval set by the concat demuxer.
22409 The default value of the select expression is "1".
22411 @subsection Examples
22415 Select all frames in input:
22420 The example above is the same as:
22432 Select only I-frames:
22434 select='eq(pict_type\,I)'
22438 Select one frame every 100:
22440 select='not(mod(n\,100))'
22444 Select only frames contained in the 10-20 time interval:
22446 select=between(t\,10\,20)
22450 Select only I-frames contained in the 10-20 time interval:
22452 select=between(t\,10\,20)*eq(pict_type\,I)
22456 Select frames with a minimum distance of 10 seconds:
22458 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
22462 Use aselect to select only audio frames with samples number > 100:
22464 aselect='gt(samples_n\,100)'
22468 Create a mosaic of the first scenes:
22470 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
22473 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
22477 Send even and odd frames to separate outputs, and compose them:
22479 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
22483 Select useful frames from an ffconcat file which is using inpoints and
22484 outpoints but where the source files are not intra frame only.
22486 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
22490 @section sendcmd, asendcmd
22492 Send commands to filters in the filtergraph.
22494 These filters read commands to be sent to other filters in the
22497 @code{sendcmd} must be inserted between two video filters,
22498 @code{asendcmd} must be inserted between two audio filters, but apart
22499 from that they act the same way.
22501 The specification of commands can be provided in the filter arguments
22502 with the @var{commands} option, or in a file specified by the
22503 @var{filename} option.
22505 These filters accept the following options:
22508 Set the commands to be read and sent to the other filters.
22510 Set the filename of the commands to be read and sent to the other
22514 @subsection Commands syntax
22516 A commands description consists of a sequence of interval
22517 specifications, comprising a list of commands to be executed when a
22518 particular event related to that interval occurs. The occurring event
22519 is typically the current frame time entering or leaving a given time
22522 An interval is specified by the following syntax:
22524 @var{START}[-@var{END}] @var{COMMANDS};
22527 The time interval is specified by the @var{START} and @var{END} times.
22528 @var{END} is optional and defaults to the maximum time.
22530 The current frame time is considered within the specified interval if
22531 it is included in the interval [@var{START}, @var{END}), that is when
22532 the time is greater or equal to @var{START} and is lesser than
22535 @var{COMMANDS} consists of a sequence of one or more command
22536 specifications, separated by ",", relating to that interval. The
22537 syntax of a command specification is given by:
22539 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
22542 @var{FLAGS} is optional and specifies the type of events relating to
22543 the time interval which enable sending the specified command, and must
22544 be a non-null sequence of identifier flags separated by "+" or "|" and
22545 enclosed between "[" and "]".
22547 The following flags are recognized:
22550 The command is sent when the current frame timestamp enters the
22551 specified interval. In other words, the command is sent when the
22552 previous frame timestamp was not in the given interval, and the
22556 The command is sent when the current frame timestamp leaves the
22557 specified interval. In other words, the command is sent when the
22558 previous frame timestamp was in the given interval, and the
22562 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
22565 @var{TARGET} specifies the target of the command, usually the name of
22566 the filter class or a specific filter instance name.
22568 @var{COMMAND} specifies the name of the command for the target filter.
22570 @var{ARG} is optional and specifies the optional list of argument for
22571 the given @var{COMMAND}.
22573 Between one interval specification and another, whitespaces, or
22574 sequences of characters starting with @code{#} until the end of line,
22575 are ignored and can be used to annotate comments.
22577 A simplified BNF description of the commands specification syntax
22580 @var{COMMAND_FLAG} ::= "enter" | "leave"
22581 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
22582 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
22583 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
22584 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
22585 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
22588 @subsection Examples
22592 Specify audio tempo change at second 4:
22594 asendcmd=c='4.0 atempo tempo 1.5',atempo
22598 Target a specific filter instance:
22600 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
22604 Specify a list of drawtext and hue commands in a file.
22606 # show text in the interval 5-10
22607 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
22608 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
22610 # desaturate the image in the interval 15-20
22611 15.0-20.0 [enter] hue s 0,
22612 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
22614 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
22616 # apply an exponential saturation fade-out effect, starting from time 25
22617 25 [enter] hue s exp(25-t)
22620 A filtergraph allowing to read and process the above command list
22621 stored in a file @file{test.cmd}, can be specified with:
22623 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
22628 @section setpts, asetpts
22630 Change the PTS (presentation timestamp) of the input frames.
22632 @code{setpts} works on video frames, @code{asetpts} on audio frames.
