1 @chapter Filtering Introduction
2 @c man begin FILTERING INTRODUCTION
4 Filtering in FFmpeg is enabled through the libavfilter library.
6 In libavfilter, a filter can have multiple inputs and multiple
8 To illustrate the sorts of things that are possible, we consider the
13 input --> split ---------------------> overlay --> output
16 +-----> crop --> vflip -------+
19 This filtergraph splits the input stream in two streams, then sends one
20 stream through the crop filter and the vflip filter, before merging it
21 back with the other stream by overlaying it on top. You can use the
22 following command to achieve this:
25 ffmpeg -i INPUT -vf "split [main][tmp]; [tmp] crop=iw:ih/2:0:0, vflip [flip]; [main][flip] overlay=0:H/2" OUTPUT
28 The result will be that the top half of the video is mirrored
29 onto the bottom half of the output video.
31 Filters in the same linear chain are separated by commas, and distinct
32 linear chains of filters are separated by semicolons. In our example,
33 @var{crop,vflip} are in one linear chain, @var{split} and
34 @var{overlay} are separately in another. The points where the linear
35 chains join are labelled by names enclosed in square brackets. In the
36 example, the split filter generates two outputs that are associated to
37 the labels @var{[main]} and @var{[tmp]}.
39 The stream sent to the second output of @var{split}, labelled as
40 @var{[tmp]}, is processed through the @var{crop} filter, which crops
41 away the lower half part of the video, and then vertically flipped. The
42 @var{overlay} filter takes in input the first unchanged output of the
43 split filter (which was labelled as @var{[main]}), and overlay on its
44 lower half the output generated by the @var{crop,vflip} filterchain.
46 Some filters take in input a list of parameters: they are specified
47 after the filter name and an equal sign, and are separated from each other
50 There exist so-called @var{source filters} that do not have an
51 audio/video input, and @var{sink filters} that will not have audio/video
54 @c man end FILTERING INTRODUCTION
57 @c man begin GRAPH2DOT
59 The @file{graph2dot} program included in the FFmpeg @file{tools}
60 directory can be used to parse a filtergraph description and issue a
61 corresponding textual representation in the dot language.
68 to see how to use @file{graph2dot}.
70 You can then pass the dot description to the @file{dot} program (from
71 the graphviz suite of programs) and obtain a graphical representation
74 For example the sequence of commands:
76 echo @var{GRAPH_DESCRIPTION} | \
77 tools/graph2dot -o graph.tmp && \
78 dot -Tpng graph.tmp -o graph.png && \
82 can be used to create and display an image representing the graph
83 described by the @var{GRAPH_DESCRIPTION} string. Note that this string must be
84 a complete self-contained graph, with its inputs and outputs explicitly defined.
85 For example if your command line is of the form:
87 ffmpeg -i infile -vf scale=640:360 outfile
89 your @var{GRAPH_DESCRIPTION} string will need to be of the form:
91 nullsrc,scale=640:360,nullsink
93 you may also need to set the @var{nullsrc} parameters and add a @var{format}
94 filter in order to simulate a specific input file.
98 @chapter Filtergraph description
99 @c man begin FILTERGRAPH DESCRIPTION
101 A filtergraph is a directed graph of connected filters. It can contain
102 cycles, and there can be multiple links between a pair of
103 filters. Each link has one input pad on one side connecting it to one
104 filter from which it takes its input, and one output pad on the other
105 side connecting it to one filter accepting its output.
107 Each filter in a filtergraph is an instance of a filter class
108 registered in the application, which defines the features and the
109 number of input and output pads of the filter.
111 A filter with no input pads is called a "source", and a filter with no
112 output pads is called a "sink".
114 @anchor{Filtergraph syntax}
115 @section Filtergraph syntax
117 A filtergraph has a textual representation, which is recognized by the
118 @option{-filter}/@option{-vf}/@option{-af} and
119 @option{-filter_complex} options in @command{ffmpeg} and
120 @option{-vf}/@option{-af} in @command{ffplay}, and by the
121 @code{avfilter_graph_parse_ptr()} function defined in
122 @file{libavfilter/avfilter.h}.
124 A filterchain consists of a sequence of connected filters, each one
125 connected to the previous one in the sequence. A filterchain is
126 represented by a list of ","-separated filter descriptions.
128 A filtergraph consists of a sequence of filterchains. A sequence of
129 filterchains is represented by a list of ";"-separated filterchain
132 A filter is represented by a string of the form:
133 [@var{in_link_1}]...[@var{in_link_N}]@var{filter_name}@@@var{id}=@var{arguments}[@var{out_link_1}]...[@var{out_link_M}]
135 @var{filter_name} is the name of the filter class of which the
136 described filter is an instance of, and has to be the name of one of
137 the filter classes registered in the program optionally followed by "@@@var{id}".
138 The name of the filter class is optionally followed by a string
141 @var{arguments} is a string which contains the parameters used to
142 initialize the filter instance. It may have one of two forms:
146 A ':'-separated list of @var{key=value} pairs.
149 A ':'-separated list of @var{value}. In this case, the keys are assumed to be
150 the option names in the order they are declared. E.g. the @code{fade} filter
151 declares three options in this order -- @option{type}, @option{start_frame} and
152 @option{nb_frames}. Then the parameter list @var{in:0:30} means that the value
153 @var{in} is assigned to the option @option{type}, @var{0} to
154 @option{start_frame} and @var{30} to @option{nb_frames}.
157 A ':'-separated list of mixed direct @var{value} and long @var{key=value}
158 pairs. The direct @var{value} must precede the @var{key=value} pairs, and
159 follow the same constraints order of the previous point. The following
160 @var{key=value} pairs can be set in any preferred order.
164 If the option value itself is a list of items (e.g. the @code{format} filter
165 takes a list of pixel formats), the items in the list are usually separated by
168 The list of arguments can be quoted using the character @samp{'} as initial
169 and ending mark, and the character @samp{\} for escaping the characters
170 within the quoted text; otherwise the argument string is considered
171 terminated when the next special character (belonging to the set
172 @samp{[]=;,}) is encountered.
174 The name and arguments of the filter are optionally preceded and
175 followed by a list of link labels.
176 A link label allows one to name a link and associate it to a filter output
177 or input pad. The preceding labels @var{in_link_1}
178 ... @var{in_link_N}, are associated to the filter input pads,
179 the following labels @var{out_link_1} ... @var{out_link_M}, are
180 associated to the output pads.
182 When two link labels with the same name are found in the
183 filtergraph, a link between the corresponding input and output pad is
186 If an output pad is not labelled, it is linked by default to the first
187 unlabelled input pad of the next filter in the filterchain.
188 For example in the filterchain
190 nullsrc, split[L1], [L2]overlay, nullsink
192 the split filter instance has two output pads, and the overlay filter
193 instance two input pads. The first output pad of split is labelled
194 "L1", the first input pad of overlay is labelled "L2", and the second
195 output pad of split is linked to the second input pad of overlay,
196 which are both unlabelled.
198 In a filter description, if the input label of the first filter is not
199 specified, "in" is assumed; if the output label of the last filter is not
200 specified, "out" is assumed.
202 In a complete filterchain all the unlabelled filter input and output
203 pads must be connected. A filtergraph is considered valid if all the
204 filter input and output pads of all the filterchains are connected.
206 Libavfilter will automatically insert @ref{scale} filters where format
207 conversion is required. It is possible to specify swscale flags
208 for those automatically inserted scalers by prepending
209 @code{sws_flags=@var{flags};}
210 to the filtergraph description.
212 Here is a BNF description of the filtergraph syntax:
214 @var{NAME} ::= sequence of alphanumeric characters and '_'
215 @var{FILTER_NAME} ::= @var{NAME}["@@"@var{NAME}]
216 @var{LINKLABEL} ::= "[" @var{NAME} "]"
217 @var{LINKLABELS} ::= @var{LINKLABEL} [@var{LINKLABELS}]
218 @var{FILTER_ARGUMENTS} ::= sequence of chars (possibly quoted)
219 @var{FILTER} ::= [@var{LINKLABELS}] @var{FILTER_NAME} ["=" @var{FILTER_ARGUMENTS}] [@var{LINKLABELS}]
220 @var{FILTERCHAIN} ::= @var{FILTER} [,@var{FILTERCHAIN}]
221 @var{FILTERGRAPH} ::= [sws_flags=@var{flags};] @var{FILTERCHAIN} [;@var{FILTERGRAPH}]
224 @anchor{filtergraph escaping}
225 @section Notes on filtergraph escaping
227 Filtergraph description composition entails several levels of
228 escaping. See @ref{quoting_and_escaping,,the "Quoting and escaping"
229 section in the ffmpeg-utils(1) manual,ffmpeg-utils} for more
230 information about the employed escaping procedure.
232 A first level escaping affects the content of each filter option
233 value, which may contain the special character @code{:} used to
234 separate values, or one of the escaping characters @code{\'}.
236 A second level escaping affects the whole filter description, which
237 may contain the escaping characters @code{\'} or the special
238 characters @code{[],;} used by the filtergraph description.
240 Finally, when you specify a filtergraph on a shell commandline, you
241 need to perform a third level escaping for the shell special
242 characters contained within it.
244 For example, consider the following string to be embedded in
245 the @ref{drawtext} filter description @option{text} value:
247 this is a 'string': may contain one, or more, special characters
250 This string contains the @code{'} special escaping character, and the
251 @code{:} special character, so it needs to be escaped in this way:
253 text=this is a \'string\'\: may contain one, or more, special characters
256 A second level of escaping is required when embedding the filter
257 description in a filtergraph description, in order to escape all the
258 filtergraph special characters. Thus the example above becomes:
260 drawtext=text=this is a \\\'string\\\'\\: may contain one\, or more\, special characters
262 (note that in addition to the @code{\'} escaping special characters,
263 also @code{,} needs to be escaped).
265 Finally an additional level of escaping is needed when writing the
266 filtergraph description in a shell command, which depends on the
267 escaping rules of the adopted shell. For example, assuming that
268 @code{\} is special and needs to be escaped with another @code{\}, the
269 previous string will finally result in:
271 -vf "drawtext=text=this is a \\\\\\'string\\\\\\'\\\\: may contain one\\, or more\\, special characters"
274 @chapter Timeline editing
276 Some filters support a generic @option{enable} option. For the filters
277 supporting timeline editing, this option can be set to an expression which is
278 evaluated before sending a frame to the filter. If the evaluation is non-zero,
279 the filter will be enabled, otherwise the frame will be sent unchanged to the
280 next filter in the filtergraph.
282 The expression accepts the following values:
285 timestamp expressed in seconds, NAN if the input timestamp is unknown
288 sequential number of the input frame, starting from 0
291 the position in the file of the input frame, NAN if unknown
295 width and height of the input frame if video
298 Additionally, these filters support an @option{enable} command that can be used
299 to re-define the expression.
301 Like any other filtering option, the @option{enable} option follows the same
304 For example, to enable a blur filter (@ref{smartblur}) from 10 seconds to 3
305 minutes, and a @ref{curves} filter starting at 3 seconds:
307 smartblur = enable='between(t,10,3*60)',
308 curves = enable='gte(t,3)' : preset=cross_process
311 See @code{ffmpeg -filters} to view which filters have timeline support.
313 @c man end FILTERGRAPH DESCRIPTION
316 @chapter Changing options at runtime with a command
318 Some options can be changed during the operation of the filter using
319 a command. These options are marked 'T' on the output of
320 @command{ffmpeg} @option{-h filter=<name of filter>}.
321 The name of the command is the name of the option and the argument is
325 @chapter Options for filters with several inputs (framesync)
326 @c man begin OPTIONS FOR FILTERS WITH SEVERAL INPUTS
328 Some filters with several inputs support a common set of options.
329 These options can only be set by name, not with the short notation.
333 The action to take when EOF is encountered on the secondary input; it accepts
334 one of the following values:
338 Repeat the last frame (the default).
342 Pass the main input through.
346 If set to 1, force the output to terminate when the shortest input
347 terminates. Default value is 0.
350 If set to 1, force the filter to extend the last frame of secondary streams
351 until the end of the primary stream. A value of 0 disables this behavior.
355 @c man end OPTIONS FOR FILTERS WITH SEVERAL INPUTS
357 @chapter Audio Filters
358 @c man begin AUDIO FILTERS
360 When you configure your FFmpeg build, you can disable any of the
361 existing filters using @code{--disable-filters}.
362 The configure output will show the audio filters included in your
365 Below is a description of the currently available audio filters.
369 A compressor is mainly used to reduce the dynamic range of a signal.
370 Especially modern music is mostly compressed at a high ratio to
371 improve the overall loudness. It's done to get the highest attention
372 of a listener, "fatten" the sound and bring more "power" to the track.
373 If a signal is compressed too much it may sound dull or "dead"
374 afterwards or it may start to "pump" (which could be a powerful effect
375 but can also destroy a track completely).
376 The right compression is the key to reach a professional sound and is
377 the high art of mixing and mastering. Because of its complex settings
378 it may take a long time to get the right feeling for this kind of effect.
380 Compression is done by detecting the volume above a chosen level
381 @code{threshold} and dividing it by the factor set with @code{ratio}.
382 So if you set the threshold to -12dB and your signal reaches -6dB a ratio
383 of 2:1 will result in a signal at -9dB. Because an exact manipulation of
384 the signal would cause distortion of the waveform the reduction can be
385 levelled over the time. This is done by setting "Attack" and "Release".
386 @code{attack} determines how long the signal has to rise above the threshold
387 before any reduction will occur and @code{release} sets the time the signal
388 has to fall below the threshold to reduce the reduction again. Shorter signals
389 than the chosen attack time will be left untouched.
390 The overall reduction of the signal can be made up afterwards with the
391 @code{makeup} setting. So compressing the peaks of a signal about 6dB and
392 raising the makeup to this level results in a signal twice as loud than the
393 source. To gain a softer entry in the compression the @code{knee} flattens the
394 hard edge at the threshold in the range of the chosen decibels.
396 The filter accepts the following options:
400 Set input gain. Default is 1. Range is between 0.015625 and 64.
403 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
404 Default is @code{downward}.
407 If a signal of stream rises above this level it will affect the gain
409 By default it is 0.125. Range is between 0.00097563 and 1.
412 Set a ratio by which the signal is reduced. 1:2 means that if the level
413 rose 4dB above the threshold, it will be only 2dB above after the reduction.
414 Default is 2. Range is between 1 and 20.
417 Amount of milliseconds the signal has to rise above the threshold before gain
418 reduction starts. Default is 20. Range is between 0.01 and 2000.
421 Amount of milliseconds the signal has to fall below the threshold before
422 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
425 Set the amount by how much signal will be amplified after processing.
426 Default is 1. Range is from 1 to 64.
429 Curve the sharp knee around the threshold to enter gain reduction more softly.
430 Default is 2.82843. Range is between 1 and 8.
433 Choose if the @code{average} level between all channels of input stream
434 or the louder(@code{maximum}) channel of input stream affects the
435 reduction. Default is @code{average}.
438 Should the exact signal be taken in case of @code{peak} or an RMS one in case
439 of @code{rms}. Default is @code{rms} which is mostly smoother.
442 How much to use compressed signal in output. Default is 1.
443 Range is between 0 and 1.
448 This filter supports the all above options as @ref{commands}.
451 Simple audio dynamic range compression/expansion filter.
453 The filter accepts the following options:
457 Set contrast. Default is 33. Allowed range is between 0 and 100.
462 Copy the input audio source unchanged to the output. This is mainly useful for
467 Apply cross fade from one input audio stream to another input audio stream.
468 The cross fade is applied for specified duration near the end of first stream.
470 The filter accepts the following options:
474 Specify the number of samples for which the cross fade effect has to last.
475 At the end of the cross fade effect the first input audio will be completely
476 silent. Default is 44100.
479 Specify the duration of the cross fade effect. See
480 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
481 for the accepted syntax.
482 By default the duration is determined by @var{nb_samples}.
483 If set this option is used instead of @var{nb_samples}.
486 Should first stream end overlap with second stream start. Default is enabled.
489 Set curve for cross fade transition for first stream.
492 Set curve for cross fade transition for second stream.
494 For description of available curve types see @ref{afade} filter description.
501 Cross fade from one input to another:
503 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
507 Cross fade from one input to another but without overlapping:
509 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
514 Split audio stream into several bands.
516 This filter splits audio stream into two or more frequency ranges.
517 Summing all streams back will give flat output.
519 The filter accepts the following options:
523 Set split frequencies. Those must be positive and increasing.
526 Set filter order for each band split. This controls filter roll-off or steepness
527 of filter transfer function.
528 Available values are:
553 Default is @var{4th}.
556 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
559 Set output gain for each band. Default value is 1 for all bands.
566 Split input audio stream into two bands (low and high) with split frequency of 1500 Hz,
567 each band will be in separate stream:
569 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
573 Same as above, but with higher filter order:
575 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500:order=8th[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
579 Same as above, but also with additional middle band (frequencies between 1500 and 8000):
581 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500 8000:order=8th[LOW][MID][HIGH]' -map '[LOW]' low.wav -map '[MID]' mid.wav -map '[HIGH]' high.wav
587 Reduce audio bit resolution.
589 This filter is bit crusher with enhanced functionality. A bit crusher
590 is used to audibly reduce number of bits an audio signal is sampled
591 with. This doesn't change the bit depth at all, it just produces the
592 effect. Material reduced in bit depth sounds more harsh and "digital".
593 This filter is able to even round to continuous values instead of discrete
595 Additionally it has a D/C offset which results in different crushing of
596 the lower and the upper half of the signal.
597 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
599 Another feature of this filter is the logarithmic mode.
600 This setting switches from linear distances between bits to logarithmic ones.
601 The result is a much more "natural" sounding crusher which doesn't gate low
602 signals for example. The human ear has a logarithmic perception,
603 so this kind of crushing is much more pleasant.
604 Logarithmic crushing is also able to get anti-aliased.
606 The filter accepts the following options:
622 Can be linear: @code{lin} or logarithmic: @code{log}.
631 Set sample reduction.
634 Enable LFO. By default disabled.
645 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
649 Remove impulsive noise from input audio.
651 Samples detected as impulsive noise are replaced by interpolated samples using
652 autoregressive modelling.
656 Set window size, in milliseconds. Allowed range is from @code{10} to
657 @code{100}. Default value is @code{55} milliseconds.
658 This sets size of window which will be processed at once.
661 Set window overlap, in percentage of window size. Allowed range is from
662 @code{50} to @code{95}. Default value is @code{75} percent.
663 Setting this to a very high value increases impulsive noise removal but makes
664 whole process much slower.
667 Set autoregression order, in percentage of window size. Allowed range is from
668 @code{0} to @code{25}. Default value is @code{2} percent. This option also
669 controls quality of interpolated samples using neighbour good samples.
672 Set threshold value. Allowed range is from @code{1} to @code{100}.
673 Default value is @code{2}.
674 This controls the strength of impulsive noise which is going to be removed.
675 The lower value, the more samples will be detected as impulsive noise.
678 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
679 @code{10}. Default value is @code{2}.
680 If any two samples detected as noise are spaced less than this value then any
681 sample between those two samples will be also detected as noise.
686 It accepts the following values:
689 Select overlap-add method. Even not interpolated samples are slightly
690 changed with this method.
693 Select overlap-save method. Not interpolated samples remain unchanged.
696 Default value is @code{a}.
700 Remove clipped samples from input audio.
702 Samples detected as clipped are replaced by interpolated samples using
703 autoregressive modelling.
707 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
708 Default value is @code{55} milliseconds.
709 This sets size of window which will be processed at once.
712 Set window overlap, in percentage of window size. Allowed range is from @code{50}
713 to @code{95}. Default value is @code{75} percent.
716 Set autoregression order, in percentage of window size. Allowed range is from
717 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
718 quality of interpolated samples using neighbour good samples.
721 Set threshold value. Allowed range is from @code{1} to @code{100}.
722 Default value is @code{10}. Higher values make clip detection less aggressive.
725 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
726 Default value is @code{1000}. Higher values make clip detection less aggressive.
731 It accepts the following values:
734 Select overlap-add method. Even not interpolated samples are slightly changed
738 Select overlap-save method. Not interpolated samples remain unchanged.
741 Default value is @code{a}.
746 Delay one or more audio channels.
748 Samples in delayed channel are filled with silence.
750 The filter accepts the following option:
754 Set list of delays in milliseconds for each channel separated by '|'.
755 Unused delays will be silently ignored. If number of given delays is
756 smaller than number of channels all remaining channels will not be delayed.
757 If you want to delay exact number of samples, append 'S' to number.
758 If you want instead to delay in seconds, append 's' to number.
761 Use last set delay for all remaining channels. By default is disabled.
762 This option if enabled changes how option @code{delays} is interpreted.
769 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
770 the second channel (and any other channels that may be present) unchanged.
776 Delay second channel by 500 samples, the third channel by 700 samples and leave
777 the first channel (and any other channels that may be present) unchanged.
783 Delay all channels by same number of samples:
785 adelay=delays=64S:all=1
790 Remedy denormals in audio by adding extremely low-level noise.
792 This filter shall be placed before any filter that can produce denormals.
794 A description of the accepted parameters follows.
798 Set level of added noise in dB. Default is @code{-351}.
799 Allowed range is from -451 to -90.
802 Set type of added noise.
815 Default is @code{dc}.
820 This filter supports the all above options as @ref{commands}.
822 @section aderivative, aintegral
824 Compute derivative/integral of audio stream.
826 Applying both filters one after another produces original audio.
830 Apply echoing to the input audio.
832 Echoes are reflected sound and can occur naturally amongst mountains
833 (and sometimes large buildings) when talking or shouting; digital echo
834 effects emulate this behaviour and are often used to help fill out the
835 sound of a single instrument or vocal. The time difference between the
836 original signal and the reflection is the @code{delay}, and the
837 loudness of the reflected signal is the @code{decay}.
838 Multiple echoes can have different delays and decays.
840 A description of the accepted parameters follows.
844 Set input gain of reflected signal. Default is @code{0.6}.
847 Set output gain of reflected signal. Default is @code{0.3}.
850 Set list of time intervals in milliseconds between original signal and reflections
851 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
852 Default is @code{1000}.
855 Set list of loudness of reflected signals separated by '|'.
856 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
857 Default is @code{0.5}.
864 Make it sound as if there are twice as many instruments as are actually playing:
866 aecho=0.8:0.88:60:0.4
870 If delay is very short, then it sounds like a (metallic) robot playing music:
876 A longer delay will sound like an open air concert in the mountains:
878 aecho=0.8:0.9:1000:0.3
882 Same as above but with one more mountain:
884 aecho=0.8:0.9:1000|1800:0.3|0.25
889 Audio emphasis filter creates or restores material directly taken from LPs or
890 emphased CDs with different filter curves. E.g. to store music on vinyl the
891 signal has to be altered by a filter first to even out the disadvantages of
892 this recording medium.
893 Once the material is played back the inverse filter has to be applied to
894 restore the distortion of the frequency response.
896 The filter accepts the following options:
906 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
907 use @code{production} mode. Default is @code{reproduction} mode.
910 Set filter type. Selects medium. Can be one of the following:
922 select Compact Disc (CD).
928 select 50µs (FM-KF).
930 select 75µs (FM-KF).
936 This filter supports the all above options as @ref{commands}.
940 Modify an audio signal according to the specified expressions.
942 This filter accepts one or more expressions (one for each channel),
943 which are evaluated and used to modify a corresponding audio signal.
945 It accepts the following parameters:
949 Set the '|'-separated expressions list for each separate channel. If
950 the number of input channels is greater than the number of
951 expressions, the last specified expression is used for the remaining
954 @item channel_layout, c
955 Set output channel layout. If not specified, the channel layout is
956 specified by the number of expressions. If set to @samp{same}, it will
957 use by default the same input channel layout.
960 Each expression in @var{exprs} can contain the following constants and functions:
964 channel number of the current expression
967 number of the evaluated sample, starting from 0
973 time of the evaluated sample expressed in seconds
976 @item nb_out_channels
977 input and output number of channels
980 the value of input channel with number @var{CH}
983 Note: this filter is slow. For faster processing you should use a
992 aeval=val(ch)/2:c=same
996 Invert phase of the second channel:
1005 Apply fade-in/out effect to input audio.
1007 A description of the accepted parameters follows.
1011 Specify the effect type, can be either @code{in} for fade-in, or
1012 @code{out} for a fade-out effect. Default is @code{in}.
1014 @item start_sample, ss
1015 Specify the number of the start sample for starting to apply the fade
1016 effect. Default is 0.
1018 @item nb_samples, ns
1019 Specify the number of samples for which the fade effect has to last. At
1020 the end of the fade-in effect the output audio will have the same
1021 volume as the input audio, at the end of the fade-out transition
1022 the output audio will be silence. Default is 44100.
1024 @item start_time, st
1025 Specify the start time of the fade effect. Default is 0.
1026 The value must be specified as a time duration; see
1027 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1028 for the accepted syntax.
1029 If set this option is used instead of @var{start_sample}.
1032 Specify the duration of the fade effect. See
1033 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1034 for the accepted syntax.
1035 At the end of the fade-in effect the output audio will have the same
1036 volume as the input audio, at the end of the fade-out transition
1037 the output audio will be silence.
1038 By default the duration is determined by @var{nb_samples}.
1039 If set this option is used instead of @var{nb_samples}.
1042 Set curve for fade transition.
1044 It accepts the following values:
1047 select triangular, linear slope (default)
1049 select quarter of sine wave
1051 select half of sine wave
1053 select exponential sine wave
1057 select inverted parabola
1071 select inverted quarter of sine wave
1073 select inverted half of sine wave
1075 select double-exponential seat
1077 select double-exponential sigmoid
1079 select logistic sigmoid
1081 select sine cardinal function
1083 select inverted sine cardinal function
1089 @subsection Commands
1091 This filter supports the all above options as @ref{commands}.
1093 @subsection Examples
1097 Fade in first 15 seconds of audio:
1099 afade=t=in:ss=0:d=15
1103 Fade out last 25 seconds of a 900 seconds audio:
1105 afade=t=out:st=875:d=25
1110 Denoise audio samples with FFT.
1112 A description of the accepted parameters follows.
1116 Set the noise reduction in dB, allowed range is 0.01 to 97.
1117 Default value is 12 dB.
1120 Set the noise floor in dB, allowed range is -80 to -20.
1121 Default value is -50 dB.
1126 It accepts the following values:
1135 Select shellac noise.
1138 Select custom noise, defined in @code{bn} option.
1140 Default value is white noise.
1144 Set custom band noise for every one of 15 bands.
1145 Bands are separated by ' ' or '|'.
1148 Set the residual floor in dB, allowed range is -80 to -20.
1149 Default value is -38 dB.
1152 Enable noise tracking. By default is disabled.
1153 With this enabled, noise floor is automatically adjusted.
1156 Enable residual tracking. By default is disabled.
1159 Set the output mode.
1161 It accepts the following values:
1164 Pass input unchanged.
1167 Pass noise filtered out.
1172 Default value is @var{o}.
1176 @subsection Commands
1178 This filter supports the following commands:
1180 @item sample_noise, sn
1181 Start or stop measuring noise profile.
1182 Syntax for the command is : "start" or "stop" string.
1183 After measuring noise profile is stopped it will be
1184 automatically applied in filtering.
1186 @item noise_reduction, nr
1187 Change noise reduction. Argument is single float number.
1188 Syntax for the command is : "@var{noise_reduction}"
1190 @item noise_floor, nf
1191 Change noise floor. Argument is single float number.
1192 Syntax for the command is : "@var{noise_floor}"
1194 @item output_mode, om
1195 Change output mode operation.
1196 Syntax for the command is : "i", "o" or "n" string.
1200 Apply arbitrary expressions to samples in frequency domain.
1204 Set frequency domain real expression for each separate channel separated
1205 by '|'. Default is "re".
1206 If the number of input channels is greater than the number of
1207 expressions, the last specified expression is used for the remaining
1211 Set frequency domain imaginary expression for each separate channel
1212 separated by '|'. Default is "im".
1214 Each expression in @var{real} and @var{imag} can contain the following
1215 constants and functions:
1222 current frequency bin number
1225 number of available bins
1228 channel number of the current expression
1237 current real part of frequency bin of current channel
1240 current imaginary part of frequency bin of current channel
1243 Return the value of real part of frequency bin at location (@var{bin},@var{channel})
1246 Return the value of imaginary part of frequency bin at location (@var{bin},@var{channel})
1250 Set window size. Allowed range is from 16 to 131072.
1251 Default is @code{4096}
1254 Set window function. Default is @code{hann}.
1257 Set window overlap. If set to 1, the recommended overlap for selected
1258 window function will be picked. Default is @code{0.75}.
1261 @subsection Examples
1265 Leave almost only low frequencies in audio:
1267 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1271 Apply robotize effect:
1273 afftfilt="real='hypot(re,im)*sin(0)':imag='hypot(re,im)*cos(0)':win_size=512:overlap=0.75"
1277 Apply whisper effect:
1279 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"
1286 Apply an arbitrary Finite Impulse Response filter.
1288 This filter is designed for applying long FIR filters,
1289 up to 60 seconds long.
1291 It can be used as component for digital crossover filters,
1292 room equalization, cross talk cancellation, wavefield synthesis,
1293 auralization, ambiophonics, ambisonics and spatialization.
1295 This filter uses the streams higher than first one as FIR coefficients.
1296 If the non-first stream holds a single channel, it will be used
1297 for all input channels in the first stream, otherwise
1298 the number of channels in the non-first stream must be same as
1299 the number of channels in the first stream.
1301 It accepts the following parameters:
1305 Set dry gain. This sets input gain.
1308 Set wet gain. This sets final output gain.
1311 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1314 Enable applying gain measured from power of IR.
1316 Set which approach to use for auto gain measurement.
1320 Do not apply any gain.
1323 select peak gain, very conservative approach. This is default value.
1326 select DC gain, limited application.
1329 select gain to noise approach, this is most popular one.
1333 Set gain to be applied to IR coefficients before filtering.
1334 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1337 Set format of IR stream. Can be @code{mono} or @code{input}.
1338 Default is @code{input}.
1341 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1342 Allowed range is 0.1 to 60 seconds.
1345 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1346 By default it is disabled.
1349 Set for which IR channel to display frequency response. By default is first channel
1350 displayed. This option is used only when @var{response} is enabled.
1353 Set video stream size. This option is used only when @var{response} is enabled.
1356 Set video stream frame rate. This option is used only when @var{response} is enabled.
1359 Set minimal partition size used for convolution. Default is @var{8192}.
1360 Allowed range is from @var{1} to @var{32768}.
1361 Lower values decreases latency at cost of higher CPU usage.
1364 Set maximal partition size used for convolution. Default is @var{8192}.
1365 Allowed range is from @var{8} to @var{32768}.
1366 Lower values may increase CPU usage.
1369 Set number of input impulse responses streams which will be switchable at runtime.
1370 Allowed range is from @var{1} to @var{32}. Default is @var{1}.
1373 Set IR stream which will be used for convolution, starting from @var{0}, should always be
1374 lower than supplied value by @code{nbirs} option. Default is @var{0}.
1375 This option can be changed at runtime via @ref{commands}.
1378 @subsection Examples
1382 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1384 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1391 Set output format constraints for the input audio. The framework will
1392 negotiate the most appropriate format to minimize conversions.
1394 It accepts the following parameters:
1397 @item sample_fmts, f
1398 A '|'-separated list of requested sample formats.
1400 @item sample_rates, r
1401 A '|'-separated list of requested sample rates.
1403 @item channel_layouts, cl
1404 A '|'-separated list of requested channel layouts.
1406 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1407 for the required syntax.
1410 If a parameter is omitted, all values are allowed.
1412 Force the output to either unsigned 8-bit or signed 16-bit stereo
1414 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1418 Apply frequency shift to input audio samples.
1420 The filter accepts the following options:
1424 Specify frequency shift. Allowed range is -INT_MAX to INT_MAX.
1425 Default value is 0.0.
1428 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
1429 Default value is 1.0.
1432 @subsection Commands
1434 This filter supports the all above options as @ref{commands}.
1438 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1439 processing reduces disturbing noise between useful signals.
1441 Gating is done by detecting the volume below a chosen level @var{threshold}
1442 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1443 floor is set via @var{range}. Because an exact manipulation of the signal
1444 would cause distortion of the waveform the reduction can be levelled over
1445 time. This is done by setting @var{attack} and @var{release}.
1447 @var{attack} determines how long the signal has to fall below the threshold
1448 before any reduction will occur and @var{release} sets the time the signal
1449 has to rise above the threshold to reduce the reduction again.
1450 Shorter signals than the chosen attack time will be left untouched.
1454 Set input level before filtering.
1455 Default is 1. Allowed range is from 0.015625 to 64.
1458 Set the mode of operation. Can be @code{upward} or @code{downward}.
1459 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
1460 will be amplified, expanding dynamic range in upward direction.
1461 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
1464 Set the level of gain reduction when the signal is below the threshold.
1465 Default is 0.06125. Allowed range is from 0 to 1.
1466 Setting this to 0 disables reduction and then filter behaves like expander.
1469 If a signal rises above this level the gain reduction is released.
1470 Default is 0.125. Allowed range is from 0 to 1.
1473 Set a ratio by which the signal is reduced.
1474 Default is 2. Allowed range is from 1 to 9000.
1477 Amount of milliseconds the signal has to rise above the threshold before gain
1479 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1482 Amount of milliseconds the signal has to fall below the threshold before the
1483 reduction is increased again. Default is 250 milliseconds.
1484 Allowed range is from 0.01 to 9000.
1487 Set amount of amplification of signal after processing.
1488 Default is 1. Allowed range is from 1 to 64.
1491 Curve the sharp knee around the threshold to enter gain reduction more softly.
1492 Default is 2.828427125. Allowed range is from 1 to 8.
1495 Choose if exact signal should be taken for detection or an RMS like one.
1496 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1499 Choose if the average level between all channels or the louder channel affects
1501 Default is @code{average}. Can be @code{average} or @code{maximum}.
1504 @subsection Commands
1506 This filter supports the all above options as @ref{commands}.
1510 Apply an arbitrary Infinite Impulse Response filter.
1512 It accepts the following parameters:
1516 Set B/numerator/zeros/reflection coefficients.
1519 Set A/denominator/poles/ladder coefficients.
1531 Set coefficients format.
1535 lattice-ladder function
1537 analog transfer function
1539 digital transfer function
1541 Z-plane zeros/poles, cartesian (default)
1543 Z-plane zeros/poles, polar radians
1545 Z-plane zeros/poles, polar degrees
1551 Set type of processing.
1563 Set filtering precision.
1567 double-precision floating-point (default)
1569 single-precision floating-point
1577 Normalize filter coefficients, by default is enabled.
1578 Enabling it will normalize magnitude response at DC to 0dB.
1581 How much to use filtered signal in output. Default is 1.
1582 Range is between 0 and 1.
1585 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1586 By default it is disabled.
1589 Set for which IR channel to display frequency response. By default is first channel
1590 displayed. This option is used only when @var{response} is enabled.
1593 Set video stream size. This option is used only when @var{response} is enabled.
1596 Coefficients in @code{tf} and @code{sf} format are separated by spaces and are in ascending
1599 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1600 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1603 Different coefficients and gains can be provided for every channel, in such case
1604 use '|' to separate coefficients or gains. Last provided coefficients will be
1605 used for all remaining channels.
1607 @subsection Examples
1611 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1613 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
1617 Same as above but in @code{zp} format:
1619 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
1623 Apply 3-rd order analog normalized Butterworth low-pass filter, using analog transfer function format:
1625 aiir=z=1.3057 0 0 0:p=1.3057 2.3892 2.1860 1:f=sf:r=d
1631 The limiter prevents an input signal from rising over a desired threshold.
1632 This limiter uses lookahead technology to prevent your signal from distorting.
1633 It means that there is a small delay after the signal is processed. Keep in mind
1634 that the delay it produces is the attack time you set.
1636 The filter accepts the following options:
1640 Set input gain. Default is 1.
1643 Set output gain. Default is 1.
1646 Don't let signals above this level pass the limiter. Default is 1.
1649 The limiter will reach its attenuation level in this amount of time in
1650 milliseconds. Default is 5 milliseconds.
1653 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1654 Default is 50 milliseconds.
1657 When gain reduction is always needed ASC takes care of releasing to an
1658 average reduction level rather than reaching a reduction of 0 in the release
1662 Select how much the release time is affected by ASC, 0 means nearly no changes
1663 in release time while 1 produces higher release times.
1666 Auto level output signal. Default is enabled.
1667 This normalizes audio back to 0dB if enabled.
1670 Depending on picked setting it is recommended to upsample input 2x or 4x times
1671 with @ref{aresample} before applying this filter.
1675 Apply a two-pole all-pass filter with central frequency (in Hz)
1676 @var{frequency}, and filter-width @var{width}.
1677 An all-pass filter changes the audio's frequency to phase relationship
1678 without changing its frequency to amplitude relationship.
1680 The filter accepts the following options:
1684 Set frequency in Hz.
1687 Set method to specify band-width of filter.
1702 Specify the band-width of a filter in width_type units.
1705 How much to use filtered signal in output. Default is 1.
1706 Range is between 0 and 1.
1709 Specify which channels to filter, by default all available are filtered.
1712 Normalize biquad coefficients, by default is disabled.
1713 Enabling it will normalize magnitude response at DC to 0dB.
1716 Set the filter order, can be 1 or 2. Default is 2.
1719 Set transform type of IIR filter.
1728 Set precison of filtering.
1731 Pick automatic sample format depending on surround filters.
1733 Always use signed 16-bit.
1735 Always use signed 32-bit.
1737 Always use float 32-bit.
1739 Always use float 64-bit.
1743 @subsection Commands
1745 This filter supports the following commands:
1748 Change allpass frequency.
1749 Syntax for the command is : "@var{frequency}"
1752 Change allpass width_type.
1753 Syntax for the command is : "@var{width_type}"
1756 Change allpass width.
1757 Syntax for the command is : "@var{width}"
1761 Syntax for the command is : "@var{mix}"
1768 The filter accepts the following options:
1772 Set the number of loops. Setting this value to -1 will result in infinite loops.
1776 Set maximal number of samples. Default is 0.
1779 Set first sample of loop. Default is 0.
1785 Merge two or more audio streams into a single multi-channel stream.
1787 The filter accepts the following options:
1792 Set the number of inputs. Default is 2.
1796 If the channel layouts of the inputs are disjoint, and therefore compatible,
1797 the channel layout of the output will be set accordingly and the channels
1798 will be reordered as necessary. If the channel layouts of the inputs are not
1799 disjoint, the output will have all the channels of the first input then all
1800 the channels of the second input, in that order, and the channel layout of
1801 the output will be the default value corresponding to the total number of
1804 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1805 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1806 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1807 first input, b1 is the first channel of the second input).
1809 On the other hand, if both input are in stereo, the output channels will be
1810 in the default order: a1, a2, b1, b2, and the channel layout will be
1811 arbitrarily set to 4.0, which may or may not be the expected value.
1813 All inputs must have the same sample rate, and format.
1815 If inputs do not have the same duration, the output will stop with the
1818 @subsection Examples
1822 Merge two mono files into a stereo stream:
1824 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1828 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1830 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
1836 Mixes multiple audio inputs into a single output.
1838 Note that this filter only supports float samples (the @var{amerge}
1839 and @var{pan} audio filters support many formats). If the @var{amix}
1840 input has integer samples then @ref{aresample} will be automatically
1841 inserted to perform the conversion to float samples.
1845 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1847 will mix 3 input audio streams to a single output with the same duration as the
1848 first input and a dropout transition time of 3 seconds.
1850 It accepts the following parameters:
1854 The number of inputs. If unspecified, it defaults to 2.
1857 How to determine the end-of-stream.
1861 The duration of the longest input. (default)
1864 The duration of the shortest input.
1867 The duration of the first input.
1871 @item dropout_transition
1872 The transition time, in seconds, for volume renormalization when an input
1873 stream ends. The default value is 2 seconds.
1876 Specify weight of each input audio stream as sequence.
1877 Each weight is separated by space. By default all inputs have same weight.
1880 @subsection Commands
1882 This filter supports the following commands:
1885 Syntax is same as option with same name.
1890 Multiply first audio stream with second audio stream and store result
1891 in output audio stream. Multiplication is done by multiplying each
1892 sample from first stream with sample at same position from second stream.
1894 With this element-wise multiplication one can create amplitude fades and
1895 amplitude modulations.
1897 @section anequalizer
1899 High-order parametric multiband equalizer for each channel.
1901 It accepts the following parameters:
1905 This option string is in format:
1906 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1907 Each equalizer band is separated by '|'.
1911 Set channel number to which equalization will be applied.
1912 If input doesn't have that channel the entry is ignored.
1915 Set central frequency for band.
1916 If input doesn't have that frequency the entry is ignored.
1919 Set band width in Hertz.
1922 Set band gain in dB.
1925 Set filter type for band, optional, can be:
1929 Butterworth, this is default.
1940 With this option activated frequency response of anequalizer is displayed
1944 Set video stream size. Only useful if curves option is activated.
1947 Set max gain that will be displayed. Only useful if curves option is activated.
1948 Setting this to a reasonable value makes it possible to display gain which is derived from
1949 neighbour bands which are too close to each other and thus produce higher gain
1950 when both are activated.
1953 Set frequency scale used to draw frequency response in video output.
1954 Can be linear or logarithmic. Default is logarithmic.
1957 Set color for each channel curve which is going to be displayed in video stream.
1958 This is list of color names separated by space or by '|'.
1959 Unrecognised or missing colors will be replaced by white color.
1962 @subsection Examples
1966 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1967 for first 2 channels using Chebyshev type 1 filter:
1969 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1973 @subsection Commands
1975 This filter supports the following commands:
1978 Alter existing filter parameters.
1979 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1981 @var{fN} is existing filter number, starting from 0, if no such filter is available
1983 @var{freq} set new frequency parameter.
1984 @var{width} set new width parameter in Hertz.
1985 @var{gain} set new gain parameter in dB.
1987 Full filter invocation with asendcmd may look like this:
1988 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1993 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1995 Each sample is adjusted by looking for other samples with similar contexts. This
1996 context similarity is defined by comparing their surrounding patches of size
1997 @option{p}. Patches are searched in an area of @option{r} around the sample.
1999 The filter accepts the following options:
2003 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
2006 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
2007 Default value is 2 milliseconds.
2010 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
2011 Default value is 6 milliseconds.
2014 Set the output mode.
2016 It accepts the following values:
2019 Pass input unchanged.
2022 Pass noise filtered out.
2027 Default value is @var{o}.
2031 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
2034 @subsection Commands
2036 This filter supports the all above options as @ref{commands}.
2039 Apply Normalized Least-Mean-Squares algorithm to the first audio stream using the second audio stream.
2041 This adaptive filter is used to mimic a desired filter by finding the filter coefficients that
2042 relate to producing the least mean square of the error signal (difference between the desired,
2043 2nd input audio stream and the actual signal, the 1st input audio stream).
2045 A description of the accepted options follows.
2058 Set the filter leakage.
2061 It accepts the following values:
2070 Pass filtered samples.
2073 Pass difference between desired and filtered samples.
2075 Default value is @var{o}.
2079 @subsection Examples
2083 One of many usages of this filter is noise reduction, input audio is filtered
2084 with same samples that are delayed by fixed amount, one such example for stereo audio is:
2086 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
2090 @subsection Commands
2092 This filter supports the same commands as options, excluding option @code{order}.
2096 Pass the audio source unchanged to the output.
2100 Pad the end of an audio stream with silence.
2102 This can be used together with @command{ffmpeg} @option{-shortest} to
2103 extend audio streams to the same length as the video stream.
2105 A description of the accepted options follows.
2109 Set silence packet size. Default value is 4096.
2112 Set the number of samples of silence to add to the end. After the
2113 value is reached, the stream is terminated. This option is mutually
2114 exclusive with @option{whole_len}.
2117 Set the minimum total number of samples in the output audio stream. If
2118 the value is longer than the input audio length, silence is added to
2119 the end, until the value is reached. This option is mutually exclusive
2120 with @option{pad_len}.
2123 Specify the duration of samples of silence to add. See
2124 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2125 for the accepted syntax. Used only if set to non-zero value.
2128 Specify the minimum total duration in the output audio stream. See
2129 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2130 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
2131 the input audio length, silence is added to the end, until the value is reached.
2132 This option is mutually exclusive with @option{pad_dur}
2135 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
2136 nor @option{whole_dur} option is set, the filter will add silence to the end of
2137 the input stream indefinitely.
2139 @subsection Examples
2143 Add 1024 samples of silence to the end of the input:
2149 Make sure the audio output will contain at least 10000 samples, pad
2150 the input with silence if required:
2152 apad=whole_len=10000
2156 Use @command{ffmpeg} to pad the audio input with silence, so that the
2157 video stream will always result the shortest and will be converted
2158 until the end in the output file when using the @option{shortest}
2161 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
2166 Add a phasing effect to the input audio.
2168 A phaser filter creates series of peaks and troughs in the frequency spectrum.
2169 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
2171 A description of the accepted parameters follows.
2175 Set input gain. Default is 0.4.
2178 Set output gain. Default is 0.74
2181 Set delay in milliseconds. Default is 3.0.
2184 Set decay. Default is 0.4.
2187 Set modulation speed in Hz. Default is 0.5.
2190 Set modulation type. Default is triangular.
2192 It accepts the following values:
2199 @section aphaseshift
2200 Apply phase shift to input audio samples.
2202 The filter accepts the following options:
2206 Specify phase shift. Allowed range is from -1.0 to 1.0.
2207 Default value is 0.0.
2210 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
2211 Default value is 1.0.
2214 @subsection Commands
2216 This filter supports the all above options as @ref{commands}.
2220 Audio pulsator is something between an autopanner and a tremolo.
2221 But it can produce funny stereo effects as well. Pulsator changes the volume
2222 of the left and right channel based on a LFO (low frequency oscillator) with
2223 different waveforms and shifted phases.
2224 This filter have the ability to define an offset between left and right
2225 channel. An offset of 0 means that both LFO shapes match each other.
2226 The left and right channel are altered equally - a conventional tremolo.
2227 An offset of 50% means that the shape of the right channel is exactly shifted
2228 in phase (or moved backwards about half of the frequency) - pulsator acts as
2229 an autopanner. At 1 both curves match again. Every setting in between moves the
2230 phase shift gapless between all stages and produces some "bypassing" sounds with
2231 sine and triangle waveforms. The more you set the offset near 1 (starting from
2232 the 0.5) the faster the signal passes from the left to the right speaker.
2234 The filter accepts the following options:
2238 Set input gain. By default it is 1. Range is [0.015625 - 64].
2241 Set output gain. By default it is 1. Range is [0.015625 - 64].
2244 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
2245 sawup or sawdown. Default is sine.
2248 Set modulation. Define how much of original signal is affected by the LFO.
2251 Set left channel offset. Default is 0. Allowed range is [0 - 1].
2254 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
2257 Set pulse width. Default is 1. Allowed range is [0 - 2].
2260 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
2263 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
2267 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
2271 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
2272 if timing is set to hz.
2278 Resample the input audio to the specified parameters, using the
2279 libswresample library. If none are specified then the filter will
2280 automatically convert between its input and output.
2282 This filter is also able to stretch/squeeze the audio data to make it match
2283 the timestamps or to inject silence / cut out audio to make it match the
2284 timestamps, do a combination of both or do neither.
2286 The filter accepts the syntax
2287 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
2288 expresses a sample rate and @var{resampler_options} is a list of
2289 @var{key}=@var{value} pairs, separated by ":". See the
2290 @ref{Resampler Options,,"Resampler Options" section in the
2291 ffmpeg-resampler(1) manual,ffmpeg-resampler}
2292 for the complete list of supported options.
2294 @subsection Examples
2298 Resample the input audio to 44100Hz:
2304 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2305 samples per second compensation:
2307 aresample=async=1000
2313 Reverse an audio clip.
2315 Warning: This filter requires memory to buffer the entire clip, so trimming
2318 @subsection Examples
2322 Take the first 5 seconds of a clip, and reverse it.
2324 atrim=end=5,areverse
2330 Reduce noise from speech using Recurrent Neural Networks.
2332 This filter accepts the following options:
2336 Set train model file to load. This option is always required.
2339 Set how much to mix filtered samples into final output.
2340 Allowed range is from -1 to 1. Default value is 1.
2341 Negative values are special, they set how much to keep filtered noise
2342 in the final filter output. Set this option to -1 to hear actual
2343 noise removed from input signal.
2346 @section asetnsamples
2348 Set the number of samples per each output audio frame.
2350 The last output packet may contain a different number of samples, as
2351 the filter will flush all the remaining samples when the input audio
2354 The filter accepts the following options:
2358 @item nb_out_samples, n
2359 Set the number of frames per each output audio frame. The number is
2360 intended as the number of samples @emph{per each channel}.
2361 Default value is 1024.
2364 If set to 1, the filter will pad the last audio frame with zeroes, so
2365 that the last frame will contain the same number of samples as the
2366 previous ones. Default value is 1.
2369 For example, to set the number of per-frame samples to 1234 and
2370 disable padding for the last frame, use:
2372 asetnsamples=n=1234:p=0
2377 Set the sample rate without altering the PCM data.
2378 This will result in a change of speed and pitch.
2380 The filter accepts the following options:
2383 @item sample_rate, r
2384 Set the output sample rate. Default is 44100 Hz.
2389 Show a line containing various information for each input audio frame.
2390 The input audio is not modified.
2392 The shown line contains a sequence of key/value pairs of the form
2393 @var{key}:@var{value}.
2395 The following values are shown in the output:
2399 The (sequential) number of the input frame, starting from 0.
2402 The presentation timestamp of the input frame, in time base units; the time base
2403 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2406 The presentation timestamp of the input frame in seconds.
2409 position of the frame in the input stream, -1 if this information in
2410 unavailable and/or meaningless (for example in case of synthetic audio)
2419 The sample rate for the audio frame.
2422 The number of samples (per channel) in the frame.
2425 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2426 audio, the data is treated as if all the planes were concatenated.
2428 @item plane_checksums
2429 A list of Adler-32 checksums for each data plane.
2433 Apply audio soft clipping.
2435 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2436 along a smooth curve, rather than the abrupt shape of hard-clipping.
2438 This filter accepts the following options:
2442 Set type of soft-clipping.
2444 It accepts the following values:
2458 Set threshold from where to start clipping. Default value is 0dB or 1.
2461 Set gain applied to output. Default value is 0dB or 1.
2464 Set additional parameter which controls sigmoid function.
2467 Set oversampling factor.
2470 @subsection Commands
2472 This filter supports the all above options as @ref{commands}.
2475 Automatic Speech Recognition
2477 This filter uses PocketSphinx for speech recognition. To enable
2478 compilation of this filter, you need to configure FFmpeg with
2479 @code{--enable-pocketsphinx}.
2481 It accepts the following options:
2485 Set sampling rate of input audio. Defaults is @code{16000}.
2486 This need to match speech models, otherwise one will get poor results.
2489 Set dictionary containing acoustic model files.
2492 Set pronunciation dictionary.
2495 Set language model file.
2498 Set language model set.
2501 Set which language model to use.
2504 Set output for log messages.
2507 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2512 Display time domain statistical information about the audio channels.
2513 Statistics are calculated and displayed for each audio channel and,
2514 where applicable, an overall figure is also given.
2516 It accepts the following option:
2519 Short window length in seconds, used for peak and trough RMS measurement.
2520 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2524 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2525 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2528 Available keys for each channel are:
2574 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2575 this @code{lavfi.astats.Overall.Peak_count}.
2577 For description what each key means read below.
2580 Set number of frame after which stats are going to be recalculated.
2581 Default is disabled.
2583 @item measure_perchannel
2584 Select the entries which need to be measured per channel. The metadata keys can
2585 be used as flags, default is @option{all} which measures everything.
2586 @option{none} disables all per channel measurement.
2588 @item measure_overall
2589 Select the entries which need to be measured overall. The metadata keys can
2590 be used as flags, default is @option{all} which measures everything.
2591 @option{none} disables all overall measurement.
2595 A description of each shown parameter follows:
2599 Mean amplitude displacement from zero.
2602 Minimal sample level.
2605 Maximal sample level.
2607 @item Min difference
2608 Minimal difference between two consecutive samples.
2610 @item Max difference
2611 Maximal difference between two consecutive samples.
2613 @item Mean difference
2614 Mean difference between two consecutive samples.
2615 The average of each difference between two consecutive samples.
2617 @item RMS difference
2618 Root Mean Square difference between two consecutive samples.
2622 Standard peak and RMS level measured in dBFS.
2626 Peak and trough values for RMS level measured over a short window.
2629 Standard ratio of peak to RMS level (note: not in dB).
2632 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2633 (i.e. either @var{Min level} or @var{Max level}).
2636 Number of occasions (not the number of samples) that the signal attained either
2637 @var{Min level} or @var{Max level}.
2639 @item Noise floor dB
2640 Minimum local peak measured in dBFS over a short window.
2642 @item Noise floor count
2643 Number of occasions (not the number of samples) that the signal attained
2647 Overall bit depth of audio. Number of bits used for each sample.
2650 Measured dynamic range of audio in dB.
2652 @item Zero crossings
2653 Number of points where the waveform crosses the zero level axis.
2655 @item Zero crossings rate
2656 Rate of Zero crossings and number of audio samples.
2660 Boost subwoofer frequencies.
2662 The filter accepts the following options:
2666 Set dry gain, how much of original signal is kept. Allowed range is from 0 to 1.
2667 Default value is 0.7.
2670 Set wet gain, how much of filtered signal is kept. Allowed range is from 0 to 1.
2671 Default value is 0.7.
2674 Set delay line decay gain value. Allowed range is from 0 to 1.
2675 Default value is 0.7.
2678 Set delay line feedback gain value. Allowed range is from 0 to 1.
2679 Default value is 0.9.
2682 Set cutoff frequency in Hertz. Allowed range is 50 to 900.
2683 Default value is 100.
2686 Set slope amount for cutoff frequency. Allowed range is 0.0001 to 1.
2687 Default value is 0.5.
2690 Set delay. Allowed range is from 1 to 100.
2691 Default value is 20.
2694 @subsection Commands
2696 This filter supports the all above options as @ref{commands}.
2699 Cut subwoofer frequencies.
2701 This filter allows to set custom, steeper
2702 roll off than highpass filter, and thus is able to more attenuate
2703 frequency content in stop-band.
2705 The filter accepts the following options:
2709 Set cutoff frequency in Hertz. Allowed range is 2 to 200.
2710 Default value is 20.
2713 Set filter order. Available values are from 3 to 20.
2714 Default value is 10.
2717 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2720 @subsection Commands
2722 This filter supports the all above options as @ref{commands}.
2725 Cut super frequencies.
2727 The filter accepts the following options:
2731 Set cutoff frequency in Hertz. Allowed range is 20000 to 192000.
2732 Default value is 20000.
2735 Set filter order. Available values are from 3 to 20.
2736 Default value is 10.
2739 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2742 @subsection Commands
2744 This filter supports the all above options as @ref{commands}.
2747 Apply high order Butterworth band-pass filter.
2749 The filter accepts the following options:
2753 Set center frequency in Hertz. Allowed range is 2 to 999999.
2754 Default value is 1000.
2757 Set filter order. Available values are from 4 to 20.
2761 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2764 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2767 @subsection Commands
2769 This filter supports the all above options as @ref{commands}.
2772 Apply high order Butterworth band-stop filter.
2774 The filter accepts the following options:
2778 Set center frequency in Hertz. Allowed range is 2 to 999999.
2779 Default value is 1000.
2782 Set filter order. Available values are from 4 to 20.
2786 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2789 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2792 @subsection Commands
2794 This filter supports the all above options as @ref{commands}.
2800 The filter accepts exactly one parameter, the audio tempo. If not
2801 specified then the filter will assume nominal 1.0 tempo. Tempo must
2802 be in the [0.5, 100.0] range.
2804 Note that tempo greater than 2 will skip some samples rather than
2805 blend them in. If for any reason this is a concern it is always
2806 possible to daisy-chain several instances of atempo to achieve the
2807 desired product tempo.
2809 @subsection Examples
2813 Slow down audio to 80% tempo:
2819 To speed up audio to 300% tempo:
2825 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2827 atempo=sqrt(3),atempo=sqrt(3)
2831 @subsection Commands
2833 This filter supports the following commands:
2836 Change filter tempo scale factor.
2837 Syntax for the command is : "@var{tempo}"
2842 Trim the input so that the output contains one continuous subpart of the input.
2844 It accepts the following parameters:
2847 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2848 sample with the timestamp @var{start} will be the first sample in the output.
2851 Specify time of the first audio sample that will be dropped, i.e. the
2852 audio sample immediately preceding the one with the timestamp @var{end} will be
2853 the last sample in the output.
2856 Same as @var{start}, except this option sets the start timestamp in samples
2860 Same as @var{end}, except this option sets the end timestamp in samples instead
2864 The maximum duration of the output in seconds.
2867 The number of the first sample that should be output.
2870 The number of the first sample that should be dropped.
2873 @option{start}, @option{end}, and @option{duration} are expressed as time
2874 duration specifications; see
2875 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2877 Note that the first two sets of the start/end options and the @option{duration}
2878 option look at the frame timestamp, while the _sample options simply count the
2879 samples that pass through the filter. So start/end_pts and start/end_sample will
2880 give different results when the timestamps are wrong, inexact or do not start at
2881 zero. Also note that this filter does not modify the timestamps. If you wish
2882 to have the output timestamps start at zero, insert the asetpts filter after the
2885 If multiple start or end options are set, this filter tries to be greedy and
2886 keep all samples that match at least one of the specified constraints. To keep
2887 only the part that matches all the constraints at once, chain multiple atrim
2890 The defaults are such that all the input is kept. So it is possible to set e.g.
2891 just the end values to keep everything before the specified time.
2896 Drop everything except the second minute of input:
2898 ffmpeg -i INPUT -af atrim=60:120
2902 Keep only the first 1000 samples:
2904 ffmpeg -i INPUT -af atrim=end_sample=1000
2909 @section axcorrelate
2910 Calculate normalized cross-correlation between two input audio streams.
2912 Resulted samples are always between -1 and 1 inclusive.
2913 If result is 1 it means two input samples are highly correlated in that selected segment.
2914 Result 0 means they are not correlated at all.
2915 If result is -1 it means two input samples are out of phase, which means they cancel each
2918 The filter accepts the following options:
2922 Set size of segment over which cross-correlation is calculated.
2923 Default is 256. Allowed range is from 2 to 131072.
2926 Set algorithm for cross-correlation. Can be @code{slow} or @code{fast}.
2927 Default is @code{slow}. Fast algorithm assumes mean values over any given segment
2928 are always zero and thus need much less calculations to make.
2929 This is generally not true, but is valid for typical audio streams.
2932 @subsection Examples
2936 Calculate correlation between channels in stereo audio stream:
2938 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
2944 Apply a two-pole Butterworth band-pass filter with central
2945 frequency @var{frequency}, and (3dB-point) band-width width.
2946 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2947 instead of the default: constant 0dB peak gain.
2948 The filter roll off at 6dB per octave (20dB per decade).
2950 The filter accepts the following options:
2954 Set the filter's central frequency. Default is @code{3000}.
2957 Constant skirt gain if set to 1. Defaults to 0.
2960 Set method to specify band-width of filter.
2975 Specify the band-width of a filter in width_type units.
2978 How much to use filtered signal in output. Default is 1.
2979 Range is between 0 and 1.
2982 Specify which channels to filter, by default all available are filtered.
2985 Normalize biquad coefficients, by default is disabled.
2986 Enabling it will normalize magnitude response at DC to 0dB.
2989 Set transform type of IIR filter.
2998 Set precison of filtering.
3001 Pick automatic sample format depending on surround filters.
3003 Always use signed 16-bit.
3005 Always use signed 32-bit.
3007 Always use float 32-bit.
3009 Always use float 64-bit.
3013 @subsection Commands
3015 This filter supports the following commands:
3018 Change bandpass frequency.
3019 Syntax for the command is : "@var{frequency}"
3022 Change bandpass width_type.
3023 Syntax for the command is : "@var{width_type}"
3026 Change bandpass width.
3027 Syntax for the command is : "@var{width}"
3030 Change bandpass mix.
3031 Syntax for the command is : "@var{mix}"
3036 Apply a two-pole Butterworth band-reject filter with central
3037 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
3038 The filter roll off at 6dB per octave (20dB per decade).
3040 The filter accepts the following options:
3044 Set the filter's central frequency. Default is @code{3000}.
3047 Set method to specify band-width of filter.
3062 Specify the band-width of a filter in width_type units.
3065 How much to use filtered signal in output. Default is 1.
3066 Range is between 0 and 1.
3069 Specify which channels to filter, by default all available are filtered.
3072 Normalize biquad coefficients, by default is disabled.
3073 Enabling it will normalize magnitude response at DC to 0dB.
3076 Set transform type of IIR filter.
3085 Set precison of filtering.
3088 Pick automatic sample format depending on surround filters.
3090 Always use signed 16-bit.
3092 Always use signed 32-bit.
3094 Always use float 32-bit.
3096 Always use float 64-bit.
3100 @subsection Commands
3102 This filter supports the following commands:
3105 Change bandreject frequency.
3106 Syntax for the command is : "@var{frequency}"
3109 Change bandreject width_type.
3110 Syntax for the command is : "@var{width_type}"
3113 Change bandreject width.
3114 Syntax for the command is : "@var{width}"
3117 Change bandreject mix.
3118 Syntax for the command is : "@var{mix}"
3121 @section bass, lowshelf
3123 Boost or cut the bass (lower) frequencies of the audio using a two-pole
3124 shelving filter with a response similar to that of a standard
3125 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
3127 The filter accepts the following options:
3131 Give the gain at 0 Hz. Its useful range is about -20
3132 (for a large cut) to +20 (for a large boost).
3133 Beware of clipping when using a positive gain.
3136 Set the filter's central frequency and so can be used
3137 to extend or reduce the frequency range to be boosted or cut.
