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.
7103 Set sigma for 1st plane, 2nd plane or 3rd plane. Default is 32767.
7104 Valid range is from 0 to 32767.
7105 This options controls weight for each pixel in radius defined by size.
7106 Default value means every pixel have same weight.
7107 Setting this option to 0 effectively disables filtering.
7110 @subsection Commands
7111 This filter supports same @ref{commands} as options except option @code{s}.
7112 The command accepts the same syntax of the corresponding option.
7116 Apply average blur filter.
7118 The filter accepts the following options:
7122 Set horizontal radius size.
7125 Set which planes to filter. By default all planes are filtered.
7128 Set vertical radius size, if zero it will be same as @code{sizeX}.
7129 Default is @code{0}.
7132 @subsection Commands
7133 This filter supports same commands as options.
7134 The command accepts the same syntax of the corresponding option.
7136 If the specified expression is not valid, it is kept at its current
7141 Compute the bounding box for the non-black pixels in the input frame
7144 This filter computes the bounding box containing all the pixels with a
7145 luminance value greater than the minimum allowed value.
7146 The parameters describing the bounding box are printed on the filter
7149 The filter accepts the following option:
7153 Set the minimal luminance value. Default is @code{16}.
7156 @subsection Commands
7158 This filter supports the all above options as @ref{commands}.
7161 Apply bilateral filter, spatial smoothing while preserving edges.
7163 The filter accepts the following options:
7166 Set sigma of gaussian function to calculate spatial weight.
7167 Allowed range is 0 to 512. Default is 0.1.
7170 Set sigma of gaussian function to calculate range weight.
7171 Allowed range is 0 to 1. Default is 0.1.
7174 Set planes to filter. Default is first only.
7177 @subsection Commands
7179 This filter supports the all above options as @ref{commands}.
7181 @section bitplanenoise
7183 Show and measure bit plane noise.
7185 The filter accepts the following options:
7189 Set which plane to analyze. Default is @code{1}.
7192 Filter out noisy pixels from @code{bitplane} set above.
7193 Default is disabled.
7196 @section blackdetect
7198 Detect video intervals that are (almost) completely black. Can be
7199 useful to detect chapter transitions, commercials, or invalid
7202 The filter outputs its detection analysis to both the log as well as
7203 frame metadata. If a black segment of at least the specified minimum
7204 duration is found, a line with the start and end timestamps as well
7205 as duration is printed to the log with level @code{info}. In addition,
7206 a log line with level @code{debug} is printed per frame showing the
7207 black amount detected for that frame.
7209 The filter also attaches metadata to the first frame of a black
7210 segment with key @code{lavfi.black_start} and to the first frame
7211 after the black segment ends with key @code{lavfi.black_end}. The
7212 value is the frame's timestamp. This metadata is added regardless
7213 of the minimum duration specified.
7215 The filter accepts the following options:
7218 @item black_min_duration, d
7219 Set the minimum detected black duration expressed in seconds. It must
7220 be a non-negative floating point number.
7222 Default value is 2.0.
7224 @item picture_black_ratio_th, pic_th
7225 Set the threshold for considering a picture "black".
7226 Express the minimum value for the ratio:
7228 @var{nb_black_pixels} / @var{nb_pixels}
7231 for which a picture is considered black.
7232 Default value is 0.98.
7234 @item pixel_black_th, pix_th
7235 Set the threshold for considering a pixel "black".
7237 The threshold expresses the maximum pixel luminance value for which a
7238 pixel is considered "black". The provided value is scaled according to
7239 the following equation:
7241 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
7244 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
7245 the input video format, the range is [0-255] for YUV full-range
7246 formats and [16-235] for YUV non full-range formats.
7248 Default value is 0.10.
7251 The following example sets the maximum pixel threshold to the minimum
7252 value, and detects only black intervals of 2 or more seconds:
7254 blackdetect=d=2:pix_th=0.00
7259 Detect frames that are (almost) completely black. Can be useful to
7260 detect chapter transitions or commercials. Output lines consist of
7261 the frame number of the detected frame, the percentage of blackness,
7262 the position in the file if known or -1 and the timestamp in seconds.
7264 In order to display the output lines, you need to set the loglevel at
7265 least to the AV_LOG_INFO value.
7267 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
7268 The value represents the percentage of pixels in the picture that
7269 are below the threshold value.