22634 This filter accepts the following options:
22639 The expression which is evaluated for each frame to construct its timestamp.
22643 The expression is evaluated through the eval API and can contain the following
22647 @item FRAME_RATE, FR
22648 frame rate, only defined for constant frame-rate video
22651 The presentation timestamp in input
22654 The count of the input frame for video or the number of consumed samples,
22655 not including the current frame for audio, starting from 0.
22657 @item NB_CONSUMED_SAMPLES
22658 The number of consumed samples, not including the current frame (only
22661 @item NB_SAMPLES, S
22662 The number of samples in the current frame (only audio)
22664 @item SAMPLE_RATE, SR
22665 The audio sample rate.
22668 The PTS of the first frame.
22671 the time in seconds of the first frame
22674 State whether the current frame is interlaced.
22677 the time in seconds of the current frame
22680 original position in the file of the frame, or undefined if undefined
22681 for the current frame
22684 The previous input PTS.
22687 previous input time in seconds
22690 The previous output PTS.
22693 previous output time in seconds
22696 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
22700 The wallclock (RTC) time at the start of the movie in microseconds.
22703 The timebase of the input timestamps.
22707 @subsection Examples
22711 Start counting PTS from zero
22713 setpts=PTS-STARTPTS
22717 Apply fast motion effect:
22723 Apply slow motion effect:
22729 Set fixed rate of 25 frames per second:
22735 Set fixed rate 25 fps with some jitter:
22737 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
22741 Apply an offset of 10 seconds to the input PTS:
22747 Generate timestamps from a "live source" and rebase onto the current timebase:
22749 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
22753 Generate timestamps by counting samples:
22762 Force color range for the output video frame.
22764 The @code{setrange} filter marks the color range property for the
22765 output frames. It does not change the input frame, but only sets the
22766 corresponding property, which affects how the frame is treated by
22769 The filter accepts the following options:
22774 Available values are:
22778 Keep the same color range property.
22780 @item unspecified, unknown
22781 Set the color range as unspecified.
22783 @item limited, tv, mpeg
22784 Set the color range as limited.
22786 @item full, pc, jpeg
22787 Set the color range as full.
22791 @section settb, asettb
22793 Set the timebase to use for the output frames timestamps.
22794 It is mainly useful for testing timebase configuration.
22796 It accepts the following parameters:
22801 The expression which is evaluated into the output timebase.
22805 The value for @option{tb} is an arithmetic expression representing a
22806 rational. The expression can contain the constants "AVTB" (the default
22807 timebase), "intb" (the input timebase) and "sr" (the sample rate,
22808 audio only). Default value is "intb".
22810 @subsection Examples
22814 Set the timebase to 1/25:
22820 Set the timebase to 1/10:
22826 Set the timebase to 1001/1000:
22832 Set the timebase to 2*intb:
22838 Set the default timebase value:
22845 Convert input audio to a video output representing frequency spectrum
22846 logarithmically using Brown-Puckette constant Q transform algorithm with
22847 direct frequency domain coefficient calculation (but the transform itself
22848 is not really constant Q, instead the Q factor is actually variable/clamped),
22849 with musical tone scale, from E0 to D#10.
22851 The filter accepts the following options:
22855 Specify the video size for the output. It must be even. For the syntax of this option,
22856 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22857 Default value is @code{1920x1080}.
22860 Set the output frame rate. Default value is @code{25}.
22863 Set the bargraph height. It must be even. Default value is @code{-1} which
22864 computes the bargraph height automatically.
22867 Set the axis height. It must be even. Default value is @code{-1} which computes
22868 the axis height automatically.