3138 The default value is @code{100} Hz.
3141 Set method to specify band-width of filter.
3156 Determine how steep is the filter's shelf transition.
3159 Set number of poles. Default is 2.
3162 How much to use filtered signal in output. Default is 1.
3163 Range is between 0 and 1.
3166 Specify which channels to filter, by default all available are filtered.
3169 Normalize biquad coefficients, by default is disabled.
3170 Enabling it will normalize magnitude response at DC to 0dB.
3173 Set transform type of IIR filter.
3182 Set precison of filtering.
3185 Pick automatic sample format depending on surround filters.
3187 Always use signed 16-bit.
3189 Always use signed 32-bit.
3191 Always use float 32-bit.
3193 Always use float 64-bit.
3197 @subsection Commands
3199 This filter supports the following commands:
3202 Change bass frequency.
3203 Syntax for the command is : "@var{frequency}"
3206 Change bass width_type.
3207 Syntax for the command is : "@var{width_type}"
3211 Syntax for the command is : "@var{width}"
3215 Syntax for the command is : "@var{gain}"
3219 Syntax for the command is : "@var{mix}"
3224 Apply a biquad IIR filter with the given coefficients.
3225 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
3226 are the numerator and denominator coefficients respectively.
3227 and @var{channels}, @var{c} specify which channels to filter, by default all
3228 available are filtered.
3230 @subsection Commands
3232 This filter supports the following commands:
3240 Change biquad parameter.
3241 Syntax for the command is : "@var{value}"
3244 How much to use filtered signal in output. Default is 1.
3245 Range is between 0 and 1.
3248 Specify which channels to filter, by default all available are filtered.
3251 Normalize biquad coefficients, by default is disabled.
3252 Enabling it will normalize magnitude response at DC to 0dB.
3255 Set transform type of IIR filter.
3264 Set precison of filtering.
3267 Pick automatic sample format depending on surround filters.
3269 Always use signed 16-bit.
3271 Always use signed 32-bit.
3273 Always use float 32-bit.
3275 Always use float 64-bit.
3280 Bauer stereo to binaural transformation, which improves headphone listening of
3281 stereo audio records.
3283 To enable compilation of this filter you need to configure FFmpeg with
3284 @code{--enable-libbs2b}.
3286 It accepts the following parameters:
3290 Pre-defined crossfeed level.
3294 Default level (fcut=700, feed=50).
3297 Chu Moy circuit (fcut=700, feed=60).
3300 Jan Meier circuit (fcut=650, feed=95).
3305 Cut frequency (in Hz).
3314 Remap input channels to new locations.
3316 It accepts the following parameters:
3319 Map channels from input to output. The argument is a '|'-separated list of
3320 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
3321 @var{in_channel} form. @var{in_channel} can be either the name of the input
3322 channel (e.g. FL for front left) or its index in the input channel layout.
3323 @var{out_channel} is the name of the output channel or its index in the output
3324 channel layout. If @var{out_channel} is not given then it is implicitly an
3325 index, starting with zero and increasing by one for each mapping.
3327 @item channel_layout
3328 The channel layout of the output stream.
3331 If no mapping is present, the filter will implicitly map input channels to
3332 output channels, preserving indices.
3334 @subsection Examples
3338 For example, assuming a 5.1+downmix input MOV file,
3340 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
3342 will create an output WAV file tagged as stereo from the downmix channels of
3346 To fix a 5.1 WAV improperly encoded in AAC's native channel order
3348 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
3352 @section channelsplit
3354 Split each channel from an input audio stream into a separate output stream.
3356 It accepts the following parameters:
3358 @item channel_layout
3359 The channel layout of the input stream. The default is "stereo".
3361 A channel layout describing the channels to be extracted as separate output streams
3362 or "all" to extract each input channel as a separate stream. The default is "all".
3364 Choosing channels not present in channel layout in the input will result in an error.
3367 @subsection Examples
3371 For example, assuming a stereo input MP3 file,
3373 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
3375 will create an output Matroska file with two audio streams, one containing only
3376 the left channel and the other the right channel.
3379 Split a 5.1 WAV file into per-channel files:
3381 ffmpeg -i in.wav -filter_complex
3382 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
3383 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
3384 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
3389 Extract only LFE from a 5.1 WAV file:
3391 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
3392 -map '[LFE]' lfe.wav
3397 Add a chorus effect to the audio.
3399 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
3401 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
3402 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
3403 The modulation depth defines the range the modulated delay is played before or after
3404 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
3405 sound tuned around the original one, like in a chorus where some vocals are slightly
3408 It accepts the following parameters:
3411 Set input gain. Default is 0.4.
3414 Set output gain. Default is 0.4.
3417 Set delays. A typical delay is around 40ms to 60ms.
3429 @subsection Examples
3435 chorus=0.7:0.9:55:0.4:0.25:2
3441 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
3445 Fuller sounding chorus with three delays:
3447 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
3452 Compress or expand the audio's dynamic range.
3454 It accepts the following parameters:
3460 A list of times in seconds for each channel over which the instantaneous level
3461 of the input signal is averaged to determine its volume. @var{attacks} refers to
3462 increase of volume and @var{decays} refers to decrease of volume. For most
3463 situations, the attack time (response to the audio getting louder) should be
3464 shorter than the decay time, because the human ear is more sensitive to sudden
3465 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
3466 a typical value for decay is 0.8 seconds.
3467 If specified number of attacks & decays is lower than number of channels, the last
3468 set attack/decay will be used for all remaining channels.
3471 A list of points for the transfer function, specified in dB relative to the
3472 maximum possible signal amplitude. Each key points list must be defined using
3473 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
3474 @code{x0/y0 x1/y1 x2/y2 ....}
3476 The input values must be in strictly increasing order but the transfer function
3477 does not have to be monotonically rising. The point @code{0/0} is assumed but
3478 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
3479 function are @code{-70/-70|-60/-20|1/0}.
3482 Set the curve radius in dB for all joints. It defaults to 0.01.
3485 Set the additional gain in dB to be applied at all points on the transfer
3486 function. This allows for easy adjustment of the overall gain.
3490 Set an initial volume, in dB, to be assumed for each channel when filtering
3491 starts. This permits the user to supply a nominal level initially, so that, for
3492 example, a very large gain is not applied to initial signal levels before the
3493 companding has begun to operate. A typical value for audio which is initially
3494 quiet is -90 dB. It defaults to 0.
3497 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
3498 delayed before being fed to the volume adjuster. Specifying a delay
3499 approximately equal to the attack/decay times allows the filter to effectively
3500 operate in predictive rather than reactive mode. It defaults to 0.
3504 @subsection Examples
3508 Make music with both quiet and loud passages suitable for listening to in a
3511 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
3514 Another example for audio with whisper and explosion parts:
3516 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
3520 A noise gate for when the noise is at a lower level than the signal:
3522 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
3526 Here is another noise gate, this time for when the noise is at a higher level
3527 than the signal (making it, in some ways, similar to squelch):
3529 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3533 2:1 compression starting at -6dB:
3535 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3539 2:1 compression starting at -9dB:
3541 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3545 2:1 compression starting at -12dB:
3547 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3551 2:1 compression starting at -18dB:
3553 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3557 3:1 compression starting at -15dB:
3559 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3565 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3571 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
3575 Hard limiter at -6dB:
3577 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3581 Hard limiter at -12dB:
3583 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3587 Hard noise gate at -35 dB:
3589 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3595 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3599 @section compensationdelay
3601 Compensation Delay Line is a metric based delay to compensate differing
3602 positions of microphones or speakers.
3604 For example, you have recorded guitar with two microphones placed in
3605 different locations. Because the front of sound wave has fixed speed in
3606 normal conditions, the phasing of microphones can vary and depends on
3607 their location and interposition. The best sound mix can be achieved when
3608 these microphones are in phase (synchronized). Note that a distance of
3609 ~30 cm between microphones makes one microphone capture the signal in
3610 antiphase to the other microphone. That makes the final mix sound moody.
3611 This filter helps to solve phasing problems by adding different delays
3612 to each microphone track and make them synchronized.
3614 The best result can be reached when you take one track as base and
3615 synchronize other tracks one by one with it.
3616 Remember that synchronization/delay tolerance depends on sample rate, too.
3617 Higher sample rates will give more tolerance.
3619 The filter accepts the following parameters:
3623 Set millimeters distance. This is compensation distance for fine tuning.
3627 Set cm distance. This is compensation distance for tightening distance setup.
3631 Set meters distance. This is compensation distance for hard distance setup.
3635 Set dry amount. Amount of unprocessed (dry) signal.
3639 Set wet amount. Amount of processed (wet) signal.
3643 Set temperature in degrees Celsius. This is the temperature of the environment.
3648 Apply headphone crossfeed filter.
3650 Crossfeed is the process of blending the left and right channels of stereo
3652 It is mainly used to reduce extreme stereo separation of low frequencies.
3654 The intent is to produce more speaker like sound to the listener.
3656 The filter accepts the following options:
3660 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3661 This sets gain of low shelf filter for side part of stereo image.
3662 Default is -6dB. Max allowed is -30db when strength is set to 1.
3665 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3666 This sets cut off frequency of low shelf filter. Default is cut off near
3667 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3670 Set curve slope of low shelf filter. Default is 0.5.
3671 Allowed range is from 0.01 to 1.
3674 Set input gain. Default is 0.9.
3677 Set output gain. Default is 1.
3680 @subsection Commands
3682 This filter supports the all above options as @ref{commands}.
3684 @section crystalizer
3685 Simple algorithm for audio noise sharpening.
3687 This filter linearly increases differences betweeen each audio sample.
3689 The filter accepts the following options:
3693 Sets the intensity of effect (default: 2.0). Must be in range between -10.0 to 0
3694 (unchanged sound) to 10.0 (maximum effect).
3695 To inverse filtering use negative value.
3698 Enable clipping. By default is enabled.
3701 @subsection Commands
3703 This filter supports the all above options as @ref{commands}.
3706 Apply a DC shift to the audio.
3708 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3709 in the recording chain) from the audio. The effect of a DC offset is reduced
3710 headroom and hence volume. The @ref{astats} filter can be used to determine if
3711 a signal has a DC offset.
3715 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3719 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3720 used to prevent clipping.
3725 Apply de-essing to the audio samples.
3729 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3733 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3737 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3741 Set the output mode.
3743 It accepts the following values:
3746 Pass input unchanged.
3749 Pass ess filtered out.
3754 Default value is @var{o}.
3760 Measure audio dynamic range.
3762 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3763 is found in transition material. And anything less that 8 have very poor dynamics
3764 and is very compressed.
3766 The filter accepts the following options:
3770 Set window length in seconds used to split audio into segments of equal length.
3771 Default is 3 seconds.
3775 Dynamic Audio Normalizer.
3777 This filter applies a certain amount of gain to the input audio in order
3778 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3779 contrast to more "simple" normalization algorithms, the Dynamic Audio
3780 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3781 This allows for applying extra gain to the "quiet" sections of the audio
3782 while avoiding distortions or clipping the "loud" sections. In other words:
3783 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3784 sections, in the sense that the volume of each section is brought to the
3785 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3786 this goal *without* applying "dynamic range compressing". It will retain 100%
3787 of the dynamic range *within* each section of the audio file.
3791 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3792 Default is 500 milliseconds.
3793 The Dynamic Audio Normalizer processes the input audio in small chunks,
3794 referred to as frames. This is required, because a peak magnitude has no
3795 meaning for just a single sample value. Instead, we need to determine the
3796 peak magnitude for a contiguous sequence of sample values. While a "standard"
3797 normalizer would simply use the peak magnitude of the complete file, the
3798 Dynamic Audio Normalizer determines the peak magnitude individually for each
3799 frame. The length of a frame is specified in milliseconds. By default, the
3800 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3801 been found to give good results with most files.
3802 Note that the exact frame length, in number of samples, will be determined
3803 automatically, based on the sampling rate of the individual input audio file.
3806 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3807 number. Default is 31.
3808 Probably the most important parameter of the Dynamic Audio Normalizer is the
3809 @code{window size} of the Gaussian smoothing filter. The filter's window size
3810 is specified in frames, centered around the current frame. For the sake of
3811 simplicity, this must be an odd number. Consequently, the default value of 31
3812 takes into account the current frame, as well as the 15 preceding frames and
3813 the 15 subsequent frames. Using a larger window results in a stronger
3814 smoothing effect and thus in less gain variation, i.e. slower gain
3815 adaptation. Conversely, using a smaller window results in a weaker smoothing
3816 effect and thus in more gain variation, i.e. faster gain adaptation.
3817 In other words, the more you increase this value, the more the Dynamic Audio
3818 Normalizer will behave like a "traditional" normalization filter. On the
3819 contrary, the more you decrease this value, the more the Dynamic Audio
3820 Normalizer will behave like a dynamic range compressor.
3823 Set the target peak value. This specifies the highest permissible magnitude
3824 level for the normalized audio input. This filter will try to approach the
3825 target peak magnitude as closely as possible, but at the same time it also
3826 makes sure that the normalized signal will never exceed the peak magnitude.
3827 A frame's maximum local gain factor is imposed directly by the target peak
3828 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3829 It is not recommended to go above this value.
3832 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3833 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3834 factor for each input frame, i.e. the maximum gain factor that does not
3835 result in clipping or distortion. The maximum gain factor is determined by
3836 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3837 additionally bounds the frame's maximum gain factor by a predetermined
3838 (global) maximum gain factor. This is done in order to avoid excessive gain
3839 factors in "silent" or almost silent frames. By default, the maximum gain
3840 factor is 10.0, For most inputs the default value should be sufficient and
3841 it usually is not recommended to increase this value. Though, for input
3842 with an extremely low overall volume level, it may be necessary to allow even
3843 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3844 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3845 Instead, a "sigmoid" threshold function will be applied. This way, the
3846 gain factors will smoothly approach the threshold value, but never exceed that
3850 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3851 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3852 This means that the maximum local gain factor for each frame is defined
3853 (only) by the frame's highest magnitude sample. This way, the samples can
3854 be amplified as much as possible without exceeding the maximum signal
3855 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3856 Normalizer can also take into account the frame's root mean square,
3857 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3858 determine the power of a time-varying signal. It is therefore considered
3859 that the RMS is a better approximation of the "perceived loudness" than
3860 just looking at the signal's peak magnitude. Consequently, by adjusting all
3861 frames to a constant RMS value, a uniform "perceived loudness" can be
3862 established. If a target RMS value has been specified, a frame's local gain
3863 factor is defined as the factor that would result in exactly that RMS value.
3864 Note, however, that the maximum local gain factor is still restricted by the
3865 frame's highest magnitude sample, in order to prevent clipping.
3868 Enable channels coupling. By default is enabled.
3869 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3870 amount. This means the same gain factor will be applied to all channels, i.e.
3871 the maximum possible gain factor is determined by the "loudest" channel.
3872 However, in some recordings, it may happen that the volume of the different
3873 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3874 In this case, this option can be used to disable the channel coupling. This way,
3875 the gain factor will be determined independently for each channel, depending
3876 only on the individual channel's highest magnitude sample. This allows for
3877 harmonizing the volume of the different channels.
3880 Enable DC bias correction. By default is disabled.
3881 An audio signal (in the time domain) is a sequence of sample values.
3882 In the Dynamic Audio Normalizer these sample values are represented in the
3883 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3884 audio signal, or "waveform", should be centered around the zero point.
3885 That means if we calculate the mean value of all samples in a file, or in a
3886 single frame, then the result should be 0.0 or at least very close to that
3887 value. If, however, there is a significant deviation of the mean value from
3888 0.0, in either positive or negative direction, this is referred to as a
3889 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3890 Audio Normalizer provides optional DC bias correction.
3891 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3892 the mean value, or "DC correction" offset, of each input frame and subtract
3893 that value from all of the frame's sample values which ensures those samples
3894 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3895 boundaries, the DC correction offset values will be interpolated smoothly
3896 between neighbouring frames.
3898 @item altboundary, b
3899 Enable alternative boundary mode. By default is disabled.
3900 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3901 around each frame. This includes the preceding frames as well as the
3902 subsequent frames. However, for the "boundary" frames, located at the very
3903 beginning and at the very end of the audio file, not all neighbouring
3904 frames are available. In particular, for the first few frames in the audio
3905 file, the preceding frames are not known. And, similarly, for the last few
3906 frames in the audio file, the subsequent frames are not known. Thus, the
3907 question arises which gain factors should be assumed for the missing frames
3908 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3909 to deal with this situation. The default boundary mode assumes a gain factor
3910 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3911 "fade out" at the beginning and at the end of the input, respectively.
3914 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3915 By default, the Dynamic Audio Normalizer does not apply "traditional"
3916 compression. This means that signal peaks will not be pruned and thus the
3917 full dynamic range will be retained within each local neighbourhood. However,
3918 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3919 normalization algorithm with a more "traditional" compression.
3920 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3921 (thresholding) function. If (and only if) the compression feature is enabled,
3922 all input frames will be processed by a soft knee thresholding function prior
3923 to the actual normalization process. Put simply, the thresholding function is
3924 going to prune all samples whose magnitude exceeds a certain threshold value.
3925 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3926 value. Instead, the threshold value will be adjusted for each individual
3928 In general, smaller parameters result in stronger compression, and vice versa.
3929 Values below 3.0 are not recommended, because audible distortion may appear.
3932 Set the target threshold value. This specifies the lowest permissible
3933 magnitude level for the audio input which will be normalized.
3934 If input frame volume is above this value frame will be normalized.
3935 Otherwise frame may not be normalized at all. The default value is set
3936 to 0, which means all input frames will be normalized.
3937 This option is mostly useful if digital noise is not wanted to be amplified.
3940 @subsection Commands
3942 This filter supports the all above options as @ref{commands}.
3946 Make audio easier to listen to on headphones.
3948 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3949 so that when listened to on headphones the stereo image is moved from
3950 inside your head (standard for headphones) to outside and in front of
3951 the listener (standard for speakers).
3957 Apply a two-pole peaking equalisation (EQ) filter. With this
3958 filter, the signal-level at and around a selected frequency can
3959 be increased or decreased, whilst (unlike bandpass and bandreject
3960 filters) that at all other frequencies is unchanged.
3962 In order to produce complex equalisation curves, this filter can
3963 be given several times, each with a different central frequency.
3965 The filter accepts the following options:
3969 Set the filter's central frequency in Hz.
3972 Set method to specify band-width of filter.
3987 Specify the band-width of a filter in width_type units.
3990 Set the required gain or attenuation in dB.
3991 Beware of clipping when using a positive gain.
3994 How much to use filtered signal in output. Default is 1.
3995 Range is between 0 and 1.
3998 Specify which channels to filter, by default all available are filtered.
4001 Normalize biquad coefficients, by default is disabled.
4002 Enabling it will normalize magnitude response at DC to 0dB.
4005 Set transform type of IIR filter.
4014 Set precison of filtering.
4017 Pick automatic sample format depending on surround filters.
4019 Always use signed 16-bit.
4021 Always use signed 32-bit.
4023 Always use float 32-bit.
4025 Always use float 64-bit.
4029 @subsection Examples
4032 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
4034 equalizer=f=1000:t=h:width=200:g=-10
4038 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
4040 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
4044 @subsection Commands
4046 This filter supports the following commands:
4049 Change equalizer frequency.
4050 Syntax for the command is : "@var{frequency}"
4053 Change equalizer width_type.
4054 Syntax for the command is : "@var{width_type}"
4057 Change equalizer width.
4058 Syntax for the command is : "@var{width}"
4061 Change equalizer gain.
4062 Syntax for the command is : "@var{gain}"
4065 Change equalizer mix.
4066 Syntax for the command is : "@var{mix}"
4069 @section extrastereo
4071 Linearly increases the difference between left and right channels which
4072 adds some sort of "live" effect to playback.
4074 The filter accepts the following options:
4078 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
4079 (average of both channels), with 1.0 sound will be unchanged, with
4080 -1.0 left and right channels will be swapped.
4083 Enable clipping. By default is enabled.
4086 @subsection Commands
4088 This filter supports the all above options as @ref{commands}.
4090 @section firequalizer
4091 Apply FIR Equalization using arbitrary frequency response.
4093 The filter accepts the following option:
4097 Set gain curve equation (in dB). The expression can contain variables:
4100 the evaluated frequency
4104 channel number, set to 0 when multichannels evaluation is disabled
4106 channel id, see libavutil/channel_layout.h, set to the first channel id when
4107 multichannels evaluation is disabled
4111 channel_layout, see libavutil/channel_layout.h
4116 @item gain_interpolate(f)
4117 interpolate gain on frequency f based on gain_entry
4118 @item cubic_interpolate(f)
4119 same as gain_interpolate, but smoother
4121 This option is also available as command. Default is @code{gain_interpolate(f)}.
4124 Set gain entry for gain_interpolate function. The expression can
4128 store gain entry at frequency f with value g
4130 This option is also available as command.
4133 Set filter delay in seconds. Higher value means more accurate.
4134 Default is @code{0.01}.
4137 Set filter accuracy in Hz. Lower value means more accurate.
4138 Default is @code{5}.
4141 Set window function. Acceptable values are:
4144 rectangular window, useful when gain curve is already smooth
4146 hann window (default)
4152 3-terms continuous 1st derivative nuttall window
4154 minimum 3-terms discontinuous nuttall window
4156 4-terms continuous 1st derivative nuttall window
4158 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
4160 blackman-harris window
4166 If enabled, use fixed number of audio samples. This improves speed when
4167 filtering with large delay. Default is disabled.
4170 Enable multichannels evaluation on gain. Default is disabled.
4173 Enable zero phase mode by subtracting timestamp to compensate delay.
4174 Default is disabled.
4177 Set scale used by gain. Acceptable values are:
4180 linear frequency, linear gain
4182 linear frequency, logarithmic (in dB) gain (default)
4184 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
4186 logarithmic frequency, logarithmic gain
4190 Set file for dumping, suitable for gnuplot.
4193 Set scale for dumpfile. Acceptable values are same with scale option.
4197 Enable 2-channel convolution using complex FFT. This improves speed significantly.
4198 Default is disabled.
4201 Enable minimum phase impulse response. Default is disabled.
4204 @subsection Examples
4209 firequalizer=gain='if(lt(f,1000), 0, -INF)'
4212 lowpass at 1000 Hz with gain_entry:
4214 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
4217 custom equalization:
4219 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
4222 higher delay with zero phase to compensate delay:
4224 firequalizer=delay=0.1:fixed=on:zero_phase=on
4227 lowpass on left channel, highpass on right channel:
4229 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
4230 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
4235 Apply a flanging effect to the audio.
4237 The filter accepts the following options:
4241 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
4244 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
4247 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
4251 Set percentage of delayed signal mixed with original. Range from 0 to 100.
4252 Default value is 71.
4255 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
4258 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
4259 Default value is @var{sinusoidal}.
4262 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
4263 Default value is 25.
4266 Set delay-line interpolation, @var{linear} or @var{quadratic}.
4267 Default is @var{linear}.
4271 Apply Haas effect to audio.
4273 Note that this makes most sense to apply on mono signals.
4274 With this filter applied to mono signals it give some directionality and
4275 stretches its stereo image.
4277 The filter accepts the following options:
4281 Set input level. By default is @var{1}, or 0dB
4284 Set output level. By default is @var{1}, or 0dB.
4287 Set gain applied to side part of signal. By default is @var{1}.
4290 Set kind of middle source. Can be one of the following:
4300 Pick middle part signal of stereo image.
4303 Pick side part signal of stereo image.
4307 Change middle phase. By default is disabled.
4310 Set left channel delay. By default is @var{2.05} milliseconds.
4313 Set left channel balance. By default is @var{-1}.
4316 Set left channel gain. By default is @var{1}.
4319 Change left phase. By default is disabled.
4322 Set right channel delay. By defaults is @var{2.12} milliseconds.
4325 Set right channel balance. By default is @var{1}.
4328 Set right channel gain. By default is @var{1}.
4331 Change right phase. By default is enabled.
4336 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
4337 embedded HDCD codes is expanded into a 20-bit PCM stream.
4339 The filter supports the Peak Extend and Low-level Gain Adjustment features
4340 of HDCD, and detects the Transient Filter flag.
4343 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
4346 When using the filter with wav, note the default encoding for wav is 16-bit,
4347 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
4348 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
4350 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
4351 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
4354 The filter accepts the following options:
4357 @item disable_autoconvert
4358 Disable any automatic format conversion or resampling in the filter graph.
4360 @item process_stereo
4361 Process the stereo channels together. If target_gain does not match between
4362 channels, consider it invalid and use the last valid target_gain.
4365 Set the code detect timer period in ms.
4368 Always extend peaks above -3dBFS even if PE isn't signaled.
4371 Replace audio with a solid tone and adjust the amplitude to signal some
4372 specific aspect of the decoding process. The output file can be loaded in
4373 an audio editor alongside the original to aid analysis.
4375 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
4382 Gain adjustment level at each sample
4384 Samples where peak extend occurs
4386 Samples where the code detect timer is active
4388 Samples where the target gain does not match between channels
4394 Apply head-related transfer functions (HRTFs) to create virtual
4395 loudspeakers around the user for binaural listening via headphones.
4396 The HRIRs are provided via additional streams, for each channel
4397 one stereo input stream is needed.
4399 The filter accepts the following options:
4403 Set mapping of input streams for convolution.
4404 The argument is a '|'-separated list of channel names in order as they
4405 are given as additional stream inputs for filter.
4406 This also specify number of input streams. Number of input streams
4407 must be not less than number of channels in first stream plus one.
4410 Set gain applied to audio. Value is in dB. Default is 0.
4413 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4414 processing audio in time domain which is slow.
4415 @var{freq} is processing audio in frequency domain which is fast.
4416 Default is @var{freq}.
4419 Set custom gain for LFE channels. Value is in dB. Default is 0.
4422 Set size of frame in number of samples which will be processed at once.
4423 Default value is @var{1024}. Allowed range is from 1024 to 96000.
4426 Set format of hrir stream.
4427 Default value is @var{stereo}. Alternative value is @var{multich}.
4428 If value is set to @var{stereo}, number of additional streams should
4429 be greater or equal to number of input channels in first input stream.
4430 Also each additional stream should have stereo number of channels.
4431 If value is set to @var{multich}, number of additional streams should
4432 be exactly one. Also number of input channels of additional stream
4433 should be equal or greater than twice number of channels of first input
4437 @subsection Examples
4441 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4442 each amovie filter use stereo file with IR coefficients as input.
4443 The files give coefficients for each position of virtual loudspeaker:
4446 -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"
4451 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4452 but now in @var{multich} @var{hrir} format.
4454 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"
4461 Apply a high-pass filter with 3dB point frequency.
4462 The filter can be either single-pole, or double-pole (the default).
4463 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4465 The filter accepts the following options:
4469 Set frequency in Hz. Default is 3000.
4472 Set number of poles. Default is 2.
4475 Set method to specify band-width of filter.
4490 Specify the band-width of a filter in width_type units.
4491 Applies only to double-pole filter.
4492 The default is 0.707q and gives a Butterworth response.
4495 How much to use filtered signal in output. Default is 1.
4496 Range is between 0 and 1.
4499 Specify which channels to filter, by default all available are filtered.
4502 Normalize biquad coefficients, by default is disabled.
4503 Enabling it will normalize magnitude response at DC to 0dB.
4506 Set transform type of IIR filter.
4515 Set precison of filtering.
4518 Pick automatic sample format depending on surround filters.
4520 Always use signed 16-bit.
4522 Always use signed 32-bit.
4524 Always use float 32-bit.
4526 Always use float 64-bit.
4530 @subsection Commands
4532 This filter supports the following commands:
4535 Change highpass frequency.
4536 Syntax for the command is : "@var{frequency}"
4539 Change highpass width_type.
4540 Syntax for the command is : "@var{width_type}"
4543 Change highpass width.
4544 Syntax for the command is : "@var{width}"
4547 Change highpass mix.
4548 Syntax for the command is : "@var{mix}"
4553 Join multiple input streams into one multi-channel stream.
4555 It accepts the following parameters:
4559 The number of input streams. It defaults to 2.
4561 @item channel_layout
4562 The desired output channel layout. It defaults to stereo.
4565 Map channels from inputs to output. The argument is a '|'-separated list of
4566 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4567 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4568 can be either the name of the input channel (e.g. FL for front left) or its
4569 index in the specified input stream. @var{out_channel} is the name of the output
4573 The filter will attempt to guess the mappings when they are not specified
4574 explicitly. It does so by first trying to find an unused matching input channel
4575 and if that fails it picks the first unused input channel.
4577 Join 3 inputs (with properly set channel layouts):
4579 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4582 Build a 5.1 output from 6 single-channel streams:
4584 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4585 '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'
4591 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4593 To enable compilation of this filter you need to configure FFmpeg with
4594 @code{--enable-ladspa}.
4598 Specifies the name of LADSPA plugin library to load. If the environment
4599 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4600 each one of the directories specified by the colon separated list in
4601 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4602 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4603 @file{/usr/lib/ladspa/}.
4606 Specifies the plugin within the library. Some libraries contain only
4607 one plugin, but others contain many of them. If this is not set filter
4608 will list all available plugins within the specified library.
4611 Set the '|' separated list of controls which are zero or more floating point
4612 values that determine the behavior of the loaded plugin (for example delay,
4614 Controls need to be defined using the following syntax:
4615 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4616 @var{valuei} is the value set on the @var{i}-th control.
4617 Alternatively they can be also defined using the following syntax:
4618 @var{value0}|@var{value1}|@var{value2}|..., where
4619 @var{valuei} is the value set on the @var{i}-th control.
4620 If @option{controls} is set to @code{help}, all available controls and
4621 their valid ranges are printed.
4623 @item sample_rate, s
4624 Specify the sample rate, default to 44100. Only used if plugin have
4628 Set the number of samples per channel per each output frame, default
4629 is 1024. Only used if plugin have zero inputs.
4632 Set the minimum duration of the sourced audio. See
4633 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4634 for the accepted syntax.
4635 Note that the resulting duration may be greater than the specified duration,
4636 as the generated audio is always cut at the end of a complete frame.
4637 If not specified, or the expressed duration is negative, the audio is
4638 supposed to be generated forever.
4639 Only used if plugin have zero inputs.
4642 Enable latency compensation, by default is disabled.
4643 Only used if plugin have inputs.
4646 @subsection Examples
4650 List all available plugins within amp (LADSPA example plugin) library:
4656 List all available controls and their valid ranges for @code{vcf_notch}
4657 plugin from @code{VCF} library:
4659 ladspa=f=vcf:p=vcf_notch:c=help
4663 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4666 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4670 Add reverberation to the audio using TAP-plugins
4671 (Tom's Audio Processing plugins):
4673 ladspa=file=tap_reverb:tap_reverb
4677 Generate white noise, with 0.2 amplitude:
4679 ladspa=file=cmt:noise_source_white:c=c0=.2
4683 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4684 @code{C* Audio Plugin Suite} (CAPS) library:
4686 ladspa=file=caps:Click:c=c1=20'
4690 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4692 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4696 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4697 @code{SWH Plugins} collection:
4699 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4703 Attenuate low frequencies using Multiband EQ from Steve Harris
4704 @code{SWH Plugins} collection:
4706 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4710 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4713 ladspa=caps:Narrower
4717 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4719 ladspa=caps:White:.2
4723 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4725 ladspa=caps:Fractal:c=c1=1
4729 Dynamic volume normalization using @code{VLevel} plugin:
4731 ladspa=vlevel-ladspa:vlevel_mono
4735 @subsection Commands
4737 This filter supports the following commands:
4740 Modify the @var{N}-th control value.
4742 If the specified value is not valid, it is ignored and prior one is kept.
4747 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4748 Support for both single pass (livestreams, files) and double pass (files) modes.
4749 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4750 detect true peaks, the audio stream will be upsampled to 192 kHz.
4751 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4753 The filter accepts the following options:
4757 Set integrated loudness target.
4758 Range is -70.0 - -5.0. Default value is -24.0.
4761 Set loudness range target.
4762 Range is 1.0 - 20.0. Default value is 7.0.
4765 Set maximum true peak.
4766 Range is -9.0 - +0.0. Default value is -2.0.
4768 @item measured_I, measured_i
4769 Measured IL of input file.
4770 Range is -99.0 - +0.0.
4772 @item measured_LRA, measured_lra
4773 Measured LRA of input file.
4774 Range is 0.0 - 99.0.
4776 @item measured_TP, measured_tp
4777 Measured true peak of input file.
4778 Range is -99.0 - +99.0.
4780 @item measured_thresh
4781 Measured threshold of input file.
4782 Range is -99.0 - +0.0.
4785 Set offset gain. Gain is applied before the true-peak limiter.
4786 Range is -99.0 - +99.0. Default is +0.0.
4789 Normalize by linearly scaling the source audio.
4790 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4791 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4792 be lower than source LRA and the change in integrated loudness shouldn't
4793 result in a true peak which exceeds the target TP. If any of these
4794 conditions aren't met, normalization mode will revert to @var{dynamic}.
4795 Options are @code{true} or @code{false}. Default is @code{true}.
4798 Treat mono input files as "dual-mono". If a mono file is intended for playback
4799 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4800 If set to @code{true}, this option will compensate for this effect.
4801 Multi-channel input files are not affected by this option.
4802 Options are true or false. Default is false.
4805 Set print format for stats. Options are summary, json, or none.
4806 Default value is none.
4811 Apply a low-pass filter with 3dB point frequency.
4812 The filter can be either single-pole or double-pole (the default).
4813 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4815 The filter accepts the following options:
4819 Set frequency in Hz. Default is 500.
4822 Set number of poles. Default is 2.
4825 Set method to specify band-width of filter.
4840 Specify the band-width of a filter in width_type units.
4841 Applies only to double-pole filter.
4842 The default is 0.707q and gives a Butterworth response.
4845 How much to use filtered signal in output. Default is 1.
4846 Range is between 0 and 1.
4849 Specify which channels to filter, by default all available are filtered.
4852 Normalize biquad coefficients, by default is disabled.
4853 Enabling it will normalize magnitude response at DC to 0dB.
4856 Set transform type of IIR filter.
4865 Set precison of filtering.
4868 Pick automatic sample format depending on surround filters.
4870 Always use signed 16-bit.
4872 Always use signed 32-bit.
4874 Always use float 32-bit.
4876 Always use float 64-bit.
4880 @subsection Examples
4883 Lowpass only LFE channel, it LFE is not present it does nothing:
4889 @subsection Commands
4891 This filter supports the following commands:
4894 Change lowpass frequency.
4895 Syntax for the command is : "@var{frequency}"
4898 Change lowpass width_type.
4899 Syntax for the command is : "@var{width_type}"
4902 Change lowpass width.
4903 Syntax for the command is : "@var{width}"
4907 Syntax for the command is : "@var{mix}"
4912 Load a LV2 (LADSPA Version 2) plugin.
4914 To enable compilation of this filter you need to configure FFmpeg with
4915 @code{--enable-lv2}.
4919 Specifies the plugin URI. You may need to escape ':'.
4922 Set the '|' separated list of controls which are zero or more floating point
4923 values that determine the behavior of the loaded plugin (for example delay,
4925 If @option{controls} is set to @code{help}, all available controls and
4926 their valid ranges are printed.
4928 @item sample_rate, s
4929 Specify the sample rate, default to 44100. Only used if plugin have
4933 Set the number of samples per channel per each output frame, default
4934 is 1024. Only used if plugin have zero inputs.
4937 Set the minimum duration of the sourced audio. See
4938 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4939 for the accepted syntax.
4940 Note that the resulting duration may be greater than the specified duration,
4941 as the generated audio is always cut at the end of a complete frame.
4942 If not specified, or the expressed duration is negative, the audio is
4943 supposed to be generated forever.
4944 Only used if plugin have zero inputs.
4947 @subsection Examples
4951 Apply bass enhancer plugin from Calf:
4953 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4957 Apply vinyl plugin from Calf:
4959 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4963 Apply bit crusher plugin from ArtyFX:
4965 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4970 Multiband Compress or expand the audio's dynamic range.
4972 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4973 This is akin to the crossover of a loudspeaker, and results in flat frequency
4974 response when absent compander action.
4976 It accepts the following parameters:
4980 This option syntax is:
4981 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4982 For explanation of each item refer to compand filter documentation.
4988 Mix channels with specific gain levels. The filter accepts the output
4989 channel layout followed by a set of channels definitions.
4991 This filter is also designed to efficiently remap the channels of an audio
4994 The filter accepts parameters of the form:
4995 "@var{l}|@var{outdef}|@var{outdef}|..."
4999 output channel layout or number of channels
5002 output channel specification, of the form:
5003 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
5006 output channel to define, either a channel name (FL, FR, etc.) or a channel
5007 number (c0, c1, etc.)
5010 multiplicative coefficient for the channel, 1 leaving the volume unchanged
5013 input channel to use, see out_name for details; it is not possible to mix
5014 named and numbered input channels
5017 If the `=' in a channel specification is replaced by `<', then the gains for
5018 that specification will be renormalized so that the total is 1, thus
5019 avoiding clipping noise.
5021 @subsection Mixing examples
5023 For example, if you want to down-mix from stereo to mono, but with a bigger
5024 factor for the left channel:
5026 pan=1c|c0=0.9*c0+0.1*c1
5029 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
5030 7-channels surround:
5032 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
5035 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
5036 that should be preferred (see "-ac" option) unless you have very specific
5039 @subsection Remapping examples
5041 The channel remapping will be effective if, and only if:
5044 @item gain coefficients are zeroes or ones,
5045 @item only one input per channel output,
5048 If all these conditions are satisfied, the filter will notify the user ("Pure
5049 channel mapping detected"), and use an optimized and lossless method to do the
5052 For example, if you have a 5.1 source and want a stereo audio stream by
5053 dropping the extra channels:
5055 pan="stereo| c0=FL | c1=FR"
5058 Given the same source, you can also switch front left and front right channels
5059 and keep the input channel layout:
5061 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
5064 If the input is a stereo audio stream, you can mute the front left channel (and
5065 still keep the stereo channel layout) with:
5070 Still with a stereo audio stream input, you can copy the right channel in both
5071 front left and right:
5073 pan="stereo| c0=FR | c1=FR"
5078 ReplayGain scanner filter. This filter takes an audio stream as an input and
5079 outputs it unchanged.
5080 At end of filtering it displays @code{track_gain} and @code{track_peak}.
5084 Convert the audio sample format, sample rate and channel layout. It is
5085 not meant to be used directly.
5088 Apply time-stretching and pitch-shifting with librubberband.
5090 To enable compilation of this filter, you need to configure FFmpeg with
5091 @code{--enable-librubberband}.
5093 The filter accepts the following options:
5097 Set tempo scale factor.
5100 Set pitch scale factor.
5103 Set transients detector.
5104 Possible values are:
5113 Possible values are:
5122 Possible values are:
5129 Set processing window size.
5130 Possible values are:
5139 Possible values are:
5146 Enable formant preservation when shift pitching.
5147 Possible values are:
5155 Possible values are:
5164 Possible values are:
5171 @subsection Commands
5173 This filter supports the following commands:
5176 Change filter tempo scale factor.
5177 Syntax for the command is : "@var{tempo}"
5180 Change filter pitch scale factor.
5181 Syntax for the command is : "@var{pitch}"
5184 @section sidechaincompress
5186 This filter acts like normal compressor but has the ability to compress
5187 detected signal using second input signal.
5188 It needs two input streams and returns one output stream.
5189 First input stream will be processed depending on second stream signal.
5190 The filtered signal then can be filtered with other filters in later stages of
5191 processing. See @ref{pan} and @ref{amerge} filter.
5193 The filter accepts the following options:
5197 Set input gain. Default is 1. Range is between 0.015625 and 64.
5200 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
5201 Default is @code{downward}.
5204 If a signal of second stream raises above this level it will affect the gain
5205 reduction of first stream.
5206 By default is 0.125. Range is between 0.00097563 and 1.
5209 Set a ratio about which the signal is reduced. 1:2 means that if the level
5210 raised 4dB above the threshold, it will be only 2dB above after the reduction.
5211 Default is 2. Range is between 1 and 20.
5214 Amount of milliseconds the signal has to rise above the threshold before gain
5215 reduction starts. Default is 20. Range is between 0.01 and 2000.
5218 Amount of milliseconds the signal has to fall below the threshold before
5219 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
5222 Set the amount by how much signal will be amplified after processing.
5223 Default is 1. Range is from 1 to 64.
5226 Curve the sharp knee around the threshold to enter gain reduction more softly.
5227 Default is 2.82843. Range is between 1 and 8.
5230 Choose if the @code{average} level between all channels of side-chain stream
5231 or the louder(@code{maximum}) channel of side-chain stream affects the
5232 reduction. Default is @code{average}.
5235 Should the exact signal be taken in case of @code{peak} or an RMS one in case
5236 of @code{rms}. Default is @code{rms} which is mainly smoother.
5239 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
5242 How much to use compressed signal in output. Default is 1.
5243 Range is between 0 and 1.
5246 @subsection Commands
5248 This filter supports the all above options as @ref{commands}.
5250 @subsection Examples
5254 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
5255 depending on the signal of 2nd input and later compressed signal to be
5256 merged with 2nd input:
5258 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
5262 @section sidechaingate
5264 A sidechain gate acts like a normal (wideband) gate but has the ability to
5265 filter the detected signal before sending it to the gain reduction stage.
5266 Normally a gate uses the full range signal to detect a level above the
5268 For example: If you cut all lower frequencies from your sidechain signal
5269 the gate will decrease the volume of your track only if not enough highs
5270 appear. With this technique you are able to reduce the resonation of a
5271 natural drum or remove "rumbling" of muted strokes from a heavily distorted
5273 It needs two input streams and returns one output stream.
5274 First input stream will be processed depending on second stream signal.
5276 The filter accepts the following options:
5280 Set input level before filtering.
5281 Default is 1. Allowed range is from 0.015625 to 64.
5284 Set the mode of operation. Can be @code{upward} or @code{downward}.
5285 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
5286 will be amplified, expanding dynamic range in upward direction.
5287 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
5290 Set the level of gain reduction when the signal is below the threshold.
5291 Default is 0.06125. Allowed range is from 0 to 1.
5292 Setting this to 0 disables reduction and then filter behaves like expander.
5295 If a signal rises above this level the gain reduction is released.
5296 Default is 0.125. Allowed range is from 0 to 1.
5299 Set a ratio about which the signal is reduced.
5300 Default is 2. Allowed range is from 1 to 9000.
5303 Amount of milliseconds the signal has to rise above the threshold before gain
5305 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
5308 Amount of milliseconds the signal has to fall below the threshold before the
5309 reduction is increased again. Default is 250 milliseconds.
5310 Allowed range is from 0.01 to 9000.
5313 Set amount of amplification of signal after processing.
5314 Default is 1. Allowed range is from 1 to 64.
5317 Curve the sharp knee around the threshold to enter gain reduction more softly.
5318 Default is 2.828427125. Allowed range is from 1 to 8.
5321 Choose if exact signal should be taken for detection or an RMS like one.
5322 Default is rms. Can be peak or rms.
5325 Choose if the average level between all channels or the louder channel affects
5327 Default is average. Can be average or maximum.
5330 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
5333 @subsection Commands
5335 This filter supports the all above options as @ref{commands}.
5337 @section silencedetect
5339 Detect silence in an audio stream.
5341 This filter logs a message when it detects that the input audio volume is less
5342 or equal to a noise tolerance value for a duration greater or equal to the
5343 minimum detected noise duration.
5345 The printed times and duration are expressed in seconds. The
5346 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
5347 is set on the first frame whose timestamp equals or exceeds the detection
5348 duration and it contains the timestamp of the first frame of the silence.
5350 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
5351 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
5352 keys are set on the first frame after the silence. If @option{mono} is
5353 enabled, and each channel is evaluated separately, the @code{.X}
5354 suffixed keys are used, and @code{X} corresponds to the channel number.
5356 The filter accepts the following options:
5360 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
5361 specified value) or amplitude ratio. Default is -60dB, or 0.001.
5364 Set silence duration until notification (default is 2 seconds). See
5365 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5366 for the accepted syntax.
5369 Process each channel separately, instead of combined. By default is disabled.
5372 @subsection Examples
5376 Detect 5 seconds of silence with -50dB noise tolerance:
5378 silencedetect=n=-50dB:d=5
5382 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
5383 tolerance in @file{silence.mp3}:
5385 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
5389 @section silenceremove
5391 Remove silence from the beginning, middle or end of the audio.
5393 The filter accepts the following options:
5397 This value is used to indicate if audio should be trimmed at beginning of
5398 the audio. A value of zero indicates no silence should be trimmed from the
5399 beginning. When specifying a non-zero value, it trims audio up until it
5400 finds non-silence. Normally, when trimming silence from beginning of audio
5401 the @var{start_periods} will be @code{1} but it can be increased to higher
5402 values to trim all audio up to specific count of non-silence periods.
5403 Default value is @code{0}.
5405 @item start_duration
5406 Specify the amount of time that non-silence must be detected before it stops
5407 trimming audio. By increasing the duration, bursts of noises can be treated
5408 as silence and trimmed off. Default value is @code{0}.
5410 @item start_threshold
5411 This indicates what sample value should be treated as silence. For digital
5412 audio, a value of @code{0} may be fine but for audio recorded from analog,
5413 you may wish to increase the value to account for background noise.
5414 Can be specified in dB (in case "dB" is appended to the specified value)
5415 or amplitude ratio. Default value is @code{0}.
5418 Specify max duration of silence at beginning that will be kept after
5419 trimming. Default is 0, which is equal to trimming all samples detected
5423 Specify mode of detection of silence end in start of multi-channel audio.
5424 Can be @var{any} or @var{all}. Default is @var{any}.
5425 With @var{any}, any sample that is detected as non-silence will cause
5426 stopped trimming of silence.
5427 With @var{all}, only if all channels are detected as non-silence will cause
5428 stopped trimming of silence.
5431 Set the count for trimming silence from the end of audio.
5432 To remove silence from the middle of a file, specify a @var{stop_periods}
5433 that is negative. This value is then treated as a positive value and is
5434 used to indicate the effect should restart processing as specified by
5435 @var{start_periods}, making it suitable for removing periods of silence
5436 in the middle of the audio.
5437 Default value is @code{0}.
5440 Specify a duration of silence that must exist before audio is not copied any
5441 more. By specifying a higher duration, silence that is wanted can be left in
5443 Default value is @code{0}.
5445 @item stop_threshold
5446 This is the same as @option{start_threshold} but for trimming silence from
5448 Can be specified in dB (in case "dB" is appended to the specified value)
5449 or amplitude ratio. Default value is @code{0}.
5452 Specify max duration of silence at end that will be kept after
5453 trimming. Default is 0, which is equal to trimming all samples detected
5457 Specify mode of detection of silence start in end of multi-channel audio.
5458 Can be @var{any} or @var{all}. Default is @var{any}.
5459 With @var{any}, any sample that is detected as non-silence will cause
5460 stopped trimming of silence.
5461 With @var{all}, only if all channels are detected as non-silence will cause
5462 stopped trimming of silence.
5465 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
5466 and works better with digital silence which is exactly 0.
5467 Default value is @code{rms}.
5470 Set duration in number of seconds used to calculate size of window in number
5471 of samples for detecting silence.
5472 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
5475 @subsection Examples
5479 The following example shows how this filter can be used to start a recording
5480 that does not contain the delay at the start which usually occurs between
5481 pressing the record button and the start of the performance:
5483 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
5487 Trim all silence encountered from beginning to end where there is more than 1
5488 second of silence in audio:
5490 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
5494 Trim all digital silence samples, using peak detection, from beginning to end
5495 where there is more than 0 samples of digital silence in audio and digital
5496 silence is detected in all channels at same positions in stream:
5498 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
5504 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
5505 loudspeakers around the user for binaural listening via headphones (audio
5506 formats up to 9 channels supported).
5507 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
5508 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
5509 Austrian Academy of Sciences.
5511 To enable compilation of this filter you need to configure FFmpeg with
5512 @code{--enable-libmysofa}.
5514 The filter accepts the following options:
5518 Set the SOFA file used for rendering.
5521 Set gain applied to audio. Value is in dB. Default is 0.
5524 Set rotation of virtual loudspeakers in deg. Default is 0.
5527 Set elevation of virtual speakers in deg. Default is 0.
5530 Set distance in meters between loudspeakers and the listener with near-field
5531 HRTFs. Default is 1.
5534 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
5535 processing audio in time domain which is slow.
5536 @var{freq} is processing audio in frequency domain which is fast.
5537 Default is @var{freq}.
5540 Set custom positions of virtual loudspeakers. Syntax for this option is:
5541 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
5542 Each virtual loudspeaker is described with short channel name following with
5543 azimuth and elevation in degrees.
5544 Each virtual loudspeaker description is separated by '|'.
5545 For example to override front left and front right channel positions use:
5546 'speakers=FL 45 15|FR 345 15'.
5547 Descriptions with unrecognised channel names are ignored.
5550 Set custom gain for LFE channels. Value is in dB. Default is 0.
5553 Set custom frame size in number of samples. Default is 1024.
5554 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5555 is set to @var{freq}.
5558 Should all IRs be normalized upon importing SOFA file.
5559 By default is enabled.
5562 Should nearest IRs be interpolated with neighbor IRs if exact position
5563 does not match. By default is disabled.
5566 Minphase all IRs upon loading of SOFA file. By default is disabled.
5569 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5572 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5575 @subsection Examples
5579 Using ClubFritz6 sofa file:
5581 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5585 Using ClubFritz12 sofa file and bigger radius with small rotation:
5587 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5591 Similar as above but with custom speaker positions for front left, front right, back left and back right
5592 and also with custom gain:
5594 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5601 This filter expands or compresses each half-cycle of audio samples
5602 (local set of samples all above or all below zero and between two nearest zero crossings) depending
5603 on threshold value, so audio reaches target peak value under conditions controlled by below options.
5605 The filter accepts the following options:
5609 Set the expansion target peak value. This specifies the highest allowed absolute amplitude
5610 level for the normalized audio input. Default value is 0.95. Allowed range is from 0.0 to 1.0.
5613 Set the maximum expansion factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5614 This option controls maximum local half-cycle of samples expansion. The maximum expansion
5615 would be such that local peak value reaches target peak value but never to surpass it and that
5616 ratio between new and previous peak value does not surpass this option value.
5618 @item compression, c
5619 Set the maximum compression factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5620 This option controls maximum local half-cycle of samples compression. This option is used
5621 only if @option{threshold} option is set to value greater than 0.0, then in such cases
5622 when local peak is lower or same as value set by @option{threshold} all samples belonging to
5623 that peak's half-cycle will be compressed by current compression factor.
5626 Set the threshold value. Default value is 0.0. Allowed range is from 0.0 to 1.0.
5627 This option specifies which half-cycles of samples will be compressed and which will be expanded.
5628 Any half-cycle samples with their local peak value below or same as this option value will be
5629 compressed by current compression factor, otherwise, if greater than threshold value they will be
5630 expanded with expansion factor so that it could reach peak target value but never surpass it.
5633 Set the expansion raising amount per each half-cycle of samples. Default value is 0.001.
5634 Allowed range is from 0.0 to 1.0. This controls how fast expansion factor is raised per
5635 each new half-cycle until it reaches @option{expansion} value.
5636 Setting this options too high may lead to distortions.
5639 Set the compression raising amount per each half-cycle of samples. Default value is 0.001.
5640 Allowed range is from 0.0 to 1.0. This controls how fast compression factor is raised per
5641 each new half-cycle until it reaches @option{compression} value.
5644 Specify which channels to filter, by default all available channels are filtered.
5647 Enable inverted filtering, by default is disabled. This inverts interpretation of @option{threshold}
5648 option. When enabled any half-cycle of samples with their local peak value below or same as
5649 @option{threshold} option will be expanded otherwise it will be compressed.
5652 Link channels when calculating gain applied to each filtered channel sample, by default is disabled.
5653 When disabled each filtered channel gain calculation is independent, otherwise when this option
5654 is enabled the minimum of all possible gains for each filtered channel is used.
5657 @subsection Commands
5659 This filter supports the all above options as @ref{commands}.
5661 @section stereotools
5663 This filter has some handy utilities to manage stereo signals, for converting
5664 M/S stereo recordings to L/R signal while having control over the parameters
5665 or spreading the stereo image of master track.
5667 The filter accepts the following options:
5671 Set input level before filtering for both channels. Defaults is 1.
5672 Allowed range is from 0.015625 to 64.
5675 Set output level after filtering for both channels. Defaults is 1.
5676 Allowed range is from 0.015625 to 64.
5679 Set input balance between both channels. Default is 0.
5680 Allowed range is from -1 to 1.
5683 Set output balance between both channels. Default is 0.
5684 Allowed range is from -1 to 1.
5687 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5688 clipping. Disabled by default.
5691 Mute the left channel. Disabled by default.
5694 Mute the right channel. Disabled by default.
5697 Change the phase of the left channel. Disabled by default.
5700 Change the phase of the right channel. Disabled by default.
5703 Set stereo mode. Available values are:
5707 Left/Right to Left/Right, this is default.
5710 Left/Right to Mid/Side.
5713 Mid/Side to Left/Right.
5716 Left/Right to Left/Left.
5719 Left/Right to Right/Right.
5722 Left/Right to Left + Right.
5725 Left/Right to Right/Left.
5728 Mid/Side to Left/Left.
5731 Mid/Side to Right/Right.
5734 Mid/Side to Right/Left.
5737 Left/Right to Left - Right.
5741 Set level of side signal. Default is 1.
5742 Allowed range is from 0.015625 to 64.
5745 Set balance of side signal. Default is 0.
5746 Allowed range is from -1 to 1.
5749 Set level of the middle signal. Default is 1.
5750 Allowed range is from 0.015625 to 64.
5753 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5756 Set stereo base between mono and inversed channels. Default is 0.
5757 Allowed range is from -1 to 1.
5760 Set delay in milliseconds how much to delay left from right channel and
5761 vice versa. Default is 0. Allowed range is from -20 to 20.
5764 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5767 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5769 @item bmode_in, bmode_out
5770 Set balance mode for balance_in/balance_out option.
5772 Can be one of the following:
5776 Classic balance mode. Attenuate one channel at time.
5777 Gain is raised up to 1.
5780 Similar as classic mode above but gain is raised up to 2.
5783 Equal power distribution, from -6dB to +6dB range.
5787 @subsection Commands
5789 This filter supports the all above options as @ref{commands}.
5791 @subsection Examples
5795 Apply karaoke like effect:
5797 stereotools=mlev=0.015625
5801 Convert M/S signal to L/R:
5803 "stereotools=mode=ms>lr"
5807 @section stereowiden
5809 This filter enhance the stereo effect by suppressing signal common to both
5810 channels and by delaying the signal of left into right and vice versa,
5811 thereby widening the stereo effect.
5813 The filter accepts the following options:
5817 Time in milliseconds of the delay of left signal into right and vice versa.
5818 Default is 20 milliseconds.
5821 Amount of gain in delayed signal into right and vice versa. Gives a delay
5822 effect of left signal in right output and vice versa which gives widening
5823 effect. Default is 0.3.
5826 Cross feed of left into right with inverted phase. This helps in suppressing
5827 the mono. If the value is 1 it will cancel all the signal common to both
5828 channels. Default is 0.3.
5831 Set level of input signal of original channel. Default is 0.8.
5834 @subsection Commands
5836 This filter supports the all above options except @code{delay} as @ref{commands}.
5838 @section superequalizer
5839 Apply 18 band equalizer.
5841 The filter accepts the following options:
5848 Set 131Hz band gain.
5850 Set 185Hz band gain.
5852 Set 262Hz band gain.
5854 Set 370Hz band gain.
5856 Set 523Hz band gain.
5858 Set 740Hz band gain.
5860 Set 1047Hz band gain.
5862 Set 1480Hz band gain.
5864 Set 2093Hz band gain.
5866 Set 2960Hz band gain.
5868 Set 4186Hz band gain.
5870 Set 5920Hz band gain.
5872 Set 8372Hz band gain.
5874 Set 11840Hz band gain.
5876 Set 16744Hz band gain.
5878 Set 20000Hz band gain.
5882 Apply audio surround upmix filter.
5884 This filter allows to produce multichannel output from audio stream.
5886 The filter accepts the following options:
5890 Set output channel layout. By default, this is @var{5.1}.
5892 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5893 for the required syntax.
5896 Set input channel layout. By default, this is @var{stereo}.
5898 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5899 for the required syntax.
5902 Set input volume level. By default, this is @var{1}.
5905 Set output volume level. By default, this is @var{1}.
5908 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5911 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5914 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5917 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5918 In @var{add} mode, LFE channel is created from input audio and added to output.
5919 In @var{sub} mode, LFE channel is created from input audio and added to output but
5920 also all non-LFE output channels are subtracted with output LFE channel.
5923 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5924 Default is @var{90}.
5927 Set front center input volume. By default, this is @var{1}.
5930 Set front center output volume. By default, this is @var{1}.
5933 Set front left input volume. By default, this is @var{1}.
5936 Set front left output volume. By default, this is @var{1}.
5939 Set front right input volume. By default, this is @var{1}.
5942 Set front right output volume. By default, this is @var{1}.
5945 Set side left input volume. By default, this is @var{1}.
5948 Set side left output volume. By default, this is @var{1}.
5951 Set side right input volume. By default, this is @var{1}.
5954 Set side right output volume. By default, this is @var{1}.
5957 Set back left input volume. By default, this is @var{1}.
5960 Set back left output volume. By default, this is @var{1}.
5963 Set back right input volume. By default, this is @var{1}.
5966 Set back right output volume. By default, this is @var{1}.
5969 Set back center input volume. By default, this is @var{1}.
5972 Set back center output volume. By default, this is @var{1}.
5975 Set LFE input volume. By default, this is @var{1}.
5978 Set LFE output volume. By default, this is @var{1}.
5981 Set spread usage of stereo image across X axis for all channels.
5984 Set spread usage of stereo image across Y axis for all channels.
5986 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5987 Set spread usage of stereo image across X axis for each channel.
5989 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5990 Set spread usage of stereo image across Y axis for each channel.
5993 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5996 Set window function.
5998 It accepts the following values:
6021 Default is @code{hann}.
6024 Set window overlap. If set to 1, the recommended overlap for selected
6025 window function will be picked. Default is @code{0.5}.
6028 @section treble, highshelf
6030 Boost or cut treble (upper) frequencies of the audio using a two-pole
6031 shelving filter with a response similar to that of a standard
6032 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
6034 The filter accepts the following options:
6038 Give the gain at whichever is the lower of ~22 kHz and the
6039 Nyquist frequency. Its useful range is about -20 (for a large cut)
6040 to +20 (for a large boost). Beware of clipping when using a positive gain.
6043 Set the filter's central frequency and so can be used
6044 to extend or reduce the frequency range to be boosted or cut.
6045 The default value is @code{3000} Hz.
6048 Set method to specify band-width of filter.
6063 Determine how steep is the filter's shelf transition.
6066 Set number of poles. Default is 2.
6069 How much to use filtered signal in output. Default is 1.
6070 Range is between 0 and 1.
6073 Specify which channels to filter, by default all available are filtered.
6076 Normalize biquad coefficients, by default is disabled.
6077 Enabling it will normalize magnitude response at DC to 0dB.
6080 Set transform type of IIR filter.
6089 Set precison of filtering.
6092 Pick automatic sample format depending on surround filters.
6094 Always use signed 16-bit.
6096 Always use signed 32-bit.
6098 Always use float 32-bit.
6100 Always use float 64-bit.
6104 @subsection Commands
6106 This filter supports the following commands:
6109 Change treble frequency.
6110 Syntax for the command is : "@var{frequency}"
6113 Change treble width_type.
6114 Syntax for the command is : "@var{width_type}"
6117 Change treble width.
6118 Syntax for the command is : "@var{width}"
6122 Syntax for the command is : "@var{gain}"
6126 Syntax for the command is : "@var{mix}"
6131 Sinusoidal amplitude modulation.
6133 The filter accepts the following options:
6137 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
6138 (20 Hz or lower) will result in a tremolo effect.
6139 This filter may also be used as a ring modulator by specifying
6140 a modulation frequency higher than 20 Hz.
6141 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6144 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6145 Default value is 0.5.
6150 Sinusoidal phase modulation.
6152 The filter accepts the following options:
6156 Modulation frequency in Hertz.
6157 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6160 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6161 Default value is 0.5.
6166 Adjust the input audio volume.
6168 It accepts the following parameters:
6172 Set audio volume expression.
6174 Output values are clipped to the maximum value.
6176 The output audio volume is given by the relation:
6178 @var{output_volume} = @var{volume} * @var{input_volume}
6181 The default value for @var{volume} is "1.0".
6184 This parameter represents the mathematical precision.
6186 It determines which input sample formats will be allowed, which affects the
6187 precision of the volume scaling.
6191 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
6193 32-bit floating-point; this limits input sample format to FLT. (default)
6195 64-bit floating-point; this limits input sample format to DBL.
6199 Choose the behaviour on encountering ReplayGain side data in input frames.
6203 Remove ReplayGain side data, ignoring its contents (the default).
6206 Ignore ReplayGain side data, but leave it in the frame.
6209 Prefer the track gain, if present.
6212 Prefer the album gain, if present.
6215 @item replaygain_preamp
6216 Pre-amplification gain in dB to apply to the selected replaygain gain.
6218 Default value for @var{replaygain_preamp} is 0.0.
6220 @item replaygain_noclip
6221 Prevent clipping by limiting the gain applied.
6223 Default value for @var{replaygain_noclip} is 1.
6226 Set when the volume expression is evaluated.
6228 It accepts the following values:
6231 only evaluate expression once during the filter initialization, or
6232 when the @samp{volume} command is sent
6235 evaluate expression for each incoming frame
6238 Default value is @samp{once}.
6241 The volume expression can contain the following parameters.
6245 frame number (starting at zero)
6248 @item nb_consumed_samples
6249 number of samples consumed by the filter
6251 number of samples in the current frame
6253 original frame position in the file
6259 PTS at start of stream
6261 time at start of stream
6267 last set volume value
6270 Note that when @option{eval} is set to @samp{once} only the
6271 @var{sample_rate} and @var{tb} variables are available, all other
6272 variables will evaluate to NAN.