7271 It accepts the following parameters:
7276 The percentage of the pixels that have to be below the threshold; it defaults to
7279 @item threshold, thresh
7280 The threshold below which a pixel value is considered black; it defaults to
7288 Blend two video frames into each other.
7290 The @code{blend} filter takes two input streams and outputs one
7291 stream, the first input is the "top" layer and second input is
7292 "bottom" layer. By default, the output terminates when the longest input terminates.
7294 The @code{tblend} (time blend) filter takes two consecutive frames
7295 from one single stream, and outputs the result obtained by blending
7296 the new frame on top of the old frame.
7298 A description of the accepted options follows.
7306 Set blend mode for specific pixel component or all pixel components in case
7307 of @var{all_mode}. Default value is @code{normal}.
7309 Available values for component modes are:
7351 Set blend opacity for specific pixel component or all pixel components in case
7352 of @var{all_opacity}. Only used in combination with pixel component blend modes.
7359 Set blend expression for specific pixel component or all pixel components in case
7360 of @var{all_expr}. Note that related mode options will be ignored if those are set.
7362 The expressions can use the following variables:
7366 The sequential number of the filtered frame, starting from @code{0}.
7370 the coordinates of the current sample
7374 the width and height of currently filtered plane
7378 Width and height scale for the plane being filtered. It is the
7379 ratio between the dimensions of the current plane to the luma plane,
7380 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
7381 the luma plane and @code{0.5,0.5} for the chroma planes.
7384 Time of the current frame, expressed in seconds.
7387 Value of pixel component at current location for first video frame (top layer).
7390 Value of pixel component at current location for second video frame (bottom layer).
7394 The @code{blend} filter also supports the @ref{framesync} options.
7396 @subsection Examples
7400 Apply transition from bottom layer to top layer in first 10 seconds:
7402 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
7406 Apply linear horizontal transition from top layer to bottom layer:
7408 blend=all_expr='A*(X/W)+B*(1-X/W)'
7412 Apply 1x1 checkerboard effect:
7414 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
7418 Apply uncover left effect:
7420 blend=all_expr='if(gte(N*SW+X,W),A,B)'
7424 Apply uncover down effect:
7426 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
7430 Apply uncover up-left effect:
7432 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
7436 Split diagonally video and shows top and bottom layer on each side:
7438 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
7442 Display differences between the current and the previous frame:
7444 tblend=all_mode=grainextract
7450 Denoise frames using Block-Matching 3D algorithm.
7452 The filter accepts the following options.
7456 Set denoising strength. Default value is 1.
7457 Allowed range is from 0 to 999.9.
7458 The denoising algorithm is very sensitive to sigma, so adjust it
7459 according to the source.
7462 Set local patch size. This sets dimensions in 2D.
7465 Set sliding step for processing blocks. Default value is 4.
7466 Allowed range is from 1 to 64.
7467 Smaller values allows processing more reference blocks and is slower.
7470 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
7471 When set to 1, no block matching is done. Larger values allows more blocks
7473 Allowed range is from 1 to 256.
7476 Set radius for search block matching. Default is 9.
7477 Allowed range is from 1 to INT32_MAX.
7480 Set step between two search locations for block matching. Default is 1.
7481 Allowed range is from 1 to 64. Smaller is slower.
7484 Set threshold of mean square error for block matching. Valid range is 0 to
7488 Set thresholding parameter for hard thresholding in 3D transformed domain.
7489 Larger values results in stronger hard-thresholding filtering in frequency
7493 Set filtering estimation mode. Can be @code{basic} or @code{final}.
7494 Default is @code{basic}.
7497 If enabled, filter will use 2nd stream for block matching.
7498 Default is disabled for @code{basic} value of @var{estim} option,
7499 and always enabled if value of @var{estim} is @code{final}.
7502 Set planes to filter. Default is all available except alpha.
7505 @subsection Examples
7509 Basic filtering with bm3d:
7511 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
7515 Same as above, but filtering only luma:
7517 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
7521 Same as above, but with both estimation modes:
7523 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
7527 Same as above, but prefilter with @ref{nlmeans} filter instead:
7529 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
7535 Apply a boxblur algorithm to the input video.
7537 It accepts the following parameters:
7541 @item luma_radius, lr
7542 @item luma_power, lp
7543 @item chroma_radius, cr
7544 @item chroma_power, cp
7545 @item alpha_radius, ar
7546 @item alpha_power, ap
7550 A description of the accepted options follows.
7553 @item luma_radius, lr
7554 @item chroma_radius, cr
7555 @item alpha_radius, ar
7556 Set an expression for the box radius in pixels used for blurring the
7557 corresponding input plane.