22871 Set the sonogram height. It must be even. Default value is @code{-1} which
22872 computes the sonogram height automatically.
22875 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
22876 instead. Default value is @code{1}.
22878 @item sono_v, volume
22879 Specify the sonogram volume expression. It can contain variables:
22882 the @var{bar_v} evaluated expression
22883 @item frequency, freq, f
22884 the frequency where it is evaluated
22885 @item timeclamp, tc
22886 the value of @var{timeclamp} option
22890 @item a_weighting(f)
22891 A-weighting of equal loudness
22892 @item b_weighting(f)
22893 B-weighting of equal loudness
22894 @item c_weighting(f)
22895 C-weighting of equal loudness.
22897 Default value is @code{16}.
22899 @item bar_v, volume2
22900 Specify the bargraph volume expression. It can contain variables:
22903 the @var{sono_v} evaluated expression
22904 @item frequency, freq, f
22905 the frequency where it is evaluated
22906 @item timeclamp, tc
22907 the value of @var{timeclamp} option
22911 @item a_weighting(f)
22912 A-weighting of equal loudness
22913 @item b_weighting(f)
22914 B-weighting of equal loudness
22915 @item c_weighting(f)
22916 C-weighting of equal loudness.
22918 Default value is @code{sono_v}.
22920 @item sono_g, gamma
22921 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
22922 higher gamma makes the spectrum having more range. Default value is @code{3}.
22923 Acceptable range is @code{[1, 7]}.
22925 @item bar_g, gamma2
22926 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
22930 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
22931 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
22933 @item timeclamp, tc
22934 Specify the transform timeclamp. At low frequency, there is trade-off between
22935 accuracy in time domain and frequency domain. If timeclamp is lower,
22936 event in time domain is represented more accurately (such as fast bass drum),
22937 otherwise event in frequency domain is represented more accurately
22938 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
22941 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
22942 limits future samples by applying asymmetric windowing in time domain, useful
22943 when low latency is required. Accepted range is @code{[0, 1]}.
22946 Specify the transform base frequency. Default value is @code{20.01523126408007475},
22947 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
22950 Specify the transform end frequency. Default value is @code{20495.59681441799654},
22951 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
22954 This option is deprecated and ignored.
22957 Specify the transform length in time domain. Use this option to control accuracy
22958 trade-off between time domain and frequency domain at every frequency sample.
22959 It can contain variables:
22961 @item frequency, freq, f
22962 the frequency where it is evaluated
22963 @item timeclamp, tc
22964 the value of @var{timeclamp} option.
22966 Default value is @code{384*tc/(384+tc*f)}.
22969 Specify the transform count for every video frame. Default value is @code{6}.
22970 Acceptable range is @code{[1, 30]}.
22973 Specify the transform count for every single pixel. Default value is @code{0},
22974 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
22977 Specify font file for use with freetype to draw the axis. If not specified,
22978 use embedded font. Note that drawing with font file or embedded font is not
22979 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
22983 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
22984 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
22988 Specify font color expression. This is arithmetic expression that should return
22989 integer value 0xRRGGBB. It can contain variables:
22991 @item frequency, freq, f
22992 the frequency where it is evaluated
22993 @item timeclamp, tc
22994 the value of @var{timeclamp} option
22999 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
23000 @item r(x), g(x), b(x)
23001 red, green, and blue value of intensity x.
23003 Default value is @code{st(0, (midi(f)-59.5)/12);
23004 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
23005 r(1-ld(1)) + b(ld(1))}.
23008 Specify image file to draw the axis. This option override @var{fontfile} and
23009 @var{fontcolor} option.