6274 @subsection Commands
6276 This filter supports the following commands:
6279 Modify the volume expression.
6280 The command accepts the same syntax of the corresponding option.
6282 If the specified expression is not valid, it is kept at its current
6286 @subsection Examples
6290 Halve the input audio volume:
6294 volume=volume=-6.0206dB
6297 In all the above example the named key for @option{volume} can be
6298 omitted, for example like in:
6304 Increase input audio power by 6 decibels using fixed-point precision:
6306 volume=volume=6dB:precision=fixed
6310 Fade volume after time 10 with an annihilation period of 5 seconds:
6312 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
6316 @section volumedetect
6318 Detect the volume of the input video.
6320 The filter has no parameters. The input is not modified. Statistics about
6321 the volume will be printed in the log when the input stream end is reached.
6323 In particular it will show the mean volume (root mean square), maximum
6324 volume (on a per-sample basis), and the beginning of a histogram of the
6325 registered volume values (from the maximum value to a cumulated 1/1000 of
6328 All volumes are in decibels relative to the maximum PCM value.
6330 @subsection Examples
6332 Here is an excerpt of the output:
6334 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
6335 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
6336 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
6337 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
6338 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
6339 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
6340 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
6341 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
6342 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
6348 The mean square energy is approximately -27 dB, or 10^-2.7.
6350 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
6352 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
6355 In other words, raising the volume by +4 dB does not cause any clipping,
6356 raising it by +5 dB causes clipping for 6 samples, etc.
6358 @c man end AUDIO FILTERS
6360 @chapter Audio Sources
6361 @c man begin AUDIO SOURCES
6363 Below is a description of the currently available audio sources.
6367 Buffer audio frames, and make them available to the filter chain.
6369 This source is mainly intended for a programmatic use, in particular
6370 through the interface defined in @file{libavfilter/buffersrc.h}.
6372 It accepts the following parameters:
6376 The timebase which will be used for timestamps of submitted frames. It must be
6377 either a floating-point number or in @var{numerator}/@var{denominator} form.
6380 The sample rate of the incoming audio buffers.
6383 The sample format of the incoming audio buffers.
6384 Either a sample format name or its corresponding integer representation from
6385 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
6387 @item channel_layout
6388 The channel layout of the incoming audio buffers.
6389 Either a channel layout name from channel_layout_map in
6390 @file{libavutil/channel_layout.c} or its corresponding integer representation
6391 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
6394 The number of channels of the incoming audio buffers.
6395 If both @var{channels} and @var{channel_layout} are specified, then they
6400 @subsection Examples
6403 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
6406 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
6407 Since the sample format with name "s16p" corresponds to the number
6408 6 and the "stereo" channel layout corresponds to the value 0x3, this is
6411 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
6416 Generate an audio signal specified by an expression.
6418 This source accepts in input one or more expressions (one for each
6419 channel), which are evaluated and used to generate a corresponding
6422 This source accepts the following options:
6426 Set the '|'-separated expressions list for each separate channel. In case the
6427 @option{channel_layout} option is not specified, the selected channel layout
6428 depends on the number of provided expressions. Otherwise the last
6429 specified expression is applied to the remaining output channels.
6431 @item channel_layout, c
6432 Set the channel layout. The number of channels in the specified layout
6433 must be equal to the number of specified expressions.
6436 Set the minimum duration of the sourced audio. See
6437 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6438 for the accepted syntax.
6439 Note that the resulting duration may be greater than the specified
6440 duration, as the generated audio is always cut at the end of a
6443 If not specified, or the expressed duration is negative, the audio is
6444 supposed to be generated forever.
6447 Set the number of samples per channel per each output frame,
6450 @item sample_rate, s
6451 Specify the sample rate, default to 44100.
6454 Each expression in @var{exprs} can contain the following constants:
6458 number of the evaluated sample, starting from 0
6461 time of the evaluated sample expressed in seconds, starting from 0
6468 @subsection Examples
6478 Generate a sin signal with frequency of 440 Hz, set sample rate to
6481 aevalsrc="sin(440*2*PI*t):s=8000"
6485 Generate a two channels signal, specify the channel layout (Front
6486 Center + Back Center) explicitly:
6488 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
6492 Generate white noise:
6494 aevalsrc="-2+random(0)"
6498 Generate an amplitude modulated signal:
6500 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
6504 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
6506 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
6513 Generate a FIR coefficients using frequency sampling method.
6515 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6517 The filter accepts the following options:
6521 Set number of filter coefficents in output audio stream.
6522 Default value is 1025.
6525 Set frequency points from where magnitude and phase are set.
6526 This must be in non decreasing order, and first element must be 0, while last element
6527 must be 1. Elements are separated by white spaces.
6530 Set magnitude value for every frequency point set by @option{frequency}.
6531 Number of values must be same as number of frequency points.
6532 Values are separated by white spaces.
6535 Set phase value for every frequency point set by @option{frequency}.
6536 Number of values must be same as number of frequency points.
6537 Values are separated by white spaces.
6539 @item sample_rate, r
6540 Set sample rate, default is 44100.
6543 Set number of samples per each frame. Default is 1024.
6546 Set window function. Default is blackman.
6551 The null audio source, return unprocessed audio frames. It is mainly useful
6552 as a template and to be employed in analysis / debugging tools, or as
6553 the source for filters which ignore the input data (for example the sox
6556 This source accepts the following options:
6560 @item channel_layout, cl
6562 Specifies the channel layout, and can be either an integer or a string
6563 representing a channel layout. The default value of @var{channel_layout}
6566 Check the channel_layout_map definition in
6567 @file{libavutil/channel_layout.c} for the mapping between strings and
6568 channel layout values.
6570 @item sample_rate, r
6571 Specifies the sample rate, and defaults to 44100.
6574 Set the number of samples per requested frames.
6577 Set the duration of the sourced audio. See
6578 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6579 for the accepted syntax.
6581 If not specified, or the expressed duration is negative, the audio is
6582 supposed to be generated forever.
6585 @subsection Examples
6589 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
6591 anullsrc=r=48000:cl=4
6595 Do the same operation with a more obvious syntax:
6597 anullsrc=r=48000:cl=mono
6601 All the parameters need to be explicitly defined.
6605 Synthesize a voice utterance using the libflite library.
6607 To enable compilation of this filter you need to configure FFmpeg with
6608 @code{--enable-libflite}.
6610 Note that versions of the flite library prior to 2.0 are not thread-safe.
6612 The filter accepts the following options:
6617 If set to 1, list the names of the available voices and exit
6618 immediately. Default value is 0.
6621 Set the maximum number of samples per frame. Default value is 512.
6624 Set the filename containing the text to speak.
6627 Set the text to speak.
6630 Set the voice to use for the speech synthesis. Default value is
6631 @code{kal}. See also the @var{list_voices} option.
6634 @subsection Examples
6638 Read from file @file{speech.txt}, and synthesize the text using the
6639 standard flite voice:
6641 flite=textfile=speech.txt
6645 Read the specified text selecting the @code{slt} voice:
6647 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6651 Input text to ffmpeg:
6653 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6657 Make @file{ffplay} speak the specified text, using @code{flite} and
6658 the @code{lavfi} device:
6660 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
6664 For more information about libflite, check:
6665 @url{http://www.festvox.org/flite/}
6669 Generate a noise audio signal.
6671 The filter accepts the following options:
6674 @item sample_rate, r
6675 Specify the sample rate. Default value is 48000 Hz.
6678 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
6682 Specify the duration of the generated audio stream. Not specifying this option
6683 results in noise with an infinite length.
6685 @item color, colour, c
6686 Specify the color of noise. Available noise colors are white, pink, brown,
6687 blue, violet and velvet. Default color is white.
6690 Specify a value used to seed the PRNG.
6693 Set the number of samples per each output frame, default is 1024.
6696 @subsection Examples
6701 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6703 anoisesrc=d=60:c=pink:r=44100:a=0.5
6709 Generate odd-tap Hilbert transform FIR coefficients.
6711 The resulting stream can be used with @ref{afir} filter for phase-shifting
6712 the signal by 90 degrees.
6714 This is used in many matrix coding schemes and for analytic signal generation.
6715 The process is often written as a multiplication by i (or j), the imaginary unit.
6717 The filter accepts the following options:
6721 @item sample_rate, s
6722 Set sample rate, default is 44100.
6725 Set length of FIR filter, default is 22051.
6728 Set number of samples per each frame.
6731 Set window function to be used when generating FIR coefficients.
6736 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6738 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6740 The filter accepts the following options:
6743 @item sample_rate, r
6744 Set sample rate, default is 44100.
6747 Set number of samples per each frame. Default is 1024.
6750 Set high-pass frequency. Default is 0.
6753 Set low-pass frequency. Default is 0.
6754 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6755 is higher than 0 then filter will create band-pass filter coefficients,
6756 otherwise band-reject filter coefficients.
6759 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6762 Set Kaiser window beta.
6765 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6768 Enable rounding, by default is disabled.
6771 Set number of taps for high-pass filter.
6774 Set number of taps for low-pass filter.
6779 Generate an audio signal made of a sine wave with amplitude 1/8.
6781 The audio signal is bit-exact.
6783 The filter accepts the following options:
6788 Set the carrier frequency. Default is 440 Hz.
6790 @item beep_factor, b
6791 Enable a periodic beep every second with frequency @var{beep_factor} times
6792 the carrier frequency. Default is 0, meaning the beep is disabled.
6794 @item sample_rate, r
6795 Specify the sample rate, default is 44100.
6798 Specify the duration of the generated audio stream.
6800 @item samples_per_frame
6801 Set the number of samples per output frame.
6803 The expression can contain the following constants:
6807 The (sequential) number of the output audio frame, starting from 0.
6810 The PTS (Presentation TimeStamp) of the output audio frame,
6811 expressed in @var{TB} units.
6814 The PTS of the output audio frame, expressed in seconds.
6817 The timebase of the output audio frames.
6820 Default is @code{1024}.
6823 @subsection Examples
6828 Generate a simple 440 Hz sine wave:
6834 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6838 sine=frequency=220:beep_factor=4:duration=5
6842 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6845 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6849 @c man end AUDIO SOURCES
6851 @chapter Audio Sinks
6852 @c man begin AUDIO SINKS
6854 Below is a description of the currently available audio sinks.
6856 @section abuffersink
6858 Buffer audio frames, and make them available to the end of filter chain.
6860 This sink is mainly intended for programmatic use, in particular
6861 through the interface defined in @file{libavfilter/buffersink.h}
6862 or the options system.
6864 It accepts a pointer to an AVABufferSinkContext structure, which
6865 defines the incoming buffers' formats, to be passed as the opaque
6866 parameter to @code{avfilter_init_filter} for initialization.
6869 Null audio sink; do absolutely nothing with the input audio. It is
6870 mainly useful as a template and for use in analysis / debugging
6873 @c man end AUDIO SINKS
6875 @chapter Video Filters
6876 @c man begin VIDEO FILTERS
6878 When you configure your FFmpeg build, you can disable any of the
6879 existing filters using @code{--disable-filters}.
6880 The configure output will show the video filters included in your
6883 Below is a description of the currently available video filters.
6887 Mark a region of interest in a video frame.
6889 The frame data is passed through unchanged, but metadata is attached
6890 to the frame indicating regions of interest which can affect the
6891 behaviour of later encoding. Multiple regions can be marked by
6892 applying the filter multiple times.
6896 Region distance in pixels from the left edge of the frame.
6898 Region distance in pixels from the top edge of the frame.
6900 Region width in pixels.
6902 Region height in pixels.
6904 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6905 and may contain the following variables:
6908 Width of the input frame.
6910 Height of the input frame.
6914 Quantisation offset to apply within the region.
6916 This must be a real value in the range -1 to +1. A value of zero
6917 indicates no quality change. A negative value asks for better quality
6918 (less quantisation), while a positive value asks for worse quality
6919 (greater quantisation).
6921 The range is calibrated so that the extreme values indicate the
6922 largest possible offset - if the rest of the frame is encoded with the
6923 worst possible quality, an offset of -1 indicates that this region
6924 should be encoded with the best possible quality anyway. Intermediate
6925 values are then interpolated in some codec-dependent way.
6927 For example, in 10-bit H.264 the quantisation parameter varies between
6928 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6929 this region should be encoded with a QP around one-tenth of the full
6930 range better than the rest of the frame. So, if most of the frame
6931 were to be encoded with a QP of around 30, this region would get a QP
6932 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6933 An extreme value of -1 would indicate that this region should be
6934 encoded with the best possible quality regardless of the treatment of
6935 the rest of the frame - that is, should be encoded at a QP of -12.
6937 If set to true, remove any existing regions of interest marked on the
6938 frame before adding the new one.
6941 @subsection Examples
6945 Mark the centre quarter of the frame as interesting.
6947 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6950 Mark the 100-pixel-wide region on the left edge of the frame as very
6951 uninteresting (to be encoded at much lower quality than the rest of
6954 addroi=0:0:100:ih:+1/5
6958 @section alphaextract
6960 Extract the alpha component from the input as a grayscale video. This
6961 is especially useful with the @var{alphamerge} filter.
6965 Add or replace the alpha component of the primary input with the
6966 grayscale value of a second input. This is intended for use with
6967 @var{alphaextract} to allow the transmission or storage of frame
6968 sequences that have alpha in a format that doesn't support an alpha
6971 For example, to reconstruct full frames from a normal YUV-encoded video
6972 and a separate video created with @var{alphaextract}, you might use:
6974 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6979 Amplify differences between current pixel and pixels of adjacent frames in
6980 same pixel location.
6982 This filter accepts the following options:
6986 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6987 For example radius of 3 will instruct filter to calculate average of 7 frames.
6990 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6993 Set threshold for difference amplification. Any difference greater or equal to
6994 this value will not alter source pixel. Default is 10.
6995 Allowed range is from 0 to 65535.
6998 Set tolerance for difference amplification. Any difference lower to
6999 this value will not alter source pixel. Default is 0.
7000 Allowed range is from 0 to 65535.
7003 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
7004 This option controls maximum possible value that will decrease source pixel value.
7007 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
7008 This option controls maximum possible value that will increase source pixel value.
7011 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
7014 @subsection Commands
7016 This filter supports the following @ref{commands} that corresponds to option of same name:
7028 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
7029 and libavformat to work. On the other hand, it is limited to ASS (Advanced
7030 Substation Alpha) subtitles files.
7032 This filter accepts the following option in addition to the common options from
7033 the @ref{subtitles} filter:
7037 Set the shaping engine
7039 Available values are:
7042 The default libass shaping engine, which is the best available.
7044 Fast, font-agnostic shaper that can do only substitutions
7046 Slower shaper using OpenType for substitutions and positioning
7049 The default is @code{auto}.
7053 Apply an Adaptive Temporal Averaging Denoiser to the video input.
7055 The filter accepts the following options:
7059 Set threshold A for 1st plane. Default is 0.02.
7060 Valid range is 0 to 0.3.
7063 Set threshold B for 1st plane. Default is 0.04.
7064 Valid range is 0 to 5.
7067 Set threshold A for 2nd plane. Default is 0.02.
7068 Valid range is 0 to 0.3.
7071 Set threshold B for 2nd plane. Default is 0.04.
7072 Valid range is 0 to 5.
7075 Set threshold A for 3rd plane. Default is 0.02.
7076 Valid range is 0 to 0.3.
7079 Set threshold B for 3rd plane. Default is 0.04.
7080 Valid range is 0 to 5.
7082 Threshold A is designed to react on abrupt changes in the input signal and
7083 threshold B is designed to react on continuous changes in the input signal.
7086 Set number of frames filter will use for averaging. Default is 9. Must be odd
7087 number in range [5, 129].
7090 Set what planes of frame filter will use for averaging. Default is all.
7093 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
7094 Alternatively can be set to @code{s} serial.
7096 Parallel can be faster then serial, while other way around is never true.
7097 Parallel will abort early on first change being greater then thresholds, while serial
7098 will continue processing other side of frames if they are equal or below thresholds.
7101 @subsection Commands
7102 This filter supports same @ref{commands} as options except option @code{s}.
7103 The command accepts the same syntax of the corresponding option.
7107 Apply average blur filter.
7109 The filter accepts the following options:
7113 Set horizontal radius size.
7116 Set which planes to filter. By default all planes are filtered.
7119 Set vertical radius size, if zero it will be same as @code{sizeX}.
7120 Default is @code{0}.
7123 @subsection Commands
7124 This filter supports same commands as options.
7125 The command accepts the same syntax of the corresponding option.
7127 If the specified expression is not valid, it is kept at its current
7132 Compute the bounding box for the non-black pixels in the input frame
7135 This filter computes the bounding box containing all the pixels with a
7136 luminance value greater than the minimum allowed value.
7137 The parameters describing the bounding box are printed on the filter
7140 The filter accepts the following option:
7144 Set the minimal luminance value. Default is @code{16}.
7147 @subsection Commands
7149 This filter supports the all above options as @ref{commands}.
7152 Apply bilateral filter, spatial smoothing while preserving edges.
7154 The filter accepts the following options:
7157 Set sigma of gaussian function to calculate spatial weight.
7158 Allowed range is 0 to 512. Default is 0.1.
7161 Set sigma of gaussian function to calculate range weight.
7162 Allowed range is 0 to 1. Default is 0.1.
7165 Set planes to filter. Default is first only.
7168 @subsection Commands
7170 This filter supports the all above options as @ref{commands}.
7172 @section bitplanenoise
7174 Show and measure bit plane noise.
7176 The filter accepts the following options:
7180 Set which plane to analyze. Default is @code{1}.
7183 Filter out noisy pixels from @code{bitplane} set above.
7184 Default is disabled.
7187 @section blackdetect
7189 Detect video intervals that are (almost) completely black. Can be
7190 useful to detect chapter transitions, commercials, or invalid
7193 The filter outputs its detection analysis to both the log as well as
7194 frame metadata. If a black segment of at least the specified minimum
7195 duration is found, a line with the start and end timestamps as well
7196 as duration is printed to the log with level @code{info}. In addition,
7197 a log line with level @code{debug} is printed per frame showing the
7198 black amount detected for that frame.
7200 The filter also attaches metadata to the first frame of a black
7201 segment with key @code{lavfi.black_start} and to the first frame
7202 after the black segment ends with key @code{lavfi.black_end}. The
7203 value is the frame's timestamp. This metadata is added regardless
7204 of the minimum duration specified.
7206 The filter accepts the following options:
7209 @item black_min_duration, d
7210 Set the minimum detected black duration expressed in seconds. It must
7211 be a non-negative floating point number.
7213 Default value is 2.0.
7215 @item picture_black_ratio_th, pic_th
7216 Set the threshold for considering a picture "black".
7217 Express the minimum value for the ratio:
7219 @var{nb_black_pixels} / @var{nb_pixels}
7222 for which a picture is considered black.
7223 Default value is 0.98.
7225 @item pixel_black_th, pix_th
7226 Set the threshold for considering a pixel "black".
7228 The threshold expresses the maximum pixel luminance value for which a
7229 pixel is considered "black". The provided value is scaled according to
7230 the following equation:
7232 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
7235 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
7236 the input video format, the range is [0-255] for YUV full-range
7237 formats and [16-235] for YUV non full-range formats.
7239 Default value is 0.10.
7242 The following example sets the maximum pixel threshold to the minimum
7243 value, and detects only black intervals of 2 or more seconds:
7245 blackdetect=d=2:pix_th=0.00
7250 Detect frames that are (almost) completely black. Can be useful to
7251 detect chapter transitions or commercials. Output lines consist of
7252 the frame number of the detected frame, the percentage of blackness,
7253 the position in the file if known or -1 and the timestamp in seconds.
7255 In order to display the output lines, you need to set the loglevel at
7256 least to the AV_LOG_INFO value.
7258 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
7259 The value represents the percentage of pixels in the picture that
7260 are below the threshold value.
7262 It accepts the following parameters:
7267 The percentage of the pixels that have to be below the threshold; it defaults to
7270 @item threshold, thresh
7271 The threshold below which a pixel value is considered black; it defaults to
7279 Blend two video frames into each other.
7281 The @code{blend} filter takes two input streams and outputs one
7282 stream, the first input is the "top" layer and second input is
7283 "bottom" layer. By default, the output terminates when the longest input terminates.
7285 The @code{tblend} (time blend) filter takes two consecutive frames
7286 from one single stream, and outputs the result obtained by blending
7287 the new frame on top of the old frame.
7289 A description of the accepted options follows.
7297 Set blend mode for specific pixel component or all pixel components in case
7298 of @var{all_mode}. Default value is @code{normal}.
7300 Available values for component modes are:
7342 Set blend opacity for specific pixel component or all pixel components in case
7343 of @var{all_opacity}. Only used in combination with pixel component blend modes.
7350 Set blend expression for specific pixel component or all pixel components in case
7351 of @var{all_expr}. Note that related mode options will be ignored if those are set.
7353 The expressions can use the following variables:
7357 The sequential number of the filtered frame, starting from @code{0}.
7361 the coordinates of the current sample
7365 the width and height of currently filtered plane
7369 Width and height scale for the plane being filtered. It is the
7370 ratio between the dimensions of the current plane to the luma plane,
7371 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
7372 the luma plane and @code{0.5,0.5} for the chroma planes.
7375 Time of the current frame, expressed in seconds.
7378 Value of pixel component at current location for first video frame (top layer).
7381 Value of pixel component at current location for second video frame (bottom layer).
7385 The @code{blend} filter also supports the @ref{framesync} options.
7387 @subsection Examples
7391 Apply transition from bottom layer to top layer in first 10 seconds:
7393 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
7397 Apply linear horizontal transition from top layer to bottom layer:
7399 blend=all_expr='A*(X/W)+B*(1-X/W)'
7403 Apply 1x1 checkerboard effect:
7405 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
7409 Apply uncover left effect:
7411 blend=all_expr='if(gte(N*SW+X,W),A,B)'
7415 Apply uncover down effect:
7417 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
7421 Apply uncover up-left effect:
7423 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
7427 Split diagonally video and shows top and bottom layer on each side:
7429 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
7433 Display differences between the current and the previous frame:
7435 tblend=all_mode=grainextract
7441 Denoise frames using Block-Matching 3D algorithm.
7443 The filter accepts the following options.
7447 Set denoising strength. Default value is 1.
7448 Allowed range is from 0 to 999.9.
7449 The denoising algorithm is very sensitive to sigma, so adjust it
7450 according to the source.
7453 Set local patch size. This sets dimensions in 2D.
7456 Set sliding step for processing blocks. Default value is 4.
7457 Allowed range is from 1 to 64.
7458 Smaller values allows processing more reference blocks and is slower.
7461 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
7462 When set to 1, no block matching is done. Larger values allows more blocks
7464 Allowed range is from 1 to 256.
7467 Set radius for search block matching. Default is 9.
7468 Allowed range is from 1 to INT32_MAX.
7471 Set step between two search locations for block matching. Default is 1.
7472 Allowed range is from 1 to 64. Smaller is slower.
7475 Set threshold of mean square error for block matching. Valid range is 0 to
7479 Set thresholding parameter for hard thresholding in 3D transformed domain.
7480 Larger values results in stronger hard-thresholding filtering in frequency
7484 Set filtering estimation mode. Can be @code{basic} or @code{final}.
7485 Default is @code{basic}.
7488 If enabled, filter will use 2nd stream for block matching.
7489 Default is disabled for @code{basic} value of @var{estim} option,
7490 and always enabled if value of @var{estim} is @code{final}.
7493 Set planes to filter. Default is all available except alpha.
7496 @subsection Examples
7500 Basic filtering with bm3d:
7502 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
7506 Same as above, but filtering only luma:
7508 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
7512 Same as above, but with both estimation modes:
7514 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
7518 Same as above, but prefilter with @ref{nlmeans} filter instead:
7520 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
7526 Apply a boxblur algorithm to the input video.
7528 It accepts the following parameters:
7532 @item luma_radius, lr
7533 @item luma_power, lp
7534 @item chroma_radius, cr
7535 @item chroma_power, cp
7536 @item alpha_radius, ar
7537 @item alpha_power, ap
7541 A description of the accepted options follows.
7544 @item luma_radius, lr
7545 @item chroma_radius, cr
7546 @item alpha_radius, ar
7547 Set an expression for the box radius in pixels used for blurring the
7548 corresponding input plane.
7550 The radius value must be a non-negative number, and must not be
7551 greater than the value of the expression @code{min(w,h)/2} for the
7552 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
7555 Default value for @option{luma_radius} is "2". If not specified,
7556 @option{chroma_radius} and @option{alpha_radius} default to the
7557 corresponding value set for @option{luma_radius}.
7559 The expressions can contain the following constants:
7563 The input width and height in pixels.
7567 The input chroma image width and height in pixels.
7571 The horizontal and vertical chroma subsample values. For example, for the
7572 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
7575 @item luma_power, lp
7576 @item chroma_power, cp
7577 @item alpha_power, ap
7578 Specify how many times the boxblur filter is applied to the
7579 corresponding plane.
7581 Default value for @option{luma_power} is 2. If not specified,
7582 @option{chroma_power} and @option{alpha_power} default to the
7583 corresponding value set for @option{luma_power}.
7585 A value of 0 will disable the effect.
7588 @subsection Examples
7592 Apply a boxblur filter with the luma, chroma, and alpha radii
7595 boxblur=luma_radius=2:luma_power=1
7600 Set the luma radius to 2, and alpha and chroma radius to 0:
7602 boxblur=2:1:cr=0:ar=0
7606 Set the luma and chroma radii to a fraction of the video dimension:
7608 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
7614 Deinterlace the input video ("bwdif" stands for "Bob Weaver
7615 Deinterlacing Filter").
7617 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
7618 interpolation algorithms.
7619 It accepts the following parameters:
7623 The interlacing mode to adopt. It accepts one of the following values:
7627 Output one frame for each frame.
7629 Output one frame for each field.
7632 The default value is @code{send_field}.
7635 The picture field parity assumed for the input interlaced video. It accepts one
7636 of the following values:
7640 Assume the top field is first.
7642 Assume the bottom field is first.
7644 Enable automatic detection of field parity.
7647 The default value is @code{auto}.
7648 If the interlacing is unknown or the decoder does not export this information,
7649 top field first will be assumed.
7652 Specify which frames to deinterlace. Accepts one of the following
7657 Deinterlace all frames.
7659 Only deinterlace frames marked as interlaced.
7662 The default value is @code{all}.
7667 Apply Contrast Adaptive Sharpen filter to video stream.
7669 The filter accepts the following options:
7673 Set the sharpening strength. Default value is 0.
7676 Set planes to filter. Default value is to filter all
7677 planes except alpha plane.
7680 @subsection Commands
7681 This filter supports same @ref{commands} as options.
7684 Remove all color information for all colors except for certain one.
7686 The filter accepts the following options:
7690 The color which will not be replaced with neutral chroma.
7693 Similarity percentage with the above color.
7694 0.01 matches only the exact key color, while 1.0 matches everything.
7698 0.0 makes pixels either fully gray, or not gray at all.
7699 Higher values result in more preserved color.
7702 Signals that the color passed is already in YUV instead of RGB.
7704 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7705 This can be used to pass exact YUV values as hexadecimal numbers.
7708 @subsection Commands
7709 This filter supports same @ref{commands} as options.
7710 The command accepts the same syntax of the corresponding option.
7712 If the specified expression is not valid, it is kept at its current
7716 YUV colorspace color/chroma keying.
7718 The filter accepts the following options:
7722 The color which will be replaced with transparency.
7725 Similarity percentage with the key color.
7727 0.01 matches only the exact key color, while 1.0 matches everything.
7732 0.0 makes pixels either fully transparent, or not transparent at all.
7734 Higher values result in semi-transparent pixels, with a higher transparency
7735 the more similar the pixels color is to the key color.
7738 Signals that the color passed is already in YUV instead of RGB.
7740 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7741 This can be used to pass exact YUV values as hexadecimal numbers.
7744 @subsection Commands
7745 This filter supports same @ref{commands} as options.
7746 The command accepts the same syntax of the corresponding option.
7748 If the specified expression is not valid, it is kept at its current
7751 @subsection Examples
7755 Make every green pixel in the input image transparent:
7757 ffmpeg -i input.png -vf chromakey=green out.png
7761 Overlay a greenscreen-video on top of a static black background.
7763 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
7768 Reduce chrominance noise.
7770 The filter accepts the following options:
7774 Set threshold for averaging chrominance values.
7775 Sum of absolute difference of Y, U and V pixel components of current
7776 pixel and neighbour pixels lower than this threshold will be used in
7777 averaging. Luma component is left unchanged and is copied to output.
7778 Default value is 30. Allowed range is from 1 to 200.
7781 Set horizontal radius of rectangle used for averaging.
7782 Allowed range is from 1 to 100. Default value is 5.
7785 Set vertical radius of rectangle used for averaging.
7786 Allowed range is from 1 to 100. Default value is 5.
7789 Set horizontal step when averaging. Default value is 1.
7790 Allowed range is from 1 to 50.
7791 Mostly useful to speed-up filtering.
7794 Set vertical step when averaging. Default value is 1.
7795 Allowed range is from 1 to 50.
7796 Mostly useful to speed-up filtering.
7799 Set Y threshold for averaging chrominance values.
7800 Set finer control for max allowed difference between Y components
7801 of current pixel and neigbour pixels.
7802 Default value is 200. Allowed range is from 1 to 200.
7805 Set U threshold for averaging chrominance values.
7806 Set finer control for max allowed difference between U components
7807 of current pixel and neigbour pixels.
7808 Default value is 200. Allowed range is from 1 to 200.
7811 Set V threshold for averaging chrominance values.
7812 Set finer control for max allowed difference between V components
7813 of current pixel and neigbour pixels.
7814 Default value is 200. Allowed range is from 1 to 200.
7817 @subsection Commands
7818 This filter supports same @ref{commands} as options.
7819 The command accepts the same syntax of the corresponding option.
7821 @section chromashift
7822 Shift chroma pixels horizontally and/or vertically.
7824 The filter accepts the following options:
7827 Set amount to shift chroma-blue horizontally.
7829 Set amount to shift chroma-blue vertically.
7831 Set amount to shift chroma-red horizontally.
7833 Set amount to shift chroma-red vertically.
7835 Set edge mode, can be @var{smear}, default, or @var{warp}.
7838 @subsection Commands
7840 This filter supports the all above options as @ref{commands}.
7844 Display CIE color diagram with pixels overlaid onto it.
7846 The filter accepts the following options:
7861 @item uhdtv, rec2020
7875 Set what gamuts to draw.
7877 See @code{system} option for available values.
7880 Set ciescope size, by default set to 512.
7883 Set intensity used to map input pixel values to CIE diagram.
7886 Set contrast used to draw tongue colors that are out of active color system gamut.
7889 Correct gamma displayed on scope, by default enabled.
7892 Show white point on CIE diagram, by default disabled.
7895 Set input gamma. Used only with XYZ input color space.
7900 Visualize information exported by some codecs.
7902 Some codecs can export information through frames using side-data or other
7903 means. For example, some MPEG based codecs export motion vectors through the
7904 @var{export_mvs} flag in the codec @option{flags2} option.
7906 The filter accepts the following option:
7910 Set motion vectors to visualize.
7912 Available flags for @var{mv} are:
7916 forward predicted MVs of P-frames
7918 forward predicted MVs of B-frames
7920 backward predicted MVs of B-frames
7924 Display quantization parameters using the chroma planes.
7927 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7929 Available flags for @var{mv_type} are:
7933 forward predicted MVs
7935 backward predicted MVs
7938 @item frame_type, ft
7939 Set frame type to visualize motion vectors of.
7941 Available flags for @var{frame_type} are:
7945 intra-coded frames (I-frames)
7947 predicted frames (P-frames)
7949 bi-directionally predicted frames (B-frames)
7953 @subsection Examples
7957 Visualize forward predicted MVs of all frames using @command{ffplay}:
7959 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7963 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7965 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7969 @section colorbalance
7970 Modify intensity of primary colors (red, green and blue) of input frames.
7972 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7973 regions for the red-cyan, green-magenta or blue-yellow balance.
7975 A positive adjustment value shifts the balance towards the primary color, a negative
7976 value towards the complementary color.
7978 The filter accepts the following options:
7984 Adjust red, green and blue shadows (darkest pixels).
7989 Adjust red, green and blue midtones (medium pixels).
7994 Adjust red, green and blue highlights (brightest pixels).
7996 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7999 Preserve lightness when changing color balance. Default is disabled.
8002 @subsection Examples
8006 Add red color cast to shadows:
8012 @subsection Commands
8014 This filter supports the all above options as @ref{commands}.
8016 @section colorchannelmixer
8018 Adjust video input frames by re-mixing color channels.
8020 This filter modifies a color channel by adding the values associated to
8021 the other channels of the same pixels. For example if the value to
8022 modify is red, the output value will be:
8024 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
8027 The filter accepts the following options:
8034 Adjust contribution of input red, green, blue and alpha channels for output red channel.
8035 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
8041 Adjust contribution of input red, green, blue and alpha channels for output green channel.
8042 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
8048 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
8049 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
8055 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
8056 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
8058 Allowed ranges for options are @code{[-2.0, 2.0]}.
8061 @subsection Examples
8065 Convert source to grayscale:
8067 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
8070 Simulate sepia tones:
8072 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
8076 @subsection Commands
8078 This filter supports the all above options as @ref{commands}.
8081 RGB colorspace color keying.
8083 The filter accepts the following options:
8087 The color which will be replaced with transparency.
8090 Similarity percentage with the key color.
8092 0.01 matches only the exact key color, while 1.0 matches everything.
8097 0.0 makes pixels either fully transparent, or not transparent at all.
8099 Higher values result in semi-transparent pixels, with a higher transparency
8100 the more similar the pixels color is to the key color.
8103 @subsection Examples
8107 Make every green pixel in the input image transparent:
8109 ffmpeg -i input.png -vf colorkey=green out.png
8113 Overlay a greenscreen-video on top of a static background image.
8115 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
8119 @subsection Commands
8120 This filter supports same @ref{commands} as options.
8121 The command accepts the same syntax of the corresponding option.
8123 If the specified expression is not valid, it is kept at its current
8127 Remove all color information for all RGB colors except for certain one.
8129 The filter accepts the following options:
8133 The color which will not be replaced with neutral gray.
8136 Similarity percentage with the above color.
8137 0.01 matches only the exact key color, while 1.0 matches everything.
8140 Blend percentage. 0.0 makes pixels fully gray.
8141 Higher values result in more preserved color.
8144 @subsection Commands
8145 This filter supports same @ref{commands} as options.
8146 The command accepts the same syntax of the corresponding option.
8148 If the specified expression is not valid, it is kept at its current
8151 @section colorlevels
8153 Adjust video input frames using levels.
8155 The filter accepts the following options:
8162 Adjust red, green, blue and alpha input black point.
8163 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
8169 Adjust red, green, blue and alpha input white point.
8170 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
8172 Input levels are used to lighten highlights (bright tones), darken shadows
8173 (dark tones), change the balance of bright and dark tones.
8179 Adjust red, green, blue and alpha output black point.
8180 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
8186 Adjust red, green, blue and alpha output white point.
8187 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
8189 Output levels allows manual selection of a constrained output level range.
8192 @subsection Examples
8196 Make video output darker:
8198 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
8204 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
8208 Make video output lighter:
8210 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
8214 Increase brightness:
8216 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
8220 @subsection Commands
8222 This filter supports the all above options as @ref{commands}.
8224 @section colormatrix
8226 Convert color matrix.
8228 The filter accepts the following options:
8233 Specify the source and destination color matrix. Both values must be
8236 The accepted values are:
8264 For example to convert from BT.601 to SMPTE-240M, use the command:
8266 colormatrix=bt601:smpte240m
8271 Convert colorspace, transfer characteristics or color primaries.
8272 Input video needs to have an even size.
8274 The filter accepts the following options:
8279 Specify all color properties at once.
8281 The accepted values are:
8311 Specify output colorspace.
8313 The accepted values are:
8322 BT.470BG or BT.601-6 625
8325 SMPTE-170M or BT.601-6 525
8334 BT.2020 with non-constant luminance
8340 Specify output transfer characteristics.
8342 The accepted values are:
8354 Constant gamma of 2.2
8357 Constant gamma of 2.8
8360 SMPTE-170M, BT.601-6 625 or BT.601-6 525
8378 BT.2020 for 10-bits content
8381 BT.2020 for 12-bits content
8387 Specify output color primaries.
8389 The accepted values are:
8398 BT.470BG or BT.601-6 625
8401 SMPTE-170M or BT.601-6 525
8425 Specify output color range.
8427 The accepted values are:
8430 TV (restricted) range
8433 MPEG (restricted) range
8444 Specify output color format.
8446 The accepted values are:
8449 YUV 4:2:0 planar 8-bits
8452 YUV 4:2:0 planar 10-bits
8455 YUV 4:2:0 planar 12-bits
8458 YUV 4:2:2 planar 8-bits
8461 YUV 4:2:2 planar 10-bits
8464 YUV 4:2:2 planar 12-bits
8467 YUV 4:4:4 planar 8-bits
8470 YUV 4:4:4 planar 10-bits
8473 YUV 4:4:4 planar 12-bits
8478 Do a fast conversion, which skips gamma/primary correction. This will take
8479 significantly less CPU, but will be mathematically incorrect. To get output
8480 compatible with that produced by the colormatrix filter, use fast=1.
8483 Specify dithering mode.
8485 The accepted values are:
8491 Floyd-Steinberg dithering
8495 Whitepoint adaptation mode.
8497 The accepted values are:
8500 Bradford whitepoint adaptation
8503 von Kries whitepoint adaptation
8506 identity whitepoint adaptation (i.e. no whitepoint adaptation)
8510 Override all input properties at once. Same accepted values as @ref{all}.
8513 Override input colorspace. Same accepted values as @ref{space}.
8516 Override input color primaries. Same accepted values as @ref{primaries}.
8519 Override input transfer characteristics. Same accepted values as @ref{trc}.
8522 Override input color range. Same accepted values as @ref{range}.
8526 The filter converts the transfer characteristics, color space and color
8527 primaries to the specified user values. The output value, if not specified,
8528 is set to a default value based on the "all" property. If that property is
8529 also not specified, the filter will log an error. The output color range and
8530 format default to the same value as the input color range and format. The
8531 input transfer characteristics, color space, color primaries and color range
8532 should be set on the input data. If any of these are missing, the filter will
8533 log an error and no conversion will take place.
8535 For example to convert the input to SMPTE-240M, use the command:
8537 colorspace=smpte240m
8540 @section convolution
8542 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
8544 The filter accepts the following options:
8551 Set matrix for each plane.
8552 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
8553 and from 1 to 49 odd number of signed integers in @var{row} mode.
8559 Set multiplier for calculated value for each plane.
8560 If unset or 0, it will be sum of all matrix elements.
8566 Set bias for each plane. This value is added to the result of the multiplication.
8567 Useful for making the overall image brighter or darker. Default is 0.0.
8573 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
8574 Default is @var{square}.
8577 @subsection Commands
8579 This filter supports the all above options as @ref{commands}.
8581 @subsection Examples
8587 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"
8593 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"
8599 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"
8605 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"
8609 Apply laplacian edge detector which includes diagonals:
8611 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"
8617 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"
8623 Apply 2D convolution of video stream in frequency domain using second stream
8626 The filter accepts the following options:
8630 Set which planes to process.
8633 Set which impulse video frames will be processed, can be @var{first}
8634 or @var{all}. Default is @var{all}.
8637 The @code{convolve} filter also supports the @ref{framesync} options.
8641 Copy the input video source unchanged to the output. This is mainly useful for
8646 Video filtering on GPU using Apple's CoreImage API on OSX.
8648 Hardware acceleration is based on an OpenGL context. Usually, this means it is
8649 processed by video hardware. However, software-based OpenGL implementations
8650 exist which means there is no guarantee for hardware processing. It depends on
8653 There are many filters and image generators provided by Apple that come with a
8654 large variety of options. The filter has to be referenced by its name along
8657 The coreimage filter accepts the following options:
8660 List all available filters and generators along with all their respective
8661 options as well as possible minimum and maximum values along with the default
8668 Specify all filters by their respective name and options.
8669 Use @var{list_filters} to determine all valid filter names and options.
8670 Numerical options are specified by a float value and are automatically clamped
8671 to their respective value range. Vector and color options have to be specified
8672 by a list of space separated float values. Character escaping has to be done.
8673 A special option name @code{default} is available to use default options for a
8676 It is required to specify either @code{default} or at least one of the filter options.
8677 All omitted options are used with their default values.
8678 The syntax of the filter string is as follows:
8680 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
8684 Specify a rectangle where the output of the filter chain is copied into the
8685 input image. It is given by a list of space separated float values:
8687 output_rect=x\ y\ width\ height
8689 If not given, the output rectangle equals the dimensions of the input image.
8690 The output rectangle is automatically cropped at the borders of the input
8691 image. Negative values are valid for each component.
8693 output_rect=25\ 25\ 100\ 100
8697 Several filters can be chained for successive processing without GPU-HOST
8698 transfers allowing for fast processing of complex filter chains.
8699 Currently, only filters with zero (generators) or exactly one (filters) input
8700 image and one output image are supported. Also, transition filters are not yet
8703 Some filters generate output images with additional padding depending on the
8704 respective filter kernel. The padding is automatically removed to ensure the
8705 filter output has the same size as the input image.
8707 For image generators, the size of the output image is determined by the
8708 previous output image of the filter chain or the input image of the whole
8709 filterchain, respectively. The generators do not use the pixel information of
8710 this image to generate their output. However, the generated output is
8711 blended onto this image, resulting in partial or complete coverage of the
8714 The @ref{coreimagesrc} video source can be used for generating input images
8715 which are directly fed into the filter chain. By using it, providing input
8716 images by another video source or an input video is not required.
8718 @subsection Examples
8723 List all filters available:
8725 coreimage=list_filters=true
8729 Use the CIBoxBlur filter with default options to blur an image:
8731 coreimage=filter=CIBoxBlur@@default
8735 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8736 its center at 100x100 and a radius of 50 pixels:
8738 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8742 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8743 given as complete and escaped command-line for Apple's standard bash shell:
8745 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8751 Cover a rectangular object
8753 It accepts the following options:
8757 Filepath of the optional cover image, needs to be in yuv420.
8762 It accepts the following values:
8765 cover it by the supplied image
8767 cover it by interpolating the surrounding pixels
8770 Default value is @var{blur}.
8773 @subsection Examples
8777 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8779 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8785 Crop the input video to given dimensions.
8787 It accepts the following parameters:
8791 The width of the output video. It defaults to @code{iw}.
8792 This expression is evaluated only once during the filter
8793 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8796 The height of the output video. It defaults to @code{ih}.
8797 This expression is evaluated only once during the filter
8798 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8801 The horizontal position, in the input video, of the left edge of the output
8802 video. It defaults to @code{(in_w-out_w)/2}.
8803 This expression is evaluated per-frame.
8806 The vertical position, in the input video, of the top edge of the output video.
8807 It defaults to @code{(in_h-out_h)/2}.
8808 This expression is evaluated per-frame.
8811 If set to 1 will force the output display aspect ratio
8812 to be the same of the input, by changing the output sample aspect
8813 ratio. It defaults to 0.
8816 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8817 width/height/x/y as specified and will not be rounded to nearest smaller value.
8821 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8822 expressions containing the following constants:
8827 The computed values for @var{x} and @var{y}. They are evaluated for
8832 The input width and height.
8836 These are the same as @var{in_w} and @var{in_h}.
8840 The output (cropped) width and height.
8844 These are the same as @var{out_w} and @var{out_h}.
8847 same as @var{iw} / @var{ih}
8850 input sample aspect ratio
8853 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8857 horizontal and vertical chroma subsample values. For example for the
8858 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8861 The number of the input frame, starting from 0.
8864 the position in the file of the input frame, NAN if unknown
8867 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8871 The expression for @var{out_w} may depend on the value of @var{out_h},
8872 and the expression for @var{out_h} may depend on @var{out_w}, but they
8873 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8874 evaluated after @var{out_w} and @var{out_h}.
8876 The @var{x} and @var{y} parameters specify the expressions for the
8877 position of the top-left corner of the output (non-cropped) area. They
8878 are evaluated for each frame. If the evaluated value is not valid, it
8879 is approximated to the nearest valid value.
8881 The expression for @var{x} may depend on @var{y}, and the expression
8882 for @var{y} may depend on @var{x}.
8884 @subsection Examples
8888 Crop area with size 100x100 at position (12,34).
8893 Using named options, the example above becomes:
8895 crop=w=100:h=100:x=12:y=34
8899 Crop the central input area with size 100x100:
8905 Crop the central input area with size 2/3 of the input video:
8907 crop=2/3*in_w:2/3*in_h
8911 Crop the input video central square:
8918 Delimit the rectangle with the top-left corner placed at position
8919 100:100 and the right-bottom corner corresponding to the right-bottom
8920 corner of the input image.
8922 crop=in_w-100:in_h-100:100:100
8926 Crop 10 pixels from the left and right borders, and 20 pixels from
8927 the top and bottom borders
8929 crop=in_w-2*10:in_h-2*20
8933 Keep only the bottom right quarter of the input image:
8935 crop=in_w/2:in_h/2:in_w/2:in_h/2
8939 Crop height for getting Greek harmony:
8941 crop=in_w:1/PHI*in_w
8945 Apply trembling effect:
8947 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)
8951 Apply erratic camera effect depending on timestamp:
8953 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)"
8957 Set x depending on the value of y:
8959 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8963 @subsection Commands
8965 This filter supports the following commands:
8971 Set width/height of the output video and the horizontal/vertical position
8973 The command accepts the same syntax of the corresponding option.
8975 If the specified expression is not valid, it is kept at its current
8981 Auto-detect the crop size.
8983 It calculates the necessary cropping parameters and prints the
8984 recommended parameters via the logging system. The detected dimensions
8985 correspond to the non-black area of the input video.
8987 It accepts the following parameters:
8992 Set higher black value threshold, which can be optionally specified
8993 from nothing (0) to everything (255 for 8-bit based formats). An intensity
8994 value greater to the set value is considered non-black. It defaults to 24.
8995 You can also specify a value between 0.0 and 1.0 which will be scaled depending
8996 on the bitdepth of the pixel format.
8999 The value which the width/height should be divisible by. It defaults to
9000 16. The offset is automatically adjusted to center the video. Use 2 to
9001 get only even dimensions (needed for 4:2:2 video). 16 is best when
9002 encoding to most video codecs.
9005 Set the number of initial frames for which evaluation is skipped.
9006 Default is 2. Range is 0 to INT_MAX.
9008 @item reset_count, reset
9009 Set the counter that determines after how many frames cropdetect will
9010 reset the previously detected largest video area and start over to
9011 detect the current optimal crop area. Default value is 0.
9013 This can be useful when channel logos distort the video area. 0
9014 indicates 'never reset', and returns the largest area encountered during
9021 Delay video filtering until a given wallclock timestamp. The filter first
9022 passes on @option{preroll} amount of frames, then it buffers at most
9023 @option{buffer} amount of frames and waits for the cue. After reaching the cue
9024 it forwards the buffered frames and also any subsequent frames coming in its
9027 The filter can be used synchronize the output of multiple ffmpeg processes for
9028 realtime output devices like decklink. By putting the delay in the filtering
9029 chain and pre-buffering frames the process can pass on data to output almost
9030 immediately after the target wallclock timestamp is reached.
9032 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
9038 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
9041 The duration of content to pass on as preroll expressed in seconds. Default is 0.
9044 The maximum duration of content to buffer before waiting for the cue expressed
9045 in seconds. Default is 0.
9052 Apply color adjustments using curves.
9054 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
9055 component (red, green and blue) has its values defined by @var{N} key points
9056 tied from each other using a smooth curve. The x-axis represents the pixel
9057 values from the input frame, and the y-axis the new pixel values to be set for
9060 By default, a component curve is defined by the two points @var{(0;0)} and
9061 @var{(1;1)}. This creates a straight line where each original pixel value is
9062 "adjusted" to its own value, which means no change to the image.
9064 The filter allows you to redefine these two points and add some more. A new
9065 curve (using a natural cubic spline interpolation) will be define to pass
9066 smoothly through all these new coordinates. The new defined points needs to be
9067 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
9068 be in the @var{[0;1]} interval. If the computed curves happened to go outside
9069 the vector spaces, the values will be clipped accordingly.
9071 The filter accepts the following options:
9075 Select one of the available color presets. This option can be used in addition
9076 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
9077 options takes priority on the preset values.
9078 Available presets are:
9081 @item color_negative
9084 @item increase_contrast
9086 @item linear_contrast
9087 @item medium_contrast
9089 @item strong_contrast
9092 Default is @code{none}.
9094 Set the master key points. These points will define a second pass mapping. It
9095 is sometimes called a "luminance" or "value" mapping. It can be used with
9096 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
9097 post-processing LUT.
9099 Set the key points for the red component.
9101 Set the key points for the green component.
9103 Set the key points for the blue component.
9105 Set the key points for all components (not including master).
9106 Can be used in addition to the other key points component
9107 options. In this case, the unset component(s) will fallback on this
9108 @option{all} setting.
9110 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
9112 Save Gnuplot script of the curves in specified file.
9115 To avoid some filtergraph syntax conflicts, each key points list need to be
9116 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
9118 @subsection Examples
9122 Increase slightly the middle level of blue:
9124 curves=blue='0/0 0.5/0.58 1/1'
9130 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'
9132 Here we obtain the following coordinates for each components:
9135 @code{(0;0.11) (0.42;0.51) (1;0.95)}
9137 @code{(0;0) (0.50;0.48) (1;1)}
9139 @code{(0;0.22) (0.49;0.44) (1;0.80)}
9143 The previous example can also be achieved with the associated built-in preset:
9145 curves=preset=vintage
9155 Use a Photoshop preset and redefine the points of the green component:
9157 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
9161 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
9162 and @command{gnuplot}:
9164 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
9165 gnuplot -p /tmp/curves.plt
9171 Video data analysis filter.
9173 This filter shows hexadecimal pixel values of part of video.
9175 The filter accepts the following options:
9179 Set output video size.
9182 Set x offset from where to pick pixels.
9185 Set y offset from where to pick pixels.
9188 Set scope mode, can be one of the following:
9191 Draw hexadecimal pixel values with white color on black background.
9194 Draw hexadecimal pixel values with input video pixel color on black
9198 Draw hexadecimal pixel values on color background picked from input video,
9199 the text color is picked in such way so its always visible.
9203 Draw rows and columns numbers on left and top of video.
9206 Set background opacity.
9209 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
9213 Apply Directional blur filter.
9215 The filter accepts the following options:
9219 Set angle of directional blur. Default is @code{45}.
9222 Set radius of directional blur. Default is @code{5}.
9225 Set which planes to filter. By default all planes are filtered.
9228 @subsection Commands
9229 This filter supports same @ref{commands} as options.
9230 The command accepts the same syntax of the corresponding option.
9232 If the specified expression is not valid, it is kept at its current
9237 Denoise frames using 2D DCT (frequency domain filtering).
9239 This filter is not designed for real time.
9241 The filter accepts the following options:
9245 Set the noise sigma constant.
9247 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
9248 coefficient (absolute value) below this threshold with be dropped.
9250 If you need a more advanced filtering, see @option{expr}.
9252 Default is @code{0}.
9255 Set number overlapping pixels for each block. Since the filter can be slow, you
9256 may want to reduce this value, at the cost of a less effective filter and the
9257 risk of various artefacts.
9259 If the overlapping value doesn't permit processing the whole input width or
9260 height, a warning will be displayed and according borders won't be denoised.
9262 Default value is @var{blocksize}-1, which is the best possible setting.
9265 Set the coefficient factor expression.
9267 For each coefficient of a DCT block, this expression will be evaluated as a
9268 multiplier value for the coefficient.
9270 If this is option is set, the @option{sigma} option will be ignored.
9272 The absolute value of the coefficient can be accessed through the @var{c}
9276 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
9277 @var{blocksize}, which is the width and height of the processed blocks.
9279 The default value is @var{3} (8x8) and can be raised to @var{4} for a
9280 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
9281 on the speed processing. Also, a larger block size does not necessarily means a
9285 @subsection Examples
9287 Apply a denoise with a @option{sigma} of @code{4.5}:
9292 The same operation can be achieved using the expression system:
9294 dctdnoiz=e='gte(c, 4.5*3)'
9297 Violent denoise using a block size of @code{16x16}:
9304 Remove banding artifacts from input video.
9305 It works by replacing banded pixels with average value of referenced pixels.
9307 The filter accepts the following options:
9314 Set banding detection threshold for each plane. Default is 0.02.
9315 Valid range is 0.00003 to 0.5.
9316 If difference between current pixel and reference pixel is less than threshold,
9317 it will be considered as banded.
9320 Banding detection range in pixels. Default is 16. If positive, random number
9321 in range 0 to set value will be used. If negative, exact absolute value
9323 The range defines square of four pixels around current pixel.
9326 Set direction in radians from which four pixel will be compared. If positive,
9327 random direction from 0 to set direction will be picked. If negative, exact of
9328 absolute value will be picked. For example direction 0, -PI or -2*PI radians
9329 will pick only pixels on same row and -PI/2 will pick only pixels on same
9333 If enabled, current pixel is compared with average value of all four
9334 surrounding pixels. The default is enabled. If disabled current pixel is
9335 compared with all four surrounding pixels. The pixel is considered banded
9336 if only all four differences with surrounding pixels are less than threshold.
9339 If enabled, current pixel is changed if and only if all pixel components are banded,
9340 e.g. banding detection threshold is triggered for all color components.
9341 The default is disabled.
9346 Remove blocking artifacts from input video.
9348 The filter accepts the following options:
9352 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
9353 This controls what kind of deblocking is applied.
9356 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
9362 Set blocking detection thresholds. Allowed range is 0 to 1.
9363 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
9364 Using higher threshold gives more deblocking strength.
9365 Setting @var{alpha} controls threshold detection at exact edge of block.
9366 Remaining options controls threshold detection near the edge. Each one for
9367 below/above or left/right. Setting any of those to @var{0} disables
9371 Set planes to filter. Default is to filter all available planes.
9374 @subsection Examples
9378 Deblock using weak filter and block size of 4 pixels.
9380 deblock=filter=weak:block=4
9384 Deblock using strong filter, block size of 4 pixels and custom thresholds for
9385 deblocking more edges.
9387 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
9391 Similar as above, but filter only first plane.
9393 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
9397 Similar as above, but filter only second and third plane.
9399 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
9406 Drop duplicated frames at regular intervals.
9408 The filter accepts the following options:
9412 Set the number of frames from which one will be dropped. Setting this to
9413 @var{N} means one frame in every batch of @var{N} frames will be dropped.
9414 Default is @code{5}.
9417 Set the threshold for duplicate detection. If the difference metric for a frame
9418 is less than or equal to this value, then it is declared as duplicate. Default
9422 Set scene change threshold. Default is @code{15}.
9426 Set the size of the x and y-axis blocks used during metric calculations.
9427 Larger blocks give better noise suppression, but also give worse detection of
9428 small movements. Must be a power of two. Default is @code{32}.
9431 Mark main input as a pre-processed input and activate clean source input
9432 stream. This allows the input to be pre-processed with various filters to help
9433 the metrics calculation while keeping the frame selection lossless. When set to
9434 @code{1}, the first stream is for the pre-processed input, and the second
9435 stream is the clean source from where the kept frames are chosen. Default is
9439 Set whether or not chroma is considered in the metric calculations. Default is
9445 Apply 2D deconvolution of video stream in frequency domain using second stream
9448 The filter accepts the following options:
9452 Set which planes to process.
9455 Set which impulse video frames will be processed, can be @var{first}
9456 or @var{all}. Default is @var{all}.
9459 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
9460 and height are not same and not power of 2 or if stream prior to convolving
9464 The @code{deconvolve} filter also supports the @ref{framesync} options.
9468 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
9470 It accepts the following options:
9474 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
9475 @var{rainbows} for cross-color reduction.
9478 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
9481 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
9484 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
9487 Set temporal chroma threshold. Lower values increases reduction of cross-color.
9492 Apply deflate effect to the video.
9494 This filter replaces the pixel by the local(3x3) average by taking into account
9495 only values lower than the pixel.
9497 It accepts the following options:
9504 Limit the maximum change for each plane, default is 65535.
9505 If 0, plane will remain unchanged.
9508 @subsection Commands
9510 This filter supports the all above options as @ref{commands}.
9514 Remove temporal frame luminance variations.
9516 It accepts the following options:
9520 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
9523 Set averaging mode to smooth temporal luminance variations.
9525 Available values are:
9550 Do not actually modify frame. Useful when one only wants metadata.
9555 Remove judder produced by partially interlaced telecined content.
9557 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
9558 source was partially telecined content then the output of @code{pullup,dejudder}
9559 will have a variable frame rate. May change the recorded frame rate of the
9560 container. Aside from that change, this filter will not affect constant frame
9563 The option available in this filter is:
9567 Specify the length of the window over which the judder repeats.
9569 Accepts any integer greater than 1. Useful values are:
9573 If the original was telecined from 24 to 30 fps (Film to NTSC).
9576 If the original was telecined from 25 to 30 fps (PAL to NTSC).
9579 If a mixture of the two.
9582 The default is @samp{4}.
9587 Suppress a TV station logo by a simple interpolation of the surrounding
9588 pixels. Just set a rectangle covering the logo and watch it disappear
9589 (and sometimes something even uglier appear - your mileage may vary).
9591 It accepts the following parameters:
9596 Specify the top left corner coordinates of the logo. They must be
9601 Specify the width and height of the logo to clear. They must be
9605 Specify the thickness of the fuzzy edge of the rectangle (added to
9606 @var{w} and @var{h}). The default value is 1. This option is
9607 deprecated, setting higher values should no longer be necessary and
9611 When set to 1, a green rectangle is drawn on the screen to simplify
9612 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
9613 The default value is 0.
9615 The rectangle is drawn on the outermost pixels which will be (partly)
9616 replaced with interpolated values. The values of the next pixels
9617 immediately outside this rectangle in each direction will be used to
9618 compute the interpolated pixel values inside the rectangle.
9622 @subsection Examples
9626 Set a rectangle covering the area with top left corner coordinates 0,0
9627 and size 100x77, and a band of size 10:
9629 delogo=x=0:y=0:w=100:h=77:band=10
9637 Remove the rain in the input image/video by applying the derain methods based on
9638 convolutional neural networks. Supported models:
9642 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
9643 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
9646 Training as well as model generation scripts are provided in
9647 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
9649 Native model files (.model) can be generated from TensorFlow model
9650 files (.pb) by using tools/python/convert.py
9652 The filter accepts the following options:
9656 Specify which filter to use. This option accepts the following values:
9660 Derain filter. To conduct derain filter, you need to use a derain model.
9663 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
9665 Default value is @samp{derain}.
9668 Specify which DNN backend to use for model loading and execution. This option accepts
9669 the following values:
9673 Native implementation of DNN loading and execution.
9676 TensorFlow backend. To enable this backend you
9677 need to install the TensorFlow for C library (see
9678 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9679 @code{--enable-libtensorflow}
9681 Default value is @samp{native}.
9684 Set path to model file specifying network architecture and its parameters.
9685 Note that different backends use different file formats. TensorFlow and native
9686 backend can load files for only its format.
9689 It can also be finished with @ref{dnn_processing} filter.
9693 Attempt to fix small changes in horizontal and/or vertical shift. This
9694 filter helps remove camera shake from hand-holding a camera, bumping a
9695 tripod, moving on a vehicle, etc.
9697 The filter accepts the following options:
9705 Specify a rectangular area where to limit the search for motion
9707 If desired the search for motion vectors can be limited to a
9708 rectangular area of the frame defined by its top left corner, width
9709 and height. These parameters have the same meaning as the drawbox
9710 filter which can be used to visualise the position of the bounding
9713 This is useful when simultaneous movement of subjects within the frame
9714 might be confused for camera motion by the motion vector search.
9716 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
9717 then the full frame is used. This allows later options to be set
9718 without specifying the bounding box for the motion vector search.
9720 Default - search the whole frame.
9724 Specify the maximum extent of movement in x and y directions in the
9725 range 0-64 pixels. Default 16.
9728 Specify how to generate pixels to fill blanks at the edge of the
9729 frame. Available values are:
9732 Fill zeroes at blank locations
9734 Original image at blank locations
9736 Extruded edge value at blank locations
9738 Mirrored edge at blank locations
9740 Default value is @samp{mirror}.
9743 Specify the blocksize to use for motion search. Range 4-128 pixels,
9747 Specify the contrast threshold for blocks. Only blocks with more than
9748 the specified contrast (difference between darkest and lightest
9749 pixels) will be considered. Range 1-255, default 125.
9752 Specify the search strategy. Available values are:
9755 Set exhaustive search
9757 Set less exhaustive search.
9759 Default value is @samp{exhaustive}.
9762 If set then a detailed log of the motion search is written to the
9769 Remove unwanted contamination of foreground colors, caused by reflected color of
9770 greenscreen or bluescreen.
9772 This filter accepts the following options:
9776 Set what type of despill to use.
9779 Set how spillmap will be generated.
9782 Set how much to get rid of still remaining spill.
9785 Controls amount of red in spill area.
9788 Controls amount of green in spill area.
9789 Should be -1 for greenscreen.
9792 Controls amount of blue in spill area.
9793 Should be -1 for bluescreen.
9796 Controls brightness of spill area, preserving colors.
9799 Modify alpha from generated spillmap.
9802 @subsection Commands
9804 This filter supports the all above options as @ref{commands}.
9808 Apply an exact inverse of the telecine operation. It requires a predefined
9809 pattern specified using the pattern option which must be the same as that passed
9810 to the telecine filter.
9812 This filter accepts the following options:
9821 The default value is @code{top}.
9825 A string of numbers representing the pulldown pattern you wish to apply.
9826 The default value is @code{23}.
9829 A number representing position of the first frame with respect to the telecine
9830 pattern. This is to be used if the stream is cut. The default value is @code{0}.
9835 Apply dilation effect to the video.
9837 This filter replaces the pixel by the local(3x3) maximum.
9839 It accepts the following options:
9846 Limit the maximum change for each plane, default is 65535.
9847 If 0, plane will remain unchanged.
9850 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9853 Flags to local 3x3 coordinates maps like this:
9860 @subsection Commands
9862 This filter supports the all above options as @ref{commands}.
9866 Displace pixels as indicated by second and third input stream.
9868 It takes three input streams and outputs one stream, the first input is the
9869 source, and second and third input are displacement maps.