7559 The radius value must be a non-negative number, and must not be
7560 greater than the value of the expression @code{min(w,h)/2} for the
7561 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
7564 Default value for @option{luma_radius} is "2". If not specified,
7565 @option{chroma_radius} and @option{alpha_radius} default to the
7566 corresponding value set for @option{luma_radius}.
7568 The expressions can contain the following constants:
7572 The input width and height in pixels.
7576 The input chroma image width and height in pixels.
7580 The horizontal and vertical chroma subsample values. For example, for the
7581 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
7584 @item luma_power, lp
7585 @item chroma_power, cp
7586 @item alpha_power, ap
7587 Specify how many times the boxblur filter is applied to the
7588 corresponding plane.
7590 Default value for @option{luma_power} is 2. If not specified,
7591 @option{chroma_power} and @option{alpha_power} default to the
7592 corresponding value set for @option{luma_power}.
7594 A value of 0 will disable the effect.
7597 @subsection Examples
7601 Apply a boxblur filter with the luma, chroma, and alpha radii
7604 boxblur=luma_radius=2:luma_power=1
7609 Set the luma radius to 2, and alpha and chroma radius to 0:
7611 boxblur=2:1:cr=0:ar=0
7615 Set the luma and chroma radii to a fraction of the video dimension:
7617 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
7623 Deinterlace the input video ("bwdif" stands for "Bob Weaver
7624 Deinterlacing Filter").
7626 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
7627 interpolation algorithms.
7628 It accepts the following parameters:
7632 The interlacing mode to adopt. It accepts one of the following values:
7636 Output one frame for each frame.
7638 Output one frame for each field.
7641 The default value is @code{send_field}.
7644 The picture field parity assumed for the input interlaced video. It accepts one
7645 of the following values:
7649 Assume the top field is first.
7651 Assume the bottom field is first.
7653 Enable automatic detection of field parity.
7656 The default value is @code{auto}.
7657 If the interlacing is unknown or the decoder does not export this information,
7658 top field first will be assumed.
7661 Specify which frames to deinterlace. Accepts one of the following
7666 Deinterlace all frames.
7668 Only deinterlace frames marked as interlaced.
7671 The default value is @code{all}.
7676 Apply Contrast Adaptive Sharpen filter to video stream.
7678 The filter accepts the following options:
7682 Set the sharpening strength. Default value is 0.
7685 Set planes to filter. Default value is to filter all
7686 planes except alpha plane.
7689 @subsection Commands
7690 This filter supports same @ref{commands} as options.
7693 Remove all color information for all colors except for certain one.
7695 The filter accepts the following options:
7699 The color which will not be replaced with neutral chroma.
7702 Similarity percentage with the above color.
7703 0.01 matches only the exact key color, while 1.0 matches everything.
7707 0.0 makes pixels either fully gray, or not gray at all.
7708 Higher values result in more preserved color.
7711 Signals that the color passed is already in YUV instead of RGB.
7713 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7714 This can be used to pass exact YUV values as hexadecimal numbers.
7717 @subsection Commands
7718 This filter supports same @ref{commands} as options.
7719 The command accepts the same syntax of the corresponding option.
7721 If the specified expression is not valid, it is kept at its current
7725 YUV colorspace color/chroma keying.
7727 The filter accepts the following options:
7731 The color which will be replaced with transparency.
7734 Similarity percentage with the key color.
7736 0.01 matches only the exact key color, while 1.0 matches everything.
7741 0.0 makes pixels either fully transparent, or not transparent at all.
7743 Higher values result in semi-transparent pixels, with a higher transparency
7744 the more similar the pixels color is to the key color.
7747 Signals that the color passed is already in YUV instead of RGB.
7749 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7750 This can be used to pass exact YUV values as hexadecimal numbers.
7753 @subsection Commands
7754 This filter supports same @ref{commands} as options.
7755 The command accepts the same syntax of the corresponding option.
7757 If the specified expression is not valid, it is kept at its current
7760 @subsection Examples
7764 Make every green pixel in the input image transparent:
7766 ffmpeg -i input.png -vf chromakey=green out.png
7770 Overlay a greenscreen-video on top of a static black background.
7772 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
7777 Reduce chrominance noise.
7779 The filter accepts the following options:
7783 Set threshold for averaging chrominance values.
7784 Sum of absolute difference of Y, U and V pixel components of current
7785 pixel and neighbour pixels lower than this threshold will be used in
7786 averaging. Luma component is left unchanged and is copied to output.