23012 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
23013 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
23014 Default value is @code{1}.
23017 Set colorspace. The accepted values are:
23020 Unspecified (default)
23029 BT.470BG or BT.601-6 625
23032 SMPTE-170M or BT.601-6 525
23038 BT.2020 with non-constant luminance
23043 Set spectrogram color scheme. This is list of floating point values with format
23044 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
23045 The default is @code{1|0.5|0|0|0.5|1}.
23049 @subsection Examples
23053 Playing audio while showing the spectrum:
23055 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
23059 Same as above, but with frame rate 30 fps:
23061 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
23065 Playing at 1280x720:
23067 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
23071 Disable sonogram display:
23077 A1 and its harmonics: A1, A2, (near)E3, A3:
23079 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),
23080 asplit[a][out1]; [a] showcqt [out0]'
23084 Same as above, but with more accuracy in frequency domain:
23086 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),
23087 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
23093 bar_v=10:sono_v=bar_v*a_weighting(f)
23097 Custom gamma, now spectrum is linear to the amplitude.
23103 Custom tlength equation:
23105 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)))'
23109 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
23111 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
23115 Custom font using fontconfig:
23117 font='Courier New,Monospace,mono|bold'
23121 Custom frequency range with custom axis using image file:
23123 axisfile=myaxis.png:basefreq=40:endfreq=10000
23129 Convert input audio to video output representing the audio power spectrum.
23130 Audio amplitude is on Y-axis while frequency is on X-axis.
23132 The filter accepts the following options:
23136 Specify size of video. For the syntax of this option, check the
23137 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23138 Default is @code{1024x512}.
23142 This set how each frequency bin will be represented.
23144 It accepts the following values:
23150 Default is @code{bar}.
23153 Set amplitude scale.
23155 It accepts the following values:
23169 Default is @code{log}.
23172 Set frequency scale.
23174 It accepts the following values:
23183 Reverse logarithmic scale.
23185 Default is @code{lin}.
23188 Set window size. Allowed range is from 16 to 65536.
23190 Default is @code{2048}
23193 Set windowing function.
23195 It accepts the following values:
23218 Default is @code{hanning}.
23221 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
23222 which means optimal overlap for selected window function will be picked.
23225 Set time averaging. Setting this to 0 will display current maximal peaks.
23226 Default is @code{1}, which means time averaging is disabled.
23229 Specify list of colors separated by space or by '|' which will be used to
23230 draw channel frequencies. Unrecognized or missing colors will be replaced
23234 Set channel display mode.
23236 It accepts the following values:
23241 Default is @code{combined}.
23244 Set minimum amplitude used in @code{log} amplitude scaler.
23248 @section showspatial
23250 Convert stereo input audio to a video output, representing the spatial relationship
23251 between two channels.
23253 The filter accepts the following options:
23257 Specify the video size for the output. For the syntax of this option, check the
23258 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23259 Default value is @code{512x512}.
23262 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
23265 Set window function.
23267 It accepts the following values:
23292 Default value is @code{hann}.
23295 Set ratio of overlap window. Default value is @code{0.5}.
23296 When value is @code{1} overlap is set to recommended size for specific
23297 window function currently used.
23300 @anchor{showspectrum}
23301 @section showspectrum
23303 Convert input audio to a video output, representing the audio frequency
23306 The filter accepts the following options:
23310 Specify the video size for the output. For the syntax of this option, check the
23311 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23312 Default value is @code{640x512}.
23315 Specify how the spectrum should slide along the window.
23317 It accepts the following values:
23320 the samples start again on the left when they reach the right
23322 the samples scroll from right to left
23324 frames are only produced when the samples reach the right
23326 the samples scroll from left to right
23329 Default value is @code{replace}.
23332 Specify display mode.
23334 It accepts the following values:
23337 all channels are displayed in the same row
23339 all channels are displayed in separate rows
23342 Default value is @samp{combined}.
23345 Specify display color mode.
23347 It accepts the following values:
23350 each channel is displayed in a separate color
23352 each channel is displayed using the same color scheme
23354 each channel is displayed using the rainbow color scheme
23356 each channel is displayed using the moreland color scheme
23358 each channel is displayed using the nebulae color scheme
23360 each channel is displayed using the fire color scheme
23362 each channel is displayed using the fiery color scheme
23364 each channel is displayed using the fruit color scheme
23366 each channel is displayed using the cool color scheme
23368 each channel is displayed using the magma color scheme
23370 each channel is displayed using the green color scheme
23372 each channel is displayed using the viridis color scheme
23374 each channel is displayed using the plasma color scheme
23376 each channel is displayed using the cividis color scheme
23378 each channel is displayed using the terrain color scheme
23381 Default value is @samp{channel}.
23384 Specify scale used for calculating intensity color values.
23386 It accepts the following values:
23391 square root, default
23402 Default value is @samp{sqrt}.
23405 Specify frequency scale.
23407 It accepts the following values:
23415 Default value is @samp{lin}.
23418 Set saturation modifier for displayed colors. Negative values provide
23419 alternative color scheme. @code{0} is no saturation at all.