9871 The second input specifies how much to displace pixels along the
9872 x-axis, while the third input specifies how much to displace pixels
9874 If one of displacement map streams terminates, last frame from that
9875 displacement map will be used.
9877 Note that once generated, displacements maps can be reused over and over again.
9879 A description of the accepted options follows.
9883 Set displace behavior for pixels that are out of range.
9885 Available values are:
9888 Missing pixels are replaced by black pixels.
9891 Adjacent pixels will spread out to replace missing pixels.
9894 Out of range pixels are wrapped so they point to pixels of other side.
9897 Out of range pixels will be replaced with mirrored pixels.
9899 Default is @samp{smear}.
9903 @subsection Examples
9907 Add ripple effect to rgb input of video size hd720:
9909 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
9913 Add wave effect to rgb input of video size hd720:
9915 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
9919 @anchor{dnn_processing}
9920 @section dnn_processing
9922 Do image processing with deep neural networks. It works together with another filter
9923 which converts the pixel format of the Frame to what the dnn network requires.
9925 The filter accepts the following options:
9929 Specify which DNN backend to use for model loading and execution. This option accepts
9930 the following values:
9934 Native implementation of DNN loading and execution.
9937 TensorFlow backend. To enable this backend you
9938 need to install the TensorFlow for C library (see
9939 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9940 @code{--enable-libtensorflow}
9943 OpenVINO backend. To enable this backend you
9944 need to build and install the OpenVINO for C library (see
9945 @url{https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md}) and configure FFmpeg with
9946 @code{--enable-libopenvino} (--extra-cflags=-I... --extra-ldflags=-L... might
9947 be needed if the header files and libraries are not installed into system path)
9951 Default value is @samp{native}.
9954 Set path to model file specifying network architecture and its parameters.
9955 Note that different backends use different file formats. TensorFlow, OpenVINO and native
9956 backend can load files for only its format.
9958 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9961 Set the input name of the dnn network.
9964 Set the output name of the dnn network.
9967 use DNN async execution if set (default: set),
9968 roll back to sync execution if the backend does not support async.
9972 @subsection Examples
9976 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
9978 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
9982 Halve the pixel value of the frame with format gray32f:
9984 ffmpeg -i input.jpg -vf format=grayf32,dnn_processing=model=halve_gray_float.model:input=dnn_in:output=dnn_out:dnn_backend=native -y out.native.png
9988 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
9990 ./ffmpeg -i 480p.jpg -vf format=yuv420p,scale=w=iw*2:h=ih*2,dnn_processing=dnn_backend=tensorflow:model=srcnn.pb:input=x:output=y -y srcnn.jpg
9994 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
9996 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
10003 Draw a colored box on the input image.
10005 It accepts the following parameters:
10010 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
10014 The expressions which specify the width and height of the box; if 0 they are interpreted as
10015 the input width and height. It defaults to 0.
10018 Specify the color of the box to write. For the general syntax of this option,
10019 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10020 value @code{invert} is used, the box edge color is the same as the
10021 video with inverted luma.
10024 The expression which sets the thickness of the box edge.
10025 A value of @code{fill} will create a filled box. Default value is @code{3}.
10027 See below for the list of accepted constants.
10030 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
10031 will overwrite the video's color and alpha pixels.
10032 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
10035 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10036 following constants:
10040 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10044 horizontal and vertical chroma subsample values. For example for the
10045 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10049 The input width and height.
10052 The input sample aspect ratio.
10056 The x and y offset coordinates where the box is drawn.
10060 The width and height of the drawn box.
10063 The thickness of the drawn box.
10065 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10066 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10070 @subsection Examples
10074 Draw a black box around the edge of the input image:
10080 Draw a box with color red and an opacity of 50%:
10082 drawbox=10:20:200:60:red@@0.5
10085 The previous example can be specified as:
10087 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
10091 Fill the box with pink color:
10093 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
10097 Draw a 2-pixel red 2.40:1 mask:
10099 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
10103 @subsection Commands
10104 This filter supports same commands as options.
10105 The command accepts the same syntax of the corresponding option.
10107 If the specified expression is not valid, it is kept at its current
10112 Draw a graph using input video metadata.
10114 It accepts the following parameters:
10118 Set 1st frame metadata key from which metadata values will be used to draw a graph.
10121 Set 1st foreground color expression.
10124 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
10127 Set 2nd foreground color expression.
10130 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
10133 Set 3rd foreground color expression.
10136 Set 4th frame metadata key from which metadata values will be used to draw a graph.
10139 Set 4th foreground color expression.
10142 Set minimal value of metadata value.
10145 Set maximal value of metadata value.
10148 Set graph background color. Default is white.
10153 Available values for mode is:
10160 Default is @code{line}.
10165 Available values for slide is:
10168 Draw new frame when right border is reached.
10171 Replace old columns with new ones.
10174 Scroll from right to left.
10177 Scroll from left to right.
10180 Draw single picture.
10183 Default is @code{frame}.
10186 Set size of graph video. For the syntax of this option, check the
10187 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
10188 The default value is @code{900x256}.
10191 Set the output frame rate. Default value is @code{25}.
10193 The foreground color expressions can use the following variables:
10196 Minimal value of metadata value.
10199 Maximal value of metadata value.
10202 Current metadata key value.
10205 The color is defined as 0xAABBGGRR.
10208 Example using metadata from @ref{signalstats} filter:
10210 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
10213 Example using metadata from @ref{ebur128} filter:
10215 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
10220 Draw a grid on the input image.
10222 It accepts the following parameters:
10227 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
10231 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
10232 input width and height, respectively, minus @code{thickness}, so image gets
10233 framed. Default to 0.
10236 Specify the color of the grid. For the general syntax of this option,
10237 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10238 value @code{invert} is used, the grid color is the same as the
10239 video with inverted luma.
10242 The expression which sets the thickness of the grid line. Default value is @code{1}.
10244 See below for the list of accepted constants.
10247 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
10248 will overwrite the video's color and alpha pixels.
10249 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
10252 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10253 following constants:
10257 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10261 horizontal and vertical chroma subsample values. For example for the
10262 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10266 The input grid cell width and height.
10269 The input sample aspect ratio.
10273 The x and y coordinates of some point of grid intersection (meant to configure offset).
10277 The width and height of the drawn cell.
10280 The thickness of the drawn cell.
10282 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10283 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10287 @subsection Examples
10291 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
10293 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
10297 Draw a white 3x3 grid with an opacity of 50%:
10299 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
10303 @subsection Commands
10304 This filter supports same commands as options.
10305 The command accepts the same syntax of the corresponding option.
10307 If the specified expression is not valid, it is kept at its current
10313 Draw a text string or text from a specified file on top of a video, using the
10314 libfreetype library.
10316 To enable compilation of this filter, you need to configure FFmpeg with
10317 @code{--enable-libfreetype}.
10318 To enable default font fallback and the @var{font} option you need to
10319 configure FFmpeg with @code{--enable-libfontconfig}.
10320 To enable the @var{text_shaping} option, you need to configure FFmpeg with
10321 @code{--enable-libfribidi}.
10325 It accepts the following parameters:
10330 Used to draw a box around text using the background color.
10331 The value must be either 1 (enable) or 0 (disable).
10332 The default value of @var{box} is 0.
10335 Set the width of the border to be drawn around the box using @var{boxcolor}.
10336 The default value of @var{boxborderw} is 0.
10339 The color to be used for drawing box around text. For the syntax of this
10340 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10342 The default value of @var{boxcolor} is "white".
10345 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
10346 The default value of @var{line_spacing} is 0.
10349 Set the width of the border to be drawn around the text using @var{bordercolor}.
10350 The default value of @var{borderw} is 0.
10353 Set the color to be used for drawing border around text. For the syntax of this
10354 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10356 The default value of @var{bordercolor} is "black".
10359 Select how the @var{text} is expanded. Can be either @code{none},
10360 @code{strftime} (deprecated) or
10361 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
10365 Set a start time for the count. Value is in microseconds. Only applied
10366 in the deprecated strftime expansion mode. To emulate in normal expansion
10367 mode use the @code{pts} function, supplying the start time (in seconds)
10368 as the second argument.
10371 If true, check and fix text coords to avoid clipping.
10374 The color to be used for drawing fonts. For the syntax of this option, check
10375 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10377 The default value of @var{fontcolor} is "black".
10379 @item fontcolor_expr
10380 String which is expanded the same way as @var{text} to obtain dynamic
10381 @var{fontcolor} value. By default this option has empty value and is not
10382 processed. When this option is set, it overrides @var{fontcolor} option.
10385 The font family to be used for drawing text. By default Sans.
10388 The font file to be used for drawing text. The path must be included.
10389 This parameter is mandatory if the fontconfig support is disabled.
10392 Draw the text applying alpha blending. The value can
10393 be a number between 0.0 and 1.0.
10394 The expression accepts the same variables @var{x, y} as well.
10395 The default value is 1.
10396 Please see @var{fontcolor_expr}.
10399 The font size to be used for drawing text.
10400 The default value of @var{fontsize} is 16.
10403 If set to 1, attempt to shape the text (for example, reverse the order of
10404 right-to-left text and join Arabic characters) before drawing it.
10405 Otherwise, just draw the text exactly as given.
10406 By default 1 (if supported).
10408 @item ft_load_flags
10409 The flags to be used for loading the fonts.
10411 The flags map the corresponding flags supported by libfreetype, and are
10412 a combination of the following values:
10419 @item vertical_layout
10420 @item force_autohint
10423 @item ignore_global_advance_width
10425 @item ignore_transform
10427 @item linear_design
10431 Default value is "default".
10433 For more information consult the documentation for the FT_LOAD_*
10437 The color to be used for drawing a shadow behind the drawn text. For the
10438 syntax of this option, check the @ref{color syntax,,"Color" section in the
10439 ffmpeg-utils manual,ffmpeg-utils}.
10441 The default value of @var{shadowcolor} is "black".
10445 The x and y offsets for the text shadow position with respect to the
10446 position of the text. They can be either positive or negative
10447 values. The default value for both is "0".
10450 The starting frame number for the n/frame_num variable. The default value
10454 The size in number of spaces to use for rendering the tab.
10455 Default value is 4.
10458 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
10459 format. It can be used with or without text parameter. @var{timecode_rate}
10460 option must be specified.
10462 @item timecode_rate, rate, r
10463 Set the timecode frame rate (timecode only). Value will be rounded to nearest
10464 integer. Minimum value is "1".
10465 Drop-frame timecode is supported for frame rates 30 & 60.
10468 If set to 1, the output of the timecode option will wrap around at 24 hours.
10469 Default is 0 (disabled).
10472 The text string to be drawn. The text must be a sequence of UTF-8
10473 encoded characters.
10474 This parameter is mandatory if no file is specified with the parameter
10478 A text file containing text to be drawn. The text must be a sequence
10479 of UTF-8 encoded characters.
10481 This parameter is mandatory if no text string is specified with the
10482 parameter @var{text}.
10484 If both @var{text} and @var{textfile} are specified, an error is thrown.
10487 If set to 1, the @var{textfile} will be reloaded before each frame.
10488 Be sure to update it atomically, or it may be read partially, or even fail.
10492 The expressions which specify the offsets where text will be drawn
10493 within the video frame. They are relative to the top/left border of the
10496 The default value of @var{x} and @var{y} is "0".
10498 See below for the list of accepted constants and functions.
10501 The parameters for @var{x} and @var{y} are expressions containing the
10502 following constants and functions:
10506 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
10510 horizontal and vertical chroma subsample values. For example for the
10511 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10514 the height of each text line
10522 @item max_glyph_a, ascent
10523 the maximum distance from the baseline to the highest/upper grid
10524 coordinate used to place a glyph outline point, for all the rendered
10526 It is a positive value, due to the grid's orientation with the Y axis
10529 @item max_glyph_d, descent
10530 the maximum distance from the baseline to the lowest grid coordinate
10531 used to place a glyph outline point, for all the rendered glyphs.
10532 This is a negative value, due to the grid's orientation, with the Y axis
10536 maximum glyph height, that is the maximum height for all the glyphs
10537 contained in the rendered text, it is equivalent to @var{ascent} -
10541 maximum glyph width, that is the maximum width for all the glyphs
10542 contained in the rendered text
10545 the number of input frame, starting from 0
10547 @item rand(min, max)
10548 return a random number included between @var{min} and @var{max}
10551 The input sample aspect ratio.
10554 timestamp expressed in seconds, NAN if the input timestamp is unknown
10557 the height of the rendered text
10560 the width of the rendered text
10564 the x and y offset coordinates where the text is drawn.
10566 These parameters allow the @var{x} and @var{y} expressions to refer
10567 to each other, so you can for example specify @code{y=x/dar}.
10570 A one character description of the current frame's picture type.
10573 The current packet's position in the input file or stream
10574 (in bytes, from the start of the input). A value of -1 indicates
10575 this info is not available.
10578 The current packet's duration, in seconds.
10581 The current packet's size (in bytes).
10584 @anchor{drawtext_expansion}
10585 @subsection Text expansion
10587 If @option{expansion} is set to @code{strftime},
10588 the filter recognizes strftime() sequences in the provided text and
10589 expands them accordingly. Check the documentation of strftime(). This
10590 feature is deprecated.
10592 If @option{expansion} is set to @code{none}, the text is printed verbatim.
10594 If @option{expansion} is set to @code{normal} (which is the default),
10595 the following expansion mechanism is used.
10597 The backslash character @samp{\}, followed by any character, always expands to
10598 the second character.
10600 Sequences of the form @code{%@{...@}} are expanded. The text between the
10601 braces is a function name, possibly followed by arguments separated by ':'.
10602 If the arguments contain special characters or delimiters (':' or '@}'),
10603 they should be escaped.
10605 Note that they probably must also be escaped as the value for the
10606 @option{text} option in the filter argument string and as the filter
10607 argument in the filtergraph description, and possibly also for the shell,
10608 that makes up to four levels of escaping; using a text file avoids these
10611 The following functions are available:
10616 The expression evaluation result.
10618 It must take one argument specifying the expression to be evaluated,
10619 which accepts the same constants and functions as the @var{x} and
10620 @var{y} values. Note that not all constants should be used, for
10621 example the text size is not known when evaluating the expression, so
10622 the constants @var{text_w} and @var{text_h} will have an undefined
10625 @item expr_int_format, eif
10626 Evaluate the expression's value and output as formatted integer.
10628 The first argument is the expression to be evaluated, just as for the @var{expr} function.
10629 The second argument specifies the output format. Allowed values are @samp{x},
10630 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
10631 @code{printf} function.
10632 The third parameter is optional and sets the number of positions taken by the output.
10633 It can be used to add padding with zeros from the left.
10636 The time at which the filter is running, expressed in UTC.
10637 It can accept an argument: a strftime() format string.
10640 The time at which the filter is running, expressed in the local time zone.
10641 It can accept an argument: a strftime() format string.
10644 Frame metadata. Takes one or two arguments.
10646 The first argument is mandatory and specifies the metadata key.
10648 The second argument is optional and specifies a default value, used when the
10649 metadata key is not found or empty.
10651 Available metadata can be identified by inspecting entries
10652 starting with TAG included within each frame section
10653 printed by running @code{ffprobe -show_frames}.
10655 String metadata generated in filters leading to
10656 the drawtext filter are also available.
10659 The frame number, starting from 0.
10662 A one character description of the current picture type.
10665 The timestamp of the current frame.
10666 It can take up to three arguments.
10668 The first argument is the format of the timestamp; it defaults to @code{flt}
10669 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
10670 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
10671 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
10672 @code{localtime} stands for the timestamp of the frame formatted as
10673 local time zone time.
10675 The second argument is an offset added to the timestamp.
10677 If the format is set to @code{hms}, a third argument @code{24HH} may be
10678 supplied to present the hour part of the formatted timestamp in 24h format
10681 If the format is set to @code{localtime} or @code{gmtime},
10682 a third argument may be supplied: a strftime() format string.
10683 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
10686 @subsection Commands
10688 This filter supports altering parameters via commands:
10691 Alter existing filter parameters.
10693 Syntax for the argument is the same as for filter invocation, e.g.
10696 fontsize=56:fontcolor=green:text='Hello World'
10699 Full filter invocation with sendcmd would look like this:
10702 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
10706 If the entire argument can't be parsed or applied as valid values then the filter will
10707 continue with its existing parameters.
10709 @subsection Examples
10713 Draw "Test Text" with font FreeSerif, using the default values for the
10714 optional parameters.
10717 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
10721 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
10722 and y=50 (counting from the top-left corner of the screen), text is
10723 yellow with a red box around it. Both the text and the box have an
10727 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
10728 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
10731 Note that the double quotes are not necessary if spaces are not used
10732 within the parameter list.
10735 Show the text at the center of the video frame:
10737 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
10741 Show the text at a random position, switching to a new position every 30 seconds:
10743 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)"
10747 Show a text line sliding from right to left in the last row of the video
10748 frame. The file @file{LONG_LINE} is assumed to contain a single line
10751 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
10755 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
10757 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
10761 Draw a single green letter "g", at the center of the input video.
10762 The glyph baseline is placed at half screen height.
10764 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
10768 Show text for 1 second every 3 seconds:
10770 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
10774 Use fontconfig to set the font. Note that the colons need to be escaped.
10776 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
10780 Draw "Test Text" with font size dependent on height of the video.
10782 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
10786 Print the date of a real-time encoding (see strftime(3)):
10788 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
10792 Show text fading in and out (appearing/disappearing):
10795 DS=1.0 # display start
10796 DE=10.0 # display end
10797 FID=1.5 # fade in duration
10798 FOD=5 # fade out duration
10799 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 @}"
10803 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
10804 and the @option{fontsize} value are included in the @option{y} offset.
10806 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
10807 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
10811 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
10812 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
10813 must have option @option{-export_path_metadata 1} for the special metadata fields
10814 to be available for filters.
10816 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
10821 For more information about libfreetype, check:
10822 @url{http://www.freetype.org/}.
10824 For more information about fontconfig, check:
10825 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
10827 For more information about libfribidi, check:
10828 @url{http://fribidi.org/}.
10830 @section edgedetect
10832 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
10834 The filter accepts the following options:
10839 Set low and high threshold values used by the Canny thresholding
10842 The high threshold selects the "strong" edge pixels, which are then
10843 connected through 8-connectivity with the "weak" edge pixels selected
10844 by the low threshold.
10846 @var{low} and @var{high} threshold values must be chosen in the range
10847 [0,1], and @var{low} should be lesser or equal to @var{high}.
10849 Default value for @var{low} is @code{20/255}, and default value for @var{high}
10853 Define the drawing mode.
10857 Draw white/gray wires on black background.
10860 Mix the colors to create a paint/cartoon effect.
10863 Apply Canny edge detector on all selected planes.
10865 Default value is @var{wires}.
10868 Select planes for filtering. By default all available planes are filtered.
10871 @subsection Examples
10875 Standard edge detection with custom values for the hysteresis thresholding:
10877 edgedetect=low=0.1:high=0.4
10881 Painting effect without thresholding:
10883 edgedetect=mode=colormix:high=0
10889 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10891 For each input image, the filter will compute the optimal mapping from
10892 the input to the output given the codebook length, that is the number
10893 of distinct output colors.
10895 This filter accepts the following options.
10898 @item codebook_length, l
10899 Set codebook length. The value must be a positive integer, and
10900 represents the number of distinct output colors. Default value is 256.
10903 Set the maximum number of iterations to apply for computing the optimal
10904 mapping. The higher the value the better the result and the higher the
10905 computation time. Default value is 1.
10908 Set a random seed, must be an integer included between 0 and
10909 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10910 will try to use a good random seed on a best effort basis.
10913 Set pal8 output pixel format. This option does not work with codebook
10914 length greater than 256.
10919 Measure graylevel entropy in histogram of color channels of video frames.
10921 It accepts the following parameters:
10925 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10927 @var{diff} mode measures entropy of histogram delta values, absolute differences
10928 between neighbour histogram values.
10932 Set brightness, contrast, saturation and approximate gamma adjustment.
10934 The filter accepts the following options:
10938 Set the contrast expression. The value must be a float value in range
10939 @code{-1000.0} to @code{1000.0}. The default value is "1".
10942 Set the brightness expression. The value must be a float value in
10943 range @code{-1.0} to @code{1.0}. The default value is "0".
10946 Set the saturation expression. The value must be a float in
10947 range @code{0.0} to @code{3.0}. The default value is "1".
10950 Set the gamma expression. The value must be a float in range
10951 @code{0.1} to @code{10.0}. The default value is "1".
10954 Set the gamma expression for red. The value must be a float in
10955 range @code{0.1} to @code{10.0}. The default value is "1".
10958 Set the gamma expression for green. The value must be a float in range
10959 @code{0.1} to @code{10.0}. The default value is "1".
10962 Set the gamma expression for blue. The value must be a float in range
10963 @code{0.1} to @code{10.0}. The default value is "1".
10966 Set the gamma weight expression. It can be used to reduce the effect
10967 of a high gamma value on bright image areas, e.g. keep them from
10968 getting overamplified and just plain white. The value must be a float
10969 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
10970 gamma correction all the way down while @code{1.0} leaves it at its
10971 full strength. Default is "1".
10974 Set when the expressions for brightness, contrast, saturation and
10975 gamma expressions are evaluated.
10977 It accepts the following values:
10980 only evaluate expressions once during the filter initialization or
10981 when a command is processed
10984 evaluate expressions for each incoming frame
10987 Default value is @samp{init}.
10990 The expressions accept the following parameters:
10993 frame count of the input frame starting from 0
10996 byte position of the corresponding packet in the input file, NAN if
11000 frame rate of the input video, NAN if the input frame rate is unknown
11003 timestamp expressed in seconds, NAN if the input timestamp is unknown
11006 @subsection Commands
11007 The filter supports the following commands:
11011 Set the contrast expression.
11014 Set the brightness expression.
11017 Set the saturation expression.
11020 Set the gamma expression.
11023 Set the gamma_r expression.
11026 Set gamma_g expression.
11029 Set gamma_b expression.
11032 Set gamma_weight expression.
11034 The command accepts the same syntax of the corresponding option.
11036 If the specified expression is not valid, it is kept at its current
11043 Apply erosion effect to the video.
11045 This filter replaces the pixel by the local(3x3) minimum.
11047 It accepts the following options:
11054 Limit the maximum change for each plane, default is 65535.
11055 If 0, plane will remain unchanged.
11058 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
11061 Flags to local 3x3 coordinates maps like this:
11068 @subsection Commands
11070 This filter supports the all above options as @ref{commands}.
11074 Deinterlace the input video ("estdif" stands for "Edge Slope
11075 Tracing Deinterlacing Filter").
11077 Spatial only filter that uses edge slope tracing algorithm
11078 to interpolate missing lines.
11079 It accepts the following parameters:
11083 The interlacing mode to adopt. It accepts one of the following values:
11087 Output one frame for each frame.
11089 Output one frame for each field.
11092 The default value is @code{field}.
11095 The picture field parity assumed for the input interlaced video. It accepts one
11096 of the following values:
11100 Assume the top field is first.
11102 Assume the bottom field is first.
11104 Enable automatic detection of field parity.
11107 The default value is @code{auto}.
11108 If the interlacing is unknown or the decoder does not export this information,
11109 top field first will be assumed.
11112 Specify which frames to deinterlace. Accepts one of the following
11117 Deinterlace all frames.
11119 Only deinterlace frames marked as interlaced.
11122 The default value is @code{all}.
11125 Specify the search radius for edge slope tracing. Default value is 1.
11126 Allowed range is from 1 to 15.
11129 Specify the search radius for best edge matching. Default value is 2.
11130 Allowed range is from 0 to 15.
11133 @subsection Commands
11134 This filter supports same @ref{commands} as options.
11136 @section extractplanes
11138 Extract color channel components from input video stream into
11139 separate grayscale video streams.
11141 The filter accepts the following option:
11145 Set plane(s) to extract.
11147 Available values for planes are:
11158 Choosing planes not available in the input will result in an error.
11159 That means you cannot select @code{r}, @code{g}, @code{b} planes
11160 with @code{y}, @code{u}, @code{v} planes at same time.
11163 @subsection Examples
11167 Extract luma, u and v color channel component from input video frame
11168 into 3 grayscale outputs:
11170 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
11176 Apply a fade-in/out effect to the input video.
11178 It accepts the following parameters:
11182 The effect type can be either "in" for a fade-in, or "out" for a fade-out
11184 Default is @code{in}.
11186 @item start_frame, s
11187 Specify the number of the frame to start applying the fade
11188 effect at. Default is 0.
11191 The number of frames that the fade effect lasts. At the end of the
11192 fade-in effect, the output video will have the same intensity as the input video.
11193 At the end of the fade-out transition, the output video will be filled with the
11194 selected @option{color}.
11198 If set to 1, fade only alpha channel, if one exists on the input.
11199 Default value is 0.
11201 @item start_time, st
11202 Specify the timestamp (in seconds) of the frame to start to apply the fade
11203 effect. If both start_frame and start_time are specified, the fade will start at
11204 whichever comes last. Default is 0.
11207 The number of seconds for which the fade effect has to last. At the end of the
11208 fade-in effect the output video will have the same intensity as the input video,
11209 at the end of the fade-out transition the output video will be filled with the
11210 selected @option{color}.
11211 If both duration and nb_frames are specified, duration is used. Default is 0
11212 (nb_frames is used by default).
11215 Specify the color of the fade. Default is "black".
11218 @subsection Examples
11222 Fade in the first 30 frames of video:
11227 The command above is equivalent to:
11233 Fade out the last 45 frames of a 200-frame video:
11236 fade=type=out:start_frame=155:nb_frames=45
11240 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
11242 fade=in:0:25, fade=out:975:25
11246 Make the first 5 frames yellow, then fade in from frame 5-24:
11248 fade=in:5:20:color=yellow
11252 Fade in alpha over first 25 frames of video:
11254 fade=in:0:25:alpha=1
11258 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
11260 fade=t=in:st=5.5:d=0.5
11266 Denoise frames using 3D FFT (frequency domain filtering).
11268 The filter accepts the following options:
11272 Set the noise sigma constant. This sets denoising strength.
11273 Default value is 1. Allowed range is from 0 to 30.
11274 Using very high sigma with low overlap may give blocking artifacts.
11277 Set amount of denoising. By default all detected noise is reduced.
11278 Default value is 1. Allowed range is from 0 to 1.
11281 Set size of block, Default is 4, can be 3, 4, 5 or 6.
11282 Actual size of block in pixels is 2 to power of @var{block}, so by default
11283 block size in pixels is 2^4 which is 16.
11286 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
11289 Set number of previous frames to use for denoising. By default is set to 0.
11292 Set number of next frames to to use for denoising. By default is set to 0.
11295 Set planes which will be filtered, by default are all available filtered
11300 Apply arbitrary expressions to samples in frequency domain
11304 Adjust the dc value (gain) of the luma plane of the image. The filter
11305 accepts an integer value in range @code{0} to @code{1000}. The default
11306 value is set to @code{0}.
11309 Adjust the dc value (gain) of the 1st chroma plane of the image. The
11310 filter accepts an integer value in range @code{0} to @code{1000}. The
11311 default value is set to @code{0}.
11314 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
11315 filter accepts an integer value in range @code{0} to @code{1000}. The
11316 default value is set to @code{0}.
11319 Set the frequency domain weight expression for the luma plane.
11322 Set the frequency domain weight expression for the 1st chroma plane.
11325 Set the frequency domain weight expression for the 2nd chroma plane.
11328 Set when the expressions are evaluated.
11330 It accepts the following values:
11333 Only evaluate expressions once during the filter initialization.
11336 Evaluate expressions for each incoming frame.
11339 Default value is @samp{init}.
11341 The filter accepts the following variables:
11344 The coordinates of the current sample.
11348 The width and height of the image.
11351 The number of input frame, starting from 0.
11354 @subsection Examples
11360 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
11366 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
11372 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
11378 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
11385 Extract a single field from an interlaced image using stride
11386 arithmetic to avoid wasting CPU time. The output frames are marked as
11389 The filter accepts the following options:
11393 Specify whether to extract the top (if the value is @code{0} or
11394 @code{top}) or the bottom field (if the value is @code{1} or
11400 Create new frames by copying the top and bottom fields from surrounding frames
11401 supplied as numbers by the hint file.
11405 Set file containing hints: absolute/relative frame numbers.
11407 There must be one line for each frame in a clip. Each line must contain two
11408 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
11409 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
11410 is current frame number for @code{absolute} mode or out of [-1, 1] range
11411 for @code{relative} mode. First number tells from which frame to pick up top
11412 field and second number tells from which frame to pick up bottom field.
11414 If optionally followed by @code{+} output frame will be marked as interlaced,
11415 else if followed by @code{-} output frame will be marked as progressive, else
11416 it will be marked same as input frame.
11417 If optionally followed by @code{t} output frame will use only top field, or in
11418 case of @code{b} it will use only bottom field.
11419 If line starts with @code{#} or @code{;} that line is skipped.
11422 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
11425 Example of first several lines of @code{hint} file for @code{relative} mode:
11427 0,0 - # first frame
11428 1,0 - # second frame, use third's frame top field and second's frame bottom field
11429 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
11444 @section fieldmatch
11446 Field matching filter for inverse telecine. It is meant to reconstruct the
11447 progressive frames from a telecined stream. The filter does not drop duplicated
11448 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
11449 followed by a decimation filter such as @ref{decimate} in the filtergraph.
11451 The separation of the field matching and the decimation is notably motivated by
11452 the possibility of inserting a de-interlacing filter fallback between the two.
11453 If the source has mixed telecined and real interlaced content,
11454 @code{fieldmatch} will not be able to match fields for the interlaced parts.
11455 But these remaining combed frames will be marked as interlaced, and thus can be
11456 de-interlaced by a later filter such as @ref{yadif} before decimation.
11458 In addition to the various configuration options, @code{fieldmatch} can take an
11459 optional second stream, activated through the @option{ppsrc} option. If
11460 enabled, the frames reconstruction will be based on the fields and frames from
11461 this second stream. This allows the first input to be pre-processed in order to
11462 help the various algorithms of the filter, while keeping the output lossless
11463 (assuming the fields are matched properly). Typically, a field-aware denoiser,
11464 or brightness/contrast adjustments can help.
11466 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
11467 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
11468 which @code{fieldmatch} is based on. While the semantic and usage are very
11469 close, some behaviour and options names can differ.
11471 The @ref{decimate} filter currently only works for constant frame rate input.
11472 If your input has mixed telecined (30fps) and progressive content with a lower
11473 framerate like 24fps use the following filterchain to produce the necessary cfr
11474 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
11476 The filter accepts the following options:
11480 Specify the assumed field order of the input stream. Available values are:
11484 Auto detect parity (use FFmpeg's internal parity value).
11486 Assume bottom field first.
11488 Assume top field first.
11491 Note that it is sometimes recommended not to trust the parity announced by the
11494 Default value is @var{auto}.
11497 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
11498 sense that it won't risk creating jerkiness due to duplicate frames when
11499 possible, but if there are bad edits or blended fields it will end up
11500 outputting combed frames when a good match might actually exist. On the other
11501 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
11502 but will almost always find a good frame if there is one. The other values are
11503 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
11504 jerkiness and creating duplicate frames versus finding good matches in sections
11505 with bad edits, orphaned fields, blended fields, etc.
11507 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
11509 Available values are:
11513 2-way matching (p/c)
11515 2-way matching, and trying 3rd match if still combed (p/c + n)
11517 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
11519 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
11520 still combed (p/c + n + u/b)
11522 3-way matching (p/c/n)
11524 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
11525 detected as combed (p/c/n + u/b)
11528 The parenthesis at the end indicate the matches that would be used for that
11529 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
11532 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
11535 Default value is @var{pc_n}.
11538 Mark the main input stream as a pre-processed input, and enable the secondary
11539 input stream as the clean source to pick the fields from. See the filter
11540 introduction for more details. It is similar to the @option{clip2} feature from
11543 Default value is @code{0} (disabled).
11546 Set the field to match from. It is recommended to set this to the same value as
11547 @option{order} unless you experience matching failures with that setting. In
11548 certain circumstances changing the field that is used to match from can have a
11549 large impact on matching performance. Available values are:
11553 Automatic (same value as @option{order}).
11555 Match from the bottom field.
11557 Match from the top field.
11560 Default value is @var{auto}.
11563 Set whether or not chroma is included during the match comparisons. In most
11564 cases it is recommended to leave this enabled. You should set this to @code{0}
11565 only if your clip has bad chroma problems such as heavy rainbowing or other
11566 artifacts. Setting this to @code{0} could also be used to speed things up at
11567 the cost of some accuracy.
11569 Default value is @code{1}.
11573 These define an exclusion band which excludes the lines between @option{y0} and
11574 @option{y1} from being included in the field matching decision. An exclusion
11575 band can be used to ignore subtitles, a logo, or other things that may
11576 interfere with the matching. @option{y0} sets the starting scan line and
11577 @option{y1} sets the ending line; all lines in between @option{y0} and
11578 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
11579 @option{y0} and @option{y1} to the same value will disable the feature.
11580 @option{y0} and @option{y1} defaults to @code{0}.
11583 Set the scene change detection threshold as a percentage of maximum change on
11584 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
11585 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
11586 @option{scthresh} is @code{[0.0, 100.0]}.
11588 Default value is @code{12.0}.
11591 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
11592 account the combed scores of matches when deciding what match to use as the
11593 final match. Available values are:
11597 No final matching based on combed scores.
11599 Combed scores are only used when a scene change is detected.
11601 Use combed scores all the time.
11604 Default is @var{sc}.
11607 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
11608 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
11609 Available values are:
11613 No forced calculation.
11615 Force p/c/n calculations.
11617 Force p/c/n/u/b calculations.
11620 Default value is @var{none}.
11623 This is the area combing threshold used for combed frame detection. This
11624 essentially controls how "strong" or "visible" combing must be to be detected.
11625 Larger values mean combing must be more visible and smaller values mean combing
11626 can be less visible or strong and still be detected. Valid settings are from
11627 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
11628 be detected as combed). This is basically a pixel difference value. A good
11629 range is @code{[8, 12]}.
11631 Default value is @code{9}.
11634 Sets whether or not chroma is considered in the combed frame decision. Only
11635 disable this if your source has chroma problems (rainbowing, etc.) that are
11636 causing problems for the combed frame detection with chroma enabled. Actually,
11637 using @option{chroma}=@var{0} is usually more reliable, except for the case
11638 where there is chroma only combing in the source.
11640 Default value is @code{0}.
11644 Respectively set the x-axis and y-axis size of the window used during combed
11645 frame detection. This has to do with the size of the area in which
11646 @option{combpel} pixels are required to be detected as combed for a frame to be
11647 declared combed. See the @option{combpel} parameter description for more info.
11648 Possible values are any number that is a power of 2 starting at 4 and going up
11651 Default value is @code{16}.
11654 The number of combed pixels inside any of the @option{blocky} by
11655 @option{blockx} size blocks on the frame for the frame to be detected as
11656 combed. While @option{cthresh} controls how "visible" the combing must be, this
11657 setting controls "how much" combing there must be in any localized area (a
11658 window defined by the @option{blockx} and @option{blocky} settings) on the
11659 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
11660 which point no frames will ever be detected as combed). This setting is known
11661 as @option{MI} in TFM/VFM vocabulary.
11663 Default value is @code{80}.
11666 @anchor{p/c/n/u/b meaning}
11667 @subsection p/c/n/u/b meaning
11669 @subsubsection p/c/n
11671 We assume the following telecined stream:
11674 Top fields: 1 2 2 3 4
11675 Bottom fields: 1 2 3 4 4
11678 The numbers correspond to the progressive frame the fields relate to. Here, the
11679 first two frames are progressive, the 3rd and 4th are combed, and so on.
11681 When @code{fieldmatch} is configured to run a matching from bottom
11682 (@option{field}=@var{bottom}) this is how this input stream get transformed:
11687 B 1 2 3 4 4 <-- matching reference
11696 As a result of the field matching, we can see that some frames get duplicated.
11697 To perform a complete inverse telecine, you need to rely on a decimation filter
11698 after this operation. See for instance the @ref{decimate} filter.
11700 The same operation now matching from top fields (@option{field}=@var{top})
11705 T 1 2 2 3 4 <-- matching reference
11715 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
11716 basically, they refer to the frame and field of the opposite parity:
11719 @item @var{p} matches the field of the opposite parity in the previous frame
11720 @item @var{c} matches the field of the opposite parity in the current frame
11721 @item @var{n} matches the field of the opposite parity in the next frame
11726 The @var{u} and @var{b} matching are a bit special in the sense that they match
11727 from the opposite parity flag. In the following examples, we assume that we are
11728 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
11729 'x' is placed above and below each matched fields.
11731 With bottom matching (@option{field}=@var{bottom}):
11736 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11737 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11745 With top matching (@option{field}=@var{top}):
11750 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11751 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11759 @subsection Examples
11761 Simple IVTC of a top field first telecined stream:
11763 fieldmatch=order=tff:combmatch=none, decimate
11766 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
11768 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
11771 @section fieldorder
11773 Transform the field order of the input video.
11775 It accepts the following parameters:
11780 The output field order. Valid values are @var{tff} for top field first or @var{bff}
11781 for bottom field first.
11784 The default value is @samp{tff}.
11786 The transformation is done by shifting the picture content up or down
11787 by one line, and filling the remaining line with appropriate picture content.
11788 This method is consistent with most broadcast field order converters.
11790 If the input video is not flagged as being interlaced, or it is already
11791 flagged as being of the required output field order, then this filter does
11792 not alter the incoming video.
11794 It is very useful when converting to or from PAL DV material,
11795 which is bottom field first.
11799 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
11802 @section fifo, afifo
11804 Buffer input images and send them when they are requested.
11806 It is mainly useful when auto-inserted by the libavfilter
11809 It does not take parameters.
11811 @section fillborders
11813 Fill borders of the input video, without changing video stream dimensions.
11814 Sometimes video can have garbage at the four edges and you may not want to
11815 crop video input to keep size multiple of some number.
11817 This filter accepts the following options:
11821 Number of pixels to fill from left border.
11824 Number of pixels to fill from right border.
11827 Number of pixels to fill from top border.
11830 Number of pixels to fill from bottom border.
11835 It accepts the following values:
11838 fill pixels using outermost pixels
11841 fill pixels using mirroring (half sample symmetric)
11844 fill pixels with constant value
11847 fill pixels using reflecting (whole sample symmetric)
11850 fill pixels using wrapping
11853 fade pixels to constant value
11856 Default is @var{smear}.
11859 Set color for pixels in fixed or fade mode. Default is @var{black}.
11862 @subsection Commands
11863 This filter supports same @ref{commands} as options.
11864 The command accepts the same syntax of the corresponding option.
11866 If the specified expression is not valid, it is kept at its current
11871 Find a rectangular object
11873 It accepts the following options:
11877 Filepath of the object image, needs to be in gray8.
11880 Detection threshold, default is 0.5.
11883 Number of mipmaps, default is 3.
11885 @item xmin, ymin, xmax, ymax
11886 Specifies the rectangle in which to search.
11889 @subsection Examples
11893 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
11895 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
11901 Flood area with values of same pixel components with another values.
11903 It accepts the following options:
11906 Set pixel x coordinate.
11909 Set pixel y coordinate.
11912 Set source #0 component value.
11915 Set source #1 component value.
11918 Set source #2 component value.
11921 Set source #3 component value.
11924 Set destination #0 component value.
11927 Set destination #1 component value.
11930 Set destination #2 component value.
11933 Set destination #3 component value.
11939 Convert the input video to one of the specified pixel formats.
11940 Libavfilter will try to pick one that is suitable as input to
11943 It accepts the following parameters:
11947 A '|'-separated list of pixel format names, such as
11948 "pix_fmts=yuv420p|monow|rgb24".
11952 @subsection Examples
11956 Convert the input video to the @var{yuv420p} format
11958 format=pix_fmts=yuv420p
11961 Convert the input video to any of the formats in the list
11963 format=pix_fmts=yuv420p|yuv444p|yuv410p
11970 Convert the video to specified constant frame rate by duplicating or dropping
11971 frames as necessary.
11973 It accepts the following parameters:
11977 The desired output frame rate. The default is @code{25}.
11980 Assume the first PTS should be the given value, in seconds. This allows for
11981 padding/trimming at the start of stream. By default, no assumption is made
11982 about the first frame's expected PTS, so no padding or trimming is done.
11983 For example, this could be set to 0 to pad the beginning with duplicates of
11984 the first frame if a video stream starts after the audio stream or to trim any
11985 frames with a negative PTS.
11988 Timestamp (PTS) rounding method.
11990 Possible values are:
11997 round towards -infinity
11999 round towards +infinity
12003 The default is @code{near}.
12006 Action performed when reading the last frame.
12008 Possible values are:
12011 Use same timestamp rounding method as used for other frames.
12013 Pass through last frame if input duration has not been reached yet.
12015 The default is @code{round}.
12019 Alternatively, the options can be specified as a flat string:
12020 @var{fps}[:@var{start_time}[:@var{round}]].
12022 See also the @ref{setpts} filter.
12024 @subsection Examples
12028 A typical usage in order to set the fps to 25:
12034 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
12036 fps=fps=film:round=near
12042 Pack two different video streams into a stereoscopic video, setting proper
12043 metadata on supported codecs. The two views should have the same size and
12044 framerate and processing will stop when the shorter video ends. Please note
12045 that you may conveniently adjust view properties with the @ref{scale} and
12048 It accepts the following parameters:
12052 The desired packing format. Supported values are:
12057 The views are next to each other (default).
12060 The views are on top of each other.
12063 The views are packed by line.
12066 The views are packed by column.
12069 The views are temporally interleaved.
12078 # Convert left and right views into a frame-sequential video
12079 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
12081 # Convert views into a side-by-side video with the same output resolution as the input
12082 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
12087 Change the frame rate by interpolating new video output frames from the source
12090 This filter is not designed to function correctly with interlaced media. If
12091 you wish to change the frame rate of interlaced media then you are required
12092 to deinterlace before this filter and re-interlace after this filter.
12094 A description of the accepted options follows.
12098 Specify the output frames per second. This option can also be specified
12099 as a value alone. The default is @code{50}.
12102 Specify the start of a range where the output frame will be created as a
12103 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12104 the default is @code{15}.
12107 Specify the end of a range where the output frame will be created as a
12108 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12109 the default is @code{240}.
12112 Specify the level at which a scene change is detected as a value between
12113 0 and 100 to indicate a new scene; a low value reflects a low
12114 probability for the current frame to introduce a new scene, while a higher
12115 value means the current frame is more likely to be one.
12116 The default is @code{8.2}.
12119 Specify flags influencing the filter process.
12121 Available value for @var{flags} is:
12124 @item scene_change_detect, scd
12125 Enable scene change detection using the value of the option @var{scene}.
12126 This flag is enabled by default.
12132 Select one frame every N-th frame.
12134 This filter accepts the following option:
12137 Select frame after every @code{step} frames.
12138 Allowed values are positive integers higher than 0. Default value is @code{1}.
12141 @section freezedetect
12143 Detect frozen video.
12145 This filter logs a message and sets frame metadata when it detects that the
12146 input video has no significant change in content during a specified duration.
12147 Video freeze detection calculates the mean average absolute difference of all
12148 the components of video frames and compares it to a noise floor.
12150 The printed times and duration are expressed in seconds. The
12151 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
12152 whose timestamp equals or exceeds the detection duration and it contains the
12153 timestamp of the first frame of the freeze. The
12154 @code{lavfi.freezedetect.freeze_duration} and
12155 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
12158 The filter accepts the following options:
12162 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
12163 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
12167 Set freeze duration until notification (default is 2 seconds).
12170 @section freezeframes
12172 Freeze video frames.
12174 This filter freezes video frames using frame from 2nd input.
12176 The filter accepts the following options:
12180 Set number of first frame from which to start freeze.
12183 Set number of last frame from which to end freeze.
12186 Set number of frame from 2nd input which will be used instead of replaced frames.
12192 Apply a frei0r effect to the input video.
12194 To enable the compilation of this filter, you need to install the frei0r
12195 header and configure FFmpeg with @code{--enable-frei0r}.
12197 It accepts the following parameters:
12202 The name of the frei0r effect to load. If the environment variable
12203 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
12204 directories specified by the colon-separated list in @env{FREI0R_PATH}.
12205 Otherwise, the standard frei0r paths are searched, in this order:
12206 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
12207 @file{/usr/lib/frei0r-1/}.
12209 @item filter_params
12210 A '|'-separated list of parameters to pass to the frei0r effect.
12214 A frei0r effect parameter can be a boolean (its value is either
12215 "y" or "n"), a double, a color (specified as
12216 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
12217 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
12218 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
12219 a position (specified as @var{X}/@var{Y}, where
12220 @var{X} and @var{Y} are floating point numbers) and/or a string.
12222 The number and types of parameters depend on the loaded effect. If an
12223 effect parameter is not specified, the default value is set.
12225 @subsection Examples
12229 Apply the distort0r effect, setting the first two double parameters:
12231 frei0r=filter_name=distort0r:filter_params=0.5|0.01
12235 Apply the colordistance effect, taking a color as the first parameter:
12237 frei0r=colordistance:0.2/0.3/0.4
12238 frei0r=colordistance:violet
12239 frei0r=colordistance:0x112233
12243 Apply the perspective effect, specifying the top left and top right image
12246 frei0r=perspective:0.2/0.2|0.8/0.2
12250 For more information, see
12251 @url{http://frei0r.dyne.org}
12253 @subsection Commands
12255 This filter supports the @option{filter_params} option as @ref{commands}.
12259 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
12261 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
12262 processing filter, one of them is performed once per block, not per pixel.
12263 This allows for much higher speed.
12265 The filter accepts the following options:
12269 Set quality. This option defines the number of levels for averaging. It accepts
12270 an integer in the range 4-5. Default value is @code{4}.
12273 Force a constant quantization parameter. It accepts an integer in range 0-63.
12274 If not set, the filter will use the QP from the video stream (if available).
12277 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
12278 more details but also more artifacts, while higher values make the image smoother
12279 but also blurrier. Default value is @code{0} − PSNR optimal.
12281 @item use_bframe_qp
12282 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
12283 option may cause flicker since the B-Frames have often larger QP. Default is
12284 @code{0} (not enabled).
12290 Apply Gaussian blur filter.
12292 The filter accepts the following options:
12296 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
12299 Set number of steps for Gaussian approximation. Default is @code{1}.
12302 Set which planes to filter. By default all planes are filtered.
12305 Set vertical sigma, if negative it will be same as @code{sigma}.
12306 Default is @code{-1}.
12309 @subsection Commands
12310 This filter supports same commands as options.
12311 The command accepts the same syntax of the corresponding option.
12313 If the specified expression is not valid, it is kept at its current
12318 Apply generic equation to each pixel.
12320 The filter accepts the following options:
12323 @item lum_expr, lum
12324 Set the luminance expression.
12326 Set the chrominance blue expression.
12328 Set the chrominance red expression.
12329 @item alpha_expr, a
12330 Set the alpha expression.
12332 Set the red expression.
12333 @item green_expr, g
12334 Set the green expression.
12336 Set the blue expression.
12339 The colorspace is selected according to the specified options. If one
12340 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
12341 options is specified, the filter will automatically select a YCbCr
12342 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
12343 @option{blue_expr} options is specified, it will select an RGB
12346 If one of the chrominance expression is not defined, it falls back on the other
12347 one. If no alpha expression is specified it will evaluate to opaque value.
12348 If none of chrominance expressions are specified, they will evaluate
12349 to the luminance expression.
12351 The expressions can use the following variables and functions:
12355 The sequential number of the filtered frame, starting from @code{0}.
12359 The coordinates of the current sample.
12363 The width and height of the image.
12367 Width and height scale depending on the currently filtered plane. It is the
12368 ratio between the corresponding luma plane number of pixels and the current
12369 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
12370 @code{0.5,0.5} for chroma planes.
12373 Time of the current frame, expressed in seconds.
12376 Return the value of the pixel at location (@var{x},@var{y}) of the current
12380 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
12384 Return the value of the pixel at location (@var{x},@var{y}) of the
12385 blue-difference chroma plane. Return 0 if there is no such plane.
12388 Return the value of the pixel at location (@var{x},@var{y}) of the
12389 red-difference chroma plane. Return 0 if there is no such plane.
12394 Return the value of the pixel at location (@var{x},@var{y}) of the
12395 red/green/blue component. Return 0 if there is no such component.
12398 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
12399 plane. Return 0 if there is no such plane.
12401 @item psum(x,y), lumsum(x, y), cbsum(x,y), crsum(x,y), rsum(x,y), gsum(x,y), bsum(x,y), alphasum(x,y)
12402 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
12403 sums of samples within a rectangle. See the functions without the sum postfix.
12405 @item interpolation
12406 Set one of interpolation methods:
12411 Default is bilinear.
12414 For functions, if @var{x} and @var{y} are outside the area, the value will be
12415 automatically clipped to the closer edge.
12417 Please note that this filter can use multiple threads in which case each slice
12418 will have its own expression state. If you want to use only a single expression
12419 state because your expressions depend on previous state then you should limit
12420 the number of filter threads to 1.
12422 @subsection Examples
12426 Flip the image horizontally:
12432 Generate a bidimensional sine wave, with angle @code{PI/3} and a
12433 wavelength of 100 pixels:
12435 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
12439 Generate a fancy enigmatic moving light:
12441 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
12445 Generate a quick emboss effect:
12447 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
12451 Modify RGB components depending on pixel position:
12453 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
12457 Create a radial gradient that is the same size as the input (also see
12458 the @ref{vignette} filter):
12460 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
12466 Fix the banding artifacts that are sometimes introduced into nearly flat
12467 regions by truncation to 8-bit color depth.
12468 Interpolate the gradients that should go where the bands are, and
12471 It is designed for playback only. Do not use it prior to
12472 lossy compression, because compression tends to lose the dither and
12473 bring back the bands.
12475 It accepts the following parameters:
12480 The maximum amount by which the filter will change any one pixel. This is also
12481 the threshold for detecting nearly flat regions. Acceptable values range from
12482 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
12486 The neighborhood to fit the gradient to. A larger radius makes for smoother
12487 gradients, but also prevents the filter from modifying the pixels near detailed
12488 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
12489 values will be clipped to the valid range.
12493 Alternatively, the options can be specified as a flat string:
12494 @var{strength}[:@var{radius}]
12496 @subsection Examples
12500 Apply the filter with a @code{3.5} strength and radius of @code{8}:
12506 Specify radius, omitting the strength (which will fall-back to the default
12514 @anchor{graphmonitor}
12515 @section graphmonitor
12516 Show various filtergraph stats.
12518 With this filter one can debug complete filtergraph.
12519 Especially issues with links filling with queued frames.
12521 The filter accepts the following options:
12525 Set video output size. Default is @var{hd720}.
12528 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
12531 Set output mode, can be @var{fulll} or @var{compact}.
12532 In @var{compact} mode only filters with some queued frames have displayed stats.
12535 Set flags which enable which stats are shown in video.
12537 Available values for flags are:
12540 Display number of queued frames in each link.
12542 @item frame_count_in
12543 Display number of frames taken from filter.
12545 @item frame_count_out
12546 Display number of frames given out from filter.
12549 Display current filtered frame pts.
12552 Display current filtered frame time.
12555 Display time base for filter link.
12558 Display used format for filter link.
12561 Display video size or number of audio channels in case of audio used by filter link.
12564 Display video frame rate or sample rate in case of audio used by filter link.
12567 Display link output status.
12571 Set upper limit for video rate of output stream, Default value is @var{25}.
12572 This guarantee that output video frame rate will not be higher than this value.
12576 A color constancy variation filter which estimates scene illumination via grey edge algorithm
12577 and corrects the scene colors accordingly.
12579 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
12581 The filter accepts the following options:
12585 The order of differentiation to be applied on the scene. Must be chosen in the range
12586 [0,2] and default value is 1.
12589 The Minkowski parameter to be used for calculating the Minkowski distance. Must
12590 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
12591 max value instead of calculating Minkowski distance.
12594 The standard deviation of Gaussian blur to be applied on the scene. Must be
12595 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
12596 can't be equal to 0 if @var{difford} is greater than 0.
12599 @subsection Examples
12605 greyedge=difford=1:minknorm=5:sigma=2
12611 greyedge=difford=1:minknorm=0:sigma=2
12619 Apply a Hald CLUT to a video stream.
12621 First input is the video stream to process, and second one is the Hald CLUT.
12622 The Hald CLUT input can be a simple picture or a complete video stream.
12624 The filter accepts the following options:
12628 Force termination when the shortest input terminates. Default is @code{0}.
12630 Continue applying the last CLUT after the end of the stream. A value of
12631 @code{0} disable the filter after the last frame of the CLUT is reached.
12632 Default is @code{1}.
12635 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
12636 filters share the same internals).
12638 This filter also supports the @ref{framesync} options.
12640 More information about the Hald CLUT can be found on Eskil Steenberg's website
12641 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
12643 @subsection Workflow examples
12645 @subsubsection Hald CLUT video stream
12647 Generate an identity Hald CLUT stream altered with various effects:
12649 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
12652 Note: make sure you use a lossless codec.
12654 Then use it with @code{haldclut} to apply it on some random stream:
12656 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
12659 The Hald CLUT will be applied to the 10 first seconds (duration of
12660 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
12661 to the remaining frames of the @code{mandelbrot} stream.
12663 @subsubsection Hald CLUT with preview
12665 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
12666 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
12667 biggest possible square starting at the top left of the picture. The remaining
12668 padding pixels (bottom or right) will be ignored. This area can be used to add
12669 a preview of the Hald CLUT.
12671 Typically, the following generated Hald CLUT will be supported by the
12672 @code{haldclut} filter:
12675 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
12676 pad=iw+320 [padded_clut];
12677 smptebars=s=320x256, split [a][b];
12678 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
12679 [main][b] overlay=W-320" -frames:v 1 clut.png
12682 It contains the original and a preview of the effect of the CLUT: SMPTE color
12683 bars are displayed on the right-top, and below the same color bars processed by
12686 Then, the effect of this Hald CLUT can be visualized with:
12688 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
12693 Flip the input video horizontally.
12695 For example, to horizontally flip the input video with @command{ffmpeg}:
12697 ffmpeg -i in.avi -vf "hflip" out.avi
12701 This filter applies a global color histogram equalization on a
12704 It can be used to correct video that has a compressed range of pixel
12705 intensities. The filter redistributes the pixel intensities to
12706 equalize their distribution across the intensity range. It may be
12707 viewed as an "automatically adjusting contrast filter". This filter is
12708 useful only for correcting degraded or poorly captured source
12711 The filter accepts the following options:
12715 Determine the amount of equalization to be applied. As the strength
12716 is reduced, the distribution of pixel intensities more-and-more
12717 approaches that of the input frame. The value must be a float number
12718 in the range [0,1] and defaults to 0.200.
12721 Set the maximum intensity that can generated and scale the output
12722 values appropriately. The strength should be set as desired and then
12723 the intensity can be limited if needed to avoid washing-out. The value
12724 must be a float number in the range [0,1] and defaults to 0.210.
12727 Set the antibanding level. If enabled the filter will randomly vary
12728 the luminance of output pixels by a small amount to avoid banding of
12729 the histogram. Possible values are @code{none}, @code{weak} or
12730 @code{strong}. It defaults to @code{none}.
12736 Compute and draw a color distribution histogram for the input video.
12738 The computed histogram is a representation of the color component
12739 distribution in an image.
12741 Standard histogram displays the color components distribution in an image.
12742 Displays color graph for each color component. Shows distribution of
12743 the Y, U, V, A or R, G, B components, depending on input format, in the
12744 current frame. Below each graph a color component scale meter is shown.
12746 The filter accepts the following options:
12750 Set height of level. Default value is @code{200}.
12751 Allowed range is [50, 2048].
12754 Set height of color scale. Default value is @code{12}.
12755 Allowed range is [0, 40].
12759 It accepts the following values:
12762 Per color component graphs are placed below each other.
12765 Per color component graphs are placed side by side.
12768 Presents information identical to that in the @code{parade}, except
12769 that the graphs representing color components are superimposed directly
12772 Default is @code{stack}.
12775 Set mode. Can be either @code{linear}, or @code{logarithmic}.
12776 Default is @code{linear}.
12779 Set what color components to display.
12780 Default is @code{7}.
12783 Set foreground opacity. Default is @code{0.7}.
12786 Set background opacity. Default is @code{0.5}.
12789 @subsection Examples
12794 Calculate and draw histogram:
12796 ffplay -i input -vf histogram
12804 This is a high precision/quality 3d denoise filter. It aims to reduce
12805 image noise, producing smooth images and making still images really
12806 still. It should enhance compressibility.
12808 It accepts the following optional parameters:
12812 A non-negative floating point number which specifies spatial luma strength.
12813 It defaults to 4.0.
12815 @item chroma_spatial
12816 A non-negative floating point number which specifies spatial chroma strength.
12817 It defaults to 3.0*@var{luma_spatial}/4.0.
12820 A floating point number which specifies luma temporal strength. It defaults to
12821 6.0*@var{luma_spatial}/4.0.
12824 A floating point number which specifies chroma temporal strength. It defaults to
12825 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
12828 @subsection Commands
12829 This filter supports same @ref{commands} as options.
12830 The command accepts the same syntax of the corresponding option.
12832 If the specified expression is not valid, it is kept at its current
12835 @anchor{hwdownload}
12836 @section hwdownload
12838 Download hardware frames to system memory.
12840 The input must be in hardware frames, and the output a non-hardware format.
12841 Not all formats will be supported on the output - it may be necessary to insert
12842 an additional @option{format} filter immediately following in the graph to get
12843 the output in a supported format.
12847 Map hardware frames to system memory or to another device.
12849 This filter has several different modes of operation; which one is used depends
12850 on the input and output formats:
12853 Hardware frame input, normal frame output
12855 Map the input frames to system memory and pass them to the output. If the
12856 original hardware frame is later required (for example, after overlaying
12857 something else on part of it), the @option{hwmap} filter can be used again
12858 in the next mode to retrieve it.
12860 Normal frame input, hardware frame output
12862 If the input is actually a software-mapped hardware frame, then unmap it -
12863 that is, return the original hardware frame.
12865 Otherwise, a device must be provided. Create new hardware surfaces on that
12866 device for the output, then map them back to the software format at the input
12867 and give those frames to the preceding filter. This will then act like the
12868 @option{hwupload} filter, but may be able to avoid an additional copy when
12869 the input is already in a compatible format.
12871 Hardware frame input and output
12873 A device must be supplied for the output, either directly or with the
12874 @option{derive_device} option. The input and output devices must be of
12875 different types and compatible - the exact meaning of this is
12876 system-dependent, but typically it means that they must refer to the same
12877 underlying hardware context (for example, refer to the same graphics card).
12879 If the input frames were originally created on the output device, then unmap
12880 to retrieve the original frames.
12882 Otherwise, map the frames to the output device - create new hardware frames
12883 on the output corresponding to the frames on the input.
12886 The following additional parameters are accepted:
12890 Set the frame mapping mode. Some combination of:
12893 The mapped frame should be readable.
12895 The mapped frame should be writeable.
12897 The mapping will always overwrite the entire frame.
12899 This may improve performance in some cases, as the original contents of the
12900 frame need not be loaded.
12902 The mapping must not involve any copying.
12904 Indirect mappings to copies of frames are created in some cases where either
12905 direct mapping is not possible or it would have unexpected properties.
12906 Setting this flag ensures that the mapping is direct and will fail if that is
12909 Defaults to @var{read+write} if not specified.
12911 @item derive_device @var{type}
12912 Rather than using the device supplied at initialisation, instead derive a new
12913 device of type @var{type} from the device the input frames exist on.
12916 In a hardware to hardware mapping, map in reverse - create frames in the sink
12917 and map them back to the source. This may be necessary in some cases where
12918 a mapping in one direction is required but only the opposite direction is
12919 supported by the devices being used.
12921 This option is dangerous - it may break the preceding filter in undefined
12922 ways if there are any additional constraints on that filter's output.
12923 Do not use it without fully understanding the implications of its use.
12929 Upload system memory frames to hardware surfaces.
12931 The device to upload to must be supplied when the filter is initialised. If
12932 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
12933 option or with the @option{derive_device} option. The input and output devices
12934 must be of different types and compatible - the exact meaning of this is
12935 system-dependent, but typically it means that they must refer to the same
12936 underlying hardware context (for example, refer to the same graphics card).
12938 The following additional parameters are accepted:
12941 @item derive_device @var{type}
12942 Rather than using the device supplied at initialisation, instead derive a new
12943 device of type @var{type} from the device the input frames exist on.
12946 @anchor{hwupload_cuda}
12947 @section hwupload_cuda
12949 Upload system memory frames to a CUDA device.
12951 It accepts the following optional parameters:
12955 The number of the CUDA device to use
12960 Apply a high-quality magnification filter designed for pixel art. This filter
12961 was originally created by Maxim Stepin.
12963 It accepts the following option:
12967 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
12968 @code{hq3x} and @code{4} for @code{hq4x}.
12969 Default is @code{3}.
12973 Stack input videos horizontally.
12975 All streams must be of same pixel format and of same height.
12977 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
12978 to create same output.
12980 The filter accepts the following option:
12984 Set number of input streams. Default is 2.
12987 If set to 1, force the output to terminate when the shortest input
12988 terminates. Default value is 0.
12993 Modify the hue and/or the saturation of the input.
12995 It accepts the following parameters:
12999 Specify the hue angle as a number of degrees. It accepts an expression,
13000 and defaults to "0".
13003 Specify the saturation in the [-10,10] range. It accepts an expression and
13007 Specify the hue angle as a number of radians. It accepts an
13008 expression, and defaults to "0".
13011 Specify the brightness in the [-10,10] range. It accepts an expression and
13015 @option{h} and @option{H} are mutually exclusive, and can't be
13016 specified at the same time.
13018 The @option{b}, @option{h}, @option{H} and @option{s} option values are
13019 expressions containing the following constants:
13023 frame count of the input frame starting from 0
13026 presentation timestamp of the input frame expressed in time base units
13029 frame rate of the input video, NAN if the input frame rate is unknown
13032 timestamp expressed in seconds, NAN if the input timestamp is unknown
13035 time base of the input video
13038 @subsection Examples
13042 Set the hue to 90 degrees and the saturation to 1.0:
13048 Same command but expressing the hue in radians:
13054 Rotate hue and make the saturation swing between 0
13055 and 2 over a period of 1 second:
13057 hue="H=2*PI*t: s=sin(2*PI*t)+1"
13061 Apply a 3 seconds saturation fade-in effect starting at 0:
13063 hue="s=min(t/3\,1)"
13066 The general fade-in expression can be written as:
13068 hue="s=min(0\, max((t-START)/DURATION\, 1))"
13072 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
13074 hue="s=max(0\, min(1\, (8-t)/3))"
13077 The general fade-out expression can be written as:
13079 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
13084 @subsection Commands
13086 This filter supports the following commands:
13092 Modify the hue and/or the saturation and/or brightness of the input video.
13093 The command accepts the same syntax of the corresponding option.
13095 If the specified expression is not valid, it is kept at its current
13099 @section hysteresis
13101 Grow first stream into second stream by connecting components.
13102 This makes it possible to build more robust edge masks.
13104 This filter accepts the following options:
13108 Set which planes will be processed as bitmap, unprocessed planes will be
13109 copied from first stream.
13110 By default value 0xf, all planes will be processed.
13113 Set threshold which is used in filtering. If pixel component value is higher than
13114 this value filter algorithm for connecting components is activated.
13115 By default value is 0.
13118 The @code{hysteresis} filter also supports the @ref{framesync} options.
13122 Detect video interlacing type.
13124 This filter tries to detect if the input frames are interlaced, progressive,
13125 top or bottom field first. It will also try to detect fields that are
13126 repeated between adjacent frames (a sign of telecine).
13128 Single frame detection considers only immediately adjacent frames when classifying each frame.
13129 Multiple frame detection incorporates the classification history of previous frames.
13131 The filter will log these metadata values:
13134 @item single.current_frame
13135 Detected type of current frame using single-frame detection. One of:
13136 ``tff'' (top field first), ``bff'' (bottom field first),
13137 ``progressive'', or ``undetermined''
13140 Cumulative number of frames detected as top field first using single-frame detection.
13143 Cumulative number of frames detected as top field first using multiple-frame detection.
13146 Cumulative number of frames detected as bottom field first using single-frame detection.
13148 @item multiple.current_frame
13149 Detected type of current frame using multiple-frame detection. One of:
13150 ``tff'' (top field first), ``bff'' (bottom field first),
13151 ``progressive'', or ``undetermined''
13154 Cumulative number of frames detected as bottom field first using multiple-frame detection.
13156 @item single.progressive
13157 Cumulative number of frames detected as progressive using single-frame detection.
13159 @item multiple.progressive
13160 Cumulative number of frames detected as progressive using multiple-frame detection.
13162 @item single.undetermined
13163 Cumulative number of frames that could not be classified using single-frame detection.
13165 @item multiple.undetermined
13166 Cumulative number of frames that could not be classified using multiple-frame detection.
13168 @item repeated.current_frame
13169 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
13171 @item repeated.neither
13172 Cumulative number of frames with no repeated field.
13175 Cumulative number of frames with the top field repeated from the previous frame's top field.
13177 @item repeated.bottom
13178 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
13181 The filter accepts the following options:
13185 Set interlacing threshold.
13187 Set progressive threshold.
13189 Threshold for repeated field detection.
13191 Number of frames after which a given frame's contribution to the
13192 statistics is halved (i.e., it contributes only 0.5 to its
13193 classification). The default of 0 means that all frames seen are given
13194 full weight of 1.0 forever.
13195 @item analyze_interlaced_flag
13196 When this is not 0 then idet will use the specified number of frames to determine
13197 if the interlaced flag is accurate, it will not count undetermined frames.
13198 If the flag is found to be accurate it will be used without any further
13199 computations, if it is found to be inaccurate it will be cleared without any
13200 further computations. This allows inserting the idet filter as a low computational
13201 method to clean up the interlaced flag
13206 Deinterleave or interleave fields.
13208 This filter allows one to process interlaced images fields without
13209 deinterlacing them. Deinterleaving splits the input frame into 2
13210 fields (so called half pictures). Odd lines are moved to the top
13211 half of the output image, even lines to the bottom half.
13212 You can process (filter) them independently and then re-interleave them.
13214 The filter accepts the following options:
13218 @item chroma_mode, c
13219 @item alpha_mode, a
13220 Available values for @var{luma_mode}, @var{chroma_mode} and
13221 @var{alpha_mode} are:
13227 @item deinterleave, d
13228 Deinterleave fields, placing one above the other.
13230 @item interleave, i
13231 Interleave fields. Reverse the effect of deinterleaving.
13233 Default value is @code{none}.
13235 @item luma_swap, ls
13236 @item chroma_swap, cs
13237 @item alpha_swap, as
13238 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
13241 @subsection Commands
13243 This filter supports the all above options as @ref{commands}.
13247 Apply inflate effect to the video.
13249 This filter replaces the pixel by the local(3x3) average by taking into account
13250 only values higher than the pixel.
13252 It accepts the following options:
13259 Limit the maximum change for each plane, default is 65535.
13260 If 0, plane will remain unchanged.
13263 @subsection Commands
13265 This filter supports the all above options as @ref{commands}.
13269 Simple interlacing filter from progressive contents. This interleaves upper (or
13270 lower) lines from odd frames with lower (or upper) lines from even frames,
13271 halving the frame rate and preserving image height.
13274 Original Original New Frame
13275 Frame 'j' Frame 'j+1' (tff)
13276 ========== =========== ==================
13277 Line 0 --------------------> Frame 'j' Line 0
13278 Line 1 Line 1 ----> Frame 'j+1' Line 1
13279 Line 2 ---------------------> Frame 'j' Line 2
13280 Line 3 Line 3 ----> Frame 'j+1' Line 3
13282 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
13285 It accepts the following optional parameters:
13289 This determines whether the interlaced frame is taken from the even
13290 (tff - default) or odd (bff) lines of the progressive frame.
13293 Vertical lowpass filter to avoid twitter interlacing and
13294 reduce moire patterns.
13298 Disable vertical lowpass filter
13301 Enable linear filter (default)
13304 Enable complex filter. This will slightly less reduce twitter and moire
13305 but better retain detail and subjective sharpness impression.
13312 Deinterlace input video by applying Donald Graft's adaptive kernel
13313 deinterling. Work on interlaced parts of a video to produce
13314 progressive frames.
13316 The description of the accepted parameters follows.
13320 Set the threshold which affects the filter's tolerance when
13321 determining if a pixel line must be processed. It must be an integer
13322 in the range [0,255] and defaults to 10. A value of 0 will result in
13323 applying the process on every pixels.
13326 Paint pixels exceeding the threshold value to white if set to 1.
13330 Set the fields order. Swap fields if set to 1, leave fields alone if
13334 Enable additional sharpening if set to 1. Default is 0.
13337 Enable twoway sharpening if set to 1. Default is 0.
13340 @subsection Examples
13344 Apply default values:
13346 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
13350 Enable additional sharpening:
13356 Paint processed pixels in white:
13364 Slowly update darker pixels.
13366 This filter makes short flashes of light appear longer.
13367 This filter accepts the following options:
13371 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
13374 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
13377 @section lenscorrection
13379 Correct radial lens distortion
13381 This filter can be used to correct for radial distortion as can result from the use
13382 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
13383 one can use tools available for example as part of opencv or simply trial-and-error.
13384 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
13385 and extract the k1 and k2 coefficients from the resulting matrix.
13387 Note that effectively the same filter is available in the open-source tools Krita and
13388 Digikam from the KDE project.
13390 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
13391 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
13392 brightness distribution, so you may want to use both filters together in certain
13393 cases, though you will have to take care of ordering, i.e. whether vignetting should
13394 be applied before or after lens correction.
13396 @subsection Options
13398 The filter accepts the following options:
13402 Relative x-coordinate of the focal point of the image, and thereby the center of the
13403 distortion. This value has a range [0,1] and is expressed as fractions of the image
13404 width. Default is 0.5.
13406 Relative y-coordinate of the focal point of the image, and thereby the center of the
13407 distortion. This value has a range [0,1] and is expressed as fractions of the image
13408 height. Default is 0.5.
13410 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
13411 no correction. Default is 0.
13413 Coefficient of the double quadratic correction term. This value has a range [-1,1].
13414 0 means no correction. Default is 0.
13417 The formula that generates the correction is:
13419 @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)
13421 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
13422 distances from the focal point in the source and target images, respectively.
13426 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
13428 The @code{lensfun} filter requires the camera make, camera model, and lens model
13429 to apply the lens correction. The filter will load the lensfun database and
13430 query it to find the corresponding camera and lens entries in the database. As
13431 long as these entries can be found with the given options, the filter can
13432 perform corrections on frames. Note that incomplete strings will result in the
13433 filter choosing the best match with the given options, and the filter will
13434 output the chosen camera and lens models (logged with level "info"). You must
13435 provide the make, camera model, and lens model as they are required.
13437 The filter accepts the following options:
13441 The make of the camera (for example, "Canon"). This option is required.
13444 The model of the camera (for example, "Canon EOS 100D"). This option is
13448 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
13449 option is required.
13452 The type of correction to apply. The following values are valid options:
13456 Enables fixing lens vignetting.
13459 Enables fixing lens geometry. This is the default.
13462 Enables fixing chromatic aberrations.
13465 Enables fixing lens vignetting and lens geometry.
13468 Enables fixing lens vignetting and chromatic aberrations.
13471 Enables fixing both lens geometry and chromatic aberrations.
13474 Enables all possible corrections.
13478 The focal length of the image/video (zoom; expected constant for video). For
13479 example, a 18--55mm lens has focal length range of [18--55], so a value in that
13480 range should be chosen when using that lens. Default 18.
13483 The aperture of the image/video (expected constant for video). Note that
13484 aperture is only used for vignetting correction. Default 3.5.
13486 @item focus_distance
13487 The focus distance of the image/video (expected constant for video). Note that
13488 focus distance is only used for vignetting and only slightly affects the
13489 vignetting correction process. If unknown, leave it at the default value (which
13493 The scale factor which is applied after transformation. After correction the
13494 video is no longer necessarily rectangular. This parameter controls how much of
13495 the resulting image is visible. The value 0 means that a value will be chosen
13496 automatically such that there is little or no unmapped area in the output
13497 image. 1.0 means that no additional scaling is done. Lower values may result
13498 in more of the corrected image being visible, while higher values may avoid
13499 unmapped areas in the output.
13501 @item target_geometry
13502 The target geometry of the output image/video. The following values are valid
13506 @item rectilinear (default)
13509 @item equirectangular
13510 @item fisheye_orthographic
13511 @item fisheye_stereographic
13512 @item fisheye_equisolid
13513 @item fisheye_thoby
13516 Apply the reverse of image correction (instead of correcting distortion, apply
13519 @item interpolation
13520 The type of interpolation used when correcting distortion. The following values
13525 @item linear (default)
13530 @subsection Examples
13534 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
13535 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
13539 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
13543 Apply the same as before, but only for the first 5 seconds of video.
13546 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
13553 Obtain the VMAF (Video Multi-Method Assessment Fusion)
13554 score between two input videos.
13556 The obtained VMAF score is printed through the logging system.
13558 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
13559 After installing the library it can be enabled using:
13560 @code{./configure --enable-libvmaf}.
13561 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
13563 The filter has following options:
13567 Set the model path which is to be used for SVM.
13568 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
13571 Set the file path to be used to store logs.
13574 Set the format of the log file (csv, json or xml).
13576 @item enable_transform
13577 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
13578 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
13579 Default value: @code{false}
13582 Invokes the phone model which will generate VMAF scores higher than in the
13583 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
13584 Default value: @code{false}
13587 Enables computing psnr along with vmaf.
13588 Default value: @code{false}
13591 Enables computing ssim along with vmaf.
13592 Default value: @code{false}
13595 Enables computing ms_ssim along with vmaf.
13596 Default value: @code{false}
13599 Set the pool method to be used for computing vmaf.
13600 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
13603 Set number of threads to be used when computing vmaf.
13604 Default value: @code{0}, which makes use of all available logical processors.
13607 Set interval for frame subsampling used when computing vmaf.
13608 Default value: @code{1}
13610 @item enable_conf_interval
13611 Enables confidence interval.
13612 Default value: @code{false}
13615 This filter also supports the @ref{framesync} options.
13617 @subsection Examples
13620 On the below examples the input file @file{main.mpg} being processed is
13621 compared with the reference file @file{ref.mpg}.
13624 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
13628 Example with options:
13630 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
13634 Example with options and different containers:
13636 ffmpeg -i main.mpg -i ref.mkv -lavfi "[0:v]settb=AVTB,setpts=PTS-STARTPTS[main];[1:v]settb=AVTB,setpts=PTS-STARTPTS[ref];[main][ref]libvmaf=psnr=1:log_fmt=json" -f null -
13642 Limits the pixel components values to the specified range [min, max].
13644 The filter accepts the following options:
13648 Lower bound. Defaults to the lowest allowed value for the input.
13651 Upper bound. Defaults to the highest allowed value for the input.
13654 Specify which planes will be processed. Defaults to all available.
13657 @subsection Commands
13659 This filter supports the all above options as @ref{commands}.
13665 The filter accepts the following options:
13669 Set the number of loops. Setting this value to -1 will result in infinite loops.
13673 Set maximal size in number of frames. Default is 0.
13676 Set first frame of loop. Default is 0.
13679 @subsection Examples
13683 Loop single first frame infinitely:
13685 loop=loop=-1:size=1:start=0
13689 Loop single first frame 10 times:
13691 loop=loop=10:size=1:start=0
13695 Loop 10 first frames 5 times:
13697 loop=loop=5:size=10:start=0
13703 Apply a 1D LUT to an input video.
13705 The filter accepts the following options:
13709 Set the 1D LUT file name.
13711 Currently supported formats:
13720 Select interpolation mode.
13722 Available values are:
13726 Use values from the nearest defined point.
13728 Interpolate values using the linear interpolation.
13730 Interpolate values using the cosine interpolation.
13732 Interpolate values using the cubic interpolation.
13734 Interpolate values using the spline interpolation.
13741 Apply a 3D LUT to an input video.
13743 The filter accepts the following options:
13747 Set the 3D LUT file name.
13749 Currently supported formats:
13763 Select interpolation mode.
13765 Available values are:
13769 Use values from the nearest defined point.
13771 Interpolate values using the 8 points defining a cube.
13773 Interpolate values using a tetrahedron.
13779 Turn certain luma values into transparency.
13781 The filter accepts the following options:
13785 Set the luma which will be used as base for transparency.
13786 Default value is @code{0}.
13789 Set the range of luma values to be keyed out.
13790 Default value is @code{0.01}.
13793 Set the range of softness. Default value is @code{0}.
13794 Use this to control gradual transition from zero to full transparency.
13797 @subsection Commands
13798 This filter supports same @ref{commands} as options.
13799 The command accepts the same syntax of the corresponding option.
13801 If the specified expression is not valid, it is kept at its current
13804 @section lut, lutrgb, lutyuv
13806 Compute a look-up table for binding each pixel component input value
13807 to an output value, and apply it to the input video.
13809 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
13810 to an RGB input video.
13812 These filters accept the following parameters:
13815 set first pixel component expression
13817 set second pixel component expression
13819 set third pixel component expression
13821 set fourth pixel component expression, corresponds to the alpha component
13824 set red component expression
13826 set green component expression
13828 set blue component expression
13830 alpha component expression
13833 set Y/luminance component expression
13835 set U/Cb component expression
13837 set V/Cr component expression
13840 Each of them specifies the expression to use for computing the lookup table for
13841 the corresponding pixel component values.
13843 The exact component associated to each of the @var{c*} options depends on the
13846 The @var{lut} filter requires either YUV or RGB pixel formats in input,
13847 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
13849 The expressions can contain the following constants and functions:
13854 The input width and height.
13857 The input value for the pixel component.
13860 The input value, clipped to the @var{minval}-@var{maxval} range.
13863 The maximum value for the pixel component.
13866 The minimum value for the pixel component.
13869 The negated value for the pixel component value, clipped to the
13870 @var{minval}-@var{maxval} range; it corresponds to the expression
13871 "maxval-clipval+minval".
13874 The computed value in @var{val}, clipped to the
13875 @var{minval}-@var{maxval} range.
13877 @item gammaval(gamma)
13878 The computed gamma correction value of the pixel component value,
13879 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
13881 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
13885 All expressions default to "val".
13887 @subsection Examples
13891 Negate input video:
13893 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
13894 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
13897 The above is the same as:
13899 lutrgb="r=negval:g=negval:b=negval"
13900 lutyuv="y=negval:u=negval:v=negval"
13910 Remove chroma components, turning the video into a graytone image:
13912 lutyuv="u=128:v=128"
13916 Apply a luma burning effect:
13922 Remove green and blue components:
13928 Set a constant alpha channel value on input:
13930 format=rgba,lutrgb=a="maxval-minval/2"
13934 Correct luminance gamma by a factor of 0.5:
13936 lutyuv=y=gammaval(0.5)
13940 Discard least significant bits of luma:
13942 lutyuv=y='bitand(val, 128+64+32)'
13946 Technicolor like effect:
13948 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
13952 @section lut2, tlut2
13954 The @code{lut2} filter takes two input streams and outputs one
13957 The @code{tlut2} (time lut2) filter takes two consecutive frames
13958 from one single stream.
13960 This filter accepts the following parameters:
13963 set first pixel component expression
13965 set second pixel component expression
13967 set third pixel component expression
13969 set fourth pixel component expression, corresponds to the alpha component
13972 set output bit depth, only available for @code{lut2} filter. By default is 0,
13973 which means bit depth is automatically picked from first input format.
13976 The @code{lut2} filter also supports the @ref{framesync} options.
13978 Each of them specifies the expression to use for computing the lookup table for
13979 the corresponding pixel component values.
13981 The exact component associated to each of the @var{c*} options depends on the
13984 The expressions can contain the following constants:
13989 The input width and height.
13992 The first input value for the pixel component.
13995 The second input value for the pixel component.
13998 The first input video bit depth.
14001 The second input video bit depth.
14004 All expressions default to "x".
14006 @subsection Examples
14010 Highlight differences between two RGB video streams:
14012 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)'
14016 Highlight differences between two YUV video streams:
14018 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)'
14022 Show max difference between two video streams:
14024 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)))'
14028 @section maskedclamp
14030 Clamp the first input stream with the second input and third input stream.
14032 Returns the value of first stream to be between second input
14033 stream - @code{undershoot} and third input stream + @code{overshoot}.
14035 This filter accepts the following options:
14038 Default value is @code{0}.
14041 Default value is @code{0}.
14044 Set which planes will be processed as bitmap, unprocessed planes will be
14045 copied from first stream.
14046 By default value 0xf, all planes will be processed.
14049 @subsection Commands
14051 This filter supports the all above options as @ref{commands}.
14055 Merge the second and third input stream into output stream using absolute differences
14056 between second input stream and first input stream and absolute difference between
14057 third input stream and first input stream. The picked value will be from second input
14058 stream if second absolute difference is greater than first one or from third input stream
14061 This filter accepts the following options:
14064 Set which planes will be processed as bitmap, unprocessed planes will be
14065 copied from first stream.
14066 By default value 0xf, all planes will be processed.
14069 @subsection Commands
14071 This filter supports the all above options as @ref{commands}.
14073 @section maskedmerge
14075 Merge the first input stream with the second input stream using per pixel
14076 weights in the third input stream.
14078 A value of 0 in the third stream pixel component means that pixel component
14079 from first stream is returned unchanged, while maximum value (eg. 255 for
14080 8-bit videos) means that pixel component from second stream is returned
14081 unchanged. Intermediate values define the amount of merging between both
14082 input stream's pixel components.
14084 This filter accepts the following options:
14087 Set which planes will be processed as bitmap, unprocessed planes will be
14088 copied from first stream.
14089 By default value 0xf, all planes will be processed.
14094 Merge the second and third input stream into output stream using absolute differences
14095 between second input stream and first input stream and absolute difference between
14096 third input stream and first input stream. The picked value will be from second input
14097 stream if second absolute difference is less than first one or from third input stream
14100 This filter accepts the following options:
14103 Set which planes will be processed as bitmap, unprocessed planes will be
14104 copied from first stream.
14105 By default value 0xf, all planes will be processed.
14108 @subsection Commands
14110 This filter supports the all above options as @ref{commands}.
14112 @section maskedthreshold
14113 Pick pixels comparing absolute difference of two video streams with fixed
14116 If absolute difference between pixel component of first and second video
14117 stream is equal or lower than user supplied threshold than pixel component
14118 from first video stream is picked, otherwise pixel component from second
14119 video stream is picked.
14121 This filter accepts the following options:
14124 Set threshold used when picking pixels from absolute difference from two input
14128 Set which planes will be processed as bitmap, unprocessed planes will be
14129 copied from second stream.
14130 By default value 0xf, all planes will be processed.
14133 @subsection Commands
14135 This filter supports the all above options as @ref{commands}.
14138 Create mask from input video.
14140 For example it is useful to create motion masks after @code{tblend} filter.
14142 This filter accepts the following options:
14146 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
14149 Set high threshold. Any pixel component higher than this value will be set to max value
14150 allowed for current pixel format.
14153 Set planes to filter, by default all available planes are filtered.
14156 Fill all frame pixels with this value.
14159 Set max average pixel value for frame. If sum of all pixel components is higher that this
14160 average, output frame will be completely filled with value set by @var{fill} option.
14161 Typically useful for scene changes when used in combination with @code{tblend} filter.
14166 Apply motion-compensation deinterlacing.
14168 It needs one field per frame as input and must thus be used together
14169 with yadif=1/3 or equivalent.
14171 This filter accepts the following options:
14174 Set the deinterlacing mode.
14176 It accepts one of the following values:
14181 use iterative motion estimation
14183 like @samp{slow}, but use multiple reference frames.
14185 Default value is @samp{fast}.
14188 Set the picture field parity assumed for the input video. It must be
14189 one of the following values:
14193 assume top field first
14195 assume bottom field first
14198 Default value is @samp{bff}.
14201 Set per-block quantization parameter (QP) used by the internal
14204 Higher values should result in a smoother motion vector field but less
14205 optimal individual vectors. Default value is 1.
14210 Pick median pixel from certain rectangle defined by radius.
14212 This filter accepts the following options:
14216 Set horizontal radius size. Default value is @code{1}.
14217 Allowed range is integer from 1 to 127.
14220 Set which planes to process. Default is @code{15}, which is all available planes.
14223 Set vertical radius size. Default value is @code{0}.
14224 Allowed range is integer from 0 to 127.
14225 If it is 0, value will be picked from horizontal @code{radius} option.
14228 Set median percentile. Default value is @code{0.5}.
14229 Default value of @code{0.5} will pick always median values, while @code{0} will pick
14230 minimum values, and @code{1} maximum values.
14233 @subsection Commands
14234 This filter supports same @ref{commands} as options.
14235 The command accepts the same syntax of the corresponding option.
14237 If the specified expression is not valid, it is kept at its current
14240 @section mergeplanes
14242 Merge color channel components from several video streams.
14244 The filter accepts up to 4 input streams, and merge selected input
14245 planes to the output video.
14247 This filter accepts the following options:
14250 Set input to output plane mapping. Default is @code{0}.
14252 The mappings is specified as a bitmap. It should be specified as a
14253 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
14254 mapping for the first plane of the output stream. 'A' sets the number of
14255 the input stream to use (from 0 to 3), and 'a' the plane number of the
14256 corresponding input to use (from 0 to 3). The rest of the mappings is
14257 similar, 'Bb' describes the mapping for the output stream second
14258 plane, 'Cc' describes the mapping for the output stream third plane and
14259 'Dd' describes the mapping for the output stream fourth plane.
14262 Set output pixel format. Default is @code{yuva444p}.
14265 @subsection Examples
14269 Merge three gray video streams of same width and height into single video stream:
14271 [a0][a1][a2]mergeplanes=0x001020:yuv444p
14275 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
14277 [a0][a1]mergeplanes=0x00010210:yuva444p
14281 Swap Y and A plane in yuva444p stream:
14283 format=yuva444p,mergeplanes=0x03010200:yuva444p
14287 Swap U and V plane in yuv420p stream:
14289 format=yuv420p,mergeplanes=0x000201:yuv420p
14293 Cast a rgb24 clip to yuv444p:
14295 format=rgb24,mergeplanes=0x000102:yuv444p
14301 Estimate and export motion vectors using block matching algorithms.
14302 Motion vectors are stored in frame side data to be used by other filters.
14304 This filter accepts the following options:
14307 Specify the motion estimation method. Accepts one of the following values:
14311 Exhaustive search algorithm.
14313 Three step search algorithm.
14315 Two dimensional logarithmic search algorithm.
14317 New three step search algorithm.
14319 Four step search algorithm.
14321 Diamond search algorithm.
14323 Hexagon-based search algorithm.
14325 Enhanced predictive zonal search algorithm.
14327 Uneven multi-hexagon search algorithm.
14329 Default value is @samp{esa}.
14332 Macroblock size. Default @code{16}.
14335 Search parameter. Default @code{7}.
14338 @section midequalizer
14340 Apply Midway Image Equalization effect using two video streams.
14342 Midway Image Equalization adjusts a pair of images to have the same
14343 histogram, while maintaining their dynamics as much as possible. It's
14344 useful for e.g. matching exposures from a pair of stereo cameras.
14346 This filter has two inputs and one output, which must be of same pixel format, but
14347 may be of different sizes. The output of filter is first input adjusted with
14348 midway histogram of both inputs.
14350 This filter accepts the following option:
14354 Set which planes to process. Default is @code{15}, which is all available planes.
14357 @section minterpolate
14359 Convert the video to specified frame rate using motion interpolation.
14361 This filter accepts the following options:
14364 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}.
14367 Motion interpolation mode. Following values are accepted:
14370 Duplicate previous or next frame for interpolating new ones.
14372 Blend source frames. Interpolated frame is mean of previous and next frames.
14374 Motion compensated interpolation. Following options are effective when this mode is selected:
14378 Motion compensation mode. Following values are accepted:
14381 Overlapped block motion compensation.
14383 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
14385 Default mode is @samp{obmc}.
14388 Motion estimation mode. Following values are accepted:
14391 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
14393 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
14395 Default mode is @samp{bilat}.
14398 The algorithm to be used for motion estimation. Following values are accepted:
14401 Exhaustive search algorithm.
14403 Three step search algorithm.
14405 Two dimensional logarithmic search algorithm.
14407 New three step search algorithm.
14409 Four step search algorithm.
14411 Diamond search algorithm.
14413 Hexagon-based search algorithm.
14415 Enhanced predictive zonal search algorithm.
14417 Uneven multi-hexagon search algorithm.
14419 Default algorithm is @samp{epzs}.
14422 Macroblock size. Default @code{16}.
14425 Motion estimation search parameter. Default @code{32}.
14428 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).
14433 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:
14436 Disable scene change detection.
14438 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
14440 Default method is @samp{fdiff}.
14442 @item scd_threshold
14443 Scene change detection threshold. Default is @code{10.}.
14448 Mix several video input streams into one video stream.
14450 A description of the accepted options follows.
14454 The number of inputs. If unspecified, it defaults to 2.
14457 Specify weight of each input video stream as sequence.
14458 Each weight is separated by space. If number of weights
14459 is smaller than number of @var{frames} last specified
14460 weight will be used for all remaining unset weights.
14463 Specify scale, if it is set it will be multiplied with sum
14464 of each weight multiplied with pixel values to give final destination
14465 pixel value. By default @var{scale} is auto scaled to sum of weights.
14468 Specify how end of stream is determined.
14471 The duration of the longest input. (default)
14474 The duration of the shortest input.
14477 The duration of the first input.
14481 @section mpdecimate
14483 Drop frames that do not differ greatly from the previous frame in
14484 order to reduce frame rate.
14486 The main use of this filter is for very-low-bitrate encoding
14487 (e.g. streaming over dialup modem), but it could in theory be used for
14488 fixing movies that were inverse-telecined incorrectly.
14490 A description of the accepted options follows.
14494 Set the maximum number of consecutive frames which can be dropped (if
14495 positive), or the minimum interval between dropped frames (if
14496 negative). If the value is 0, the frame is dropped disregarding the
14497 number of previous sequentially dropped frames.
14499 Default value is 0.
14504 Set the dropping threshold values.
14506 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
14507 represent actual pixel value differences, so a threshold of 64
14508 corresponds to 1 unit of difference for each pixel, or the same spread
14509 out differently over the block.
14511 A frame is a candidate for dropping if no 8x8 blocks differ by more
14512 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
14513 meaning the whole image) differ by more than a threshold of @option{lo}.
14515 Default value for @option{hi} is 64*12, default value for @option{lo} is
14516 64*5, and default value for @option{frac} is 0.33.
14522 Negate (invert) the input video.
14524 It accepts the following option:
14529 With value 1, it negates the alpha component, if present. Default value is 0.
14535 Denoise frames using Non-Local Means algorithm.
14537 Each pixel is adjusted by looking for other pixels with similar contexts. This
14538 context similarity is defined by comparing their surrounding patches of size
14539 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
14542 Note that the research area defines centers for patches, which means some
14543 patches will be made of pixels outside that research area.
14545 The filter accepts the following options.
14549 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
14552 Set patch size. Default is 7. Must be odd number in range [0, 99].
14555 Same as @option{p} but for chroma planes.
14557 The default value is @var{0} and means automatic.
14560 Set research size. Default is 15. Must be odd number in range [0, 99].
14563 Same as @option{r} but for chroma planes.
14565 The default value is @var{0} and means automatic.
14570 Deinterlace video using neural network edge directed interpolation.
14572 This filter accepts the following options:
14576 Mandatory option, without binary file filter can not work.
14577 Currently file can be found here:
14578 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
14581 Set which frames to deinterlace, by default it is @code{all}.
14582 Can be @code{all} or @code{interlaced}.
14585 Set mode of operation.
14587 Can be one of the following:
14591 Use frame flags, both fields.
14593 Use frame flags, single field.
14595 Use top field only.
14597 Use bottom field only.
14599 Use both fields, top first.
14601 Use both fields, bottom first.
14605 Set which planes to process, by default filter process all frames.
14608 Set size of local neighborhood around each pixel, used by the predictor neural
14611 Can be one of the following:
14624 Set the number of neurons in predictor neural network.
14625 Can be one of the following:
14636 Controls the number of different neural network predictions that are blended
14637 together to compute the final output value. Can be @code{fast}, default or
14641 Set which set of weights to use in the predictor.
14642 Can be one of the following:
14646 weights trained to minimize absolute error
14648 weights trained to minimize squared error
14652 Controls whether or not the prescreener neural network is used to decide
14653 which pixels should be processed by the predictor neural network and which
14654 can be handled by simple cubic interpolation.
14655 The prescreener is trained to know whether cubic interpolation will be
14656 sufficient for a pixel or whether it should be predicted by the predictor nn.
14657 The computational complexity of the prescreener nn is much less than that of
14658 the predictor nn. Since most pixels can be handled by cubic interpolation,
14659 using the prescreener generally results in much faster processing.
14660 The prescreener is pretty accurate, so the difference between using it and not
14661 using it is almost always unnoticeable.
14663 Can be one of the following:
14671 Default is @code{new}.
14674 Set various debugging flags.
14679 Force libavfilter not to use any of the specified pixel formats for the
14680 input to the next filter.
14682 It accepts the following parameters:
14686 A '|'-separated list of pixel format names, such as
14687 pix_fmts=yuv420p|monow|rgb24".
14691 @subsection Examples
14695 Force libavfilter to use a format different from @var{yuv420p} for the
14696 input to the vflip filter:
14698 noformat=pix_fmts=yuv420p,vflip
14702 Convert the input video to any of the formats not contained in the list:
14704 noformat=yuv420p|yuv444p|yuv410p
14710 Add noise on video input frame.
14712 The filter accepts the following options:
14720 Set noise seed for specific pixel component or all pixel components in case
14721 of @var{all_seed}. Default value is @code{123457}.
14723 @item all_strength, alls
14724 @item c0_strength, c0s
14725 @item c1_strength, c1s
14726 @item c2_strength, c2s
14727 @item c3_strength, c3s
14728 Set noise strength for specific pixel component or all pixel components in case
14729 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
14731 @item all_flags, allf
14732 @item c0_flags, c0f
14733 @item c1_flags, c1f
14734 @item c2_flags, c2f
14735 @item c3_flags, c3f
14736 Set pixel component flags or set flags for all components if @var{all_flags}.
14737 Available values for component flags are:
14740 averaged temporal noise (smoother)
14742 mix random noise with a (semi)regular pattern
14744 temporal noise (noise pattern changes between frames)
14746 uniform noise (gaussian otherwise)
14750 @subsection Examples
14752 Add temporal and uniform noise to input video:
14754 noise=alls=20:allf=t+u
14759 Normalize RGB video (aka histogram stretching, contrast stretching).
14760 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
14762 For each channel of each frame, the filter computes the input range and maps
14763 it linearly to the user-specified output range. The output range defaults
14764 to the full dynamic range from pure black to pure white.
14766 Temporal smoothing can be used on the input range to reduce flickering (rapid
14767 changes in brightness) caused when small dark or bright objects enter or leave
14768 the scene. This is similar to the auto-exposure (automatic gain control) on a
14769 video camera, and, like a video camera, it may cause a period of over- or
14770 under-exposure of the video.
14772 The R,G,B channels can be normalized independently, which may cause some
14773 color shifting, or linked together as a single channel, which prevents
14774 color shifting. Linked normalization preserves hue. Independent normalization
14775 does not, so it can be used to remove some color casts. Independent and linked
14776 normalization can be combined in any ratio.
14778 The normalize filter accepts the following options:
14783 Colors which define the output range. The minimum input value is mapped to
14784 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
14785 The defaults are black and white respectively. Specifying white for
14786 @var{blackpt} and black for @var{whitept} will give color-inverted,
14787 normalized video. Shades of grey can be used to reduce the dynamic range
14788 (contrast). Specifying saturated colors here can create some interesting
14792 The number of previous frames to use for temporal smoothing. The input range
14793 of each channel is smoothed using a rolling average over the current frame
14794 and the @var{smoothing} previous frames. The default is 0 (no temporal
14798 Controls the ratio of independent (color shifting) channel normalization to
14799 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
14800 independent. Defaults to 1.0 (fully independent).
14803 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
14804 expensive no-op. Defaults to 1.0 (full strength).
14808 @subsection Commands
14809 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
14810 The command accepts the same syntax of the corresponding option.
14812 If the specified expression is not valid, it is kept at its current
14815 @subsection Examples
14817 Stretch video contrast to use the full dynamic range, with no temporal
14818 smoothing; may flicker depending on the source content:
14820 normalize=blackpt=black:whitept=white:smoothing=0
14823 As above, but with 50 frames of temporal smoothing; flicker should be
14824 reduced, depending on the source content:
14826 normalize=blackpt=black:whitept=white:smoothing=50
14829 As above, but with hue-preserving linked channel normalization:
14831 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
14834 As above, but with half strength:
14836 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
14839 Map the darkest input color to red, the brightest input color to cyan:
14841 normalize=blackpt=red:whitept=cyan
14846 Pass the video source unchanged to the output.
14849 Optical Character Recognition
14851 This filter uses Tesseract for optical character recognition. To enable
14852 compilation of this filter, you need to configure FFmpeg with
14853 @code{--enable-libtesseract}.
14855 It accepts the following options:
14859 Set datapath to tesseract data. Default is to use whatever was
14860 set at installation.
14863 Set language, default is "eng".
14866 Set character whitelist.
14869 Set character blacklist.
14872 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
14873 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
14877 Apply a video transform using libopencv.
14879 To enable this filter, install the libopencv library and headers and
14880 configure FFmpeg with @code{--enable-libopencv}.
14882 It accepts the following parameters:
14887 The name of the libopencv filter to apply.
14889 @item filter_params
14890 The parameters to pass to the libopencv filter. If not specified, the default
14891 values are assumed.
14895 Refer to the official libopencv documentation for more precise
14897 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
14899 Several libopencv filters are supported; see the following subsections.
14904 Dilate an image by using a specific structuring element.
14905 It corresponds to the libopencv function @code{cvDilate}.
14907 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
14909 @var{struct_el} represents a structuring element, and has the syntax:
14910 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
14912 @var{cols} and @var{rows} represent the number of columns and rows of
14913 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
14914 point, and @var{shape} the shape for the structuring element. @var{shape}
14915 must be "rect", "cross", "ellipse", or "custom".
14917 If the value for @var{shape} is "custom", it must be followed by a
14918 string of the form "=@var{filename}". The file with name
14919 @var{filename} is assumed to represent a binary image, with each
14920 printable character corresponding to a bright pixel. When a custom
14921 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
14922 or columns and rows of the read file are assumed instead.
14924 The default value for @var{struct_el} is "3x3+0x0/rect".
14926 @var{nb_iterations} specifies the number of times the transform is
14927 applied to the image, and defaults to 1.
14931 # Use the default values
14934 # Dilate using a structuring element with a 5x5 cross, iterating two times
14935 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
14937 # Read the shape from the file diamond.shape, iterating two times.
14938 # The file diamond.shape may contain a pattern of characters like this
14944 # The specified columns and rows are ignored
14945 # but the anchor point coordinates are not
14946 ocv=dilate:0x0+2x2/custom=diamond.shape|2
14951 Erode an image by using a specific structuring element.
14952 It corresponds to the libopencv function @code{cvErode}.
14954 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
14955 with the same syntax and semantics as the @ref{dilate} filter.
14959 Smooth the input video.
14961 The filter takes the following parameters:
14962 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
14964 @var{type} is the type of smooth filter to apply, and must be one of
14965 the following values: "blur", "blur_no_scale", "median", "gaussian",
14966 or "bilateral". The default value is "gaussian".
14968 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
14969 depends on the smooth type. @var{param1} and
14970 @var{param2} accept integer positive values or 0. @var{param3} and
14971 @var{param4} accept floating point values.
14973 The default value for @var{param1} is 3. The default value for the
14974 other parameters is 0.
14976 These parameters correspond to the parameters assigned to the
14977 libopencv function @code{cvSmooth}.
14979 @section oscilloscope
14981 2D Video Oscilloscope.
14983 Useful to measure spatial impulse, step responses, chroma delays, etc.
14985 It accepts the following parameters:
14989 Set scope center x position.
14992 Set scope center y position.
14995 Set scope size, relative to frame diagonal.
14998 Set scope tilt/rotation.
15004 Set trace center x position.
15007 Set trace center y position.
15010 Set trace width, relative to width of frame.
15013 Set trace height, relative to height of frame.
15016 Set which components to trace. By default it traces first three components.
15019 Draw trace grid. By default is enabled.
15022 Draw some statistics. By default is enabled.
15025 Draw scope. By default is enabled.
15028 @subsection Commands
15029 This filter supports same @ref{commands} as options.
15030 The command accepts the same syntax of the corresponding option.
15032 If the specified expression is not valid, it is kept at its current
15035 @subsection Examples
15039 Inspect full first row of video frame.
15041 oscilloscope=x=0.5:y=0:s=1
15045 Inspect full last row of video frame.
15047 oscilloscope=x=0.5:y=1:s=1
15051 Inspect full 5th line of video frame of height 1080.
15053 oscilloscope=x=0.5:y=5/1080:s=1
15057 Inspect full last column of video frame.
15059 oscilloscope=x=1:y=0.5:s=1:t=1
15067 Overlay one video on top of another.
15069 It takes two inputs and has one output. The first input is the "main"
15070 video on which the second input is overlaid.
15072 It accepts the following parameters:
15074 A description of the accepted options follows.
15079 Set the expression for the x and y coordinates of the overlaid video
15080 on the main video. Default value is "0" for both expressions. In case
15081 the expression is invalid, it is set to a huge value (meaning that the
15082 overlay will not be displayed within the output visible area).
15085 See @ref{framesync}.
15088 Set when the expressions for @option{x}, and @option{y} are evaluated.
15090 It accepts the following values:
15093 only evaluate expressions once during the filter initialization or
15094 when a command is processed
15097 evaluate expressions for each incoming frame
15100 Default value is @samp{frame}.
15103 See @ref{framesync}.
15106 Set the format for the output video.
15108 It accepts the following values:
15111 force YUV420 output
15114 force YUV420p10 output
15117 force YUV422 output
15120 force YUV422p10 output
15123 force YUV444 output
15126 force packed RGB output
15129 force planar RGB output
15132 automatically pick format
15135 Default value is @samp{yuv420}.
15138 See @ref{framesync}.
15141 Set format of alpha of the overlaid video, it can be @var{straight} or
15142 @var{premultiplied}. Default is @var{straight}.
15145 The @option{x}, and @option{y} expressions can contain the following
15151 The main input width and height.
15155 The overlay input width and height.
15159 The computed values for @var{x} and @var{y}. They are evaluated for
15164 horizontal and vertical chroma subsample values of the output
15165 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
15169 the number of input frame, starting from 0
15172 the position in the file of the input frame, NAN if unknown
15175 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
15179 This filter also supports the @ref{framesync} options.
15181 Note that the @var{n}, @var{pos}, @var{t} variables are available only
15182 when evaluation is done @emph{per frame}, and will evaluate to NAN
15183 when @option{eval} is set to @samp{init}.
15185 Be aware that frames are taken from each input video in timestamp
15186 order, hence, if their initial timestamps differ, it is a good idea
15187 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
15188 have them begin in the same zero timestamp, as the example for
15189 the @var{movie} filter does.
15191 You can chain together more overlays but you should test the
15192 efficiency of such approach.
15194 @subsection Commands
15196 This filter supports the following commands:
15200 Modify the x and y of the overlay input.
15201 The command accepts the same syntax of the corresponding option.
15203 If the specified expression is not valid, it is kept at its current
15207 @subsection Examples
15211 Draw the overlay at 10 pixels from the bottom right corner of the main
15214 overlay=main_w-overlay_w-10:main_h-overlay_h-10
15217 Using named options the example above becomes:
15219 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
15223 Insert a transparent PNG logo in the bottom left corner of the input,
15224 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
15226 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
15230 Insert 2 different transparent PNG logos (second logo on bottom
15231 right corner) using the @command{ffmpeg} tool:
15233 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
15237 Add a transparent color layer on top of the main video; @code{WxH}
15238 must specify the size of the main input to the overlay filter:
15240 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
15244 Play an original video and a filtered version (here with the deshake
15245 filter) side by side using the @command{ffplay} tool:
15247 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
15250 The above command is the same as:
15252 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
15256 Make a sliding overlay appearing from the left to the right top part of the
15257 screen starting since time 2:
15259 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
15263 Compose output by putting two input videos side to side:
15265 ffmpeg -i left.avi -i right.avi -filter_complex "
15266 nullsrc=size=200x100 [background];
15267 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
15268 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
15269 [background][left] overlay=shortest=1 [background+left];
15270 [background+left][right] overlay=shortest=1:x=100 [left+right]
15275 Mask 10-20 seconds of a video by applying the delogo filter to a section
15277 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
15278 -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]'
15283 Chain several overlays in cascade:
15285 nullsrc=s=200x200 [bg];
15286 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
15287 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
15288 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
15289 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
15290 [in3] null, [mid2] overlay=100:100 [out0]
15295 @anchor{overlay_cuda}
15296 @section overlay_cuda
15298 Overlay one video on top of another.
15300 This is the CUDA variant of the @ref{overlay} filter.
15301 It only accepts CUDA frames. The underlying input pixel formats have to match.
15303 It takes two inputs and has one output. The first input is the "main"
15304 video on which the second input is overlaid.
15306 It accepts the following parameters:
15311 Set the x and y coordinates of the overlaid video on the main video.
15312 Default value is "0" for both expressions.
15315 See @ref{framesync}.
15318 See @ref{framesync}.
15321 See @ref{framesync}.
15325 This filter also supports the @ref{framesync} options.
15329 Apply Overcomplete Wavelet denoiser.
15331 The filter accepts the following options:
15337 Larger depth values will denoise lower frequency components more, but
15338 slow down filtering.
15340 Must be an int in the range 8-16, default is @code{8}.
15342 @item luma_strength, ls
15345 Must be a double value in the range 0-1000, default is @code{1.0}.
15347 @item chroma_strength, cs
15348 Set chroma strength.
15350 Must be a double value in the range 0-1000, default is @code{1.0}.
15356 Add paddings to the input image, and place the original input at the
15357 provided @var{x}, @var{y} coordinates.
15359 It accepts the following parameters:
15364 Specify an expression for the size of the output image with the
15365 paddings added. If the value for @var{width} or @var{height} is 0, the
15366 corresponding input size is used for the output.
15368 The @var{width} expression can reference the value set by the
15369 @var{height} expression, and vice versa.
15371 The default value of @var{width} and @var{height} is 0.
15375 Specify the offsets to place the input image at within the padded area,
15376 with respect to the top/left border of the output image.
15378 The @var{x} expression can reference the value set by the @var{y}
15379 expression, and vice versa.
15381 The default value of @var{x} and @var{y} is 0.
15383 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
15384 so the input image is centered on the padded area.
15387 Specify the color of the padded area. For the syntax of this option,
15388 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
15389 manual,ffmpeg-utils}.
15391 The default value of @var{color} is "black".
15394 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
15396 It accepts the following values:
15400 Only evaluate expressions once during the filter initialization or when
15401 a command is processed.
15404 Evaluate expressions for each incoming frame.
15408 Default value is @samp{init}.
15411 Pad to aspect instead to a resolution.
15415 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
15416 options are expressions containing the following constants:
15421 The input video width and height.
15425 These are the same as @var{in_w} and @var{in_h}.
15429 The output width and height (the size of the padded area), as
15430 specified by the @var{width} and @var{height} expressions.
15434 These are the same as @var{out_w} and @var{out_h}.
15438 The x and y offsets as specified by the @var{x} and @var{y}
15439 expressions, or NAN if not yet specified.
15442 same as @var{iw} / @var{ih}
15445 input sample aspect ratio
15448 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
15452 The horizontal and vertical chroma subsample values. For example for the
15453 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15456 @subsection Examples
15460 Add paddings with the color "violet" to the input video. The output video
15461 size is 640x480, and the top-left corner of the input video is placed at
15464 pad=640:480:0:40:violet
15467 The example above is equivalent to the following command:
15469 pad=width=640:height=480:x=0:y=40:color=violet
15473 Pad the input to get an output with dimensions increased by 3/2,
15474 and put the input video at the center of the padded area:
15476 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
15480 Pad the input to get a squared output with size equal to the maximum
15481 value between the input width and height, and put the input video at
15482 the center of the padded area:
15484 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
15488 Pad the input to get a final w/h ratio of 16:9:
15490 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
15494 In case of anamorphic video, in order to set the output display aspect
15495 correctly, it is necessary to use @var{sar} in the expression,
15496 according to the relation:
15498 (ih * X / ih) * sar = output_dar
15499 X = output_dar / sar
15502 Thus the previous example needs to be modified to:
15504 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
15508 Double the output size and put the input video in the bottom-right
15509 corner of the output padded area:
15511 pad="2*iw:2*ih:ow-iw:oh-ih"
15515 @anchor{palettegen}
15516 @section palettegen
15518 Generate one palette for a whole video stream.
15520 It accepts the following options:
15524 Set the maximum number of colors to quantize in the palette.
15525 Note: the palette will still contain 256 colors; the unused palette entries
15528 @item reserve_transparent
15529 Create a palette of 255 colors maximum and reserve the last one for
15530 transparency. Reserving the transparency color is useful for GIF optimization.
15531 If not set, the maximum of colors in the palette will be 256. You probably want
15532 to disable this option for a standalone image.
15535 @item transparency_color
15536 Set the color that will be used as background for transparency.
15539 Set statistics mode.
15541 It accepts the following values:
15544 Compute full frame histograms.
15546 Compute histograms only for the part that differs from previous frame. This
15547 might be relevant to give more importance to the moving part of your input if
15548 the background is static.
15550 Compute new histogram for each frame.
15553 Default value is @var{full}.
15556 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
15557 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
15558 color quantization of the palette. This information is also visible at
15559 @var{info} logging level.
15561 @subsection Examples
15565 Generate a representative palette of a given video using @command{ffmpeg}:
15567 ffmpeg -i input.mkv -vf palettegen palette.png
15571 @section paletteuse
15573 Use a palette to downsample an input video stream.
15575 The filter takes two inputs: one video stream and a palette. The palette must
15576 be a 256 pixels image.
15578 It accepts the following options:
15582 Select dithering mode. Available algorithms are:
15585 Ordered 8x8 bayer dithering (deterministic)
15587 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
15588 Note: this dithering is sometimes considered "wrong" and is included as a
15590 @item floyd_steinberg
15591 Floyd and Steingberg dithering (error diffusion)
15593 Frankie Sierra dithering v2 (error diffusion)
15595 Frankie Sierra dithering v2 "Lite" (error diffusion)
15598 Default is @var{sierra2_4a}.
15601 When @var{bayer} dithering is selected, this option defines the scale of the
15602 pattern (how much the crosshatch pattern is visible). A low value means more
15603 visible pattern for less banding, and higher value means less visible pattern
15604 at the cost of more banding.
15606 The option must be an integer value in the range [0,5]. Default is @var{2}.
15609 If set, define the zone to process
15613 Only the changing rectangle will be reprocessed. This is similar to GIF
15614 cropping/offsetting compression mechanism. This option can be useful for speed
15615 if only a part of the image is changing, and has use cases such as limiting the
15616 scope of the error diffusal @option{dither} to the rectangle that bounds the
15617 moving scene (it leads to more deterministic output if the scene doesn't change
15618 much, and as a result less moving noise and better GIF compression).
15621 Default is @var{none}.
15624 Take new palette for each output frame.
15626 @item alpha_threshold
15627 Sets the alpha threshold for transparency. Alpha values above this threshold
15628 will be treated as completely opaque, and values below this threshold will be
15629 treated as completely transparent.
15631 The option must be an integer value in the range [0,255]. Default is @var{128}.
15634 @subsection Examples
15638 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
15639 using @command{ffmpeg}:
15641 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
15645 @section perspective
15647 Correct perspective of video not recorded perpendicular to the screen.
15649 A description of the accepted parameters follows.
15660 Set coordinates expression for top left, top right, bottom left and bottom right corners.
15661 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
15662 If the @code{sense} option is set to @code{source}, then the specified points will be sent
15663 to the corners of the destination. If the @code{sense} option is set to @code{destination},
15664 then the corners of the source will be sent to the specified coordinates.
15666 The expressions can use the following variables:
15671 the width and height of video frame.
15675 Output frame count.
15678 @item interpolation
15679 Set interpolation for perspective correction.
15681 It accepts the following values:
15687 Default value is @samp{linear}.
15690 Set interpretation of coordinate options.
15692 It accepts the following values:
15696 Send point in the source specified by the given coordinates to
15697 the corners of the destination.
15699 @item 1, destination
15701 Send the corners of the source to the point in the destination specified
15702 by the given coordinates.
15704 Default value is @samp{source}.
15708 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
15710 It accepts the following values:
15713 only evaluate expressions once during the filter initialization or
15714 when a command is processed
15717 evaluate expressions for each incoming frame
15720 Default value is @samp{init}.
15725 Delay interlaced video by one field time so that the field order changes.
15727 The intended use is to fix PAL movies that have been captured with the
15728 opposite field order to the film-to-video transfer.
15730 A description of the accepted parameters follows.
15736 It accepts the following values:
15739 Capture field order top-first, transfer bottom-first.
15740 Filter will delay the bottom field.
15743 Capture field order bottom-first, transfer top-first.
15744 Filter will delay the top field.
15747 Capture and transfer with the same field order. This mode only exists
15748 for the documentation of the other options to refer to, but if you
15749 actually select it, the filter will faithfully do nothing.
15752 Capture field order determined automatically by field flags, transfer
15754 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
15755 basis using field flags. If no field information is available,
15756 then this works just like @samp{u}.
15759 Capture unknown or varying, transfer opposite.
15760 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
15761 analyzing the images and selecting the alternative that produces best
15762 match between the fields.
15765 Capture top-first, transfer unknown or varying.
15766 Filter selects among @samp{t} and @samp{p} using image analysis.
15769 Capture bottom-first, transfer unknown or varying.
15770 Filter selects among @samp{b} and @samp{p} using image analysis.
15773 Capture determined by field flags, transfer unknown or varying.
15774 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
15775 image analysis. If no field information is available, then this works just
15776 like @samp{U}. This is the default mode.
15779 Both capture and transfer unknown or varying.
15780 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
15784 @subsection Commands
15786 This filter supports the all above options as @ref{commands}.
15788 @section photosensitivity
15789 Reduce various flashes in video, so to help users with epilepsy.
15791 It accepts the following options:
15794 Set how many frames to use when filtering. Default is 30.
15797 Set detection threshold factor. Default is 1.
15801 Set how many pixels to skip when sampling frames. Default is 1.
15802 Allowed range is from 1 to 1024.
15805 Leave frames unchanged. Default is disabled.
15808 @section pixdesctest
15810 Pixel format descriptor test filter, mainly useful for internal
15811 testing. The output video should be equal to the input video.
15815 format=monow, pixdesctest
15818 can be used to test the monowhite pixel format descriptor definition.
15822 Display sample values of color channels. Mainly useful for checking color
15823 and levels. Minimum supported resolution is 640x480.
15825 The filters accept the following options:
15829 Set scope X position, relative offset on X axis.
15832 Set scope Y position, relative offset on Y axis.
15841 Set window opacity. This window also holds statistics about pixel area.
15844 Set window X position, relative offset on X axis.
15847 Set window Y position, relative offset on Y axis.
15852 Enable the specified chain of postprocessing subfilters using libpostproc. This
15853 library should be automatically selected with a GPL build (@code{--enable-gpl}).
15854 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
15855 Each subfilter and some options have a short and a long name that can be used
15856 interchangeably, i.e. dr/dering are the same.
15858 The filters accept the following options:
15862 Set postprocessing subfilters string.
15865 All subfilters share common options to determine their scope:
15869 Honor the quality commands for this subfilter.
15872 Do chrominance filtering, too (default).
15875 Do luminance filtering only (no chrominance).
15878 Do chrominance filtering only (no luminance).
15881 These options can be appended after the subfilter name, separated by a '|'.
15883 Available subfilters are:
15886 @item hb/hdeblock[|difference[|flatness]]
15887 Horizontal deblocking filter
15890 Difference factor where higher values mean more deblocking (default: @code{32}).
15892 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15895 @item vb/vdeblock[|difference[|flatness]]
15896 Vertical deblocking filter
15899 Difference factor where higher values mean more deblocking (default: @code{32}).
15901 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15904 @item ha/hadeblock[|difference[|flatness]]
15905 Accurate horizontal deblocking filter
15908 Difference factor where higher values mean more deblocking (default: @code{32}).
15910 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15913 @item va/vadeblock[|difference[|flatness]]
15914 Accurate vertical deblocking filter
15917 Difference factor where higher values mean more deblocking (default: @code{32}).
15919 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15923 The horizontal and vertical deblocking filters share the difference and
15924 flatness values so you cannot set different horizontal and vertical
15928 @item h1/x1hdeblock
15929 Experimental horizontal deblocking filter
15931 @item v1/x1vdeblock
15932 Experimental vertical deblocking filter
15937 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
15940 larger -> stronger filtering
15942 larger -> stronger filtering
15944 larger -> stronger filtering
15947 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
15950 Stretch luminance to @code{0-255}.
15953 @item lb/linblenddeint
15954 Linear blend deinterlacing filter that deinterlaces the given block by
15955 filtering all lines with a @code{(1 2 1)} filter.
15957 @item li/linipoldeint
15958 Linear interpolating deinterlacing filter that deinterlaces the given block by
15959 linearly interpolating every second line.
15961 @item ci/cubicipoldeint
15962 Cubic interpolating deinterlacing filter deinterlaces the given block by
15963 cubically interpolating every second line.
15965 @item md/mediandeint
15966 Median deinterlacing filter that deinterlaces the given block by applying a
15967 median filter to every second line.
15969 @item fd/ffmpegdeint
15970 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
15971 second line with a @code{(-1 4 2 4 -1)} filter.
15974 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
15975 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
15977 @item fq/forceQuant[|quantizer]
15978 Overrides the quantizer table from the input with the constant quantizer you
15986 Default pp filter combination (@code{hb|a,vb|a,dr|a})
15989 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
15992 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
15995 @subsection Examples
15999 Apply horizontal and vertical deblocking, deringing and automatic
16000 brightness/contrast:
16006 Apply default filters without brightness/contrast correction:
16012 Apply default filters and temporal denoiser:
16014 pp=default/tmpnoise|1|2|3
16018 Apply deblocking on luminance only, and switch vertical deblocking on or off
16019 automatically depending on available CPU time:
16026 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
16027 similar to spp = 6 with 7 point DCT, where only the center sample is
16030 The filter accepts the following options:
16034 Force a constant quantization parameter. It accepts an integer in range
16035 0 to 63. If not set, the filter will use the QP from the video stream
16039 Set thresholding mode. Available modes are:
16043 Set hard thresholding.
16045 Set soft thresholding (better de-ringing effect, but likely blurrier).
16047 Set medium thresholding (good results, default).
16051 @section premultiply
16052 Apply alpha premultiply effect to input video stream using first plane
16053 of second stream as alpha.
16055 Both streams must have same dimensions and same pixel format.
16057 The filter accepts the following option:
16061 Set which planes will be processed, unprocessed planes will be copied.
16062 By default value 0xf, all planes will be processed.
16065 Do not require 2nd input for processing, instead use alpha plane from input stream.
16069 Apply prewitt operator to input video stream.
16071 The filter accepts the following option:
16075 Set which planes will be processed, unprocessed planes will be copied.
16076 By default value 0xf, all planes will be processed.
16079 Set value which will be multiplied with filtered result.
16082 Set value which will be added to filtered result.
16085 @subsection Commands
16087 This filter supports the all above options as @ref{commands}.
16089 @section pseudocolor
16091 Alter frame colors in video with pseudocolors.
16093 This filter accepts the following options:
16097 set pixel first component expression
16100 set pixel second component expression
16103 set pixel third component expression
16106 set pixel fourth component expression, corresponds to the alpha component
16109 set component to use as base for altering colors
16112 Each of them specifies the expression to use for computing the lookup table for
16113 the corresponding pixel component values.
16115 The expressions can contain the following constants and functions:
16120 The input width and height.
16123 The input value for the pixel component.
16125 @item ymin, umin, vmin, amin
16126 The minimum allowed component value.
16128 @item ymax, umax, vmax, amax
16129 The maximum allowed component value.
16132 All expressions default to "val".
16134 @subsection Examples
16138 Change too high luma values to gradient:
16140 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'"
16146 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
16147 Ratio) between two input videos.
16149 This filter takes in input two input videos, the first input is
16150 considered the "main" source and is passed unchanged to the
16151 output. The second input is used as a "reference" video for computing
16154 Both video inputs must have the same resolution and pixel format for
16155 this filter to work correctly. Also it assumes that both inputs
16156 have the same number of frames, which are compared one by one.
16158 The obtained average PSNR is printed through the logging system.
16160 The filter stores the accumulated MSE (mean squared error) of each
16161 frame, and at the end of the processing it is averaged across all frames
16162 equally, and the following formula is applied to obtain the PSNR:
16165 PSNR = 10*log10(MAX^2/MSE)
16168 Where MAX is the average of the maximum values of each component of the
16171 The description of the accepted parameters follows.
16174 @item stats_file, f
16175 If specified the filter will use the named file to save the PSNR of
16176 each individual frame. When filename equals "-" the data is sent to
16179 @item stats_version
16180 Specifies which version of the stats file format to use. Details of
16181 each format are written below.
16182 Default value is 1.
16184 @item stats_add_max
16185 Determines whether the max value is output to the stats log.
16186 Default value is 0.
16187 Requires stats_version >= 2. If this is set and stats_version < 2,
16188 the filter will return an error.
16191 This filter also supports the @ref{framesync} options.
16193 The file printed if @var{stats_file} is selected, contains a sequence of
16194 key/value pairs of the form @var{key}:@var{value} for each compared
16197 If a @var{stats_version} greater than 1 is specified, a header line precedes
16198 the list of per-frame-pair stats, with key value pairs following the frame
16199 format with the following parameters:
16202 @item psnr_log_version
16203 The version of the log file format. Will match @var{stats_version}.
16206 A comma separated list of the per-frame-pair parameters included in
16210 A description of each shown per-frame-pair parameter follows:
16214 sequential number of the input frame, starting from 1
16217 Mean Square Error pixel-by-pixel average difference of the compared
16218 frames, averaged over all the image components.
16220 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
16221 Mean Square Error pixel-by-pixel average difference of the compared
16222 frames for the component specified by the suffix.
16224 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
16225 Peak Signal to Noise ratio of the compared frames for the component
16226 specified by the suffix.
16228 @item max_avg, max_y, max_u, max_v
16229 Maximum allowed value for each channel, and average over all
16233 @subsection Examples
16238 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
16239 [main][ref] psnr="stats_file=stats.log" [out]
16242 On this example the input file being processed is compared with the
16243 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
16244 is stored in @file{stats.log}.
16247 Another example with different containers:
16249 ffmpeg -i main.mpg -i ref.mkv -lavfi "[0:v]settb=AVTB,setpts=PTS-STARTPTS[main];[1:v]settb=AVTB,setpts=PTS-STARTPTS[ref];[main][ref]psnr" -f null -
16256 Pulldown reversal (inverse telecine) filter, capable of handling mixed
16257 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
16260 The pullup filter is designed to take advantage of future context in making
16261 its decisions. This filter is stateless in the sense that it does not lock
16262 onto a pattern to follow, but it instead looks forward to the following
16263 fields in order to identify matches and rebuild progressive frames.
16265 To produce content with an even framerate, insert the fps filter after
16266 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
16267 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
16269 The filter accepts the following options:
16276 These options set the amount of "junk" to ignore at the left, right, top, and
16277 bottom of the image, respectively. Left and right are in units of 8 pixels,
16278 while top and bottom are in units of 2 lines.
16279 The default is 8 pixels on each side.
16282 Set the strict breaks. Setting this option to 1 will reduce the chances of
16283 filter generating an occasional mismatched frame, but it may also cause an
16284 excessive number of frames to be dropped during high motion sequences.