23420 Saturation must be in [-10.0, 10.0] range.
23421 Default value is @code{1}.
23424 Set window function.
23426 It accepts the following values:
23451 Default value is @code{hann}.
23454 Set orientation of time vs frequency axis. Can be @code{vertical} or
23455 @code{horizontal}. Default is @code{vertical}.
23458 Set ratio of overlap window. Default value is @code{0}.
23459 When value is @code{1} overlap is set to recommended size for specific
23460 window function currently used.
23463 Set scale gain for calculating intensity color values.
23464 Default value is @code{1}.
23467 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
23470 Set color rotation, must be in [-1.0, 1.0] range.
23471 Default value is @code{0}.
23474 Set start frequency from which to display spectrogram. Default is @code{0}.
23477 Set stop frequency to which to display spectrogram. Default is @code{0}.
23480 Set upper frame rate limit. Default is @code{auto}, unlimited.
23483 Draw time and frequency axes and legends. Default is disabled.
23486 The usage is very similar to the showwaves filter; see the examples in that
23489 @subsection Examples
23493 Large window with logarithmic color scaling:
23495 showspectrum=s=1280x480:scale=log
23499 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
23501 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
23502 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
23506 @section showspectrumpic
23508 Convert input audio to a single video frame, representing the audio frequency
23511 The filter accepts the following options:
23515 Specify the video size for the output. For the syntax of this option, check the
23516 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23517 Default value is @code{4096x2048}.
23520 Specify display mode.
23522 It accepts the following values:
23525 all channels are displayed in the same row
23527 all channels are displayed in separate rows
23529 Default value is @samp{combined}.
23532 Specify display color mode.
23534 It accepts the following values:
23537 each channel is displayed in a separate color
23539 each channel is displayed using the same color scheme
23541 each channel is displayed using the rainbow color scheme
23543 each channel is displayed using the moreland color scheme
23545 each channel is displayed using the nebulae color scheme
23547 each channel is displayed using the fire color scheme
23549 each channel is displayed using the fiery color scheme
23551 each channel is displayed using the fruit color scheme
23553 each channel is displayed using the cool color scheme
23555 each channel is displayed using the magma color scheme
23557 each channel is displayed using the green color scheme
23559 each channel is displayed using the viridis color scheme
23561 each channel is displayed using the plasma color scheme
23563 each channel is displayed using the cividis color scheme
23565 each channel is displayed using the terrain color scheme
23567 Default value is @samp{intensity}.
23570 Specify scale used for calculating intensity color values.
23572 It accepts the following values:
23577 square root, default
23587 Default value is @samp{log}.
23590 Specify frequency scale.
23592 It accepts the following values:
23600 Default value is @samp{lin}.
23603 Set saturation modifier for displayed colors. Negative values provide
23604 alternative color scheme. @code{0} is no saturation at all.
23605 Saturation must be in [-10.0, 10.0] range.
23606 Default value is @code{1}.
23609 Set window function.
23611 It accepts the following values:
23635 Default value is @code{hann}.
23638 Set orientation of time vs frequency axis. Can be @code{vertical} or
23639 @code{horizontal}. Default is @code{vertical}.