16285 Conversely, setting it to -1 will make filter match fields more easily.
16286 This may help processing of video where there is slight blurring between
16287 the fields, but may also cause there to be interlaced frames in the output.
16288 Default value is @code{0}.
16291 Set the metric plane to use. It accepts the following values:
16297 Use chroma blue plane.
16300 Use chroma red plane.
16303 This option may be set to use chroma plane instead of the default luma plane
16304 for doing filter's computations. This may improve accuracy on very clean
16305 source material, but more likely will decrease accuracy, especially if there
16306 is chroma noise (rainbow effect) or any grayscale video.
16307 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
16308 load and make pullup usable in realtime on slow machines.
16311 For best results (without duplicated frames in the output file) it is
16312 necessary to change the output frame rate. For example, to inverse
16313 telecine NTSC input:
16315 ffmpeg -i input -vf pullup -r 24000/1001 ...
16320 Change video quantization parameters (QP).
16322 The filter accepts the following option:
16326 Set expression for quantization parameter.
16329 The expression is evaluated through the eval API and can contain, among others,
16330 the following constants:
16334 1 if index is not 129, 0 otherwise.
16337 Sequential index starting from -129 to 128.
16340 @subsection Examples
16344 Some equation like:
16352 Flush video frames from internal cache of frames into a random order.
16353 No frame is discarded.
16354 Inspired by @ref{frei0r} nervous filter.
16358 Set size in number of frames of internal cache, in range from @code{2} to
16359 @code{512}. Default is @code{30}.
16362 Set seed for random number generator, must be an integer included between
16363 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
16364 less than @code{0}, the filter will try to use a good random seed on a
16368 @section readeia608
16370 Read closed captioning (EIA-608) information from the top lines of a video frame.
16372 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
16373 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
16374 with EIA-608 data (starting from 0). A description of each metadata value follows:
16377 @item lavfi.readeia608.X.cc
16378 The two bytes stored as EIA-608 data (printed in hexadecimal).
16380 @item lavfi.readeia608.X.line
16381 The number of the line on which the EIA-608 data was identified and read.
16384 This filter accepts the following options:
16388 Set the line to start scanning for EIA-608 data. Default is @code{0}.
16391 Set the line to end scanning for EIA-608 data. Default is @code{29}.
16394 Set the ratio of width reserved for sync code detection.
16395 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
16398 Enable checking the parity bit. In the event of a parity error, the filter will output
16399 @code{0x00} for that character. Default is false.
16402 Lowpass lines prior to further processing. Default is enabled.
16405 @subsection Commands
16407 This filter supports the all above options as @ref{commands}.
16409 @subsection Examples
16413 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
16415 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
16421 Read vertical interval timecode (VITC) information from the top lines of a
16424 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
16425 timecode value, if a valid timecode has been detected. Further metadata key
16426 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
16427 timecode data has been found or not.
16429 This filter accepts the following options:
16433 Set the maximum number of lines to scan for VITC data. If the value is set to
16434 @code{-1} the full video frame is scanned. Default is @code{45}.
16437 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
16438 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
16441 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
16442 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
16445 @subsection Examples
16449 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
16450 draw @code{--:--:--:--} as a placeholder:
16452 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
16458 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
16460 Destination pixel at position (X, Y) will be picked from source (x, y) position
16461 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
16462 value for pixel will be used for destination pixel.
16464 Xmap and Ymap input video streams must be of same dimensions. Output video stream
16465 will have Xmap/Ymap video stream dimensions.
16466 Xmap and Ymap input video streams are 16bit depth, single channel.
16470 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
16471 Default is @code{color}.
16474 Specify the color of the unmapped pixels. For the syntax of this option,
16475 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
16476 manual,ffmpeg-utils}. Default color is @code{black}.
16479 @section removegrain
16481 The removegrain filter is a spatial denoiser for progressive video.
16485 Set mode for the first plane.
16488 Set mode for the second plane.
16491 Set mode for the third plane.
16494 Set mode for the fourth plane.
16497 Range of mode is from 0 to 24. Description of each mode follows:
16501 Leave input plane unchanged. Default.
16504 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
16507 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
16510 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
16513 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
16514 This is equivalent to a median filter.
16517 Line-sensitive clipping giving the minimal change.
16520 Line-sensitive clipping, intermediate.
16523 Line-sensitive clipping, intermediate.
16526 Line-sensitive clipping, intermediate.
16529 Line-sensitive clipping on a line where the neighbours pixels are the closest.
16532 Replaces the target pixel with the closest neighbour.
16535 [1 2 1] horizontal and vertical kernel blur.
16541 Bob mode, interpolates top field from the line where the neighbours
16542 pixels are the closest.
16545 Bob mode, interpolates bottom field from the line where the neighbours
16546 pixels are the closest.
16549 Bob mode, interpolates top field. Same as 13 but with a more complicated
16550 interpolation formula.
16553 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
16554 interpolation formula.
16557 Clips the pixel with the minimum and maximum of respectively the maximum and
16558 minimum of each pair of opposite neighbour pixels.
16561 Line-sensitive clipping using opposite neighbours whose greatest distance from
16562 the current pixel is minimal.
16565 Replaces the pixel with the average of its 8 neighbours.
16568 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
16571 Clips pixels using the averages of opposite neighbour.
16574 Same as mode 21 but simpler and faster.
16577 Small edge and halo removal, but reputed useless.
16583 @section removelogo
16585 Suppress a TV station logo, using an image file to determine which
16586 pixels comprise the logo. It works by filling in the pixels that
16587 comprise the logo with neighboring pixels.
16589 The filter accepts the following options:
16593 Set the filter bitmap file, which can be any image format supported by
16594 libavformat. The width and height of the image file must match those of the
16595 video stream being processed.
16598 Pixels in the provided bitmap image with a value of zero are not
16599 considered part of the logo, non-zero pixels are considered part of
16600 the logo. If you use white (255) for the logo and black (0) for the
16601 rest, you will be safe. For making the filter bitmap, it is
16602 recommended to take a screen capture of a black frame with the logo
16603 visible, and then using a threshold filter followed by the erode
16604 filter once or twice.
16606 If needed, little splotches can be fixed manually. Remember that if
16607 logo pixels are not covered, the filter quality will be much
16608 reduced. Marking too many pixels as part of the logo does not hurt as
16609 much, but it will increase the amount of blurring needed to cover over
16610 the image and will destroy more information than necessary, and extra
16611 pixels will slow things down on a large logo.
16613 @section repeatfields
16615 This filter uses the repeat_field flag from the Video ES headers and hard repeats
16616 fields based on its value.
16620 Reverse a video clip.
16622 Warning: This filter requires memory to buffer the entire clip, so trimming
16625 @subsection Examples
16629 Take the first 5 seconds of a clip, and reverse it.
16636 Shift R/G/B/A pixels horizontally and/or vertically.
16638 The filter accepts the following options:
16641 Set amount to shift red horizontally.
16643 Set amount to shift red vertically.
16645 Set amount to shift green horizontally.
16647 Set amount to shift green vertically.
16649 Set amount to shift blue horizontally.
16651 Set amount to shift blue vertically.
16653 Set amount to shift alpha horizontally.
16655 Set amount to shift alpha vertically.
16657 Set edge mode, can be @var{smear}, default, or @var{warp}.
16660 @subsection Commands
16662 This filter supports the all above options as @ref{commands}.
16665 Apply roberts cross operator to input video stream.
16667 The filter accepts the following option:
16671 Set which planes will be processed, unprocessed planes will be copied.
16672 By default value 0xf, all planes will be processed.
16675 Set value which will be multiplied with filtered result.
16678 Set value which will be added to filtered result.
16681 @subsection Commands
16683 This filter supports the all above options as @ref{commands}.
16687 Rotate video by an arbitrary angle expressed in radians.
16689 The filter accepts the following options:
16691 A description of the optional parameters follows.
16694 Set an expression for the angle by which to rotate the input video
16695 clockwise, expressed as a number of radians. A negative value will
16696 result in a counter-clockwise rotation. By default it is set to "0".
16698 This expression is evaluated for each frame.
16701 Set the output width expression, default value is "iw".
16702 This expression is evaluated just once during configuration.
16705 Set the output height expression, default value is "ih".
16706 This expression is evaluated just once during configuration.
16709 Enable bilinear interpolation if set to 1, a value of 0 disables
16710 it. Default value is 1.
16713 Set the color used to fill the output area not covered by the rotated
16714 image. For the general syntax of this option, check the
16715 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
16716 If the special value "none" is selected then no
16717 background is printed (useful for example if the background is never shown).
16719 Default value is "black".
16722 The expressions for the angle and the output size can contain the
16723 following constants and functions:
16727 sequential number of the input frame, starting from 0. It is always NAN
16728 before the first frame is filtered.
16731 time in seconds of the input frame, it is set to 0 when the filter is
16732 configured. It is always NAN before the first frame is filtered.
16736 horizontal and vertical chroma subsample values. For example for the
16737 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16741 the input video width and height
16745 the output width and height, that is the size of the padded area as
16746 specified by the @var{width} and @var{height} expressions
16750 the minimal width/height required for completely containing the input
16751 video rotated by @var{a} radians.
16753 These are only available when computing the @option{out_w} and
16754 @option{out_h} expressions.
16757 @subsection Examples
16761 Rotate the input by PI/6 radians clockwise:
16767 Rotate the input by PI/6 radians counter-clockwise:
16773 Rotate the input by 45 degrees clockwise:
16779 Apply a constant rotation with period T, starting from an angle of PI/3:
16781 rotate=PI/3+2*PI*t/T
16785 Make the input video rotation oscillating with a period of T
16786 seconds and an amplitude of A radians:
16788 rotate=A*sin(2*PI/T*t)
16792 Rotate the video, output size is chosen so that the whole rotating
16793 input video is always completely contained in the output:
16795 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
16799 Rotate the video, reduce the output size so that no background is ever
16802 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
16806 @subsection Commands
16808 The filter supports the following commands:
16812 Set the angle expression.
16813 The command accepts the same syntax of the corresponding option.
16815 If the specified expression is not valid, it is kept at its current
16821 Apply Shape Adaptive Blur.
16823 The filter accepts the following options:
16826 @item luma_radius, lr
16827 Set luma blur filter strength, must be a value in range 0.1-4.0, default
16828 value is 1.0. A greater value will result in a more blurred image, and
16829 in slower processing.
16831 @item luma_pre_filter_radius, lpfr
16832 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
16835 @item luma_strength, ls
16836 Set luma maximum difference between pixels to still be considered, must
16837 be a value in the 0.1-100.0 range, default value is 1.0.
16839 @item chroma_radius, cr
16840 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
16841 greater value will result in a more blurred image, and in slower
16844 @item chroma_pre_filter_radius, cpfr
16845 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
16847 @item chroma_strength, cs
16848 Set chroma maximum difference between pixels to still be considered,
16849 must be a value in the -0.9-100.0 range.
16852 Each chroma option value, if not explicitly specified, is set to the
16853 corresponding luma option value.
16858 Scale (resize) the input video, using the libswscale library.
16860 The scale filter forces the output display aspect ratio to be the same
16861 of the input, by changing the output sample aspect ratio.
16863 If the input image format is different from the format requested by
16864 the next filter, the scale filter will convert the input to the
16867 @subsection Options
16868 The filter accepts the following options, or any of the options
16869 supported by the libswscale scaler.
16871 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
16872 the complete list of scaler options.
16877 Set the output video dimension expression. Default value is the input
16880 If the @var{width} or @var{w} value is 0, the input width is used for
16881 the output. If the @var{height} or @var{h} value is 0, the input height
16882 is used for the output.
16884 If one and only one of the values is -n with n >= 1, the scale filter
16885 will use a value that maintains the aspect ratio of the input image,
16886 calculated from the other specified dimension. After that it will,
16887 however, make sure that the calculated dimension is divisible by n and
16888 adjust the value if necessary.
16890 If both values are -n with n >= 1, the behavior will be identical to
16891 both values being set to 0 as previously detailed.
16893 See below for the list of accepted constants for use in the dimension
16897 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
16901 Only evaluate expressions once during the filter initialization or when a command is processed.
16904 Evaluate expressions for each incoming frame.
16908 Default value is @samp{init}.
16912 Set the interlacing mode. It accepts the following values:
16916 Force interlaced aware scaling.
16919 Do not apply interlaced scaling.
16922 Select interlaced aware scaling depending on whether the source frames
16923 are flagged as interlaced or not.
16926 Default value is @samp{0}.
16929 Set libswscale scaling flags. See
16930 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16931 complete list of values. If not explicitly specified the filter applies
16935 @item param0, param1
16936 Set libswscale input parameters for scaling algorithms that need them. See
16937 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16938 complete documentation. If not explicitly specified the filter applies
16944 Set the video size. For the syntax of this option, check the
16945 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16947 @item in_color_matrix
16948 @item out_color_matrix
16949 Set in/output YCbCr color space type.
16951 This allows the autodetected value to be overridden as well as allows forcing
16952 a specific value used for the output and encoder.
16954 If not specified, the color space type depends on the pixel format.
16960 Choose automatically.
16963 Format conforming to International Telecommunication Union (ITU)
16964 Recommendation BT.709.
16967 Set color space conforming to the United States Federal Communications
16968 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
16973 Set color space conforming to:
16977 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
16980 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
16983 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
16988 Set color space conforming to SMPTE ST 240:1999.
16991 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
16996 Set in/output YCbCr sample range.
16998 This allows the autodetected value to be overridden as well as allows forcing
16999 a specific value used for the output and encoder. If not specified, the
17000 range depends on the pixel format. Possible values:
17004 Choose automatically.
17007 Set full range (0-255 in case of 8-bit luma).
17009 @item mpeg/limited/tv
17010 Set "MPEG" range (16-235 in case of 8-bit luma).
17013 @item force_original_aspect_ratio
17014 Enable decreasing or increasing output video width or height if necessary to
17015 keep the original aspect ratio. Possible values:
17019 Scale the video as specified and disable this feature.
17022 The output video dimensions will automatically be decreased if needed.
17025 The output video dimensions will automatically be increased if needed.
17029 One useful instance of this option is that when you know a specific device's
17030 maximum allowed resolution, you can use this to limit the output video to
17031 that, while retaining the aspect ratio. For example, device A allows
17032 1280x720 playback, and your video is 1920x800. Using this option (set it to
17033 decrease) and specifying 1280x720 to the command line makes the output
17036 Please note that this is a different thing than specifying -1 for @option{w}
17037 or @option{h}, you still need to specify the output resolution for this option
17040 @item force_divisible_by
17041 Ensures that both the output dimensions, width and height, are divisible by the
17042 given integer when used together with @option{force_original_aspect_ratio}. This
17043 works similar to using @code{-n} in the @option{w} and @option{h} options.
17045 This option respects the value set for @option{force_original_aspect_ratio},
17046 increasing or decreasing the resolution accordingly. The video's aspect ratio
17047 may be slightly modified.
17049 This option can be handy if you need to have a video fit within or exceed
17050 a defined resolution using @option{force_original_aspect_ratio} but also have
17051 encoder restrictions on width or height divisibility.
17055 The values of the @option{w} and @option{h} options are expressions
17056 containing the following constants:
17061 The input width and height
17065 These are the same as @var{in_w} and @var{in_h}.
17069 The output (scaled) width and height
17073 These are the same as @var{out_w} and @var{out_h}
17076 The same as @var{iw} / @var{ih}
17079 input sample aspect ratio
17082 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
17086 horizontal and vertical input chroma subsample values. For example for the
17087 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17091 horizontal and vertical output chroma subsample values. For example for the
17092 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17095 The (sequential) number of the input frame, starting from 0.
17096 Only available with @code{eval=frame}.
17099 The presentation timestamp of the input frame, expressed as a number of
17100 seconds. Only available with @code{eval=frame}.
17103 The position (byte offset) of the frame in the input stream, or NaN if
17104 this information is unavailable and/or meaningless (for example in case of synthetic video).
17105 Only available with @code{eval=frame}.
17108 @subsection Examples
17112 Scale the input video to a size of 200x100
17117 This is equivalent to:
17128 Specify a size abbreviation for the output size:
17133 which can also be written as:
17139 Scale the input to 2x:
17141 scale=w=2*iw:h=2*ih
17145 The above is the same as:
17147 scale=2*in_w:2*in_h
17151 Scale the input to 2x with forced interlaced scaling:
17153 scale=2*iw:2*ih:interl=1
17157 Scale the input to half size:
17159 scale=w=iw/2:h=ih/2
17163 Increase the width, and set the height to the same size:
17169 Seek Greek harmony:
17176 Increase the height, and set the width to 3/2 of the height:
17178 scale=w=3/2*oh:h=3/5*ih
17182 Increase the size, making the size a multiple of the chroma
17185 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
17189 Increase the width to a maximum of 500 pixels,
17190 keeping the same aspect ratio as the input:
17192 scale=w='min(500\, iw*3/2):h=-1'
17196 Make pixels square by combining scale and setsar:
17198 scale='trunc(ih*dar):ih',setsar=1/1
17202 Make pixels square by combining scale and setsar,
17203 making sure the resulting resolution is even (required by some codecs):
17205 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
17209 @subsection Commands
17211 This filter supports the following commands:
17215 Set the output video dimension expression.
17216 The command accepts the same syntax of the corresponding option.
17218 If the specified expression is not valid, it is kept at its current
17224 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
17225 format conversion on CUDA video frames. Setting the output width and height
17226 works in the same way as for the @var{scale} filter.
17228 The following additional options are accepted:
17231 The pixel format of the output CUDA frames. If set to the string "same" (the
17232 default), the input format will be kept. Note that automatic format negotiation
17233 and conversion is not yet supported for hardware frames
17236 The interpolation algorithm used for resizing. One of the following:
17243 @item cubic2p_bspline
17244 2-parameter cubic (B=1, C=0)
17246 @item cubic2p_catmullrom
17247 2-parameter cubic (B=0, C=1/2)
17249 @item cubic2p_b05c03
17250 2-parameter cubic (B=1/2, C=3/10)
17258 @item force_original_aspect_ratio
17259 Enable decreasing or increasing output video width or height if necessary to
17260 keep the original aspect ratio. Possible values:
17264 Scale the video as specified and disable this feature.
17267 The output video dimensions will automatically be decreased if needed.
17270 The output video dimensions will automatically be increased if needed.
17274 One useful instance of this option is that when you know a specific device's
17275 maximum allowed resolution, you can use this to limit the output video to
17276 that, while retaining the aspect ratio. For example, device A allows
17277 1280x720 playback, and your video is 1920x800. Using this option (set it to
17278 decrease) and specifying 1280x720 to the command line makes the output
17281 Please note that this is a different thing than specifying -1 for @option{w}
17282 or @option{h}, you still need to specify the output resolution for this option
17285 @item force_divisible_by
17286 Ensures that both the output dimensions, width and height, are divisible by the
17287 given integer when used together with @option{force_original_aspect_ratio}. This
17288 works similar to using @code{-n} in the @option{w} and @option{h} options.
17290 This option respects the value set for @option{force_original_aspect_ratio},
17291 increasing or decreasing the resolution accordingly. The video's aspect ratio
17292 may be slightly modified.
17294 This option can be handy if you need to have a video fit within or exceed
17295 a defined resolution using @option{force_original_aspect_ratio} but also have
17296 encoder restrictions on width or height divisibility.
17302 Scale (resize) the input video, based on a reference video.
17304 See the scale filter for available options, scale2ref supports the same but
17305 uses the reference video instead of the main input as basis. scale2ref also
17306 supports the following additional constants for the @option{w} and
17307 @option{h} options:
17312 The main input video's width and height
17315 The same as @var{main_w} / @var{main_h}
17318 The main input video's sample aspect ratio
17320 @item main_dar, mdar
17321 The main input video's display aspect ratio. Calculated from
17322 @code{(main_w / main_h) * main_sar}.
17326 The main input video's horizontal and vertical chroma subsample values.
17327 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
17331 The (sequential) number of the main input frame, starting from 0.
17332 Only available with @code{eval=frame}.
17335 The presentation timestamp of the main input frame, expressed as a number of
17336 seconds. Only available with @code{eval=frame}.
17339 The position (byte offset) of the frame in the main input stream, or NaN if
17340 this information is unavailable and/or meaningless (for example in case of synthetic video).
17341 Only available with @code{eval=frame}.
17344 @subsection Examples
17348 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
17350 'scale2ref[b][a];[a][b]overlay'
17354 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
17356 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
17360 @subsection Commands
17362 This filter supports the following commands:
17366 Set the output video dimension expression.
17367 The command accepts the same syntax of the corresponding option.
17369 If the specified expression is not valid, it is kept at its current
17374 Scroll input video horizontally and/or vertically by constant speed.
17376 The filter accepts the following options:
17378 @item horizontal, h
17379 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
17380 Negative values changes scrolling direction.
17383 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
17384 Negative values changes scrolling direction.
17387 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
17390 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
17393 @subsection Commands
17395 This filter supports the following @ref{commands}:
17397 @item horizontal, h
17398 Set the horizontal scrolling speed.
17400 Set the vertical scrolling speed.
17406 Detect video scene change.
17408 This filter sets frame metadata with mafd between frame, the scene score, and
17409 forward the frame to the next filter, so they can use these metadata to detect
17410 scene change or others.
17412 In addition, this filter logs a message and sets frame metadata when it detects
17413 a scene change by @option{threshold}.
17415 @code{lavfi.scd.mafd} metadata keys are set with mafd for every frame.
17417 @code{lavfi.scd.score} metadata keys are set with scene change score for every frame
17418 to detect scene change.
17420 @code{lavfi.scd.time} metadata keys are set with current filtered frame time which
17421 detect scene change with @option{threshold}.
17423 The filter accepts the following options:
17427 Set the scene change detection threshold as a percentage of maximum change. Good
17428 values are in the @code{[8.0, 14.0]} range. The range for @option{threshold} is
17431 Default value is @code{10.}.
17434 Set the flag to pass scene change frames to the next filter. Default value is @code{0}
17435 You can enable it if you want to get snapshot of scene change frames only.
17438 @anchor{selectivecolor}
17439 @section selectivecolor
17441 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
17442 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
17443 by the "purity" of the color (that is, how saturated it already is).
17445 This filter is similar to the Adobe Photoshop Selective Color tool.
17447 The filter accepts the following options:
17450 @item correction_method
17451 Select color correction method.
17453 Available values are:
17456 Specified adjustments are applied "as-is" (added/subtracted to original pixel
17459 Specified adjustments are relative to the original component value.
17461 Default is @code{absolute}.
17463 Adjustments for red pixels (pixels where the red component is the maximum)
17465 Adjustments for yellow pixels (pixels where the blue component is the minimum)
17467 Adjustments for green pixels (pixels where the green component is the maximum)
17469 Adjustments for cyan pixels (pixels where the red component is the minimum)
17471 Adjustments for blue pixels (pixels where the blue component is the maximum)
17473 Adjustments for magenta pixels (pixels where the green component is the minimum)
17475 Adjustments for white pixels (pixels where all components are greater than 128)
17477 Adjustments for all pixels except pure black and pure white
17479 Adjustments for black pixels (pixels where all components are lesser than 128)
17481 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
17484 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
17485 4 space separated floating point adjustment values in the [-1,1] range,
17486 respectively to adjust the amount of cyan, magenta, yellow and black for the
17487 pixels of its range.
17489 @subsection Examples
17493 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
17494 increase magenta by 27% in blue areas:
17496 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
17500 Use a Photoshop selective color preset:
17502 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
17506 @anchor{separatefields}
17507 @section separatefields
17509 The @code{separatefields} takes a frame-based video input and splits
17510 each frame into its components fields, producing a new half height clip
17511 with twice the frame rate and twice the frame count.
17513 This filter use field-dominance information in frame to decide which
17514 of each pair of fields to place first in the output.
17515 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
17517 @section setdar, setsar
17519 The @code{setdar} filter sets the Display Aspect Ratio for the filter
17522 This is done by changing the specified Sample (aka Pixel) Aspect
17523 Ratio, according to the following equation:
17525 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
17528 Keep in mind that the @code{setdar} filter does not modify the pixel
17529 dimensions of the video frame. Also, the display aspect ratio set by
17530 this filter may be changed by later filters in the filterchain,
17531 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
17534 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
17535 the filter output video.
17537 Note that as a consequence of the application of this filter, the
17538 output display aspect ratio will change according to the equation
17541 Keep in mind that the sample aspect ratio set by the @code{setsar}
17542 filter may be changed by later filters in the filterchain, e.g. if
17543 another "setsar" or a "setdar" filter is applied.
17545 It accepts the following parameters:
17548 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
17549 Set the aspect ratio used by the filter.
17551 The parameter can be a floating point number string, an expression, or
17552 a string of the form @var{num}:@var{den}, where @var{num} and
17553 @var{den} are the numerator and denominator of the aspect ratio. If
17554 the parameter is not specified, it is assumed the value "0".
17555 In case the form "@var{num}:@var{den}" is used, the @code{:} character
17559 Set the maximum integer value to use for expressing numerator and
17560 denominator when reducing the expressed aspect ratio to a rational.
17561 Default value is @code{100}.
17565 The parameter @var{sar} is an expression containing
17566 the following constants:
17570 These are approximated values for the mathematical constants e
17571 (Euler's number), pi (Greek pi), and phi (the golden ratio).
17574 The input width and height.
17577 These are the same as @var{w} / @var{h}.
17580 The input sample aspect ratio.
17583 The input display aspect ratio. It is the same as
17584 (@var{w} / @var{h}) * @var{sar}.
17587 Horizontal and vertical chroma subsample values. For example, for the
17588 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17591 @subsection Examples
17596 To change the display aspect ratio to 16:9, specify one of the following:
17603 To change the sample aspect ratio to 10:11, specify:
17609 To set a display aspect ratio of 16:9, and specify a maximum integer value of
17610 1000 in the aspect ratio reduction, use the command:
17612 setdar=ratio=16/9:max=1000
17620 Force field for the output video frame.
17622 The @code{setfield} filter marks the interlace type field for the
17623 output frames. It does not change the input frame, but only sets the
17624 corresponding property, which affects how the frame is treated by
17625 following filters (e.g. @code{fieldorder} or @code{yadif}).
17627 The filter accepts the following options:
17632 Available values are:
17636 Keep the same field property.
17639 Mark the frame as bottom-field-first.
17642 Mark the frame as top-field-first.
17645 Mark the frame as progressive.
17652 Force frame parameter for the output video frame.
17654 The @code{setparams} filter marks interlace and color range for the
17655 output frames. It does not change the input frame, but only sets the
17656 corresponding property, which affects how the frame is treated by
17661 Available values are:
17665 Keep the same field property (default).
17668 Mark the frame as bottom-field-first.
17671 Mark the frame as top-field-first.
17674 Mark the frame as progressive.
17678 Available values are:
17682 Keep the same color range property (default).
17684 @item unspecified, unknown
17685 Mark the frame as unspecified color range.
17687 @item limited, tv, mpeg
17688 Mark the frame as limited range.
17690 @item full, pc, jpeg
17691 Mark the frame as full range.
17694 @item color_primaries
17695 Set the color primaries.
17696 Available values are:
17700 Keep the same color primaries property (default).
17717 Set the color transfer.
17718 Available values are:
17722 Keep the same color trc property (default).
17744 Set the colorspace.
17745 Available values are:
17749 Keep the same colorspace property (default).
17762 @item chroma-derived-nc
17763 @item chroma-derived-c
17770 Show a line containing various information for each input video frame.
17771 The input video is not modified.
17773 This filter supports the following options:
17777 Calculate checksums of each plane. By default enabled.
17780 The shown line contains a sequence of key/value pairs of the form
17781 @var{key}:@var{value}.
17783 The following values are shown in the output:
17787 The (sequential) number of the input frame, starting from 0.
17790 The Presentation TimeStamp of the input frame, expressed as a number of
17791 time base units. The time base unit depends on the filter input pad.
17794 The Presentation TimeStamp of the input frame, expressed as a number of
17798 The position of the frame in the input stream, or -1 if this information is
17799 unavailable and/or meaningless (for example in case of synthetic video).
17802 The pixel format name.
17805 The sample aspect ratio of the input frame, expressed in the form
17806 @var{num}/@var{den}.
17809 The size of the input frame. For the syntax of this option, check the
17810 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17813 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
17814 for bottom field first).
17817 This is 1 if the frame is a key frame, 0 otherwise.
17820 The picture type of the input frame ("I" for an I-frame, "P" for a
17821 P-frame, "B" for a B-frame, or "?" for an unknown type).
17822 Also refer to the documentation of the @code{AVPictureType} enum and of
17823 the @code{av_get_picture_type_char} function defined in
17824 @file{libavutil/avutil.h}.
17827 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
17829 @item plane_checksum
17830 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
17831 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
17834 The mean value of pixels in each plane of the input frame, expressed in the form
17835 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
17838 The standard deviation of pixel values in each plane of the input frame, expressed
17839 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
17843 @section showpalette
17845 Displays the 256 colors palette of each frame. This filter is only relevant for
17846 @var{pal8} pixel format frames.
17848 It accepts the following option:
17852 Set the size of the box used to represent one palette color entry. Default is
17853 @code{30} (for a @code{30x30} pixel box).
17856 @section shuffleframes
17858 Reorder and/or duplicate and/or drop video frames.
17860 It accepts the following parameters:
17864 Set the destination indexes of input frames.
17865 This is space or '|' separated list of indexes that maps input frames to output
17866 frames. Number of indexes also sets maximal value that each index may have.
17867 '-1' index have special meaning and that is to drop frame.
17870 The first frame has the index 0. The default is to keep the input unchanged.
17872 @subsection Examples
17876 Swap second and third frame of every three frames of the input:
17878 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
17882 Swap 10th and 1st frame of every ten frames of the input:
17884 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
17888 @section shufflepixels
17890 Reorder pixels in video frames.
17892 This filter accepts the following options:
17896 Set shuffle direction. Can be forward or inverse direction.
17897 Default direction is forward.
17900 Set shuffle mode. Can be horizontal, vertical or block mode.
17904 Set shuffle block_size. In case of horizontal shuffle mode only width
17905 part of size is used, and in case of vertical shuffle mode only height
17906 part of size is used.
17909 Set random seed used with shuffling pixels. Mainly useful to set to be able
17910 to reverse filtering process to get original input.
17911 For example, to reverse forward shuffle you need to use same parameters
17912 and exact same seed and to set direction to inverse.
17915 @section shuffleplanes
17917 Reorder and/or duplicate video planes.
17919 It accepts the following parameters:
17924 The index of the input plane to be used as the first output plane.
17927 The index of the input plane to be used as the second output plane.
17930 The index of the input plane to be used as the third output plane.
17933 The index of the input plane to be used as the fourth output plane.
17937 The first plane has the index 0. The default is to keep the input unchanged.
17939 @subsection Examples
17943 Swap the second and third planes of the input:
17945 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
17949 @anchor{signalstats}
17950 @section signalstats
17951 Evaluate various visual metrics that assist in determining issues associated
17952 with the digitization of analog video media.
17954 By default the filter will log these metadata values:
17958 Display the minimal Y value contained within the input frame. Expressed in
17962 Display the Y value at the 10% percentile within the input frame. Expressed in
17966 Display the average Y value within the input frame. Expressed in range of
17970 Display the Y value at the 90% percentile within the input frame. Expressed in
17974 Display the maximum Y value contained within the input frame. Expressed in
17978 Display the minimal U value contained within the input frame. Expressed in
17982 Display the U value at the 10% percentile within the input frame. Expressed in
17986 Display the average U value within the input frame. Expressed in range of
17990 Display the U value at the 90% percentile within the input frame. Expressed in
17994 Display the maximum U value contained within the input frame. Expressed in
17998 Display the minimal V value contained within the input frame. Expressed in
18002 Display the V value at the 10% percentile within the input frame. Expressed in
18006 Display the average V value within the input frame. Expressed in range of
18010 Display the V value at the 90% percentile within the input frame. Expressed in
18014 Display the maximum V value contained within the input frame. Expressed in
18018 Display the minimal saturation value contained within the input frame.
18019 Expressed in range of [0-~181.02].
18022 Display the saturation value at the 10% percentile within the input frame.
18023 Expressed in range of [0-~181.02].
18026 Display the average saturation value within the input frame. Expressed in range
18030 Display the saturation value at the 90% percentile within the input frame.
18031 Expressed in range of [0-~181.02].
18034 Display the maximum saturation value contained within the input frame.
18035 Expressed in range of [0-~181.02].
18038 Display the median value for hue within the input frame. Expressed in range of
18042 Display the average value for hue within the input frame. Expressed in range of
18046 Display the average of sample value difference between all values of the Y
18047 plane in the current frame and corresponding values of the previous input frame.
18048 Expressed in range of [0-255].
18051 Display the average of sample value difference between all values of the U
18052 plane in the current frame and corresponding values of the previous input frame.
18053 Expressed in range of [0-255].
18056 Display the average of sample value difference between all values of the V
18057 plane in the current frame and corresponding values of the previous input frame.
18058 Expressed in range of [0-255].
18061 Display bit depth of Y plane in current frame.
18062 Expressed in range of [0-16].
18065 Display bit depth of U plane in current frame.
18066 Expressed in range of [0-16].
18069 Display bit depth of V plane in current frame.
18070 Expressed in range of [0-16].
18073 The filter accepts the following options:
18079 @option{stat} specify an additional form of image analysis.
18080 @option{out} output video with the specified type of pixel highlighted.
18082 Both options accept the following values:
18086 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
18087 unlike the neighboring pixels of the same field. Examples of temporal outliers
18088 include the results of video dropouts, head clogs, or tape tracking issues.
18091 Identify @var{vertical line repetition}. Vertical line repetition includes
18092 similar rows of pixels within a frame. In born-digital video vertical line
18093 repetition is common, but this pattern is uncommon in video digitized from an
18094 analog source. When it occurs in video that results from the digitization of an
18095 analog source it can indicate concealment from a dropout compensator.
18098 Identify pixels that fall outside of legal broadcast range.
18102 Set the highlight color for the @option{out} option. The default color is
18106 @subsection Examples
18110 Output data of various video metrics:
18112 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
18116 Output specific data about the minimum and maximum values of the Y plane per frame:
18118 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
18122 Playback video while highlighting pixels that are outside of broadcast range in red.
18124 ffplay example.mov -vf signalstats="out=brng:color=red"
18128 Playback video with signalstats metadata drawn over the frame.
18130 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
18133 The contents of signalstat_drawtext.txt used in the command are:
18136 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
18137 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
18138 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
18139 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
18147 Calculates the MPEG-7 Video Signature. The filter can handle more than one
18148 input. In this case the matching between the inputs can be calculated additionally.
18149 The filter always passes through the first input. The signature of each stream can
18150 be written into a file.
18152 It accepts the following options:
18156 Enable or disable the matching process.
18158 Available values are:
18162 Disable the calculation of a matching (default).
18164 Calculate the matching for the whole video and output whether the whole video
18165 matches or only parts.
18167 Calculate only until a matching is found or the video ends. Should be faster in
18172 Set the number of inputs. The option value must be a non negative integer.
18173 Default value is 1.
18176 Set the path to which the output is written. If there is more than one input,
18177 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
18178 integer), that will be replaced with the input number. If no filename is
18179 specified, no output will be written. This is the default.
18182 Choose the output format.
18184 Available values are:
18188 Use the specified binary representation (default).
18190 Use the specified xml representation.
18194 Set threshold to detect one word as similar. The option value must be an integer
18195 greater than zero. The default value is 9000.
18198 Set threshold to detect all words as similar. The option value must be an integer
18199 greater than zero. The default value is 60000.
18202 Set threshold to detect frames as similar. The option value must be an integer
18203 greater than zero. The default value is 116.
18206 Set the minimum length of a sequence in frames to recognize it as matching
18207 sequence. The option value must be a non negative integer value.
18208 The default value is 0.
18211 Set the minimum relation, that matching frames to all frames must have.
18212 The option value must be a double value between 0 and 1. The default value is 0.5.
18215 @subsection Examples
18219 To calculate the signature of an input video and store it in signature.bin:
18221 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
18225 To detect whether two videos match and store the signatures in XML format in
18226 signature0.xml and signature1.xml:
18228 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 -
18236 Blur the input video without impacting the outlines.
18238 It accepts the following options:
18241 @item luma_radius, lr
18242 Set the luma radius. The option value must be a float number in
18243 the range [0.1,5.0] that specifies the variance of the gaussian filter
18244 used to blur the image (slower if larger). Default value is 1.0.
18246 @item luma_strength, ls
18247 Set the luma strength. The option value must be a float number
18248 in the range [-1.0,1.0] that configures the blurring. A value included
18249 in [0.0,1.0] will blur the image whereas a value included in
18250 [-1.0,0.0] will sharpen the image. Default value is 1.0.
18252 @item luma_threshold, lt
18253 Set the luma threshold used as a coefficient to determine
18254 whether a pixel should be blurred or not. The option value must be an
18255 integer in the range [-30,30]. A value of 0 will filter all the image,
18256 a value included in [0,30] will filter flat areas and a value included
18257 in [-30,0] will filter edges. Default value is 0.
18259 @item chroma_radius, cr
18260 Set the chroma radius. The option value must be a float number in
18261 the range [0.1,5.0] that specifies the variance of the gaussian filter
18262 used to blur the image (slower if larger). Default value is @option{luma_radius}.
18264 @item chroma_strength, cs
18265 Set the chroma strength. The option value must be a float number
18266 in the range [-1.0,1.0] that configures the blurring. A value included
18267 in [0.0,1.0] will blur the image whereas a value included in
18268 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
18270 @item chroma_threshold, ct
18271 Set the chroma threshold used as a coefficient to determine
18272 whether a pixel should be blurred or not. The option value must be an
18273 integer in the range [-30,30]. A value of 0 will filter all the image,
18274 a value included in [0,30] will filter flat areas and a value included
18275 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
18278 If a chroma option is not explicitly set, the corresponding luma value
18282 Apply sobel operator to input video stream.
18284 The filter accepts the following option:
18288 Set which planes will be processed, unprocessed planes will be copied.
18289 By default value 0xf, all planes will be processed.
18292 Set value which will be multiplied with filtered result.
18295 Set value which will be added to filtered result.
18298 @subsection Commands
18300 This filter supports the all above options as @ref{commands}.
18305 Apply a simple postprocessing filter that compresses and decompresses the image
18306 at several (or - in the case of @option{quality} level @code{6} - all) shifts
18307 and average the results.
18309 The filter accepts the following options:
18313 Set quality. This option defines the number of levels for averaging. It accepts
18314 an integer in the range 0-6. If set to @code{0}, the filter will have no
18315 effect. A value of @code{6} means the higher quality. For each increment of
18316 that value the speed drops by a factor of approximately 2. Default value is
18320 Force a constant quantization parameter. If not set, the filter will use the QP
18321 from the video stream (if available).
18324 Set thresholding mode. Available modes are:
18328 Set hard thresholding (default).
18330 Set soft thresholding (better de-ringing effect, but likely blurrier).
18333 @item use_bframe_qp
18334 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
18335 option may cause flicker since the B-Frames have often larger QP. Default is
18336 @code{0} (not enabled).
18339 @subsection Commands
18341 This filter supports the following commands:
18343 @item quality, level
18344 Set quality level. The value @code{max} can be used to set the maximum level,
18345 currently @code{6}.
18351 Scale the input by applying one of the super-resolution methods based on
18352 convolutional neural networks. Supported models:
18356 Super-Resolution Convolutional Neural Network model (SRCNN).
18357 See @url{https://arxiv.org/abs/1501.00092}.
18360 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
18361 See @url{https://arxiv.org/abs/1609.05158}.
18364 Training scripts as well as scripts for model file (.pb) saving can be found at
18365 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
18366 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
18368 Native model files (.model) can be generated from TensorFlow model
18369 files (.pb) by using tools/python/convert.py
18371 The filter accepts the following options:
18375 Specify which DNN backend to use for model loading and execution. This option accepts
18376 the following values:
18380 Native implementation of DNN loading and execution.
18383 TensorFlow backend. To enable this backend you
18384 need to install the TensorFlow for C library (see
18385 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
18386 @code{--enable-libtensorflow}
18389 Default value is @samp{native}.
18392 Set path to model file specifying network architecture and its parameters.
18393 Note that different backends use different file formats. TensorFlow backend
18394 can load files for both formats, while native backend can load files for only
18398 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
18399 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
18400 input upscaled using bicubic upscaling with proper scale factor.
18403 This feature can also be finished with @ref{dnn_processing} filter.
18407 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
18409 This filter takes in input two input videos, the first input is
18410 considered the "main" source and is passed unchanged to the
18411 output. The second input is used as a "reference" video for computing
18414 Both video inputs must have the same resolution and pixel format for
18415 this filter to work correctly. Also it assumes that both inputs
18416 have the same number of frames, which are compared one by one.
18418 The filter stores the calculated SSIM of each frame.
18420 The description of the accepted parameters follows.
18423 @item stats_file, f
18424 If specified the filter will use the named file to save the SSIM of
18425 each individual frame. When filename equals "-" the data is sent to
18429 The file printed if @var{stats_file} is selected, contains a sequence of
18430 key/value pairs of the form @var{key}:@var{value} for each compared
18433 A description of each shown parameter follows:
18437 sequential number of the input frame, starting from 1
18439 @item Y, U, V, R, G, B
18440 SSIM of the compared frames for the component specified by the suffix.
18443 SSIM of the compared frames for the whole frame.
18446 Same as above but in dB representation.
18449 This filter also supports the @ref{framesync} options.
18451 @subsection Examples
18456 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
18457 [main][ref] ssim="stats_file=stats.log" [out]
18460 On this example the input file being processed is compared with the
18461 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
18462 is stored in @file{stats.log}.
18465 Another example with both psnr and ssim at same time:
18467 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
18471 Another example with different containers:
18473 ffmpeg -i main.mpg -i ref.mkv -lavfi "[0:v]settb=AVTB,setpts=PTS-STARTPTS[main];[1:v]settb=AVTB,setpts=PTS-STARTPTS[ref];[main][ref]ssim" -f null -
18479 Convert between different stereoscopic image formats.
18481 The filters accept the following options:
18485 Set stereoscopic image format of input.
18487 Available values for input image formats are:
18490 side by side parallel (left eye left, right eye right)
18493 side by side crosseye (right eye left, left eye right)
18496 side by side parallel with half width resolution
18497 (left eye left, right eye right)
18500 side by side crosseye with half width resolution
18501 (right eye left, left eye right)
18505 above-below (left eye above, right eye below)
18509 above-below (right eye above, left eye below)
18513 above-below with half height resolution
18514 (left eye above, right eye below)
18518 above-below with half height resolution
18519 (right eye above, left eye below)
18522 alternating frames (left eye first, right eye second)
18525 alternating frames (right eye first, left eye second)
18528 interleaved rows (left eye has top row, right eye starts on next row)
18531 interleaved rows (right eye has top row, left eye starts on next row)
18534 interleaved columns, left eye first
18537 interleaved columns, right eye first
18539 Default value is @samp{sbsl}.
18543 Set stereoscopic image format of output.
18547 side by side parallel (left eye left, right eye right)
18550 side by side crosseye (right eye left, left eye right)
18553 side by side parallel with half width resolution
18554 (left eye left, right eye right)
18557 side by side crosseye with half width resolution
18558 (right eye left, left eye right)
18562 above-below (left eye above, right eye below)
18566 above-below (right eye above, left eye below)
18570 above-below with half height resolution
18571 (left eye above, right eye below)
18575 above-below with half height resolution
18576 (right eye above, left eye below)
18579 alternating frames (left eye first, right eye second)
18582 alternating frames (right eye first, left eye second)
18585 interleaved rows (left eye has top row, right eye starts on next row)
18588 interleaved rows (right eye has top row, left eye starts on next row)
18591 anaglyph red/blue gray
18592 (red filter on left eye, blue filter on right eye)
18595 anaglyph red/green gray
18596 (red filter on left eye, green filter on right eye)
18599 anaglyph red/cyan gray
18600 (red filter on left eye, cyan filter on right eye)
18603 anaglyph red/cyan half colored
18604 (red filter on left eye, cyan filter on right eye)
18607 anaglyph red/cyan color
18608 (red filter on left eye, cyan filter on right eye)
18611 anaglyph red/cyan color optimized with the least squares projection of dubois
18612 (red filter on left eye, cyan filter on right eye)
18615 anaglyph green/magenta gray
18616 (green filter on left eye, magenta filter on right eye)
18619 anaglyph green/magenta half colored
18620 (green filter on left eye, magenta filter on right eye)
18623 anaglyph green/magenta colored
18624 (green filter on left eye, magenta filter on right eye)
18627 anaglyph green/magenta color optimized with the least squares projection of dubois
18628 (green filter on left eye, magenta filter on right eye)
18631 anaglyph yellow/blue gray
18632 (yellow filter on left eye, blue filter on right eye)
18635 anaglyph yellow/blue half colored
18636 (yellow filter on left eye, blue filter on right eye)
18639 anaglyph yellow/blue colored
18640 (yellow filter on left eye, blue filter on right eye)
18643 anaglyph yellow/blue color optimized with the least squares projection of dubois
18644 (yellow filter on left eye, blue filter on right eye)
18647 mono output (left eye only)
18650 mono output (right eye only)
18653 checkerboard, left eye first
18656 checkerboard, right eye first
18659 interleaved columns, left eye first
18662 interleaved columns, right eye first
18668 Default value is @samp{arcd}.
18671 @subsection Examples
18675 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
18681 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
18687 @section streamselect, astreamselect
18688 Select video or audio streams.
18690 The filter accepts the following options:
18694 Set number of inputs. Default is 2.
18697 Set input indexes to remap to outputs.
18700 @subsection Commands
18702 The @code{streamselect} and @code{astreamselect} filter supports the following
18707 Set input indexes to remap to outputs.
18710 @subsection Examples
18714 Select first 5 seconds 1st stream and rest of time 2nd stream:
18716 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
18720 Same as above, but for audio:
18722 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
18729 Draw subtitles on top of input video using the libass library.
18731 To enable compilation of this filter you need to configure FFmpeg with
18732 @code{--enable-libass}. This filter also requires a build with libavcodec and
18733 libavformat to convert the passed subtitles file to ASS (Advanced Substation
18734 Alpha) subtitles format.
18736 The filter accepts the following options:
18740 Set the filename of the subtitle file to read. It must be specified.
18742 @item original_size
18743 Specify the size of the original video, the video for which the ASS file
18744 was composed. For the syntax of this option, check the
18745 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18746 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
18747 correctly scale the fonts if the aspect ratio has been changed.
18750 Set a directory path containing fonts that can be used by the filter.
18751 These fonts will be used in addition to whatever the font provider uses.
18754 Process alpha channel, by default alpha channel is untouched.
18757 Set subtitles input character encoding. @code{subtitles} filter only. Only
18758 useful if not UTF-8.
18760 @item stream_index, si
18761 Set subtitles stream index. @code{subtitles} filter only.
18764 Override default style or script info parameters of the subtitles. It accepts a
18765 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
18768 If the first key is not specified, it is assumed that the first value
18769 specifies the @option{filename}.
18771 For example, to render the file @file{sub.srt} on top of the input
18772 video, use the command:
18777 which is equivalent to:
18779 subtitles=filename=sub.srt
18782 To render the default subtitles stream from file @file{video.mkv}, use:
18784 subtitles=video.mkv
18787 To render the second subtitles stream from that file, use:
18789 subtitles=video.mkv:si=1
18792 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
18793 @code{DejaVu Serif}, use:
18795 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
18798 @section super2xsai
18800 Scale the input by 2x and smooth using the Super2xSaI (Scale and
18801 Interpolate) pixel art scaling algorithm.
18803 Useful for enlarging pixel art images without reducing sharpness.
18807 Swap two rectangular objects in video.
18809 This filter accepts the following options:
18819 Set 1st rect x coordinate.
18822 Set 1st rect y coordinate.
18825 Set 2nd rect x coordinate.
18828 Set 2nd rect y coordinate.
18830 All expressions are evaluated once for each frame.
18833 The all options are expressions containing the following constants:
18838 The input width and height.
18841 same as @var{w} / @var{h}
18844 input sample aspect ratio
18847 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
18850 The number of the input frame, starting from 0.
18853 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
18856 the position in the file of the input frame, NAN if unknown
18863 Blend successive video frames.
18869 Apply telecine process to the video.
18871 This filter accepts the following options:
18880 The default value is @code{top}.
18884 A string of numbers representing the pulldown pattern you wish to apply.
18885 The default value is @code{23}.
18889 Some typical patterns:
18894 24p: 2332 (preferred)
18901 24p: 222222222223 ("Euro pulldown")
18906 @section thistogram
18908 Compute and draw a color distribution histogram for the input video across time.
18910 Unlike @ref{histogram} video filter which only shows histogram of single input frame
18911 at certain time, this filter shows also past histograms of number of frames defined
18912 by @code{width} option.
18914 The computed histogram is a representation of the color component
18915 distribution in an image.
18917 The filter accepts the following options:
18921 Set width of single color component output. Default value is @code{0}.
18922 Value of @code{0} means width will be picked from input video.
18923 This also set number of passed histograms to keep.
18924 Allowed range is [0, 8192].
18926 @item display_mode, d
18928 It accepts the following values:
18931 Per color component graphs are placed below each other.
18934 Per color component graphs are placed side by side.
18937 Presents information identical to that in the @code{parade}, except
18938 that the graphs representing color components are superimposed directly
18941 Default is @code{stack}.
18943 @item levels_mode, m
18944 Set mode. Can be either @code{linear}, or @code{logarithmic}.
18945 Default is @code{linear}.
18947 @item components, c
18948 Set what color components to display.
18949 Default is @code{7}.
18952 Set background opacity. Default is @code{0.9}.
18955 Show envelope. Default is disabled.
18958 Set envelope color. Default is @code{gold}.
18963 Available values for slide is:
18966 Draw new frame when right border is reached.
18969 Replace old columns with new ones.
18972 Scroll from right to left.
18975 Scroll from left to right.
18978 Draw single picture.
18981 Default is @code{replace}.
18986 Apply threshold effect to video stream.
18988 This filter needs four video streams to perform thresholding.
18989 First stream is stream we are filtering.
18990 Second stream is holding threshold values, third stream is holding min values,
18991 and last, fourth stream is holding max values.
18993 The filter accepts the following option:
18997 Set which planes will be processed, unprocessed planes will be copied.
18998 By default value 0xf, all planes will be processed.
19001 For example if first stream pixel's component value is less then threshold value
19002 of pixel component from 2nd threshold stream, third stream value will picked,
19003 otherwise fourth stream pixel component value will be picked.
19005 Using color source filter one can perform various types of thresholding:
19007 @subsection Examples
19011 Binary threshold, using gray color as threshold:
19013 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
19017 Inverted binary threshold, using gray color as threshold:
19019 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
19023 Truncate binary threshold, using gray color as threshold:
19025 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
19029 Threshold to zero, using gray color as threshold:
19031 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
19035 Inverted threshold to zero, using gray color as threshold:
19037 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
19042 Select the most representative frame in a given sequence of consecutive frames.
19044 The filter accepts the following options:
19048 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
19049 will pick one of them, and then handle the next batch of @var{n} frames until
19050 the end. Default is @code{100}.
19053 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
19054 value will result in a higher memory usage, so a high value is not recommended.
19056 @subsection Examples
19060 Extract one picture each 50 frames:
19066 Complete example of a thumbnail creation with @command{ffmpeg}:
19068 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
19075 Tile several successive frames together.
19077 The @ref{untile} filter can do the reverse.
19079 The filter accepts the following options:
19084 Set the grid size (i.e. the number of lines and columns). For the syntax of
19085 this option, check the
19086 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19089 Set the maximum number of frames to render in the given area. It must be less
19090 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
19091 the area will be used.
19094 Set the outer border margin in pixels.
19097 Set the inner border thickness (i.e. the number of pixels between frames). For
19098 more advanced padding options (such as having different values for the edges),
19099 refer to the pad video filter.
19102 Specify the color of the unused area. For the syntax of this option, check the
19103 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
19104 The default value of @var{color} is "black".
19107 Set the number of frames to overlap when tiling several successive frames together.
19108 The value must be between @code{0} and @var{nb_frames - 1}.
19111 Set the number of frames to initially be empty before displaying first output frame.
19112 This controls how soon will one get first output frame.
19113 The value must be between @code{0} and @var{nb_frames - 1}.
19116 @subsection Examples
19120 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
19122 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
19124 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
19125 duplicating each output frame to accommodate the originally detected frame
19129 Display @code{5} pictures in an area of @code{3x2} frames,
19130 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
19131 mixed flat and named options:
19133 tile=3x2:nb_frames=5:padding=7:margin=2
19137 @section tinterlace
19139 Perform various types of temporal field interlacing.
19141 Frames are counted starting from 1, so the first input frame is
19144 The filter accepts the following options:
19149 Specify the mode of the interlacing. This option can also be specified
19150 as a value alone. See below for a list of values for this option.
19152 Available values are:
19156 Move odd frames into the upper field, even into the lower field,
19157 generating a double height frame at half frame rate.
19161 Frame 1 Frame 2 Frame 3 Frame 4
19163 11111 22222 33333 44444
19164 11111 22222 33333 44444
19165 11111 22222 33333 44444
19166 11111 22222 33333 44444
19180 Only output odd frames, even frames are dropped, generating a frame with
19181 unchanged height at half frame rate.
19186 Frame 1 Frame 2 Frame 3 Frame 4
19188 11111 22222 33333 44444
19189 11111 22222 33333 44444
19190 11111 22222 33333 44444
19191 11111 22222 33333 44444
19201 Only output even frames, odd frames are dropped, generating a frame with
19202 unchanged height at half frame rate.
19207 Frame 1 Frame 2 Frame 3 Frame 4
19209 11111 22222 33333 44444
19210 11111 22222 33333 44444
19211 11111 22222 33333 44444
19212 11111 22222 33333 44444
19222 Expand each frame to full height, but pad alternate lines with black,
19223 generating a frame with double height at the same input frame rate.
19228 Frame 1 Frame 2 Frame 3 Frame 4
19230 11111 22222 33333 44444
19231 11111 22222 33333 44444
19232 11111 22222 33333 44444
19233 11111 22222 33333 44444
19236 11111 ..... 33333 .....
19237 ..... 22222 ..... 44444
19238 11111 ..... 33333 .....
19239 ..... 22222 ..... 44444
19240 11111 ..... 33333 .....
19241 ..... 22222 ..... 44444
19242 11111 ..... 33333 .....
19243 ..... 22222 ..... 44444
19247 @item interleave_top, 4
19248 Interleave the upper field from odd frames with the lower field from
19249 even frames, generating a frame with unchanged height at half frame rate.
19254 Frame 1 Frame 2 Frame 3 Frame 4
19256 11111<- 22222 33333<- 44444
19257 11111 22222<- 33333 44444<-
19258 11111<- 22222 33333<- 44444
19259 11111 22222<- 33333 44444<-
19269 @item interleave_bottom, 5
19270 Interleave the lower field from odd frames with the upper field from
19271 even frames, generating a frame with unchanged height at half frame rate.
19276 Frame 1 Frame 2 Frame 3 Frame 4
19278 11111 22222<- 33333 44444<-
19279 11111<- 22222 33333<- 44444
19280 11111 22222<- 33333 44444<-
19281 11111<- 22222 33333<- 44444
19291 @item interlacex2, 6
19292 Double frame rate with unchanged height. Frames are inserted each
19293 containing the second temporal field from the previous input frame and
19294 the first temporal field from the next input frame. This mode relies on
19295 the top_field_first flag. Useful for interlaced video displays with no
19296 field synchronisation.
19301 Frame 1 Frame 2 Frame 3 Frame 4
19303 11111 22222 33333 44444
19304 11111 22222 33333 44444
19305 11111 22222 33333 44444
19306 11111 22222 33333 44444
19309 11111 22222 22222 33333 33333 44444 44444
19310 11111 11111 22222 22222 33333 33333 44444
19311 11111 22222 22222 33333 33333 44444 44444
19312 11111 11111 22222 22222 33333 33333 44444
19317 Move odd frames into the upper field, even into the lower field,
19318 generating a double height frame at same frame rate.
19323 Frame 1 Frame 2 Frame 3 Frame 4
19325 11111 22222 33333 44444
19326 11111 22222 33333 44444
19327 11111 22222 33333 44444
19328 11111 22222 33333 44444
19331 11111 33333 33333 55555
19332 22222 22222 44444 44444
19333 11111 33333 33333 55555
19334 22222 22222 44444 44444
19335 11111 33333 33333 55555
19336 22222 22222 44444 44444
19337 11111 33333 33333 55555
19338 22222 22222 44444 44444
19343 Numeric values are deprecated but are accepted for backward
19344 compatibility reasons.
19346 Default mode is @code{merge}.
19349 Specify flags influencing the filter process.
19351 Available value for @var{flags} is:
19354 @item low_pass_filter, vlpf
19355 Enable linear vertical low-pass filtering in the filter.
19356 Vertical low-pass filtering is required when creating an interlaced
19357 destination from a progressive source which contains high-frequency
19358 vertical detail. Filtering will reduce interlace 'twitter' and Moire
19361 @item complex_filter, cvlpf
19362 Enable complex vertical low-pass filtering.
19363 This will slightly less reduce interlace 'twitter' and Moire
19364 patterning but better retain detail and subjective sharpness impression.
19367 Bypass already interlaced frames, only adjust the frame rate.
19370 Vertical low-pass filtering and bypassing already interlaced frames can only be
19371 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
19376 Pick median pixels from several successive input video frames.
19378 The filter accepts the following options:
19382 Set radius of median filter.
19383 Default is 1. Allowed range is from 1 to 127.
19386 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
19389 Set median percentile. Default value is @code{0.5}.
19390 Default value of @code{0.5} will pick always median values, while @code{0} will pick
19391 minimum values, and @code{1} maximum values.
19394 @subsection Commands
19396 This filter supports all above options as @ref{commands}, excluding option @code{radius}.
19398 @section tmidequalizer
19400 Apply Temporal Midway Video Equalization effect.
19402 Midway Video Equalization adjusts a sequence of video frames to have the same
19403 histograms, while maintaining their dynamics as much as possible. It's
19404 useful for e.g. matching exposures from a video frames sequence.
19406 This filter accepts the following option:
19410 Set filtering radius. Default is @code{5}. Allowed range is from 1 to 127.
19413 Set filtering sigma. Default is @code{0.5}. This controls strength of filtering.
19414 Setting this option to 0 effectively does nothing.
19417 Set which planes to process. Default is @code{15}, which is all available planes.
19422 Mix successive video frames.
19424 A description of the accepted options follows.
19428 The number of successive frames to mix. If unspecified, it defaults to 3.
19431 Specify weight of each input video frame.
19432 Each weight is separated by space. If number of weights is smaller than
19433 number of @var{frames} last specified weight will be used for all remaining
19437 Specify scale, if it is set it will be multiplied with sum
19438 of each weight multiplied with pixel values to give final destination
19439 pixel value. By default @var{scale} is auto scaled to sum of weights.
19442 @subsection Examples
19446 Average 7 successive frames:
19448 tmix=frames=7:weights="1 1 1 1 1 1 1"
19452 Apply simple temporal convolution:
19454 tmix=frames=3:weights="-1 3 -1"
19458 Similar as above but only showing temporal differences:
19460 tmix=frames=3:weights="-1 2 -1":scale=1
19466 Tone map colors from different dynamic ranges.
19468 This filter expects data in single precision floating point, as it needs to
19469 operate on (and can output) out-of-range values. Another filter, such as
19470 @ref{zscale}, is needed to convert the resulting frame to a usable format.
19472 The tonemapping algorithms implemented only work on linear light, so input
19473 data should be linearized beforehand (and possibly correctly tagged).
19476 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
19479 @subsection Options
19480 The filter accepts the following options.
19484 Set the tone map algorithm to use.
19486 Possible values are:
19489 Do not apply any tone map, only desaturate overbright pixels.
19492 Hard-clip any out-of-range values. Use it for perfect color accuracy for
19493 in-range values, while distorting out-of-range values.
19496 Stretch the entire reference gamut to a linear multiple of the display.
19499 Fit a logarithmic transfer between the tone curves.
19502 Preserve overall image brightness with a simple curve, using nonlinear
19503 contrast, which results in flattening details and degrading color accuracy.
19506 Preserve both dark and bright details better than @var{reinhard}, at the cost
19507 of slightly darkening everything. Use it when detail preservation is more
19508 important than color and brightness accuracy.
19511 Smoothly map out-of-range values, while retaining contrast and colors for
19512 in-range material as much as possible. Use it when color accuracy is more
19513 important than detail preservation.
19519 Tune the tone mapping algorithm.
19521 This affects the following algorithms:
19527 Specifies the scale factor to use while stretching.
19531 Specifies the exponent of the function.
19535 Specify an extra linear coefficient to multiply into the signal before clipping.
19539 Specify the local contrast coefficient at the display peak.
19540 Default to 0.5, which means that in-gamut values will be about half as bright
19547 Specify the transition point from linear to mobius transform. Every value
19548 below this point is guaranteed to be mapped 1:1. The higher the value, the
19549 more accurate the result will be, at the cost of losing bright details.
19550 Default to 0.3, which due to the steep initial slope still preserves in-range
19551 colors fairly accurately.
19555 Apply desaturation for highlights that exceed this level of brightness. The
19556 higher the parameter, the more color information will be preserved. This
19557 setting helps prevent unnaturally blown-out colors for super-highlights, by
19558 (smoothly) turning into white instead. This makes images feel more natural,
19559 at the cost of reducing information about out-of-range colors.
19561 The default of 2.0 is somewhat conservative and will mostly just apply to
19562 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
19564 This option works only if the input frame has a supported color tag.
19567 Override signal/nominal/reference peak with this value. Useful when the
19568 embedded peak information in display metadata is not reliable or when tone
19569 mapping from a lower range to a higher range.
19574 Temporarily pad video frames.
19576 The filter accepts the following options:
19580 Specify number of delay frames before input video stream. Default is 0.
19583 Specify number of padding frames after input video stream.
19584 Set to -1 to pad indefinitely. Default is 0.
19587 Set kind of frames added to beginning of stream.
19588 Can be either @var{add} or @var{clone}.
19589 With @var{add} frames of solid-color are added.
19590 With @var{clone} frames are clones of first frame.
19591 Default is @var{add}.
19594 Set kind of frames added to end of stream.
19595 Can be either @var{add} or @var{clone}.
19596 With @var{add} frames of solid-color are added.
19597 With @var{clone} frames are clones of last frame.
19598 Default is @var{add}.
19600 @item start_duration, stop_duration
19601 Specify the duration of the start/stop delay. See
19602 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19603 for the accepted syntax.
19604 These options override @var{start} and @var{stop}. Default is 0.
19607 Specify the color of the padded area. For the syntax of this option,
19608 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
19609 manual,ffmpeg-utils}.
19611 The default value of @var{color} is "black".
19617 Transpose rows with columns in the input video and optionally flip it.
19619 It accepts the following parameters:
19624 Specify the transposition direction.
19626 Can assume the following values:
19628 @item 0, 4, cclock_flip
19629 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
19637 Rotate by 90 degrees clockwise, that is:
19645 Rotate by 90 degrees counterclockwise, that is:
19652 @item 3, 7, clock_flip
19653 Rotate by 90 degrees clockwise and vertically flip, that is:
19661 For values between 4-7, the transposition is only done if the input
19662 video geometry is portrait and not landscape. These values are
19663 deprecated, the @code{passthrough} option should be used instead.
19665 Numerical values are deprecated, and should be dropped in favor of
19666 symbolic constants.
19669 Do not apply the transposition if the input geometry matches the one
19670 specified by the specified value. It accepts the following values:
19673 Always apply transposition.
19675 Preserve portrait geometry (when @var{height} >= @var{width}).
19677 Preserve landscape geometry (when @var{width} >= @var{height}).
19680 Default value is @code{none}.
19683 For example to rotate by 90 degrees clockwise and preserve portrait
19686 transpose=dir=1:passthrough=portrait
19689 The command above can also be specified as:
19691 transpose=1:portrait
19694 @section transpose_npp
19696 Transpose rows with columns in the input video and optionally flip it.
19697 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
19699 It accepts the following parameters:
19704 Specify the transposition direction.
19706 Can assume the following values:
19709 Rotate by 90 degrees counterclockwise and vertically flip. (default)
19712 Rotate by 90 degrees clockwise.
19715 Rotate by 90 degrees counterclockwise.
19718 Rotate by 90 degrees clockwise and vertically flip.
19722 Do not apply the transposition if the input geometry matches the one
19723 specified by the specified value. It accepts the following values:
19726 Always apply transposition. (default)
19728 Preserve portrait geometry (when @var{height} >= @var{width}).
19730 Preserve landscape geometry (when @var{width} >= @var{height}).
19736 Trim the input so that the output contains one continuous subpart of the input.
19738 It accepts the following parameters:
19741 Specify the time of the start of the kept section, i.e. the frame with the
19742 timestamp @var{start} will be the first frame in the output.
19745 Specify the time of the first frame that will be dropped, i.e. the frame
19746 immediately preceding the one with the timestamp @var{end} will be the last
19747 frame in the output.
19750 This is the same as @var{start}, except this option sets the start timestamp
19751 in timebase units instead of seconds.
19754 This is the same as @var{end}, except this option sets the end timestamp
19755 in timebase units instead of seconds.
19758 The maximum duration of the output in seconds.
19761 The number of the first frame that should be passed to the output.
19764 The number of the first frame that should be dropped.
19767 @option{start}, @option{end}, and @option{duration} are expressed as time
19768 duration specifications; see
19769 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19770 for the accepted syntax.
19772 Note that the first two sets of the start/end options and the @option{duration}
19773 option look at the frame timestamp, while the _frame variants simply count the
19774 frames that pass through the filter. Also note that this filter does not modify
19775 the timestamps. If you wish for the output timestamps to start at zero, insert a
19776 setpts filter after the trim filter.
19778 If multiple start or end options are set, this filter tries to be greedy and
19779 keep all the frames that match at least one of the specified constraints. To keep
19780 only the part that matches all the constraints at once, chain multiple trim
19783 The defaults are such that all the input is kept. So it is possible to set e.g.
19784 just the end values to keep everything before the specified time.
19789 Drop everything except the second minute of input:
19791 ffmpeg -i INPUT -vf trim=60:120
19795 Keep only the first second:
19797 ffmpeg -i INPUT -vf trim=duration=1
19802 @section unpremultiply
19803 Apply alpha unpremultiply effect to input video stream using first plane
19804 of second stream as alpha.
19806 Both streams must have same dimensions and same pixel format.
19808 The filter accepts the following option:
19812 Set which planes will be processed, unprocessed planes will be copied.
19813 By default value 0xf, all planes will be processed.
19815 If the format has 1 or 2 components, then luma is bit 0.
19816 If the format has 3 or 4 components:
19817 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
19818 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
19819 If present, the alpha channel is always the last bit.
19822 Do not require 2nd input for processing, instead use alpha plane from input stream.
19828 Sharpen or blur the input video.
19830 It accepts the following parameters:
19833 @item luma_msize_x, lx
19834 Set the luma matrix horizontal size. It must be an odd integer between
19835 3 and 23. The default value is 5.
19837 @item luma_msize_y, ly
19838 Set the luma matrix vertical size. It must be an odd integer between 3
19839 and 23. The default value is 5.
19841 @item luma_amount, la
19842 Set the luma effect strength. It must be a floating point number, reasonable
19843 values lay between -1.5 and 1.5.
19845 Negative values will blur the input video, while positive values will
19846 sharpen it, a value of zero will disable the effect.
19848 Default value is 1.0.
19850 @item chroma_msize_x, cx
19851 Set the chroma matrix horizontal size. It must be an odd integer
19852 between 3 and 23. The default value is 5.
19854 @item chroma_msize_y, cy
19855 Set the chroma matrix vertical size. It must be an odd integer
19856 between 3 and 23. The default value is 5.
19858 @item chroma_amount, ca
19859 Set the chroma effect strength. It must be a floating point number, reasonable
19860 values lay between -1.5 and 1.5.
19862 Negative values will blur the input video, while positive values will
19863 sharpen it, a value of zero will disable the effect.
19865 Default value is 0.0.
19869 All parameters are optional and default to the equivalent of the
19870 string '5:5:1.0:5:5:0.0'.
19872 @subsection Examples
19876 Apply strong luma sharpen effect:
19878 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
19882 Apply a strong blur of both luma and chroma parameters:
19884 unsharp=7:7:-2:7:7:-2
19891 Decompose a video made of tiled images into the individual images.
19893 The frame rate of the output video is the frame rate of the input video
19894 multiplied by the number of tiles.
19896 This filter does the reverse of @ref{tile}.
19898 The filter accepts the following options:
19903 Set the grid size (i.e. the number of lines and columns). For the syntax of
19904 this option, check the
19905 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19908 @subsection Examples
19912 Produce a 1-second video from a still image file made of 25 frames stacked
19913 vertically, like an analogic film reel:
19915 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
19921 Apply ultra slow/simple postprocessing filter that compresses and decompresses
19922 the image at several (or - in the case of @option{quality} level @code{8} - all)
19923 shifts and average the results.
19925 The way this differs from the behavior of spp is that uspp actually encodes &
19926 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
19927 DCT similar to MJPEG.
19929 The filter accepts the following options:
19933 Set quality. This option defines the number of levels for averaging. It accepts
19934 an integer in the range 0-8. If set to @code{0}, the filter will have no
19935 effect. A value of @code{8} means the higher quality. For each increment of
19936 that value the speed drops by a factor of approximately 2. Default value is
19940 Force a constant quantization parameter. If not set, the filter will use the QP
19941 from the video stream (if available).
19946 Convert 360 videos between various formats.
19948 The filter accepts the following options:
19954 Set format of the input/output video.
19962 Equirectangular projection.
19967 Cubemap with 3x2/6x1/1x6 layout.
19969 Format specific options:
19974 Set padding proportion for the input/output cubemap. Values in decimals.
19981 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)
19984 Default value is @b{@samp{0}}.
19985 Maximum value is @b{@samp{0.1}}.
19989 Set fixed padding for the input/output cubemap. Values in pixels.
19991 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
19995 Set order of faces for the input/output cubemap. Choose one direction for each position.
19997 Designation of directions:
20013 Default value is @b{@samp{rludfb}}.
20017 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
20019 Designation of angles:
20022 0 degrees clockwise
20024 90 degrees clockwise
20026 180 degrees clockwise
20028 270 degrees clockwise
20031 Default value is @b{@samp{000000}}.
20035 Equi-Angular Cubemap.
20042 Format specific options:
20047 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20049 If diagonal field of view is set it overrides horizontal and vertical field of view.
20054 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20056 If diagonal field of view is set it overrides horizontal and vertical field of view.
20062 Format specific options:
20067 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20069 If diagonal field of view is set it overrides horizontal and vertical field of view.
20074 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20076 If diagonal field of view is set it overrides horizontal and vertical field of view.
20082 Facebook's 360 formats.
20085 Stereographic format.
20087 Format specific options:
20092 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20094 If diagonal field of view is set it overrides horizontal and vertical field of view.
20099 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20101 If diagonal field of view is set it overrides horizontal and vertical field of view.
20108 Ball format, gives significant distortion toward the back.
20111 Hammer-Aitoff map projection format.
20114 Sinusoidal map projection format.
20117 Fisheye projection.
20119 Format specific options:
20124 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20126 If diagonal field of view is set it overrides horizontal and vertical field of view.
20131 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20133 If diagonal field of view is set it overrides horizontal and vertical field of view.
20137 Pannini projection.
20139 Format specific options:
20142 Set output pannini parameter.
20145 Set input pannini parameter.
20149 Cylindrical projection.
20151 Format specific options:
20156 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20158 If diagonal field of view is set it overrides horizontal and vertical field of view.
20163 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20165 If diagonal field of view is set it overrides horizontal and vertical field of view.
20169 Perspective projection. @i{(output only)}
20171 Format specific options:
20174 Set perspective parameter.
20178 Tetrahedron projection.
20181 Truncated square pyramid projection.
20185 Half equirectangular projection.
20190 Format specific options:
20195 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20197 If diagonal field of view is set it overrides horizontal and vertical field of view.
20202 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20204 If diagonal field of view is set it overrides horizontal and vertical field of view.
20208 Orthographic format.
20210 Format specific options:
20215 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20217 If diagonal field of view is set it overrides horizontal and vertical field of view.
20222 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20224 If diagonal field of view is set it overrides horizontal and vertical field of view.
20228 Octahedron projection.
20232 Set interpolation method.@*
20233 @i{Note: more complex interpolation methods require much more memory to run.}
20243 Bilinear interpolation.
20245 Lagrange9 interpolation.
20248 Bicubic interpolation.
20251 Lanczos interpolation.
20254 Spline16 interpolation.
20257 Gaussian interpolation.
20259 Mitchell interpolation.
20262 Default value is @b{@samp{line}}.
20266 Set the output video resolution.
20268 Default resolution depends on formats.
20272 Set the input/output stereo format.
20283 Default value is @b{@samp{2d}} for input and output format.
20288 Set rotation for the output video. Values in degrees.
20291 Set rotation order for the output video. Choose one item for each position.
20302 Default value is @b{@samp{ypr}}.
20307 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
20311 Set if input video is flipped horizontally/vertically. Boolean values.
20314 Set if input video is transposed. Boolean value, by default disabled.
20317 Set if output video needs to be transposed. Boolean value, by default disabled.
20320 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
20323 @subsection Examples
20327 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
20329 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
20332 Extract back view of Equi-Angular Cubemap:
20334 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
20337 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
20339 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
20343 @subsection Commands
20345 This filter supports subset of above options as @ref{commands}.
20347 @section vaguedenoiser
20349 Apply a wavelet based denoiser.
20351 It transforms each frame from the video input into the wavelet domain,
20352 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
20353 the obtained coefficients. It does an inverse wavelet transform after.
20354 Due to wavelet properties, it should give a nice smoothed result, and
20355 reduced noise, without blurring picture features.
20357 This filter accepts the following options:
20361 The filtering strength. The higher, the more filtered the video will be.
20362 Hard thresholding can use a higher threshold than soft thresholding
20363 before the video looks overfiltered. Default value is 2.
20366 The filtering method the filter will use.
20368 It accepts the following values:
20371 All values under the threshold will be zeroed.
20374 All values under the threshold will be zeroed. All values above will be
20375 reduced by the threshold.
20378 Scales or nullifies coefficients - intermediary between (more) soft and
20379 (less) hard thresholding.
20382 Default is garrote.
20385 Number of times, the wavelet will decompose the picture. Picture can't
20386 be decomposed beyond a particular point (typically, 8 for a 640x480
20387 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
20390 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
20393 A list of the planes to process. By default all planes are processed.
20396 The threshold type the filter will use.
20398 It accepts the following values:
20401 Threshold used is same for all decompositions.
20404 Threshold used depends also on each decomposition coefficients.
20407 Default is universal.
20410 @section vectorscope
20412 Display 2 color component values in the two dimensional graph (which is called
20415 This filter accepts the following options:
20419 Set vectorscope mode.
20421 It accepts the following values:
20425 Gray values are displayed on graph, higher brightness means more pixels have
20426 same component color value on location in graph. This is the default mode.
20429 Gray values are displayed on graph. Surrounding pixels values which are not
20430 present in video frame are drawn in gradient of 2 color components which are
20431 set by option @code{x} and @code{y}. The 3rd color component is static.
20434 Actual color components values present in video frame are displayed on graph.
20437 Similar as color2 but higher frequency of same values @code{x} and @code{y}
20438 on graph increases value of another color component, which is luminance by
20439 default values of @code{x} and @code{y}.
20442 Actual colors present in video frame are displayed on graph. If two different
20443 colors map to same position on graph then color with higher value of component
20444 not present in graph is picked.
20447 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
20448 component picked from radial gradient.
20452 Set which color component will be represented on X-axis. Default is @code{1}.
20455 Set which color component will be represented on Y-axis. Default is @code{2}.
20458 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
20459 of color component which represents frequency of (X, Y) location in graph.
20464 No envelope, this is default.
20467 Instant envelope, even darkest single pixel will be clearly highlighted.
20470 Hold maximum and minimum values presented in graph over time. This way you
20471 can still spot out of range values without constantly looking at vectorscope.
20474 Peak and instant envelope combined together.
20478 Set what kind of graticule to draw.
20487 Set graticule opacity.
20490 Set graticule flags.
20494 Draw graticule for white point.
20497 Draw graticule for black point.
20500 Draw color points short names.
20504 Set background opacity.
20506 @item lthreshold, l
20507 Set low threshold for color component not represented on X or Y axis.
20508 Values lower than this value will be ignored. Default is 0.
20509 Note this value is multiplied with actual max possible value one pixel component
20510 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
20513 @item hthreshold, h
20514 Set high threshold for color component not represented on X or Y axis.
20515 Values higher than this value will be ignored. Default is 1.
20516 Note this value is multiplied with actual max possible value one pixel component
20517 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
20518 is 0.9 * 255 = 230.
20520 @item colorspace, c
20521 Set what kind of colorspace to use when drawing graticule.
20531 Set color tint for gray/tint vectorscope mode. By default both options are zero.
20532 This means no tint, and output will remain gray.
20535 @anchor{vidstabdetect}
20536 @section vidstabdetect
20538 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
20539 @ref{vidstabtransform} for pass 2.
20541 This filter generates a file with relative translation and rotation
20542 transform information about subsequent frames, which is then used by
20543 the @ref{vidstabtransform} filter.
20545 To enable compilation of this filter you need to configure FFmpeg with
20546 @code{--enable-libvidstab}.
20548 This filter accepts the following options:
20552 Set the path to the file used to write the transforms information.
20553 Default value is @file{transforms.trf}.
20556 Set how shaky the video is and how quick the camera is. It accepts an
20557 integer in the range 1-10, a value of 1 means little shakiness, a
20558 value of 10 means strong shakiness. Default value is 5.
20561 Set the accuracy of the detection process. It must be a value in the
20562 range 1-15. A value of 1 means low accuracy, a value of 15 means high
20563 accuracy. Default value is 15.
20566 Set stepsize of the search process. The region around minimum is
20567 scanned with 1 pixel resolution. Default value is 6.
20570 Set minimum contrast. Below this value a local measurement field is
20571 discarded. Must be a floating point value in the range 0-1. Default
20575 Set reference frame number for tripod mode.
20577 If enabled, the motion of the frames is compared to a reference frame
20578 in the filtered stream, identified by the specified number. The idea
20579 is to compensate all movements in a more-or-less static scene and keep
20580 the camera view absolutely still.
20582 If set to 0, it is disabled. The frames are counted starting from 1.
20585 Show fields and transforms in the resulting frames. It accepts an
20586 integer in the range 0-2. Default value is 0, which disables any
20590 @subsection Examples
20594 Use default values:
20600 Analyze strongly shaky movie and put the results in file
20601 @file{mytransforms.trf}:
20603 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
20607 Visualize the result of internal transformations in the resulting
20610 vidstabdetect=show=1
20614 Analyze a video with medium shakiness using @command{ffmpeg}:
20616 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
20620 @anchor{vidstabtransform}
20621 @section vidstabtransform
20623 Video stabilization/deshaking: pass 2 of 2,
20624 see @ref{vidstabdetect} for pass 1.
20626 Read a file with transform information for each frame and
20627 apply/compensate them. Together with the @ref{vidstabdetect}
20628 filter this can be used to deshake videos. See also
20629 @url{http://public.hronopik.de/vid.stab}. It is important to also use
20630 the @ref{unsharp} filter, see below.
20632 To enable compilation of this filter you need to configure FFmpeg with
20633 @code{--enable-libvidstab}.
20635 @subsection Options
20639 Set path to the file used to read the transforms. Default value is
20640 @file{transforms.trf}.
20643 Set the number of frames (value*2 + 1) used for lowpass filtering the
20644 camera movements. Default value is 10.
20646 For example a number of 10 means that 21 frames are used (10 in the
20647 past and 10 in the future) to smoothen the motion in the video. A
20648 larger value leads to a smoother video, but limits the acceleration of
20649 the camera (pan/tilt movements). 0 is a special case where a static
20650 camera is simulated.
20653 Set the camera path optimization algorithm.
20655 Accepted values are:
20658 gaussian kernel low-pass filter on camera motion (default)
20660 averaging on transformations
20664 Set maximal number of pixels to translate frames. Default value is -1,
20668 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
20669 value is -1, meaning no limit.
20672 Specify how to deal with borders that may be visible due to movement
20675 Available values are:
20678 keep image information from previous frame (default)
20680 fill the border black
20684 Invert transforms if set to 1. Default value is 0.
20687 Consider transforms as relative to previous frame if set to 1,
20688 absolute if set to 0. Default value is 0.
20691 Set percentage to zoom. A positive value will result in a zoom-in
20692 effect, a negative value in a zoom-out effect. Default value is 0 (no
20696 Set optimal zooming to avoid borders.
20698 Accepted values are:
20703 optimal static zoom value is determined (only very strong movements
20704 will lead to visible borders) (default)
20706 optimal adaptive zoom value is determined (no borders will be
20707 visible), see @option{zoomspeed}
20710 Note that the value given at zoom is added to the one calculated here.
20713 Set percent to zoom maximally each frame (enabled when
20714 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
20718 Specify type of interpolation.
20720 Available values are:
20725 linear only horizontal
20727 linear in both directions (default)
20729 cubic in both directions (slow)
20733 Enable virtual tripod mode if set to 1, which is equivalent to
20734 @code{relative=0:smoothing=0}. Default value is 0.
20736 Use also @code{tripod} option of @ref{vidstabdetect}.
20739 Increase log verbosity if set to 1. Also the detected global motions
20740 are written to the temporary file @file{global_motions.trf}. Default
20744 @subsection Examples
20748 Use @command{ffmpeg} for a typical stabilization with default values:
20750 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
20753 Note the use of the @ref{unsharp} filter which is always recommended.
20756 Zoom in a bit more and load transform data from a given file:
20758 vidstabtransform=zoom=5:input="mytransforms.trf"
20762 Smoothen the video even more:
20764 vidstabtransform=smoothing=30
20770 Flip the input video vertically.
20772 For example, to vertically flip a video with @command{ffmpeg}:
20774 ffmpeg -i in.avi -vf "vflip" out.avi
20779 Detect variable frame rate video.
20781 This filter tries to detect if the input is variable or constant frame rate.
20783 At end it will output number of frames detected as having variable delta pts,
20784 and ones with constant delta pts.
20785 If there was frames with variable delta, than it will also show min, max and
20786 average delta encountered.
20790 Boost or alter saturation.
20792 The filter accepts the following options:
20795 Set strength of boost if positive value or strength of alter if negative value.
20796 Default is 0. Allowed range is from -2 to 2.
20799 Set the red balance. Default is 1. Allowed range is from -10 to 10.
20802 Set the green balance. Default is 1. Allowed range is from -10 to 10.
20805 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
20808 Set the red luma coefficient.
20811 Set the green luma coefficient.
20814 Set the blue luma coefficient.
20817 If @code{intensity} is negative and this is set to 1, colors will change,
20818 otherwise colors will be less saturated, more towards gray.
20821 @subsection Commands
20823 This filter supports the all above options as @ref{commands}.
20828 Make or reverse a natural vignetting effect.
20830 The filter accepts the following options:
20834 Set lens angle expression as a number of radians.
20836 The value is clipped in the @code{[0,PI/2]} range.
20838 Default value: @code{"PI/5"}
20842 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
20846 Set forward/backward mode.
20848 Available modes are:
20851 The larger the distance from the central point, the darker the image becomes.
20854 The larger the distance from the central point, the brighter the image becomes.
20855 This can be used to reverse a vignette effect, though there is no automatic
20856 detection to extract the lens @option{angle} and other settings (yet). It can
20857 also be used to create a burning effect.
20860 Default value is @samp{forward}.
20863 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
20865 It accepts the following values:
20868 Evaluate expressions only once during the filter initialization.
20871 Evaluate expressions for each incoming frame. This is way slower than the
20872 @samp{init} mode since it requires all the scalers to be re-computed, but it
20873 allows advanced dynamic expressions.
20876 Default value is @samp{init}.
20879 Set dithering to reduce the circular banding effects. Default is @code{1}
20883 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
20884 Setting this value to the SAR of the input will make a rectangular vignetting
20885 following the dimensions of the video.
20887 Default is @code{1/1}.
20890 @subsection Expressions
20892 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
20893 following parameters.
20898 input width and height
20901 the number of input frame, starting from 0
20904 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
20905 @var{TB} units, NAN if undefined
20908 frame rate of the input video, NAN if the input frame rate is unknown
20911 the PTS (Presentation TimeStamp) of the filtered video frame,
20912 expressed in seconds, NAN if undefined
20915 time base of the input video
20919 @subsection Examples
20923 Apply simple strong vignetting effect:
20929 Make a flickering vignetting:
20931 vignette='PI/4+random(1)*PI/50':eval=frame
20936 @section vmafmotion
20938 Obtain the average VMAF motion score of a video.
20939 It is one of the component metrics of VMAF.
20941 The obtained average motion score is printed through the logging system.
20943 The filter accepts the following options:
20947 If specified, the filter will use the named file to save the motion score of
20948 each frame with respect to the previous frame.
20949 When filename equals "-" the data is sent to standard output.
20954 ffmpeg -i ref.mpg -vf vmafmotion -f null -
20958 Stack input videos vertically.
20960 All streams must be of same pixel format and of same width.
20962 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
20963 to create same output.
20965 The filter accepts the following options:
20969 Set number of input streams. Default is 2.
20972 If set to 1, force the output to terminate when the shortest input
20973 terminates. Default value is 0.
20978 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
20979 Deinterlacing Filter").
20981 Based on the process described by Martin Weston for BBC R&D, and
20982 implemented based on the de-interlace algorithm written by Jim
20983 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
20984 uses filter coefficients calculated by BBC R&D.
20986 This filter uses field-dominance information in frame to decide which
20987 of each pair of fields to place first in the output.
20988 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
20990 There are two sets of filter coefficients, so called "simple"
20991 and "complex". Which set of filter coefficients is used can
20992 be set by passing an optional parameter:
20996 Set the interlacing filter coefficients. Accepts one of the following values:
21000 Simple filter coefficient set.
21002 More-complex filter coefficient set.
21004 Default value is @samp{complex}.
21007 The interlacing mode to adopt. It accepts one of the following values:
21011 Output one frame for each frame.
21013 Output one frame for each field.
21016 The default value is @code{field}.
21019 The picture field parity assumed for the input interlaced video. It accepts one
21020 of the following values:
21024 Assume the top field is first.
21026 Assume the bottom field is first.
21028 Enable automatic detection of field parity.
21031 The default value is @code{auto}.
21032 If the interlacing is unknown or the decoder does not export this information,
21033 top field first will be assumed.
21036 Specify which frames to deinterlace. Accepts one of the following values:
21040 Deinterlace all frames,
21042 Only deinterlace frames marked as interlaced.
21045 Default value is @samp{all}.
21048 @subsection Commands
21049 This filter supports same @ref{commands} as options.
21052 Video waveform monitor.
21054 The waveform monitor plots color component intensity. By default luminance
21055 only. Each column of the waveform corresponds to a column of pixels in the
21058 It accepts the following options:
21062 Can be either @code{row}, or @code{column}. Default is @code{column}.
21063 In row mode, the graph on the left side represents color component value 0 and
21064 the right side represents value = 255. In column mode, the top side represents
21065 color component value = 0 and bottom side represents value = 255.
21068 Set intensity. Smaller values are useful to find out how many values of the same
21069 luminance are distributed across input rows/columns.
21070 Default value is @code{0.04}. Allowed range is [0, 1].
21073 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
21074 In mirrored mode, higher values will be represented on the left
21075 side for @code{row} mode and at the top for @code{column} mode. Default is
21076 @code{1} (mirrored).
21080 It accepts the following values:
21083 Presents information identical to that in the @code{parade}, except
21084 that the graphs representing color components are superimposed directly
21087 This display mode makes it easier to spot relative differences or similarities
21088 in overlapping areas of the color components that are supposed to be identical,
21089 such as neutral whites, grays, or blacks.
21092 Display separate graph for the color components side by side in
21093 @code{row} mode or one below the other in @code{column} mode.
21096 Display separate graph for the color components side by side in
21097 @code{column} mode or one below the other in @code{row} mode.
21099 Using this display mode makes it easy to spot color casts in the highlights
21100 and shadows of an image, by comparing the contours of the top and the bottom
21101 graphs of each waveform. Since whites, grays, and blacks are characterized
21102 by exactly equal amounts of red, green, and blue, neutral areas of the picture
21103 should display three waveforms of roughly equal width/height. If not, the
21104 correction is easy to perform by making level adjustments the three waveforms.
21106 Default is @code{stack}.
21108 @item components, c
21109 Set which color components to display. Default is 1, which means only luminance
21110 or red color component if input is in RGB colorspace. If is set for example to
21111 7 it will display all 3 (if) available color components.
21116 No envelope, this is default.
21119 Instant envelope, minimum and maximum values presented in graph will be easily
21120 visible even with small @code{step} value.
21123 Hold minimum and maximum values presented in graph across time. This way you
21124 can still spot out of range values without constantly looking at waveforms.
21127 Peak and instant envelope combined together.
21133 No filtering, this is default.
21136 Luma and chroma combined together.
21139 Similar as above, but shows difference between blue and red chroma.
21142 Similar as above, but use different colors.
21145 Similar as above, but again with different colors.
21148 Displays only chroma.
21151 Displays actual color value on waveform.
21154 Similar as above, but with luma showing frequency of chroma values.
21158 Set which graticule to display.
21162 Do not display graticule.
21165 Display green graticule showing legal broadcast ranges.
21168 Display orange graticule showing legal broadcast ranges.
21171 Display invert graticule showing legal broadcast ranges.
21175 Set graticule opacity.
21178 Set graticule flags.
21182 Draw numbers above lines. By default enabled.
21185 Draw dots instead of lines.
21189 Set scale used for displaying graticule.
21196 Default is digital.
21199 Set background opacity.
21203 Set tint for output.
21204 Only used with lowpass filter and when display is not overlay and input
21205 pixel formats are not RGB.
21208 @section weave, doubleweave
21210 The @code{weave} takes a field-based video input and join
21211 each two sequential fields into single frame, producing a new double
21212 height clip with half the frame rate and half the frame count.
21214 The @code{doubleweave} works same as @code{weave} but without
21215 halving frame rate and frame count.
21217 It accepts the following option:
21221 Set first field. Available values are:
21225 Set the frame as top-field-first.
21228 Set the frame as bottom-field-first.
21232 @subsection Examples
21236 Interlace video using @ref{select} and @ref{separatefields} filter:
21238 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
21243 Apply the xBR high-quality magnification filter which is designed for pixel
21244 art. It follows a set of edge-detection rules, see
21245 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
21247 It accepts the following option:
21251 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
21252 @code{3xBR} and @code{4} for @code{4xBR}.
21253 Default is @code{3}.
21258 Apply cross fade from one input video stream to another input video stream.
21259 The cross fade is applied for specified duration.
21261 The filter accepts the following options:
21265 Set one of available transition effects:
21313 Default transition effect is fade.
21316 Set cross fade duration in seconds.
21317 Default duration is 1 second.
21320 Set cross fade start relative to first input stream in seconds.
21321 Default offset is 0.
21324 Set expression for custom transition effect.
21326 The expressions can use the following variables and functions:
21331 The coordinates of the current sample.
21335 The width and height of the image.
21338 Progress of transition effect.
21341 Currently processed plane.
21344 Return value of first input at current location and plane.
21347 Return value of second input at current location and plane.
21353 Return the value of the pixel at location (@var{x},@var{y}) of the
21354 first/second/third/fourth component of first input.
21360 Return the value of the pixel at location (@var{x},@var{y}) of the
21361 first/second/third/fourth component of second input.
21365 @subsection Examples
21369 Cross fade from one input video to another input video, with fade transition and duration of transition
21370 of 2 seconds starting at offset of 5 seconds:
21372 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
21377 Pick median pixels from several input videos.
21379 The filter accepts the following options:
21383 Set number of inputs.
21384 Default is 3. Allowed range is from 3 to 255.
21385 If number of inputs is even number, than result will be mean value between two median values.
21388 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
21391 Set median percentile. Default value is @code{0.5}.
21392 Default value of @code{0.5} will pick always median values, while @code{0} will pick
21393 minimum values, and @code{1} maximum values.
21396 @subsection Commands
21398 This filter supports all above options as @ref{commands}, excluding option @code{inputs}.
21401 Stack video inputs into custom layout.
21403 All streams must be of same pixel format.
21405 The filter accepts the following options:
21409 Set number of input streams. Default is 2.
21412 Specify layout of inputs.
21413 This option requires the desired layout configuration to be explicitly set by the user.
21414 This sets position of each video input in output. Each input
21415 is separated by '|'.
21416 The first number represents the column, and the second number represents the row.
21417 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
21418 where X is video input from which to take width or height.
21419 Multiple values can be used when separated by '+'. In such
21420 case values are summed together.
21422 Note that if inputs are of different sizes gaps may appear, as not all of
21423 the output video frame will be filled. Similarly, videos can overlap each
21424 other if their position doesn't leave enough space for the full frame of
21427 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
21428 a layout must be set by the user.
21431 If set to 1, force the output to terminate when the shortest input
21432 terminates. Default value is 0.
21435 If set to valid color, all unused pixels will be filled with that color.
21436 By default fill is set to none, so it is disabled.
21439 @subsection Examples
21443 Display 4 inputs into 2x2 grid.
21447 input1(0, 0) | input3(w0, 0)
21448 input2(0, h0) | input4(w0, h0)
21452 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
21455 Note that if inputs are of different sizes, gaps or overlaps may occur.
21458 Display 4 inputs into 1x4 grid.
21465 input4(0, h0+h1+h2)
21469 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
21472 Note that if inputs are of different widths, unused space will appear.
21475 Display 9 inputs into 3x3 grid.
21479 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
21480 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
21481 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
21485 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
21488 Note that if inputs are of different sizes, gaps or overlaps may occur.
21491 Display 16 inputs into 4x4 grid.
21495 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
21496 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
21497 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
21498 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
21502 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|
21503 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
21506 Note that if inputs are of different sizes, gaps or overlaps may occur.
21513 Deinterlace the input video ("yadif" means "yet another deinterlacing
21516 It accepts the following parameters:
21522 The interlacing mode to adopt. It accepts one of the following values:
21525 @item 0, send_frame
21526 Output one frame for each frame.
21527 @item 1, send_field
21528 Output one frame for each field.
21529 @item 2, send_frame_nospatial
21530 Like @code{send_frame}, but it skips the spatial interlacing check.
21531 @item 3, send_field_nospatial
21532 Like @code{send_field}, but it skips the spatial interlacing check.
21535 The default value is @code{send_frame}.
21538 The picture field parity assumed for the input interlaced video. It accepts one
21539 of the following values:
21543 Assume the top field is first.
21545 Assume the bottom field is first.
21547 Enable automatic detection of field parity.
21550 The default value is @code{auto}.
21551 If the interlacing is unknown or the decoder does not export this information,
21552 top field first will be assumed.
21555 Specify which frames to deinterlace. Accepts one of the following
21560 Deinterlace all frames.
21561 @item 1, interlaced
21562 Only deinterlace frames marked as interlaced.
21565 The default value is @code{all}.
21568 @section yadif_cuda
21570 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
21571 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
21574 It accepts the following parameters:
21580 The interlacing mode to adopt. It accepts one of the following values:
21583 @item 0, send_frame
21584 Output one frame for each frame.
21585 @item 1, send_field
21586 Output one frame for each field.
21587 @item 2, send_frame_nospatial
21588 Like @code{send_frame}, but it skips the spatial interlacing check.
21589 @item 3, send_field_nospatial
21590 Like @code{send_field}, but it skips the spatial interlacing check.
21593 The default value is @code{send_frame}.
21596 The picture field parity assumed for the input interlaced video. It accepts one
21597 of the following values:
21601 Assume the top field is first.
21603 Assume the bottom field is first.
21605 Enable automatic detection of field parity.
21608 The default value is @code{auto}.
21609 If the interlacing is unknown or the decoder does not export this information,
21610 top field first will be assumed.
21613 Specify which frames to deinterlace. Accepts one of the following
21618 Deinterlace all frames.
21619 @item 1, interlaced
21620 Only deinterlace frames marked as interlaced.
21623 The default value is @code{all}.
21628 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
21629 The algorithm is described in
21630 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
21632 It accepts the following parameters:
21636 Set the window radius. Default value is 3.
21639 Set which planes to filter. Default is only the first plane.
21642 Set blur strength. Default value is 128.
21645 @subsection Commands
21646 This filter supports same @ref{commands} as options.
21650 Apply Zoom & Pan effect.
21652 This filter accepts the following options:
21656 Set the zoom expression. Range is 1-10. Default is 1.
21660 Set the x and y expression. Default is 0.
21663 Set the duration expression in number of frames.
21664 This sets for how many number of frames effect will last for
21665 single input image.
21668 Set the output image size, default is 'hd720'.
21671 Set the output frame rate, default is '25'.
21674 Each expression can contain the following constants:
21693 Output frame count.
21696 The input timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
21698 @item out_time, time, ot
21699 The output timestamp expressed in seconds.
21703 Last calculated 'x' and 'y' position from 'x' and 'y' expression
21704 for current input frame.
21708 'x' and 'y' of last output frame of previous input frame or 0 when there was
21709 not yet such frame (first input frame).
21712 Last calculated zoom from 'z' expression for current input frame.
21715 Last calculated zoom of last output frame of previous input frame.
21718 Number of output frames for current input frame. Calculated from 'd' expression
21719 for each input frame.
21722 number of output frames created for previous input frame
21725 Rational number: input width / input height
21728 sample aspect ratio
21731 display aspect ratio
21735 @subsection Examples
21739 Zoom in up to 1.5x and pan at same time to some spot near center of picture:
21741 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
21745 Zoom in up to 1.5x and pan always at center of picture:
21747 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21751 Same as above but without pausing:
21753 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21757 Zoom in 2x into center of picture only for the first second of the input video:
21759 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21766 Scale (resize) the input video, using the z.lib library:
21767 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
21768 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
21770 The zscale filter forces the output display aspect ratio to be the same
21771 as the input, by changing the output sample aspect ratio.
21773 If the input image format is different from the format requested by
21774 the next filter, the zscale filter will convert the input to the
21777 @subsection Options
21778 The filter accepts the following options.
21783 Set the output video dimension expression. Default value is the input
21786 If the @var{width} or @var{w} value is 0, the input width is used for
21787 the output. If the @var{height} or @var{h} value is 0, the input height
21788 is used for the output.
21790 If one and only one of the values is -n with n >= 1, the zscale filter
21791 will use a value that maintains the aspect ratio of the input image,
21792 calculated from the other specified dimension. After that it will,
21793 however, make sure that the calculated dimension is divisible by n and
21794 adjust the value if necessary.
21796 If both values are -n with n >= 1, the behavior will be identical to
21797 both values being set to 0 as previously detailed.
21799 See below for the list of accepted constants for use in the dimension
21803 Set the video size. For the syntax of this option, check the
21804 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21807 Set the dither type.
21809 Possible values are:
21814 @item error_diffusion
21820 Set the resize filter type.
21822 Possible values are:
21832 Default is bilinear.
21835 Set the color range.
21837 Possible values are:
21844 Default is same as input.
21847 Set the color primaries.
21849 Possible values are:
21859 Default is same as input.
21862 Set the transfer characteristics.
21864 Possible values are:
21878 Default is same as input.
21881 Set the colorspace matrix.
21883 Possible value are:
21894 Default is same as input.
21897 Set the input color range.
21899 Possible values are:
21906 Default is same as input.
21908 @item primariesin, pin
21909 Set the input color primaries.
21911 Possible values are:
21921 Default is same as input.
21923 @item transferin, tin
21924 Set the input transfer characteristics.
21926 Possible values are:
21937 Default is same as input.
21939 @item matrixin, min
21940 Set the input colorspace matrix.
21942 Possible value are:
21954 Set the output chroma location.
21956 Possible values are:
21967 @item chromalin, cin
21968 Set the input chroma location.
21970 Possible values are:
21982 Set the nominal peak luminance.
21985 The values of the @option{w} and @option{h} options are expressions
21986 containing the following constants:
21991 The input width and height
21995 These are the same as @var{in_w} and @var{in_h}.
21999 The output (scaled) width and height
22003 These are the same as @var{out_w} and @var{out_h}
22006 The same as @var{iw} / @var{ih}
22009 input sample aspect ratio
22012 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
22016 horizontal and vertical input chroma subsample values. For example for the
22017 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
22021 horizontal and vertical output chroma subsample values. For example for the
22022 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
22025 @subsection Commands
22027 This filter supports the following commands:
22031 Set the output video dimension expression.
22032 The command accepts the same syntax of the corresponding option.
22034 If the specified expression is not valid, it is kept at its current
22038 @c man end VIDEO FILTERS
22040 @chapter OpenCL Video Filters
22041 @c man begin OPENCL VIDEO FILTERS
22043 Below is a description of the currently available OpenCL video filters.
22045 To enable compilation of these filters you need to configure FFmpeg with
22046 @code{--enable-opencl}.
22048 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
22051 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
22052 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
22053 given device parameters.
22055 @item -filter_hw_device @var{name}
22056 Pass the hardware device called @var{name} to all filters in any filter graph.
22060 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
22064 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
22066 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
22070 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.
22072 @section avgblur_opencl
22074 Apply average blur filter.
22076 The filter accepts the following options:
22080 Set horizontal radius size.
22081 Range is @code{[1, 1024]} and default value is @code{1}.
22084 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22087 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
22090 @subsection Example
22094 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.
22096 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
22100 @section boxblur_opencl
22102 Apply a boxblur algorithm to the input video.
22104 It accepts the following parameters:
22108 @item luma_radius, lr
22109 @item luma_power, lp
22110 @item chroma_radius, cr
22111 @item chroma_power, cp
22112 @item alpha_radius, ar
22113 @item alpha_power, ap
22117 A description of the accepted options follows.
22120 @item luma_radius, lr
22121 @item chroma_radius, cr
22122 @item alpha_radius, ar
22123 Set an expression for the box radius in pixels used for blurring the
22124 corresponding input plane.
22126 The radius value must be a non-negative number, and must not be
22127 greater than the value of the expression @code{min(w,h)/2} for the
22128 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
22131 Default value for @option{luma_radius} is "2". If not specified,
22132 @option{chroma_radius} and @option{alpha_radius} default to the
22133 corresponding value set for @option{luma_radius}.
22135 The expressions can contain the following constants:
22139 The input width and height in pixels.
22143 The input chroma image width and height in pixels.
22147 The horizontal and vertical chroma subsample values. For example, for the
22148 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
22151 @item luma_power, lp
22152 @item chroma_power, cp
22153 @item alpha_power, ap
22154 Specify how many times the boxblur filter is applied to the
22155 corresponding plane.
22157 Default value for @option{luma_power} is 2. If not specified,
22158 @option{chroma_power} and @option{alpha_power} default to the
22159 corresponding value set for @option{luma_power}.
22161 A value of 0 will disable the effect.
22164 @subsection Examples
22166 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.
22170 Apply a boxblur filter with the luma, chroma, and alpha radius
22171 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.
22173 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
22174 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
22178 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.
22180 For the luma plane, a 2x2 box radius will be run once.
22182 For the chroma plane, a 4x4 box radius will be run 5 times.
22184 For the alpha plane, a 3x3 box radius will be run 7 times.
22186 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
22190 @section colorkey_opencl
22191 RGB colorspace color keying.
22193 The filter accepts the following options:
22197 The color which will be replaced with transparency.
22200 Similarity percentage with the key color.
22202 0.01 matches only the exact key color, while 1.0 matches everything.
22207 0.0 makes pixels either fully transparent, or not transparent at all.
22209 Higher values result in semi-transparent pixels, with a higher transparency
22210 the more similar the pixels color is to the key color.
22213 @subsection Examples
22217 Make every semi-green pixel in the input transparent with some slight blending:
22219 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
22223 @section convolution_opencl
22225 Apply convolution of 3x3, 5x5, 7x7 matrix.
22227 The filter accepts the following options:
22234 Set matrix for each plane.
22235 Matrix is sequence of 9, 25 or 49 signed numbers.
22236 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
22242 Set multiplier for calculated value for each plane.
22243 If unset or 0, it will be sum of all matrix elements.
22244 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
22250 Set bias for each plane. This value is added to the result of the multiplication.
22251 Useful for making the overall image brighter or darker.
22252 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
22256 @subsection Examples
22262 -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
22268 -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
22272 Apply edge enhance:
22274 -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
22280 -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
22284 Apply laplacian edge detector which includes diagonals:
22286 -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
22292 -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
22296 @section erosion_opencl
22298 Apply erosion effect to the video.
22300 This filter replaces the pixel by the local(3x3) minimum.
22302 It accepts the following options:
22309 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22310 If @code{0}, plane will remain unchanged.
22313 Flag which specifies the pixel to refer to.
22314 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22316 Flags to local 3x3 coordinates region centered on @code{x}:
22325 @subsection Example
22329 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.
22331 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22335 @section deshake_opencl
22336 Feature-point based video stabilization filter.
22338 The filter accepts the following options:
22342 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
22345 Whether or not additional debug info should be displayed, both in the processed output and in the console.
22347 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
22349 Viewing point matches in the output video is only supported for RGB input.
22351 Defaults to @code{0}.
22353 @item adaptive_crop
22354 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
22356 Defaults to @code{1}.
22358 @item refine_features
22359 Whether or not feature points should be refined at a sub-pixel level.
22361 This can be turned off for a slight performance gain at the cost of precision.
22363 Defaults to @code{1}.
22365 @item smooth_strength
22366 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
22368 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
22370 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
22372 Defaults to @code{0.0}.
22374 @item smooth_window_multiplier
22375 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
22377 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
22379 Acceptable values range from @code{0.1} to @code{10.0}.
22381 Larger values increase the amount of motion data available for determining how to smooth the camera path,
22382 potentially improving smoothness, but also increase latency and memory usage.
22384 Defaults to @code{2.0}.
22388 @subsection Examples
22392 Stabilize a video with a fixed, medium smoothing strength:
22394 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
22398 Stabilize a video with debugging (both in console and in rendered video):
22400 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
22404 @section dilation_opencl
22406 Apply dilation effect to the video.
22408 This filter replaces the pixel by the local(3x3) maximum.
22410 It accepts the following options:
22417 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22418 If @code{0}, plane will remain unchanged.
22421 Flag which specifies the pixel to refer to.
22422 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22424 Flags to local 3x3 coordinates region centered on @code{x}:
22433 @subsection Example
22437 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.
22439 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22443 @section nlmeans_opencl
22445 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
22447 @section overlay_opencl
22449 Overlay one video on top of another.
22451 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
22452 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
22454 The filter accepts the following options:
22459 Set the x coordinate of the overlaid video on the main video.
22460 Default value is @code{0}.
22463 Set the y coordinate of the overlaid video on the main video.
22464 Default value is @code{0}.
22468 @subsection Examples
22472 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
22474 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22477 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
22479 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22484 @section pad_opencl
22486 Add paddings to the input image, and place the original input at the
22487 provided @var{x}, @var{y} coordinates.
22489 It accepts the following options:
22494 Specify an expression for the size of the output image with the
22495 paddings added. If the value for @var{width} or @var{height} is 0, the
22496 corresponding input size is used for the output.
22498 The @var{width} expression can reference the value set by the
22499 @var{height} expression, and vice versa.
22501 The default value of @var{width} and @var{height} is 0.
22505 Specify the offsets to place the input image at within the padded area,
22506 with respect to the top/left border of the output image.
22508 The @var{x} expression can reference the value set by the @var{y}
22509 expression, and vice versa.
22511 The default value of @var{x} and @var{y} is 0.
22513 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
22514 so the input image is centered on the padded area.
22517 Specify the color of the padded area. For the syntax of this option,
22518 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
22519 manual,ffmpeg-utils}.
22522 Pad to an aspect instead to a resolution.
22525 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
22526 options are expressions containing the following constants:
22531 The input video width and height.
22535 These are the same as @var{in_w} and @var{in_h}.
22539 The output width and height (the size of the padded area), as
22540 specified by the @var{width} and @var{height} expressions.
22544 These are the same as @var{out_w} and @var{out_h}.
22548 The x and y offsets as specified by the @var{x} and @var{y}
22549 expressions, or NAN if not yet specified.
22552 same as @var{iw} / @var{ih}
22555 input sample aspect ratio
22558 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
22561 @section prewitt_opencl
22563 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
22565 The filter accepts the following option:
22569 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22572 Set value which will be multiplied with filtered result.
22573 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22576 Set value which will be added to filtered result.
22577 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22580 @subsection Example
22584 Apply the Prewitt operator with scale set to 2 and delta set to 10.
22586 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
22590 @anchor{program_opencl}
22591 @section program_opencl
22593 Filter video using an OpenCL program.
22598 OpenCL program source file.
22601 Kernel name in program.
22604 Number of inputs to the filter. Defaults to 1.
22607 Size of output frames. Defaults to the same as the first input.
22611 The @code{program_opencl} filter also supports the @ref{framesync} options.
22613 The program source file must contain a kernel function with the given name,
22614 which will be run once for each plane of the output. Each run on a plane
22615 gets enqueued as a separate 2D global NDRange with one work-item for each
22616 pixel to be generated. The global ID offset for each work-item is therefore
22617 the coordinates of a pixel in the destination image.
22619 The kernel function needs to take the following arguments:
22622 Destination image, @var{__write_only image2d_t}.
22624 This image will become the output; the kernel should write all of it.
22626 Frame index, @var{unsigned int}.
22628 This is a counter starting from zero and increasing by one for each frame.
22630 Source images, @var{__read_only image2d_t}.
22632 These are the most recent images on each input. The kernel may read from
22633 them to generate the output, but they can't be written to.
22640 Copy the input to the output (output must be the same size as the input).
22642 __kernel void copy(__write_only image2d_t destination,
22643 unsigned int index,
22644 __read_only image2d_t source)
22646 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
22648 int2 location = (int2)(get_global_id(0), get_global_id(1));
22650 float4 value = read_imagef(source, sampler, location);
22652 write_imagef(destination, location, value);
22657 Apply a simple transformation, rotating the input by an amount increasing
22658 with the index counter. Pixel values are linearly interpolated by the
22659 sampler, and the output need not have the same dimensions as the input.
22661 __kernel void rotate_image(__write_only image2d_t dst,
22662 unsigned int index,
22663 __read_only image2d_t src)
22665 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22666 CLK_FILTER_LINEAR);
22668 float angle = (float)index / 100.0f;
22670 float2 dst_dim = convert_float2(get_image_dim(dst));
22671 float2 src_dim = convert_float2(get_image_dim(src));
22673 float2 dst_cen = dst_dim / 2.0f;
22674 float2 src_cen = src_dim / 2.0f;
22676 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22678 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
22680 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
22681 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
22683 src_pos = src_pos * src_dim / dst_dim;
22685 float2 src_loc = src_pos + src_cen;
22687 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
22688 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
22689 write_imagef(dst, dst_loc, 0.5f);
22691 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
22696 Blend two inputs together, with the amount of each input used varying
22697 with the index counter.
22699 __kernel void blend_images(__write_only image2d_t dst,
22700 unsigned int index,
22701 __read_only image2d_t src1,
22702 __read_only image2d_t src2)
22704 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22705 CLK_FILTER_LINEAR);
22707 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
22709 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22710 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
22711 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
22713 float4 val1 = read_imagef(src1, sampler, src1_loc);
22714 float4 val2 = read_imagef(src2, sampler, src2_loc);
22716 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
22722 @section roberts_opencl
22723 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
22725 The filter accepts the following option:
22729 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22732 Set value which will be multiplied with filtered result.
22733 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22736 Set value which will be added to filtered result.
22737 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22740 @subsection Example
22744 Apply the Roberts cross operator with scale set to 2 and delta set to 10
22746 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
22750 @section sobel_opencl
22752 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
22754 The filter accepts the following option:
22758 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22761 Set value which will be multiplied with filtered result.
22762 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22765 Set value which will be added to filtered result.
22766 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22769 @subsection Example
22773 Apply sobel operator with scale set to 2 and delta set to 10
22775 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
22779 @section tonemap_opencl
22781 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
22783 It accepts the following parameters:
22787 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
22790 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
22793 Apply desaturation for highlights that exceed this level of brightness. The
22794 higher the parameter, the more color information will be preserved. This
22795 setting helps prevent unnaturally blown-out colors for super-highlights, by
22796 (smoothly) turning into white instead. This makes images feel more natural,
22797 at the cost of reducing information about out-of-range colors.
22799 The default value is 0.5, and the algorithm here is a little different from
22800 the cpu version tonemap currently. A setting of 0.0 disables this option.
22803 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
22804 is used to detect whether the scene has changed or not. If the distance between
22805 the current frame average brightness and the current running average exceeds
22806 a threshold value, we would re-calculate scene average and peak brightness.
22807 The default value is 0.2.
22810 Specify the output pixel format.
22812 Currently supported formats are:
22819 Set the output color range.
22821 Possible values are:
22827 Default is same as input.
22830 Set the output color primaries.
22832 Possible values are:
22838 Default is same as input.
22841 Set the output transfer characteristics.
22843 Possible values are:
22852 Set the output colorspace matrix.
22854 Possible value are:
22860 Default is same as input.
22864 @subsection Example
22868 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
22870 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
22874 @section unsharp_opencl
22876 Sharpen or blur the input video.
22878 It accepts the following parameters:
22881 @item luma_msize_x, lx
22882 Set the luma matrix horizontal size.
22883 Range is @code{[1, 23]} and default value is @code{5}.
22885 @item luma_msize_y, ly
22886 Set the luma matrix vertical size.
22887 Range is @code{[1, 23]} and default value is @code{5}.
22889 @item luma_amount, la
22890 Set the luma effect strength.
22891 Range is @code{[-10, 10]} and default value is @code{1.0}.
22893 Negative values will blur the input video, while positive values will
22894 sharpen it, a value of zero will disable the effect.
22896 @item chroma_msize_x, cx
22897 Set the chroma matrix horizontal size.
22898 Range is @code{[1, 23]} and default value is @code{5}.
22900 @item chroma_msize_y, cy
22901 Set the chroma matrix vertical size.
22902 Range is @code{[1, 23]} and default value is @code{5}.
22904 @item chroma_amount, ca
22905 Set the chroma effect strength.
22906 Range is @code{[-10, 10]} and default value is @code{0.0}.
22908 Negative values will blur the input video, while positive values will
22909 sharpen it, a value of zero will disable the effect.
22913 All parameters are optional and default to the equivalent of the
22914 string '5:5:1.0:5:5:0.0'.
22916 @subsection Examples
22920 Apply strong luma sharpen effect:
22922 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
22926 Apply a strong blur of both luma and chroma parameters:
22928 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
22932 @section xfade_opencl
22934 Cross fade two videos with custom transition effect by using OpenCL.
22936 It accepts the following options:
22940 Set one of possible transition effects.
22944 Select custom transition effect, the actual transition description
22945 will be picked from source and kernel options.
22957 Default transition is fade.
22961 OpenCL program source file for custom transition.
22964 Set name of kernel to use for custom transition from program source file.
22967 Set duration of video transition.
22970 Set time of start of transition relative to first video.
22973 The program source file must contain a kernel function with the given name,
22974 which will be run once for each plane of the output. Each run on a plane
22975 gets enqueued as a separate 2D global NDRange with one work-item for each
22976 pixel to be generated. The global ID offset for each work-item is therefore
22977 the coordinates of a pixel in the destination image.
22979 The kernel function needs to take the following arguments:
22982 Destination image, @var{__write_only image2d_t}.
22984 This image will become the output; the kernel should write all of it.
22987 First Source image, @var{__read_only image2d_t}.
22988 Second Source image, @var{__read_only image2d_t}.
22990 These are the most recent images on each input. The kernel may read from
22991 them to generate the output, but they can't be written to.
22994 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
23001 Apply dots curtain transition effect:
23003 __kernel void blend_images(__write_only image2d_t dst,
23004 __read_only image2d_t src1,
23005 __read_only image2d_t src2,
23008 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
23009 CLK_FILTER_LINEAR);
23010 int2 p = (int2)(get_global_id(0), get_global_id(1));
23011 float2 rp = (float2)(get_global_id(0), get_global_id(1));
23012 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
23015 float2 dots = (float2)(20.0, 20.0);
23016 float2 center = (float2)(0,0);
23019 float4 val1 = read_imagef(src1, sampler, p);
23020 float4 val2 = read_imagef(src2, sampler, p);
23021 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
23023 write_imagef(dst, p, next ? val1 : val2);
23029 @c man end OPENCL VIDEO FILTERS
23031 @chapter VAAPI Video Filters
23032 @c man begin VAAPI VIDEO FILTERS
23034 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
23036 To enable compilation of these filters you need to configure FFmpeg with
23037 @code{--enable-vaapi}.
23039 To use vaapi filters, you need to setup the vaapi device correctly. For more information, please read @url{https://trac.ffmpeg.org/wiki/Hardware/VAAPI}
23041 @section tonemap_vaapi
23043 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
23044 It maps the dynamic range of HDR10 content to the SDR content.
23045 It currently only accepts HDR10 as input.
23047 It accepts the following parameters:
23051 Specify the output pixel format.
23053 Currently supported formats are:
23062 Set the output color primaries.
23064 Default is same as input.
23067 Set the output transfer characteristics.
23072 Set the output colorspace matrix.
23074 Default is same as input.
23078 @subsection Example
23082 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
23084 tonemap_vaapi=format=p010:t=bt2020-10
23088 @c man end VAAPI VIDEO FILTERS
23090 @chapter Video Sources
23091 @c man begin VIDEO SOURCES
23093 Below is a description of the currently available video sources.
23097 Buffer video frames, and make them available to the filter chain.
23099 This source is mainly intended for a programmatic use, in particular
23100 through the interface defined in @file{libavfilter/buffersrc.h}.
23102 It accepts the following parameters:
23107 Specify the size (width and height) of the buffered video frames. For the
23108 syntax of this option, check the
23109 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23112 The input video width.
23115 The input video height.
23118 A string representing the pixel format of the buffered video frames.
23119 It may be a number corresponding to a pixel format, or a pixel format
23123 Specify the timebase assumed by the timestamps of the buffered frames.
23126 Specify the frame rate expected for the video stream.
23128 @item pixel_aspect, sar
23129 The sample (pixel) aspect ratio of the input video.
23132 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
23133 to the filtergraph description to specify swscale flags for automatically
23134 inserted scalers. See @ref{Filtergraph syntax}.
23136 @item hw_frames_ctx
23137 When using a hardware pixel format, this should be a reference to an
23138 AVHWFramesContext describing input frames.
23143 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
23146 will instruct the source to accept video frames with size 320x240 and
23147 with format "yuv410p", assuming 1/24 as the timestamps timebase and
23148 square pixels (1:1 sample aspect ratio).
23149 Since the pixel format with name "yuv410p" corresponds to the number 6
23150 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
23151 this example corresponds to:
23153 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
23156 Alternatively, the options can be specified as a flat string, but this
23157 syntax is deprecated:
23159 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
23163 Create a pattern generated by an elementary cellular automaton.
23165 The initial state of the cellular automaton can be defined through the
23166 @option{filename} and @option{pattern} options. If such options are
23167 not specified an initial state is created randomly.
23169 At each new frame a new row in the video is filled with the result of
23170 the cellular automaton next generation. The behavior when the whole
23171 frame is filled is defined by the @option{scroll} option.
23173 This source accepts the following options:
23177 Read the initial cellular automaton state, i.e. the starting row, from
23178 the specified file.
23179 In the file, each non-whitespace character is considered an alive
23180 cell, a newline will terminate the row, and further characters in the
23181 file will be ignored.
23184 Read the initial cellular automaton state, i.e. the starting row, from
23185 the specified string.
23187 Each non-whitespace character in the string is considered an alive
23188 cell, a newline will terminate the row, and further characters in the
23189 string will be ignored.
23192 Set the video rate, that is the number of frames generated per second.
23195 @item random_fill_ratio, ratio
23196 Set the random fill ratio for the initial cellular automaton row. It
23197 is a floating point number value ranging from 0 to 1, defaults to
23200 This option is ignored when a file or a pattern is specified.
23202 @item random_seed, seed
23203 Set the seed for filling randomly the initial row, must be an integer
23204 included between 0 and UINT32_MAX. If not specified, or if explicitly
23205 set to -1, the filter will try to use a good random seed on a best
23209 Set the cellular automaton rule, it is a number ranging from 0 to 255.
23210 Default value is 110.
23213 Set the size of the output video. For the syntax of this option, check the
23214 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23216 If @option{filename} or @option{pattern} is specified, the size is set
23217 by default to the width of the specified initial state row, and the
23218 height is set to @var{width} * PHI.
23220 If @option{size} is set, it must contain the width of the specified
23221 pattern string, and the specified pattern will be centered in the
23224 If a filename or a pattern string is not specified, the size value
23225 defaults to "320x518" (used for a randomly generated initial state).
23228 If set to 1, scroll the output upward when all the rows in the output
23229 have been already filled. If set to 0, the new generated row will be
23230 written over the top row just after the bottom row is filled.
23233 @item start_full, full
23234 If set to 1, completely fill the output with generated rows before
23235 outputting the first frame.
23236 This is the default behavior, for disabling set the value to 0.
23239 If set to 1, stitch the left and right row edges together.
23240 This is the default behavior, for disabling set the value to 0.
23243 @subsection Examples
23247 Read the initial state from @file{pattern}, and specify an output of
23250 cellauto=f=pattern:s=200x400
23254 Generate a random initial row with a width of 200 cells, with a fill
23257 cellauto=ratio=2/3:s=200x200
23261 Create a pattern generated by rule 18 starting by a single alive cell
23262 centered on an initial row with width 100:
23264 cellauto=p=@@:s=100x400:full=0:rule=18
23268 Specify a more elaborated initial pattern:
23270 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
23275 @anchor{coreimagesrc}
23276 @section coreimagesrc
23277 Video source generated on GPU using Apple's CoreImage API on OSX.
23279 This video source is a specialized version of the @ref{coreimage} video filter.
23280 Use a core image generator at the beginning of the applied filterchain to
23281 generate the content.
23283 The coreimagesrc video source accepts the following options:
23285 @item list_generators
23286 List all available generators along with all their respective options as well as
23287 possible minimum and maximum values along with the default values.
23289 list_generators=true
23293 Specify the size of the sourced video. For the syntax of this option, check the
23294 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23295 The default value is @code{320x240}.
23298 Specify the frame rate of the sourced video, as the number of frames
23299 generated per second. It has to be a string in the format
23300 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23301 number or a valid video frame rate abbreviation. The default value is
23305 Set the sample aspect ratio of the sourced video.
23308 Set the duration of the sourced video. See
23309 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23310 for the accepted syntax.
23312 If not specified, or the expressed duration is negative, the video is
23313 supposed to be generated forever.
23316 Additionally, all options of the @ref{coreimage} video filter are accepted.
23317 A complete filterchain can be used for further processing of the
23318 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
23319 and examples for details.
23321 @subsection Examples
23326 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
23327 given as complete and escaped command-line for Apple's standard bash shell:
23329 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
23331 This example is equivalent to the QRCode example of @ref{coreimage} without the
23332 need for a nullsrc video source.
23337 Generate several gradients.
23341 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23342 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23345 Set frame rate, expressed as number of frames per second. Default
23348 @item c0, c1, c2, c3, c4, c5, c6, c7
23349 Set 8 colors. Default values for colors is to pick random one.
23351 @item x0, y0, y0, y1
23352 Set gradient line source and destination points. If negative or out of range, random ones
23356 Set number of colors to use at once. Allowed range is from 2 to 8. Default value is 2.
23359 Set seed for picking gradient line points.
23362 Set the duration of the sourced video. See
23363 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23364 for the accepted syntax.
23366 If not specified, or the expressed duration is negative, the video is
23367 supposed to be generated forever.
23370 Set speed of gradients rotation.
23374 @section mandelbrot
23376 Generate a Mandelbrot set fractal, and progressively zoom towards the
23377 point specified with @var{start_x} and @var{start_y}.