23642 Set scale gain for calculating intensity color values.
23643 Default value is @code{1}.
23646 Draw time and frequency axes and legends. Default is enabled.
23649 Set color rotation, must be in [-1.0, 1.0] range.
23650 Default value is @code{0}.
23653 Set start frequency from which to display spectrogram. Default is @code{0}.
23656 Set stop frequency to which to display spectrogram. Default is @code{0}.
23659 @subsection Examples
23663 Extract an audio spectrogram of a whole audio track
23664 in a 1024x1024 picture using @command{ffmpeg}:
23666 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
23670 @section showvolume
23672 Convert input audio volume to a video output.
23674 The filter accepts the following options:
23681 Set border width, allowed range is [0, 5]. Default is 1.
23684 Set channel width, allowed range is [80, 8192]. Default is 400.
23687 Set channel height, allowed range is [1, 900]. Default is 20.
23690 Set fade, allowed range is [0, 1]. Default is 0.95.
23693 Set volume color expression.
23695 The expression can use the following variables:
23699 Current max volume of channel in dB.
23705 Current channel number, starting from 0.
23709 If set, displays channel names. Default is enabled.
23712 If set, displays volume values. Default is enabled.
23715 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
23716 default is @code{h}.
23719 Set step size, allowed range is [0, 5]. Default is 0, which means
23723 Set background opacity, allowed range is [0, 1]. Default is 0.
23726 Set metering mode, can be peak: @code{p} or rms: @code{r},
23727 default is @code{p}.
23730 Set display scale, can be linear: @code{lin} or log: @code{log},
23731 default is @code{lin}.
23735 If set to > 0., display a line for the max level
23736 in the previous seconds.
23737 default is disabled: @code{0.}
23740 The color of the max line. Use when @code{dm} option is set to > 0.
23741 default is: @code{orange}
23746 Convert input audio to a video output, representing the samples waves.
23748 The filter accepts the following options:
23752 Specify the video size for the output. For the syntax of this option, check the
23753 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23754 Default value is @code{600x240}.
23759 Available values are:
23762 Draw a point for each sample.
23765 Draw a vertical line for each sample.
23768 Draw a point for each sample and a line between them.
23771 Draw a centered vertical line for each sample.
23774 Default value is @code{point}.
23777 Set the number of samples which are printed on the same column. A
23778 larger value will decrease the frame rate. Must be a positive
23779 integer. This option can be set only if the value for @var{rate}
23780 is not explicitly specified.
23783 Set the (approximate) output frame rate. This is done by setting the
23784 option @var{n}. Default value is "25".
23786 @item split_channels
23787 Set if channels should be drawn separately or overlap. Default value is 0.
23790 Set colors separated by '|' which are going to be used for drawing of each channel.
23793 Set amplitude scale.
23795 Available values are:
23813 Set the draw mode. This is mostly useful to set for high @var{n}.
23815 Available values are:
23818 Scale pixel values for each drawn sample.
23821 Draw every sample directly.
23824 Default value is @code{scale}.
23827 @subsection Examples
23831 Output the input file audio and the corresponding video representation
23834 amovie=a.mp3,asplit[out0],showwaves[out1]
23838 Create a synthetic signal and show it with showwaves, forcing a
23839 frame rate of 30 frames per second:
23841 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
23845 @section showwavespic
23847 Convert input audio to a single video frame, representing the samples waves.
23849 The filter accepts the following options:
23853 Specify the video size for the output. For the syntax of this option, check the
23854 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23855 Default value is @code{600x240}.
23857 @item split_channels
23858 Set if channels should be drawn separately or overlap. Default value is 0.
23861 Set colors separated by '|' which are going to be used for drawing of each channel.
23864 Set amplitude scale.
23866 Available values are:
23886 Available values are:
23889 Scale pixel values for each drawn sample.
23892 Draw every sample directly.
23895 Default value is @code{scale}.
23898 @subsection Examples
23902 Extract a channel split representation of the wave form of a whole audio track
23903 in a 1024x800 picture using @command{ffmpeg}:
23905 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
23909 @section sidedata, asidedata
23911 Delete frame side data, or select frames based on it.
23913 This filter accepts the following options:
23917 Set mode of operation of the filter.
23919 Can be one of the following:
23923 Select every frame with side data of @code{type}.
23926 Delete side data of @code{type}. If @code{type} is not set, delete all side
23932 Set side data type used with all modes. Must be set for @code{select} mode. For
23933 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
23934 in @file{libavutil/frame.h}. For example, to choose
23935 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
23939 @section spectrumsynth
23941 Synthesize audio from 2 input video spectrums, first input stream represents
23942 magnitude across time and second represents phase across time.
23943 The filter will transform from frequency domain as displayed in videos back
23944 to time domain as presented in audio output.
23946 This filter is primarily created for reversing processed @ref{showspectrum}
23947 filter outputs, but can synthesize sound from other spectrograms too.
23948 But in such case results are going to be poor if the phase data is not
23949 available, because in such cases phase data need to be recreated, usually
23950 it's just recreated from random noise.
23951 For best results use gray only output (@code{channel} color mode in
23952 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
23953 @code{lin} scale for phase video. To produce phase, for 2nd video, use
23954 @code{data} option. Inputs videos should generally use @code{fullframe}
23955 slide mode as that saves resources needed for decoding video.
23957 The filter accepts the following options:
23961 Specify sample rate of output audio, the sample rate of audio from which
23962 spectrum was generated may differ.
23965 Set number of channels represented in input video spectrums.
23968 Set scale which was used when generating magnitude input spectrum.
23969 Can be @code{lin} or @code{log}. Default is @code{log}.
23972 Set slide which was used when generating inputs spectrums.
23973 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
23974 Default is @code{fullframe}.
23977 Set window function used for resynthesis.
23980 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
23981 which means optimal overlap for selected window function will be picked.
23984 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
23985 Default is @code{vertical}.
23988 @subsection Examples
23992 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
23993 then resynthesize videos back to audio with spectrumsynth:
23995 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
23996 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
23997 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
24001 @section split, asplit
24003 Split input into several identical outputs.
24005 @code{asplit} works with audio input, @code{split} with video.
24007 The filter accepts a single parameter which specifies the number of outputs. If
24008 unspecified, it defaults to 2.
24010 @subsection Examples
24014 Create two separate outputs from the same input:
24016 [in] split [out0][out1]
24020 To create 3 or more outputs, you need to specify the number of
24023 [in] asplit=3 [out0][out1][out2]
24027 Create two separate outputs from the same input, one cropped and
24030 [in] split [splitout1][splitout2];
24031 [splitout1] crop=100:100:0:0 [cropout];
24032 [splitout2] pad=200:200:100:100 [padout];
24036 Create 5 copies of the input audio with @command{ffmpeg}:
24038 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
24044 Receive commands sent through a libzmq client, and forward them to
24045 filters in the filtergraph.
24047 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
24048 must be inserted between two video filters, @code{azmq} between two
24049 audio filters. Both are capable to send messages to any filter type.
24051 To enable these filters you need to install the libzmq library and
24052 headers and configure FFmpeg with @code{--enable-libzmq}.
24054 For more information about libzmq see:
24055 @url{http://www.zeromq.org/}
24057 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
24058 receives messages sent through a network interface defined by the
24059 @option{bind_address} (or the abbreviation "@option{b}") option.
24060 Default value of this option is @file{tcp://localhost:5555}. You may
24061 want to alter this value to your needs, but do not forget to escape any
24062 ':' signs (see @ref{filtergraph escaping}).
24064 The received message must be in the form:
24066 @var{TARGET} @var{COMMAND} [@var{ARG}]
24069 @var{TARGET} specifies the target of the command, usually the name of
24070 the filter class or a specific filter instance name. The default
24071 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
24072 but you can override this by using the @samp{filter_name@@id} syntax
24073 (see @ref{Filtergraph syntax}).
24075 @var{COMMAND} specifies the name of the command for the target filter.
24077 @var{ARG} is optional and specifies the optional argument list for the
24078 given @var{COMMAND}.
24080 Upon reception, the message is processed and the corresponding command
24081 is injected into the filtergraph. Depending on the result, the filter
24082 will send a reply to the client, adopting the format:
24084 @var{ERROR_CODE} @var{ERROR_REASON}
24088 @var{MESSAGE} is optional.
24090 @subsection Examples
24092 Look at @file{tools/zmqsend} for an example of a zmq client which can
24093 be used to send commands processed by these filters.