23379 This source accepts the following options:
23384 Set the terminal pts value. Default value is 400.
23387 Set the terminal scale value.
23388 Must be a floating point value. Default value is 0.3.
23391 Set the inner coloring mode, that is the algorithm used to draw the
23392 Mandelbrot fractal internal region.
23394 It shall assume one of the following values:
23399 Show time until convergence.
23401 Set color based on point closest to the origin of the iterations.
23406 Default value is @var{mincol}.
23409 Set the bailout value. Default value is 10.0.
23412 Set the maximum of iterations performed by the rendering
23413 algorithm. Default value is 7189.
23416 Set outer coloring mode.
23417 It shall assume one of following values:
23419 @item iteration_count
23420 Set iteration count mode.
23421 @item normalized_iteration_count
23422 set normalized iteration count mode.
23424 Default value is @var{normalized_iteration_count}.
23427 Set frame rate, expressed as number of frames per second. Default
23431 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23432 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23435 Set the initial scale value. Default value is 3.0.
23438 Set the initial x position. Must be a floating point value between
23439 -100 and 100. Default value is -0.743643887037158704752191506114774.
23442 Set the initial y position. Must be a floating point value between
23443 -100 and 100. Default value is -0.131825904205311970493132056385139.
23448 Generate various test patterns, as generated by the MPlayer test filter.
23450 The size of the generated video is fixed, and is 256x256.
23451 This source is useful in particular for testing encoding features.
23453 This source accepts the following options:
23458 Specify the frame rate of the sourced video, as the number of frames
23459 generated per second. It has to be a string in the format
23460 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23461 number or a valid video frame rate abbreviation. The default value is
23465 Set the duration of the sourced video. See
23466 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23467 for the accepted syntax.
23469 If not specified, or the expressed duration is negative, the video is
23470 supposed to be generated forever.
23474 Set the number or the name of the test to perform. Supported tests are:
23488 @item max_frames, m
23489 Set the maximum number of frames generated for each test, default value is 30.
23493 Default value is "all", which will cycle through the list of all tests.
23498 mptestsrc=t=dc_luma
23501 will generate a "dc_luma" test pattern.
23503 @section frei0r_src
23505 Provide a frei0r source.
23507 To enable compilation of this filter you need to install the frei0r
23508 header and configure FFmpeg with @code{--enable-frei0r}.
23510 This source accepts the following parameters:
23515 The size of the video to generate. For the syntax of this option, check the
23516 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23519 The framerate of the generated video. It may be a string of the form
23520 @var{num}/@var{den} or a frame rate abbreviation.
23523 The name to the frei0r source to load. For more information regarding frei0r and
23524 how to set the parameters, read the @ref{frei0r} section in the video filters
23527 @item filter_params
23528 A '|'-separated list of parameters to pass to the frei0r source.
23532 For example, to generate a frei0r partik0l source with size 200x200
23533 and frame rate 10 which is overlaid on the overlay filter main input:
23535 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
23540 Generate a life pattern.
23542 This source is based on a generalization of John Conway's life game.
23544 The sourced input represents a life grid, each pixel represents a cell
23545 which can be in one of two possible states, alive or dead. Every cell
23546 interacts with its eight neighbours, which are the cells that are
23547 horizontally, vertically, or diagonally adjacent.
23549 At each interaction the grid evolves according to the adopted rule,
23550 which specifies the number of neighbor alive cells which will make a
23551 cell stay alive or born. The @option{rule} option allows one to specify
23554 This source accepts the following options:
23558 Set the file from which to read the initial grid state. In the file,
23559 each non-whitespace character is considered an alive cell, and newline
23560 is used to delimit the end of each row.
23562 If this option is not specified, the initial grid is generated
23566 Set the video rate, that is the number of frames generated per second.
23569 @item random_fill_ratio, ratio
23570 Set the random fill ratio for the initial random grid. It is a
23571 floating point number value ranging from 0 to 1, defaults to 1/PHI.
23572 It is ignored when a file is specified.
23574 @item random_seed, seed
23575 Set the seed for filling the initial random grid, must be an integer
23576 included between 0 and UINT32_MAX. If not specified, or if explicitly
23577 set to -1, the filter will try to use a good random seed on a best
23583 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
23584 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
23585 @var{NS} specifies the number of alive neighbor cells which make a
23586 live cell stay alive, and @var{NB} the number of alive neighbor cells
23587 which make a dead cell to become alive (i.e. to "born").
23588 "s" and "b" can be used in place of "S" and "B", respectively.
23590 Alternatively a rule can be specified by an 18-bits integer. The 9
23591 high order bits are used to encode the next cell state if it is alive
23592 for each number of neighbor alive cells, the low order bits specify
23593 the rule for "borning" new cells. Higher order bits encode for an
23594 higher number of neighbor cells.
23595 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
23596 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
23598 Default value is "S23/B3", which is the original Conway's game of life
23599 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
23600 cells, and will born a new cell if there are three alive cells around
23604 Set the size of the output video. For the syntax of this option, check the
23605 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23607 If @option{filename} is specified, the size is set by default to the
23608 same size of the input file. If @option{size} is set, it must contain
23609 the size specified in the input file, and the initial grid defined in
23610 that file is centered in the larger resulting area.
23612 If a filename is not specified, the size value defaults to "320x240"
23613 (used for a randomly generated initial grid).
23616 If set to 1, stitch the left and right grid edges together, and the
23617 top and bottom edges also. Defaults to 1.
23620 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
23621 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
23622 value from 0 to 255.
23625 Set the color of living (or new born) cells.
23628 Set the color of dead cells. If @option{mold} is set, this is the first color
23629 used to represent a dead cell.
23632 Set mold color, for definitely dead and moldy cells.
23634 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
23635 ffmpeg-utils manual,ffmpeg-utils}.
23638 @subsection Examples
23642 Read a grid from @file{pattern}, and center it on a grid of size
23645 life=f=pattern:s=300x300
23649 Generate a random grid of size 200x200, with a fill ratio of 2/3:
23651 life=ratio=2/3:s=200x200
23655 Specify a custom rule for evolving a randomly generated grid:
23661 Full example with slow death effect (mold) using @command{ffplay}:
23663 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
23670 @anchor{haldclutsrc}
23673 @anchor{pal100bars}
23674 @anchor{rgbtestsrc}
23676 @anchor{smptehdbars}
23679 @anchor{yuvtestsrc}
23680 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
23682 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
23684 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
23686 The @code{color} source provides an uniformly colored input.
23688 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
23689 @ref{haldclut} filter.
23691 The @code{nullsrc} source returns unprocessed video frames. It is
23692 mainly useful to be employed in analysis / debugging tools, or as the
23693 source for filters which ignore the input data.
23695 The @code{pal75bars} source generates a color bars pattern, based on
23696 EBU PAL recommendations with 75% color levels.
23698 The @code{pal100bars} source generates a color bars pattern, based on
23699 EBU PAL recommendations with 100% color levels.
23701 The @code{rgbtestsrc} source generates an RGB test pattern useful for
23702 detecting RGB vs BGR issues. You should see a red, green and blue
23703 stripe from top to bottom.
23705 The @code{smptebars} source generates a color bars pattern, based on
23706 the SMPTE Engineering Guideline EG 1-1990.
23708 The @code{smptehdbars} source generates a color bars pattern, based on
23709 the SMPTE RP 219-2002.
23711 The @code{testsrc} source generates a test video pattern, showing a
23712 color pattern, a scrolling gradient and a timestamp. This is mainly
23713 intended for testing purposes.
23715 The @code{testsrc2} source is similar to testsrc, but supports more
23716 pixel formats instead of just @code{rgb24}. This allows using it as an
23717 input for other tests without requiring a format conversion.
23719 The @code{yuvtestsrc} source generates an YUV test pattern. You should
23720 see a y, cb and cr stripe from top to bottom.
23722 The sources accept the following parameters:
23727 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
23728 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
23729 pixels to be used as identity matrix for 3D lookup tables. Each component is
23730 coded on a @code{1/(N*N)} scale.
23733 Specify the color of the source, only available in the @code{color}
23734 source. For the syntax of this option, check the
23735 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
23738 Specify the size of the sourced video. For the syntax of this option, check the
23739 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23740 The default value is @code{320x240}.
23742 This option is not available with the @code{allrgb}, @code{allyuv}, and
23743 @code{haldclutsrc} filters.
23746 Specify the frame rate of the sourced video, as the number of frames
23747 generated per second. It has to be a string in the format
23748 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23749 number or a valid video frame rate abbreviation. The default value is
23753 Set the duration of the sourced video. See
23754 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23755 for the accepted syntax.
23757 If not specified, or the expressed duration is negative, the video is
23758 supposed to be generated forever.
23760 Since the frame rate is used as time base, all frames including the last one
23761 will have their full duration. If the specified duration is not a multiple
23762 of the frame duration, it will be rounded up.
23765 Set the sample aspect ratio of the sourced video.
23768 Specify the alpha (opacity) of the background, only available in the
23769 @code{testsrc2} source. The value must be between 0 (fully transparent) and
23770 255 (fully opaque, the default).
23773 Set the number of decimals to show in the timestamp, only available in the
23774 @code{testsrc} source.
23776 The displayed timestamp value will correspond to the original
23777 timestamp value multiplied by the power of 10 of the specified
23778 value. Default value is 0.
23781 @subsection Examples
23785 Generate a video with a duration of 5.3 seconds, with size
23786 176x144 and a frame rate of 10 frames per second:
23788 testsrc=duration=5.3:size=qcif:rate=10
23792 The following graph description will generate a red source
23793 with an opacity of 0.2, with size "qcif" and a frame rate of 10
23796 color=c=red@@0.2:s=qcif:r=10
23800 If the input content is to be ignored, @code{nullsrc} can be used. The
23801 following command generates noise in the luminance plane by employing
23802 the @code{geq} filter:
23804 nullsrc=s=256x256, geq=random(1)*255:128:128
23808 @subsection Commands
23810 The @code{color} source supports the following commands:
23814 Set the color of the created image. Accepts the same syntax of the
23815 corresponding @option{color} option.
23820 Generate video using an OpenCL program.
23825 OpenCL program source file.
23828 Kernel name in program.
23831 Size of frames to generate. This must be set.
23834 Pixel format to use for the generated frames. This must be set.
23837 Number of frames generated every second. Default value is '25'.
23841 For details of how the program loading works, see the @ref{program_opencl}
23848 Generate a colour ramp by setting pixel values from the position of the pixel
23849 in the output image. (Note that this will work with all pixel formats, but
23850 the generated output will not be the same.)
23852 __kernel void ramp(__write_only image2d_t dst,
23853 unsigned int index)
23855 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23858 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
23860 write_imagef(dst, loc, val);
23865 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
23867 __kernel void sierpinski_carpet(__write_only image2d_t dst,
23868 unsigned int index)
23870 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23872 float4 value = 0.0f;
23873 int x = loc.x + index;
23874 int y = loc.y + index;
23875 while (x > 0 || y > 0) {
23876 if (x % 3 == 1 && y % 3 == 1) {
23884 write_imagef(dst, loc, value);
23890 @section sierpinski
23892 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
23894 This source accepts the following options:
23898 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23899 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23902 Set frame rate, expressed as number of frames per second. Default
23906 Set seed which is used for random panning.
23909 Set max jump for single pan destination. Allowed range is from 1 to 10000.
23912 Set fractal type, can be default @code{carpet} or @code{triangle}.
23915 @c man end VIDEO SOURCES
23917 @chapter Video Sinks
23918 @c man begin VIDEO SINKS
23920 Below is a description of the currently available video sinks.
23922 @section buffersink
23924 Buffer video frames, and make them available to the end of the filter
23927 This sink is mainly intended for programmatic use, in particular
23928 through the interface defined in @file{libavfilter/buffersink.h}
23929 or the options system.
23931 It accepts a pointer to an AVBufferSinkContext structure, which
23932 defines the incoming buffers' formats, to be passed as the opaque
23933 parameter to @code{avfilter_init_filter} for initialization.
23937 Null video sink: do absolutely nothing with the input video. It is
23938 mainly useful as a template and for use in analysis / debugging
23941 @c man end VIDEO SINKS
23943 @chapter Multimedia Filters
23944 @c man begin MULTIMEDIA FILTERS
23946 Below is a description of the currently available multimedia filters.
23950 Convert input audio to a video output, displaying the audio bit scope.
23952 The filter accepts the following options:
23956 Set frame rate, expressed as number of frames per second. Default
23960 Specify the video size for the output. For the syntax of this option, check the
23961 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23962 Default value is @code{1024x256}.
23965 Specify list of colors separated by space or by '|' which will be used to
23966 draw channels. Unrecognized or missing colors will be replaced
23970 @section adrawgraph
23971 Draw a graph using input audio metadata.
23973 See @ref{drawgraph}
23975 @section agraphmonitor
23977 See @ref{graphmonitor}.
23979 @section ahistogram
23981 Convert input audio to a video output, displaying the volume histogram.
23983 The filter accepts the following options:
23987 Specify how histogram is calculated.
23989 It accepts the following values:
23992 Use single histogram for all channels.
23994 Use separate histogram for each channel.
23996 Default is @code{single}.
23999 Set frame rate, expressed as number of frames per second. Default
24003 Specify the video size for the output. For the syntax of this option, check the
24004 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24005 Default value is @code{hd720}.
24010 It accepts the following values:
24021 reverse logarithmic
24023 Default is @code{log}.
24026 Set amplitude scale.
24028 It accepts the following values:
24035 Default is @code{log}.
24038 Set how much frames to accumulate in histogram.
24039 Default is 1. Setting this to -1 accumulates all frames.
24042 Set histogram ratio of window height.
24045 Set sonogram sliding.
24047 It accepts the following values:
24050 replace old rows with new ones.
24052 scroll from top to bottom.
24054 Default is @code{replace}.
24057 @section aphasemeter
24059 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
24060 representing mean phase of current audio frame. A video output can also be produced and is
24061 enabled by default. The audio is passed through as first output.
24063 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
24064 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
24065 and @code{1} means channels are in phase.
24067 The filter accepts the following options, all related to its video output:
24071 Set the output frame rate. Default value is @code{25}.
24074 Set the video size for the output. For the syntax of this option, check the
24075 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24076 Default value is @code{800x400}.
24081 Specify the red, green, blue contrast. Default values are @code{2},
24082 @code{7} and @code{1}.
24083 Allowed range is @code{[0, 255]}.
24086 Set color which will be used for drawing median phase. If color is
24087 @code{none} which is default, no median phase value will be drawn.
24090 Enable video output. Default is enabled.
24093 @subsection phasing detection
24095 The filter also detects out of phase and mono sequences in stereo streams.
24096 It logs the sequence start, end and duration when it lasts longer or as long as the minimum set.
24098 The filter accepts the following options for this detection:
24102 Enable mono and out of phase detection. Default is disabled.
24105 Set phase tolerance for mono detection, in amplitude ratio. Default is @code{0}.
24106 Allowed range is @code{[0, 1]}.
24109 Set angle threshold for out of phase detection, in degree. Default is @code{170}.
24110 Allowed range is @code{[90, 180]}.
24113 Set mono or out of phase duration until notification, expressed in seconds. Default is @code{2}.
24116 @subsection Examples
24120 Complete example with @command{ffmpeg} to detect 1 second of mono with 0.001 phase tolerance:
24122 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
24126 @section avectorscope
24128 Convert input audio to a video output, representing the audio vector
24131 The filter is used to measure the difference between channels of stereo
24132 audio stream. A monaural signal, consisting of identical left and right
24133 signal, results in straight vertical line. Any stereo separation is visible
24134 as a deviation from this line, creating a Lissajous figure.
24135 If the straight (or deviation from it) but horizontal line appears this
24136 indicates that the left and right channels are out of phase.
24138 The filter accepts the following options:
24142 Set the vectorscope mode.
24144 Available values are:
24147 Lissajous rotated by 45 degrees.
24150 Same as above but not rotated.
24153 Shape resembling half of circle.
24156 Default value is @samp{lissajous}.
24159 Set the video size for the output. For the syntax of this option, check the
24160 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24161 Default value is @code{400x400}.
24164 Set the output frame rate. Default value is @code{25}.
24170 Specify the red, green, blue and alpha contrast. Default values are @code{40},
24171 @code{160}, @code{80} and @code{255}.
24172 Allowed range is @code{[0, 255]}.
24178 Specify the red, green, blue and alpha fade. Default values are @code{15},
24179 @code{10}, @code{5} and @code{5}.
24180 Allowed range is @code{[0, 255]}.
24183 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
24184 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
24187 Set the vectorscope drawing mode.
24189 Available values are:
24192 Draw dot for each sample.
24195 Draw line between previous and current sample.
24198 Default value is @samp{dot}.
24201 Specify amplitude scale of audio samples.
24203 Available values are:
24219 Swap left channel axis with right channel axis.
24229 Mirror only x axis.
24232 Mirror only y axis.
24240 @subsection Examples
24244 Complete example using @command{ffplay}:
24246 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
24247 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
24251 @section bench, abench
24253 Benchmark part of a filtergraph.
24255 The filter accepts the following options:
24259 Start or stop a timer.
24261 Available values are:
24264 Get the current time, set it as frame metadata (using the key
24265 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
24268 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
24269 the input frame metadata to get the time difference. Time difference, average,
24270 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
24271 @code{min}) are then printed. The timestamps are expressed in seconds.
24275 @subsection Examples
24279 Benchmark @ref{selectivecolor} filter:
24281 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
24287 Concatenate audio and video streams, joining them together one after the
24290 The filter works on segments of synchronized video and audio streams. All
24291 segments must have the same number of streams of each type, and that will
24292 also be the number of streams at output.
24294 The filter accepts the following options:
24299 Set the number of segments. Default is 2.
24302 Set the number of output video streams, that is also the number of video
24303 streams in each segment. Default is 1.
24306 Set the number of output audio streams, that is also the number of audio
24307 streams in each segment. Default is 0.
24310 Activate unsafe mode: do not fail if segments have a different format.
24314 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
24315 @var{a} audio outputs.
24317 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
24318 segment, in the same order as the outputs, then the inputs for the second
24321 Related streams do not always have exactly the same duration, for various
24322 reasons including codec frame size or sloppy authoring. For that reason,
24323 related synchronized streams (e.g. a video and its audio track) should be
24324 concatenated at once. The concat filter will use the duration of the longest
24325 stream in each segment (except the last one), and if necessary pad shorter
24326 audio streams with silence.
24328 For this filter to work correctly, all segments must start at timestamp 0.
24330 All corresponding streams must have the same parameters in all segments; the
24331 filtering system will automatically select a common pixel format for video
24332 streams, and a common sample format, sample rate and channel layout for
24333 audio streams, but other settings, such as resolution, must be converted
24334 explicitly by the user.
24336 Different frame rates are acceptable but will result in variable frame rate
24337 at output; be sure to configure the output file to handle it.
24339 @subsection Examples
24343 Concatenate an opening, an episode and an ending, all in bilingual version
24344 (video in stream 0, audio in streams 1 and 2):
24346 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
24347 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
24348 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
24349 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
24353 Concatenate two parts, handling audio and video separately, using the
24354 (a)movie sources, and adjusting the resolution:
24356 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
24357 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
24358 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
24360 Note that a desync will happen at the stitch if the audio and video streams
24361 do not have exactly the same duration in the first file.
24365 @subsection Commands
24367 This filter supports the following commands:
24370 Close the current segment and step to the next one
24376 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
24377 level. By default, it logs a message at a frequency of 10Hz with the
24378 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
24379 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
24381 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
24382 sample format is double-precision floating point. The input stream will be converted to
24383 this specification, if needed. Users may need to insert aformat and/or aresample filters
24384 after this filter to obtain the original parameters.
24386 The filter also has a video output (see the @var{video} option) with a real
24387 time graph to observe the loudness evolution. The graphic contains the logged
24388 message mentioned above, so it is not printed anymore when this option is set,
24389 unless the verbose logging is set. The main graphing area contains the
24390 short-term loudness (3 seconds of analysis), and the gauge on the right is for
24391 the momentary loudness (400 milliseconds), but can optionally be configured
24392 to instead display short-term loudness (see @var{gauge}).
24394 The green area marks a +/- 1LU target range around the target loudness
24395 (-23LUFS by default, unless modified through @var{target}).
24397 More information about the Loudness Recommendation EBU R128 on
24398 @url{http://tech.ebu.ch/loudness}.
24400 The filter accepts the following options:
24405 Activate the video output. The audio stream is passed unchanged whether this
24406 option is set or no. The video stream will be the first output stream if
24407 activated. Default is @code{0}.
24410 Set the video size. This option is for video only. For the syntax of this
24412 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24413 Default and minimum resolution is @code{640x480}.
24416 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
24417 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
24418 other integer value between this range is allowed.
24421 Set metadata injection. If set to @code{1}, the audio input will be segmented
24422 into 100ms output frames, each of them containing various loudness information
24423 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
24425 Default is @code{0}.
24428 Force the frame logging level.
24430 Available values are:
24433 information logging level
24435 verbose logging level
24438 By default, the logging level is set to @var{info}. If the @option{video} or
24439 the @option{metadata} options are set, it switches to @var{verbose}.
24444 Available modes can be cumulated (the option is a @code{flag} type). Possible
24448 Disable any peak mode (default).
24450 Enable sample-peak mode.
24452 Simple peak mode looking for the higher sample value. It logs a message
24453 for sample-peak (identified by @code{SPK}).
24455 Enable true-peak mode.
24457 If enabled, the peak lookup is done on an over-sampled version of the input
24458 stream for better peak accuracy. It logs a message for true-peak.
24459 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
24460 This mode requires a build with @code{libswresample}.
24464 Treat mono input files as "dual mono". If a mono file is intended for playback
24465 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
24466 If set to @code{true}, this option will compensate for this effect.
24467 Multi-channel input files are not affected by this option.
24470 Set a specific pan law to be used for the measurement of dual mono files.
24471 This parameter is optional, and has a default value of -3.01dB.
24474 Set a specific target level (in LUFS) used as relative zero in the visualization.
24475 This parameter is optional and has a default value of -23LUFS as specified
24476 by EBU R128. However, material published online may prefer a level of -16LUFS
24477 (e.g. for use with podcasts or video platforms).
24480 Set the value displayed by the gauge. Valid values are @code{momentary} and s
24481 @code{shortterm}. By default the momentary value will be used, but in certain
24482 scenarios it may be more useful to observe the short term value instead (e.g.
24486 Sets the display scale for the loudness. Valid parameters are @code{absolute}
24487 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
24488 video output, not the summary or continuous log output.
24491 @subsection Examples
24495 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
24497 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
24501 Run an analysis with @command{ffmpeg}:
24503 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
24507 @section interleave, ainterleave
24509 Temporally interleave frames from several inputs.
24511 @code{interleave} works with video inputs, @code{ainterleave} with audio.
24513 These filters read frames from several inputs and send the oldest
24514 queued frame to the output.
24516 Input streams must have well defined, monotonically increasing frame
24519 In order to submit one frame to output, these filters need to enqueue
24520 at least one frame for each input, so they cannot work in case one
24521 input is not yet terminated and will not receive incoming frames.
24523 For example consider the case when one input is a @code{select} filter
24524 which always drops input frames. The @code{interleave} filter will keep
24525 reading from that input, but it will never be able to send new frames
24526 to output until the input sends an end-of-stream signal.
24528 Also, depending on inputs synchronization, the filters will drop
24529 frames in case one input receives more frames than the other ones, and
24530 the queue is already filled.
24532 These filters accept the following options:
24536 Set the number of different inputs, it is 2 by default.
24539 How to determine the end-of-stream.
24543 The duration of the longest input. (default)
24546 The duration of the shortest input.
24549 The duration of the first input.
24554 @subsection Examples
24558 Interleave frames belonging to different streams using @command{ffmpeg}:
24560 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
24564 Add flickering blur effect:
24566 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
24570 @section metadata, ametadata
24572 Manipulate frame metadata.
24574 This filter accepts the following options:
24578 Set mode of operation of the filter.
24580 Can be one of the following:
24584 If both @code{value} and @code{key} is set, select frames
24585 which have such metadata. If only @code{key} is set, select
24586 every frame that has such key in metadata.
24589 Add new metadata @code{key} and @code{value}. If key is already available
24593 Modify value of already present key.
24596 If @code{value} is set, delete only keys that have such value.
24597 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
24601 Print key and its value if metadata was found. If @code{key} is not set print all
24602 metadata values available in frame.
24606 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
24609 Set metadata value which will be used. This option is mandatory for
24610 @code{modify} and @code{add} mode.
24613 Which function to use when comparing metadata value and @code{value}.
24615 Can be one of following:
24619 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
24622 Values are interpreted as strings, returns true if metadata value starts with
24623 the @code{value} option string.
24626 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
24629 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
24632 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
24635 Values are interpreted as floats, returns true if expression from option @code{expr}
24639 Values are interpreted as strings, returns true if metadata value ends with
24640 the @code{value} option string.
24644 Set expression which is used when @code{function} is set to @code{expr}.
24645 The expression is evaluated through the eval API and can contain the following
24650 Float representation of @code{value} from metadata key.
24653 Float representation of @code{value} as supplied by user in @code{value} option.
24657 If specified in @code{print} mode, output is written to the named file. Instead of
24658 plain filename any writable url can be specified. Filename ``-'' is a shorthand
24659 for standard output. If @code{file} option is not set, output is written to the log
24660 with AV_LOG_INFO loglevel.
24663 Reduces buffering in print mode when output is written to a URL set using @var{file}.
24667 @subsection Examples
24671 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
24674 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
24677 Print silencedetect output to file @file{metadata.txt}.
24679 silencedetect,ametadata=mode=print:file=metadata.txt
24682 Direct all metadata to a pipe with file descriptor 4.
24684 metadata=mode=print:file='pipe\:4'
24688 @section perms, aperms
24690 Set read/write permissions for the output frames.
24692 These filters are mainly aimed at developers to test direct path in the
24693 following filter in the filtergraph.
24695 The filters accept the following options:
24699 Select the permissions mode.
24701 It accepts the following values:
24704 Do nothing. This is the default.
24706 Set all the output frames read-only.
24708 Set all the output frames directly writable.
24710 Make the frame read-only if writable, and writable if read-only.
24712 Set each output frame read-only or writable randomly.
24716 Set the seed for the @var{random} mode, must be an integer included between
24717 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
24718 @code{-1}, the filter will try to use a good random seed on a best effort
24722 Note: in case of auto-inserted filter between the permission filter and the
24723 following one, the permission might not be received as expected in that
24724 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
24725 perms/aperms filter can avoid this problem.
24727 @section realtime, arealtime
24729 Slow down filtering to match real time approximately.
24731 These filters will pause the filtering for a variable amount of time to
24732 match the output rate with the input timestamps.
24733 They are similar to the @option{re} option to @code{ffmpeg}.
24735 They accept the following options:
24739 Time limit for the pauses. Any pause longer than that will be considered
24740 a timestamp discontinuity and reset the timer. Default is 2 seconds.
24742 Speed factor for processing. The value must be a float larger than zero.
24743 Values larger than 1.0 will result in faster than realtime processing,
24744 smaller will slow processing down. The @var{limit} is automatically adapted
24745 accordingly. Default is 1.0.
24747 A processing speed faster than what is possible without these filters cannot
24752 @section select, aselect
24754 Select frames to pass in output.
24756 This filter accepts the following options:
24761 Set expression, which is evaluated for each input frame.
24763 If the expression is evaluated to zero, the frame is discarded.
24765 If the evaluation result is negative or NaN, the frame is sent to the
24766 first output; otherwise it is sent to the output with index
24767 @code{ceil(val)-1}, assuming that the input index starts from 0.
24769 For example a value of @code{1.2} corresponds to the output with index
24770 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
24773 Set the number of outputs. The output to which to send the selected
24774 frame is based on the result of the evaluation. Default value is 1.
24777 The expression can contain the following constants:
24781 The (sequential) number of the filtered frame, starting from 0.
24784 The (sequential) number of the selected frame, starting from 0.
24786 @item prev_selected_n
24787 The sequential number of the last selected frame. It's NAN if undefined.
24790 The timebase of the input timestamps.
24793 The PTS (Presentation TimeStamp) of the filtered video frame,
24794 expressed in @var{TB} units. It's NAN if undefined.
24797 The PTS of the filtered video frame,
24798 expressed in seconds. It's NAN if undefined.
24801 The PTS of the previously filtered video frame. It's NAN if undefined.
24803 @item prev_selected_pts
24804 The PTS of the last previously filtered video frame. It's NAN if undefined.
24806 @item prev_selected_t
24807 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
24810 The PTS of the first video frame in the video. It's NAN if undefined.
24813 The time of the first video frame in the video. It's NAN if undefined.
24815 @item pict_type @emph{(video only)}
24816 The type of the filtered frame. It can assume one of the following
24828 @item interlace_type @emph{(video only)}
24829 The frame interlace type. It can assume one of the following values:
24832 The frame is progressive (not interlaced).
24834 The frame is top-field-first.
24836 The frame is bottom-field-first.
24839 @item consumed_sample_n @emph{(audio only)}
24840 the number of selected samples before the current frame
24842 @item samples_n @emph{(audio only)}
24843 the number of samples in the current frame
24845 @item sample_rate @emph{(audio only)}
24846 the input sample rate
24849 This is 1 if the filtered frame is a key-frame, 0 otherwise.
24852 the position in the file of the filtered frame, -1 if the information
24853 is not available (e.g. for synthetic video)
24855 @item scene @emph{(video only)}
24856 value between 0 and 1 to indicate a new scene; a low value reflects a low
24857 probability for the current frame to introduce a new scene, while a higher
24858 value means the current frame is more likely to be one (see the example below)
24860 @item concatdec_select
24861 The concat demuxer can select only part of a concat input file by setting an
24862 inpoint and an outpoint, but the output packets may not be entirely contained
24863 in the selected interval. By using this variable, it is possible to skip frames
24864 generated by the concat demuxer which are not exactly contained in the selected
24867 This works by comparing the frame pts against the @var{lavf.concat.start_time}
24868 and the @var{lavf.concat.duration} packet metadata values which are also
24869 present in the decoded frames.
24871 The @var{concatdec_select} variable is -1 if the frame pts is at least
24872 start_time and either the duration metadata is missing or the frame pts is less
24873 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
24876 That basically means that an input frame is selected if its pts is within the
24877 interval set by the concat demuxer.
24881 The default value of the select expression is "1".
24883 @subsection Examples
24887 Select all frames in input:
24892 The example above is the same as:
24904 Select only I-frames:
24906 select='eq(pict_type\,I)'
24910 Select one frame every 100:
24912 select='not(mod(n\,100))'
24916 Select only frames contained in the 10-20 time interval:
24918 select=between(t\,10\,20)
24922 Select only I-frames contained in the 10-20 time interval:
24924 select=between(t\,10\,20)*eq(pict_type\,I)
24928 Select frames with a minimum distance of 10 seconds:
24930 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
24934 Use aselect to select only audio frames with samples number > 100:
24936 aselect='gt(samples_n\,100)'
24940 Create a mosaic of the first scenes:
24942 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
24945 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
24949 Send even and odd frames to separate outputs, and compose them:
24951 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
24955 Select useful frames from an ffconcat file which is using inpoints and
24956 outpoints but where the source files are not intra frame only.
24958 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
24962 @section sendcmd, asendcmd
24964 Send commands to filters in the filtergraph.
24966 These filters read commands to be sent to other filters in the
24969 @code{sendcmd} must be inserted between two video filters,
24970 @code{asendcmd} must be inserted between two audio filters, but apart
24971 from that they act the same way.
24973 The specification of commands can be provided in the filter arguments
24974 with the @var{commands} option, or in a file specified by the
24975 @var{filename} option.
24977 These filters accept the following options:
24980 Set the commands to be read and sent to the other filters.
24982 Set the filename of the commands to be read and sent to the other
24986 @subsection Commands syntax
24988 A commands description consists of a sequence of interval
24989 specifications, comprising a list of commands to be executed when a
24990 particular event related to that interval occurs. The occurring event
24991 is typically the current frame time entering or leaving a given time
24994 An interval is specified by the following syntax:
24996 @var{START}[-@var{END}] @var{COMMANDS};
24999 The time interval is specified by the @var{START} and @var{END} times.
25000 @var{END} is optional and defaults to the maximum time.
25002 The current frame time is considered within the specified interval if
25003 it is included in the interval [@var{START}, @var{END}), that is when
25004 the time is greater or equal to @var{START} and is lesser than
25007 @var{COMMANDS} consists of a sequence of one or more command
25008 specifications, separated by ",", relating to that interval. The
25009 syntax of a command specification is given by:
25011 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
25014 @var{FLAGS} is optional and specifies the type of events relating to
25015 the time interval which enable sending the specified command, and must
25016 be a non-null sequence of identifier flags separated by "+" or "|" and
25017 enclosed between "[" and "]".
25019 The following flags are recognized:
25022 The command is sent when the current frame timestamp enters the
25023 specified interval. In other words, the command is sent when the
25024 previous frame timestamp was not in the given interval, and the
25028 The command is sent when the current frame timestamp leaves the
25029 specified interval. In other words, the command is sent when the
25030 previous frame timestamp was in the given interval, and the
25034 The command @var{ARG} is interpreted as expression and result of
25035 expression is passed as @var{ARG}.
25037 The expression is evaluated through the eval API and can contain the following
25042 Original position in the file of the frame, or undefined if undefined
25043 for the current frame.
25046 The presentation timestamp in input.
25049 The count of the input frame for video or audio, starting from 0.
25052 The time in seconds of the current frame.
25055 The start time in seconds of the current command interval.
25058 The end time in seconds of the current command interval.
25061 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
25066 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
25069 @var{TARGET} specifies the target of the command, usually the name of
25070 the filter class or a specific filter instance name.
25072 @var{COMMAND} specifies the name of the command for the target filter.
25074 @var{ARG} is optional and specifies the optional list of argument for
25075 the given @var{COMMAND}.
25077 Between one interval specification and another, whitespaces, or
25078 sequences of characters starting with @code{#} until the end of line,
25079 are ignored and can be used to annotate comments.
25081 A simplified BNF description of the commands specification syntax
25084 @var{COMMAND_FLAG} ::= "enter" | "leave"
25085 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
25086 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
25087 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
25088 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
25089 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
25092 @subsection Examples
25096 Specify audio tempo change at second 4:
25098 asendcmd=c='4.0 atempo tempo 1.5',atempo
25102 Target a specific filter instance:
25104 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
25108 Specify a list of drawtext and hue commands in a file.
25110 # show text in the interval 5-10
25111 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
25112 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
25114 # desaturate the image in the interval 15-20
25115 15.0-20.0 [enter] hue s 0,
25116 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
25118 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
25120 # apply an exponential saturation fade-out effect, starting from time 25
25121 25 [enter] hue s exp(25-t)
25124 A filtergraph allowing to read and process the above command list
25125 stored in a file @file{test.cmd}, can be specified with:
25127 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
25132 @section setpts, asetpts
25134 Change the PTS (presentation timestamp) of the input frames.
25136 @code{setpts} works on video frames, @code{asetpts} on audio frames.
25138 This filter accepts the following options:
25143 The expression which is evaluated for each frame to construct its timestamp.
25147 The expression is evaluated through the eval API and can contain the following
25151 @item FRAME_RATE, FR
25152 frame rate, only defined for constant frame-rate video
25155 The presentation timestamp in input
25158 The count of the input frame for video or the number of consumed samples,
25159 not including the current frame for audio, starting from 0.
25161 @item NB_CONSUMED_SAMPLES
25162 The number of consumed samples, not including the current frame (only
25165 @item NB_SAMPLES, S
25166 The number of samples in the current frame (only audio)
25168 @item SAMPLE_RATE, SR
25169 The audio sample rate.
25172 The PTS of the first frame.
25175 the time in seconds of the first frame
25178 State whether the current frame is interlaced.
25181 the time in seconds of the current frame
25184 original position in the file of the frame, or undefined if undefined
25185 for the current frame
25188 The previous input PTS.
25191 previous input time in seconds
25194 The previous output PTS.
25197 previous output time in seconds
25200 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
25204 The wallclock (RTC) time at the start of the movie in microseconds.
25207 The timebase of the input timestamps.
25211 @subsection Examples
25215 Start counting PTS from zero
25217 setpts=PTS-STARTPTS
25221 Apply fast motion effect:
25227 Apply slow motion effect:
25233 Set fixed rate of 25 frames per second:
25239 Set fixed rate 25 fps with some jitter:
25241 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
25245 Apply an offset of 10 seconds to the input PTS:
25251 Generate timestamps from a "live source" and rebase onto the current timebase:
25253 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
25257 Generate timestamps by counting samples:
25266 Force color range for the output video frame.
25268 The @code{setrange} filter marks the color range property for the
25269 output frames. It does not change the input frame, but only sets the
25270 corresponding property, which affects how the frame is treated by
25273 The filter accepts the following options:
25278 Available values are:
25282 Keep the same color range property.
25284 @item unspecified, unknown
25285 Set the color range as unspecified.
25287 @item limited, tv, mpeg
25288 Set the color range as limited.
25290 @item full, pc, jpeg
25291 Set the color range as full.
25295 @section settb, asettb
25297 Set the timebase to use for the output frames timestamps.
25298 It is mainly useful for testing timebase configuration.
25300 It accepts the following parameters:
25305 The expression which is evaluated into the output timebase.
25309 The value for @option{tb} is an arithmetic expression representing a
25310 rational. The expression can contain the constants "AVTB" (the default
25311 timebase), "intb" (the input timebase) and "sr" (the sample rate,
25312 audio only). Default value is "intb".
25314 @subsection Examples
25318 Set the timebase to 1/25:
25324 Set the timebase to 1/10:
25330 Set the timebase to 1001/1000:
25336 Set the timebase to 2*intb:
25342 Set the default timebase value:
25349 Convert input audio to a video output representing frequency spectrum
25350 logarithmically using Brown-Puckette constant Q transform algorithm with
25351 direct frequency domain coefficient calculation (but the transform itself
25352 is not really constant Q, instead the Q factor is actually variable/clamped),
25353 with musical tone scale, from E0 to D#10.
25355 The filter accepts the following options:
25359 Specify the video size for the output. It must be even. For the syntax of this option,
25360 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25361 Default value is @code{1920x1080}.
25364 Set the output frame rate. Default value is @code{25}.
25367 Set the bargraph height. It must be even. Default value is @code{-1} which
25368 computes the bargraph height automatically.
25371 Set the axis height. It must be even. Default value is @code{-1} which computes
25372 the axis height automatically.
25375 Set the sonogram height. It must be even. Default value is @code{-1} which
25376 computes the sonogram height automatically.
25379 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
25380 instead. Default value is @code{1}.
25382 @item sono_v, volume
25383 Specify the sonogram volume expression. It can contain variables:
25386 the @var{bar_v} evaluated expression
25387 @item frequency, freq, f
25388 the frequency where it is evaluated
25389 @item timeclamp, tc
25390 the value of @var{timeclamp} option
25394 @item a_weighting(f)
25395 A-weighting of equal loudness
25396 @item b_weighting(f)
25397 B-weighting of equal loudness
25398 @item c_weighting(f)
25399 C-weighting of equal loudness.
25401 Default value is @code{16}.
25403 @item bar_v, volume2
25404 Specify the bargraph volume expression. It can contain variables:
25407 the @var{sono_v} evaluated expression
25408 @item frequency, freq, f
25409 the frequency where it is evaluated
25410 @item timeclamp, tc
25411 the value of @var{timeclamp} option
25415 @item a_weighting(f)
25416 A-weighting of equal loudness
25417 @item b_weighting(f)
25418 B-weighting of equal loudness
25419 @item c_weighting(f)
25420 C-weighting of equal loudness.
25422 Default value is @code{sono_v}.
25424 @item sono_g, gamma
25425 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
25426 higher gamma makes the spectrum having more range. Default value is @code{3}.
25427 Acceptable range is @code{[1, 7]}.
25429 @item bar_g, gamma2
25430 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
25434 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
25435 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
25437 @item timeclamp, tc
25438 Specify the transform timeclamp. At low frequency, there is trade-off between
25439 accuracy in time domain and frequency domain. If timeclamp is lower,
25440 event in time domain is represented more accurately (such as fast bass drum),
25441 otherwise event in frequency domain is represented more accurately
25442 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
25445 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
25446 limits future samples by applying asymmetric windowing in time domain, useful
25447 when low latency is required. Accepted range is @code{[0, 1]}.
25450 Specify the transform base frequency. Default value is @code{20.01523126408007475},
25451 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
25454 Specify the transform end frequency. Default value is @code{20495.59681441799654},
25455 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
25458 This option is deprecated and ignored.
25461 Specify the transform length in time domain. Use this option to control accuracy
25462 trade-off between time domain and frequency domain at every frequency sample.
25463 It can contain variables:
25465 @item frequency, freq, f
25466 the frequency where it is evaluated
25467 @item timeclamp, tc
25468 the value of @var{timeclamp} option.
25470 Default value is @code{384*tc/(384+tc*f)}.
25473 Specify the transform count for every video frame. Default value is @code{6}.
25474 Acceptable range is @code{[1, 30]}.
25477 Specify the transform count for every single pixel. Default value is @code{0},
25478 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
25481 Specify font file for use with freetype to draw the axis. If not specified,
25482 use embedded font. Note that drawing with font file or embedded font is not
25483 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
25487 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
25488 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
25492 Specify font color expression. This is arithmetic expression that should return
25493 integer value 0xRRGGBB. It can contain variables:
25495 @item frequency, freq, f
25496 the frequency where it is evaluated
25497 @item timeclamp, tc
25498 the value of @var{timeclamp} option
25503 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
25504 @item r(x), g(x), b(x)
25505 red, green, and blue value of intensity x.
25507 Default value is @code{st(0, (midi(f)-59.5)/12);
25508 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
25509 r(1-ld(1)) + b(ld(1))}.
25512 Specify image file to draw the axis. This option override @var{fontfile} and
25513 @var{fontcolor} option.
25516 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
25517 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
25518 Default value is @code{1}.
25521 Set colorspace. The accepted values are:
25524 Unspecified (default)
25533 BT.470BG or BT.601-6 625
25536 SMPTE-170M or BT.601-6 525
25542 BT.2020 with non-constant luminance
25547 Set spectrogram color scheme. This is list of floating point values with format
25548 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
25549 The default is @code{1|0.5|0|0|0.5|1}.
25553 @subsection Examples
25557 Playing audio while showing the spectrum:
25559 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
25563 Same as above, but with frame rate 30 fps:
25565 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
25569 Playing at 1280x720:
25571 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
25575 Disable sonogram display:
25581 A1 and its harmonics: A1, A2, (near)E3, A3:
25583 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),
25584 asplit[a][out1]; [a] showcqt [out0]'
25588 Same as above, but with more accuracy in frequency domain:
25590 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),
25591 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
25597 bar_v=10:sono_v=bar_v*a_weighting(f)
25601 Custom gamma, now spectrum is linear to the amplitude.
25607 Custom tlength equation:
25609 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)))'
25613 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
25615 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
25619 Custom font using fontconfig:
25621 font='Courier New,Monospace,mono|bold'
25625 Custom frequency range with custom axis using image file:
25627 axisfile=myaxis.png:basefreq=40:endfreq=10000
25633 Convert input audio to video output representing the audio power spectrum.
25634 Audio amplitude is on Y-axis while frequency is on X-axis.
25636 The filter accepts the following options:
25640 Specify size of video. For the syntax of this option, check the
25641 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25642 Default is @code{1024x512}.
25646 This set how each frequency bin will be represented.
25648 It accepts the following values:
25654 Default is @code{bar}.
25657 Set amplitude scale.
25659 It accepts the following values:
25673 Default is @code{log}.
25676 Set frequency scale.
25678 It accepts the following values:
25687 Reverse logarithmic scale.
25689 Default is @code{lin}.
25692 Set window size. Allowed range is from 16 to 65536.
25694 Default is @code{2048}
25697 Set windowing function.
25699 It accepts the following values:
25722 Default is @code{hanning}.
25725 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25726 which means optimal overlap for selected window function will be picked.
25729 Set time averaging. Setting this to 0 will display current maximal peaks.
25730 Default is @code{1}, which means time averaging is disabled.
25733 Specify list of colors separated by space or by '|' which will be used to
25734 draw channel frequencies. Unrecognized or missing colors will be replaced
25738 Set channel display mode.
25740 It accepts the following values:
25745 Default is @code{combined}.
25748 Set minimum amplitude used in @code{log} amplitude scaler.
25751 Set data display mode.
25753 It accepts the following values:
25759 Default is @code{magnitude}.
25762 @section showspatial
25764 Convert stereo input audio to a video output, representing the spatial relationship
25765 between two channels.
25767 The filter accepts the following options:
25771 Specify the video size for the output. For the syntax of this option, check the
25772 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25773 Default value is @code{512x512}.
25776 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
25779 Set window function.
25781 It accepts the following values:
25806 Default value is @code{hann}.
25809 Set ratio of overlap window. Default value is @code{0.5}.
25810 When value is @code{1} overlap is set to recommended size for specific
25811 window function currently used.
25814 @anchor{showspectrum}
25815 @section showspectrum
25817 Convert input audio to a video output, representing the audio frequency
25820 The filter accepts the following options:
25824 Specify the video size for the output. For the syntax of this option, check the
25825 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25826 Default value is @code{640x512}.
25829 Specify how the spectrum should slide along the window.
25831 It accepts the following values:
25834 the samples start again on the left when they reach the right
25836 the samples scroll from right to left
25838 frames are only produced when the samples reach the right
25840 the samples scroll from left to right
25843 Default value is @code{replace}.
25846 Specify display mode.
25848 It accepts the following values:
25851 all channels are displayed in the same row
25853 all channels are displayed in separate rows
25856 Default value is @samp{combined}.
25859 Specify display color mode.
25861 It accepts the following values:
25864 each channel is displayed in a separate color
25866 each channel is displayed using the same color scheme
25868 each channel is displayed using the rainbow color scheme
25870 each channel is displayed using the moreland color scheme
25872 each channel is displayed using the nebulae color scheme
25874 each channel is displayed using the fire color scheme
25876 each channel is displayed using the fiery color scheme
25878 each channel is displayed using the fruit color scheme
25880 each channel is displayed using the cool color scheme
25882 each channel is displayed using the magma color scheme
25884 each channel is displayed using the green color scheme
25886 each channel is displayed using the viridis color scheme
25888 each channel is displayed using the plasma color scheme
25890 each channel is displayed using the cividis color scheme
25892 each channel is displayed using the terrain color scheme
25895 Default value is @samp{channel}.
25898 Specify scale used for calculating intensity color values.
25900 It accepts the following values:
25905 square root, default
25916 Default value is @samp{sqrt}.
25919 Specify frequency scale.
25921 It accepts the following values:
25929 Default value is @samp{lin}.
25932 Set saturation modifier for displayed colors. Negative values provide
25933 alternative color scheme. @code{0} is no saturation at all.
25934 Saturation must be in [-10.0, 10.0] range.
25935 Default value is @code{1}.
25938 Set window function.
25940 It accepts the following values:
25965 Default value is @code{hann}.
25968 Set orientation of time vs frequency axis. Can be @code{vertical} or
25969 @code{horizontal}. Default is @code{vertical}.
25972 Set ratio of overlap window. Default value is @code{0}.
25973 When value is @code{1} overlap is set to recommended size for specific
25974 window function currently used.
25977 Set scale gain for calculating intensity color values.
25978 Default value is @code{1}.
25981 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
25984 Set color rotation, must be in [-1.0, 1.0] range.
25985 Default value is @code{0}.
25988 Set start frequency from which to display spectrogram. Default is @code{0}.
25991 Set stop frequency to which to display spectrogram. Default is @code{0}.
25994 Set upper frame rate limit. Default is @code{auto}, unlimited.
25997 Draw time and frequency axes and legends. Default is disabled.
26000 The usage is very similar to the showwaves filter; see the examples in that
26003 @subsection Examples
26007 Large window with logarithmic color scaling:
26009 showspectrum=s=1280x480:scale=log
26013 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
26015 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
26016 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
26020 @section showspectrumpic
26022 Convert input audio to a single video frame, representing the audio frequency
26025 The filter accepts the following options:
26029 Specify the video size for the output. For the syntax of this option, check the
26030 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26031 Default value is @code{4096x2048}.
26034 Specify display mode.
26036 It accepts the following values:
26039 all channels are displayed in the same row
26041 all channels are displayed in separate rows
26043 Default value is @samp{combined}.
26046 Specify display color mode.
26048 It accepts the following values:
26051 each channel is displayed in a separate color
26053 each channel is displayed using the same color scheme
26055 each channel is displayed using the rainbow color scheme
26057 each channel is displayed using the moreland color scheme
26059 each channel is displayed using the nebulae color scheme
26061 each channel is displayed using the fire color scheme
26063 each channel is displayed using the fiery color scheme
26065 each channel is displayed using the fruit color scheme
26067 each channel is displayed using the cool color scheme
26069 each channel is displayed using the magma color scheme
26071 each channel is displayed using the green color scheme
26073 each channel is displayed using the viridis color scheme
26075 each channel is displayed using the plasma color scheme
26077 each channel is displayed using the cividis color scheme
26079 each channel is displayed using the terrain color scheme
26081 Default value is @samp{intensity}.
26084 Specify scale used for calculating intensity color values.
26086 It accepts the following values:
26091 square root, default
26101 Default value is @samp{log}.
26104 Specify frequency scale.
26106 It accepts the following values:
26114 Default value is @samp{lin}.
26117 Set saturation modifier for displayed colors. Negative values provide
26118 alternative color scheme. @code{0} is no saturation at all.
26119 Saturation must be in [-10.0, 10.0] range.
26120 Default value is @code{1}.
26123 Set window function.
26125 It accepts the following values:
26149 Default value is @code{hann}.
26152 Set orientation of time vs frequency axis. Can be @code{vertical} or
26153 @code{horizontal}. Default is @code{vertical}.
26156 Set scale gain for calculating intensity color values.
26157 Default value is @code{1}.
26160 Draw time and frequency axes and legends. Default is enabled.
26163 Set color rotation, must be in [-1.0, 1.0] range.
26164 Default value is @code{0}.
26167 Set start frequency from which to display spectrogram. Default is @code{0}.
26170 Set stop frequency to which to display spectrogram. Default is @code{0}.
26173 @subsection Examples
26177 Extract an audio spectrogram of a whole audio track
26178 in a 1024x1024 picture using @command{ffmpeg}:
26180 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
26184 @section showvolume
26186 Convert input audio volume to a video output.
26188 The filter accepts the following options:
26195 Set border width, allowed range is [0, 5]. Default is 1.
26198 Set channel width, allowed range is [80, 8192]. Default is 400.
26201 Set channel height, allowed range is [1, 900]. Default is 20.
26204 Set fade, allowed range is [0, 1]. Default is 0.95.
26207 Set volume color expression.
26209 The expression can use the following variables:
26213 Current max volume of channel in dB.
26219 Current channel number, starting from 0.
26223 If set, displays channel names. Default is enabled.
26226 If set, displays volume values. Default is enabled.
26229 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
26230 default is @code{h}.
26233 Set step size, allowed range is [0, 5]. Default is 0, which means
26237 Set background opacity, allowed range is [0, 1]. Default is 0.
26240 Set metering mode, can be peak: @code{p} or rms: @code{r},
26241 default is @code{p}.
26244 Set display scale, can be linear: @code{lin} or log: @code{log},
26245 default is @code{lin}.
26249 If set to > 0., display a line for the max level
26250 in the previous seconds.
26251 default is disabled: @code{0.}
26254 The color of the max line. Use when @code{dm} option is set to > 0.
26255 default is: @code{orange}
26260 Convert input audio to a video output, representing the samples waves.
26262 The filter accepts the following options:
26266 Specify the video size for the output. For the syntax of this option, check the
26267 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26268 Default value is @code{600x240}.
26273 Available values are:
26276 Draw a point for each sample.
26279 Draw a vertical line for each sample.
26282 Draw a point for each sample and a line between them.
26285 Draw a centered vertical line for each sample.
26288 Default value is @code{point}.
26291 Set the number of samples which are printed on the same column. A
26292 larger value will decrease the frame rate. Must be a positive
26293 integer. This option can be set only if the value for @var{rate}
26294 is not explicitly specified.
26297 Set the (approximate) output frame rate. This is done by setting the
26298 option @var{n}. Default value is "25".
26300 @item split_channels
26301 Set if channels should be drawn separately or overlap. Default value is 0.
26304 Set colors separated by '|' which are going to be used for drawing of each channel.
26307 Set amplitude scale.
26309 Available values are:
26327 Set the draw mode. This is mostly useful to set for high @var{n}.
26329 Available values are:
26332 Scale pixel values for each drawn sample.
26335 Draw every sample directly.
26338 Default value is @code{scale}.
26341 @subsection Examples
26345 Output the input file audio and the corresponding video representation
26348 amovie=a.mp3,asplit[out0],showwaves[out1]
26352 Create a synthetic signal and show it with showwaves, forcing a
26353 frame rate of 30 frames per second:
26355 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
26359 @section showwavespic
26361 Convert input audio to a single video frame, representing the samples waves.
26363 The filter accepts the following options:
26367 Specify the video size for the output. For the syntax of this option, check the
26368 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26369 Default value is @code{600x240}.
26371 @item split_channels
26372 Set if channels should be drawn separately or overlap. Default value is 0.
26375 Set colors separated by '|' which are going to be used for drawing of each channel.
26378 Set amplitude scale.
26380 Available values are:
26400 Available values are:
26403 Scale pixel values for each drawn sample.
26406 Draw every sample directly.
26409 Default value is @code{scale}.
26412 Set the filter mode.
26414 Available values are:
26417 Use average samples values for each drawn sample.
26420 Use peak samples values for each drawn sample.
26423 Default value is @code{average}.
26426 @subsection Examples
26430 Extract a channel split representation of the wave form of a whole audio track
26431 in a 1024x800 picture using @command{ffmpeg}:
26433 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
26437 @section sidedata, asidedata
26439 Delete frame side data, or select frames based on it.
26441 This filter accepts the following options:
26445 Set mode of operation of the filter.
26447 Can be one of the following:
26451 Select every frame with side data of @code{type}.
26454 Delete side data of @code{type}. If @code{type} is not set, delete all side
26460 Set side data type used with all modes. Must be set for @code{select} mode. For
26461 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
26462 in @file{libavutil/frame.h}. For example, to choose
26463 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
26467 @section spectrumsynth
26469 Synthesize audio from 2 input video spectrums, first input stream represents
26470 magnitude across time and second represents phase across time.
26471 The filter will transform from frequency domain as displayed in videos back
26472 to time domain as presented in audio output.
26474 This filter is primarily created for reversing processed @ref{showspectrum}
26475 filter outputs, but can synthesize sound from other spectrograms too.
26476 But in such case results are going to be poor if the phase data is not
26477 available, because in such cases phase data need to be recreated, usually
26478 it's just recreated from random noise.
26479 For best results use gray only output (@code{channel} color mode in
26480 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
26481 @code{lin} scale for phase video. To produce phase, for 2nd video, use
26482 @code{data} option. Inputs videos should generally use @code{fullframe}
26483 slide mode as that saves resources needed for decoding video.
26485 The filter accepts the following options:
26489 Specify sample rate of output audio, the sample rate of audio from which
26490 spectrum was generated may differ.
26493 Set number of channels represented in input video spectrums.
26496 Set scale which was used when generating magnitude input spectrum.
26497 Can be @code{lin} or @code{log}. Default is @code{log}.
26500 Set slide which was used when generating inputs spectrums.
26501 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
26502 Default is @code{fullframe}.
26505 Set window function used for resynthesis.
26508 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
26509 which means optimal overlap for selected window function will be picked.
26512 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
26513 Default is @code{vertical}.
26516 @subsection Examples
26520 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
26521 then resynthesize videos back to audio with spectrumsynth:
26523 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
26524 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
26525 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
26529 @section split, asplit
26531 Split input into several identical outputs.
26533 @code{asplit} works with audio input, @code{split} with video.
26535 The filter accepts a single parameter which specifies the number of outputs. If
26536 unspecified, it defaults to 2.
26538 @subsection Examples
26542 Create two separate outputs from the same input:
26544 [in] split [out0][out1]
26548 To create 3 or more outputs, you need to specify the number of
26551 [in] asplit=3 [out0][out1][out2]
26555 Create two separate outputs from the same input, one cropped and
26558 [in] split [splitout1][splitout2];
26559 [splitout1] crop=100:100:0:0 [cropout];
26560 [splitout2] pad=200:200:100:100 [padout];
26564 Create 5 copies of the input audio with @command{ffmpeg}:
26566 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
26572 Receive commands sent through a libzmq client, and forward them to
26573 filters in the filtergraph.
26575 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
26576 must be inserted between two video filters, @code{azmq} between two
26577 audio filters. Both are capable to send messages to any filter type.
26579 To enable these filters you need to install the libzmq library and
26580 headers and configure FFmpeg with @code{--enable-libzmq}.
26582 For more information about libzmq see:
26583 @url{http://www.zeromq.org/}
26585 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
26586 receives messages sent through a network interface defined by the
26587 @option{bind_address} (or the abbreviation "@option{b}") option.
26588 Default value of this option is @file{tcp://localhost:5555}. You may
26589 want to alter this value to your needs, but do not forget to escape any
26590 ':' signs (see @ref{filtergraph escaping}).
26592 The received message must be in the form:
26594 @var{TARGET} @var{COMMAND} [@var{ARG}]
26597 @var{TARGET} specifies the target of the command, usually the name of
26598 the filter class or a specific filter instance name. The default
26599 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
26600 but you can override this by using the @samp{filter_name@@id} syntax
26601 (see @ref{Filtergraph syntax}).
26603 @var{COMMAND} specifies the name of the command for the target filter.
26605 @var{ARG} is optional and specifies the optional argument list for the
26606 given @var{COMMAND}.
26608 Upon reception, the message is processed and the corresponding command
26609 is injected into the filtergraph. Depending on the result, the filter
26610 will send a reply to the client, adopting the format:
26612 @var{ERROR_CODE} @var{ERROR_REASON}
26616 @var{MESSAGE} is optional.
26618 @subsection Examples
26620 Look at @file{tools/zmqsend} for an example of a zmq client which can
26621 be used to send commands processed by these filters.
26623 Consider the following filtergraph generated by @command{ffplay}.
26624 In this example the last overlay filter has an instance name. All other
26625 filters will have default instance names.
26628 ffplay -dumpgraph 1 -f lavfi "
26629 color=s=100x100:c=red [l];
26630 color=s=100x100:c=blue [r];
26631 nullsrc=s=200x100, zmq [bg];
26632 [bg][l] overlay [bg+l];
26633 [bg+l][r] overlay@@my=x=100 "
26636 To change the color of the left side of the video, the following
26637 command can be used:
26639 echo Parsed_color_0 c yellow | tools/zmqsend
26642 To change the right side:
26644 echo Parsed_color_1 c pink | tools/zmqsend
26647 To change the position of the right side:
26649 echo overlay@@my x 150 | tools/zmqsend
26653 @c man end MULTIMEDIA FILTERS
26655 @chapter Multimedia Sources
26656 @c man begin MULTIMEDIA SOURCES
26658 Below is a description of the currently available multimedia sources.
26662 This is the same as @ref{movie} source, except it selects an audio
26668 Read audio and/or video stream(s) from a movie container.
26670 It accepts the following parameters:
26674 The name of the resource to read (not necessarily a file; it can also be a
26675 device or a stream accessed through some protocol).
26677 @item format_name, f
26678 Specifies the format assumed for the movie to read, and can be either
26679 the name of a container or an input device. If not specified, the
26680 format is guessed from @var{movie_name} or by probing.
26682 @item seek_point, sp
26683 Specifies the seek point in seconds. The frames will be output
26684 starting from this seek point. The parameter is evaluated with
26685 @code{av_strtod}, so the numerical value may be suffixed by an IS
26686 postfix. The default value is "0".
26689 Specifies the streams to read. Several streams can be specified,
26690 separated by "+". The source will then have as many outputs, in the
26691 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
26692 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
26693 respectively the default (best suited) video and audio stream. Default
26694 is "dv", or "da" if the filter is called as "amovie".
26696 @item stream_index, si
26697 Specifies the index of the video stream to read. If the value is -1,
26698 the most suitable video stream will be automatically selected. The default
26699 value is "-1". Deprecated. If the filter is called "amovie", it will select
26700 audio instead of video.
26703 Specifies how many times to read the stream in sequence.
26704 If the value is 0, the stream will be looped infinitely.
26705 Default value is "1".
26707 Note that when the movie is looped the source timestamps are not
26708 changed, so it will generate non monotonically increasing timestamps.
26710 @item discontinuity
26711 Specifies the time difference between frames above which the point is
26712 considered a timestamp discontinuity which is removed by adjusting the later
26716 It allows overlaying a second video on top of the main input of
26717 a filtergraph, as shown in this graph:
26719 input -----------> deltapts0 --> overlay --> output
26722 movie --> scale--> deltapts1 -------+
26724 @subsection Examples
26728 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
26729 on top of the input labelled "in":
26731 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
26732 [in] setpts=PTS-STARTPTS [main];
26733 [main][over] overlay=16:16 [out]
26737 Read from a video4linux2 device, and overlay it on top of the input
26740 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
26741 [in] setpts=PTS-STARTPTS [main];
26742 [main][over] overlay=16:16 [out]
26746 Read the first video stream and the audio stream with id 0x81 from
26747 dvd.vob; the video is connected to the pad named "video" and the audio is
26748 connected to the pad named "audio":
26750 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
26754 @subsection Commands
26756 Both movie and amovie support the following commands:
26759 Perform seek using "av_seek_frame".
26760 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
26763 @var{stream_index}: If stream_index is -1, a default
26764 stream is selected, and @var{timestamp} is automatically converted
26765 from AV_TIME_BASE units to the stream specific time_base.
26767 @var{timestamp}: Timestamp in AVStream.time_base units
26768 or, if no stream is specified, in AV_TIME_BASE units.
26770 @var{flags}: Flags which select direction and seeking mode.
26774 Get movie duration in AV_TIME_BASE units.
26778 @c man end MULTIMEDIA SOURCES