24095 Consider the following filtergraph generated by @command{ffplay}.
24096 In this example the last overlay filter has an instance name. All other
24097 filters will have default instance names.
24100 ffplay -dumpgraph 1 -f lavfi "
24101 color=s=100x100:c=red [l];
24102 color=s=100x100:c=blue [r];
24103 nullsrc=s=200x100, zmq [bg];
24104 [bg][l] overlay [bg+l];
24105 [bg+l][r] overlay@@my=x=100 "
24108 To change the color of the left side of the video, the following
24109 command can be used:
24111 echo Parsed_color_0 c yellow | tools/zmqsend
24114 To change the right side:
24116 echo Parsed_color_1 c pink | tools/zmqsend
24119 To change the position of the right side:
24121 echo overlay@@my x 150 | tools/zmqsend
24125 @c man end MULTIMEDIA FILTERS
24127 @chapter Multimedia Sources
24128 @c man begin MULTIMEDIA SOURCES
24130 Below is a description of the currently available multimedia sources.
24134 This is the same as @ref{movie} source, except it selects an audio
24140 Read audio and/or video stream(s) from a movie container.
24142 It accepts the following parameters:
24146 The name of the resource to read (not necessarily a file; it can also be a
24147 device or a stream accessed through some protocol).
24149 @item format_name, f
24150 Specifies the format assumed for the movie to read, and can be either
24151 the name of a container or an input device. If not specified, the
24152 format is guessed from @var{movie_name} or by probing.
24154 @item seek_point, sp
24155 Specifies the seek point in seconds. The frames will be output
24156 starting from this seek point. The parameter is evaluated with
24157 @code{av_strtod}, so the numerical value may be suffixed by an IS
24158 postfix. The default value is "0".
24161 Specifies the streams to read. Several streams can be specified,
24162 separated by "+". The source will then have as many outputs, in the
24163 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
24164 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
24165 respectively the default (best suited) video and audio stream. Default
24166 is "dv", or "da" if the filter is called as "amovie".
24168 @item stream_index, si
24169 Specifies the index of the video stream to read. If the value is -1,
24170 the most suitable video stream will be automatically selected. The default
24171 value is "-1". Deprecated. If the filter is called "amovie", it will select
24172 audio instead of video.
24175 Specifies how many times to read the stream in sequence.
24176 If the value is 0, the stream will be looped infinitely.
24177 Default value is "1".
24179 Note that when the movie is looped the source timestamps are not
24180 changed, so it will generate non monotonically increasing timestamps.
24182 @item discontinuity
24183 Specifies the time difference between frames above which the point is
24184 considered a timestamp discontinuity which is removed by adjusting the later
24188 It allows overlaying a second video on top of the main input of
24189 a filtergraph, as shown in this graph:
24191 input -----------> deltapts0 --> overlay --> output
24194 movie --> scale--> deltapts1 -------+
24196 @subsection Examples
24200 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
24201 on top of the input labelled "in":
24203 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
24204 [in] setpts=PTS-STARTPTS [main];
24205 [main][over] overlay=16:16 [out]
24209 Read from a video4linux2 device, and overlay it on top of the input
24212 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
24213 [in] setpts=PTS-STARTPTS [main];
24214 [main][over] overlay=16:16 [out]
24218 Read the first video stream and the audio stream with id 0x81 from
24219 dvd.vob; the video is connected to the pad named "video" and the audio is
24220 connected to the pad named "audio":
24222 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
24226 @subsection Commands
24228 Both movie and amovie support the following commands:
24231 Perform seek using "av_seek_frame".
24232 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
24235 @var{stream_index}: If stream_index is -1, a default
24236 stream is selected, and @var{timestamp} is automatically converted
24237 from AV_TIME_BASE units to the stream specific time_base.
24239 @var{timestamp}: Timestamp in AVStream.time_base units
24240 or, if no stream is specified, in AV_TIME_BASE units.
24242 @var{flags}: Flags which select direction and seeking mode.
24246 Get movie duration in AV_TIME_BASE units.
24250 @c man end MULTIMEDIA SOURCES