7787 Default value is 30. Allowed range is from 1 to 200.
7790 Set horizontal radius of rectangle used for averaging.
7791 Allowed range is from 1 to 100. Default value is 5.
7794 Set vertical radius of rectangle used for averaging.
7795 Allowed range is from 1 to 100. Default value is 5.
7798 Set horizontal step when averaging. Default value is 1.
7799 Allowed range is from 1 to 50.
7800 Mostly useful to speed-up filtering.
7803 Set vertical step when averaging. Default value is 1.
7804 Allowed range is from 1 to 50.
7805 Mostly useful to speed-up filtering.
7808 Set Y threshold for averaging chrominance values.
7809 Set finer control for max allowed difference between Y components
7810 of current pixel and neigbour pixels.
7811 Default value is 200. Allowed range is from 1 to 200.
7814 Set U threshold for averaging chrominance values.
7815 Set finer control for max allowed difference between U components
7816 of current pixel and neigbour pixels.
7817 Default value is 200. Allowed range is from 1 to 200.
7820 Set V threshold for averaging chrominance values.
7821 Set finer control for max allowed difference between V components
7822 of current pixel and neigbour pixels.
7823 Default value is 200. Allowed range is from 1 to 200.
7826 @subsection Commands
7827 This filter supports same @ref{commands} as options.
7828 The command accepts the same syntax of the corresponding option.
7830 @section chromashift
7831 Shift chroma pixels horizontally and/or vertically.
7833 The filter accepts the following options:
7836 Set amount to shift chroma-blue horizontally.
7838 Set amount to shift chroma-blue vertically.
7840 Set amount to shift chroma-red horizontally.
7842 Set amount to shift chroma-red vertically.
7844 Set edge mode, can be @var{smear}, default, or @var{warp}.
7847 @subsection Commands
7849 This filter supports the all above options as @ref{commands}.
7853 Display CIE color diagram with pixels overlaid onto it.
7855 The filter accepts the following options:
7870 @item uhdtv, rec2020
7884 Set what gamuts to draw.
7886 See @code{system} option for available values.
7889 Set ciescope size, by default set to 512.
7892 Set intensity used to map input pixel values to CIE diagram.
7895 Set contrast used to draw tongue colors that are out of active color system gamut.
7898 Correct gamma displayed on scope, by default enabled.
7901 Show white point on CIE diagram, by default disabled.
7904 Set input gamma. Used only with XYZ input color space.
7909 Visualize information exported by some codecs.
7911 Some codecs can export information through frames using side-data or other
7912 means. For example, some MPEG based codecs export motion vectors through the
7913 @var{export_mvs} flag in the codec @option{flags2} option.
7915 The filter accepts the following option:
7919 Set motion vectors to visualize.
7921 Available flags for @var{mv} are:
7925 forward predicted MVs of P-frames
7927 forward predicted MVs of B-frames
7929 backward predicted MVs of B-frames
7933 Display quantization parameters using the chroma planes.
7936 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7938 Available flags for @var{mv_type} are:
7942 forward predicted MVs
7944 backward predicted MVs
7947 @item frame_type, ft
7948 Set frame type to visualize motion vectors of.
7950 Available flags for @var{frame_type} are:
7954 intra-coded frames (I-frames)
7956 predicted frames (P-frames)
7958 bi-directionally predicted frames (B-frames)
7962 @subsection Examples
7966 Visualize forward predicted MVs of all frames using @command{ffplay}:
7968 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7972 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7974 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7978 @section colorbalance
7979 Modify intensity of primary colors (red, green and blue) of input frames.
7981 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7982 regions for the red-cyan, green-magenta or blue-yellow balance.
7984 A positive adjustment value shifts the balance towards the primary color, a negative
7985 value towards the complementary color.
7987 The filter accepts the following options:
7993 Adjust red, green and blue shadows (darkest pixels).
7998 Adjust red, green and blue midtones (medium pixels).
8003 Adjust red, green and blue highlights (brightest pixels).
8005 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
8008 Preserve lightness when changing color balance. Default is disabled.
8011 @subsection Examples
8015 Add red color cast to shadows:
8021 @subsection Commands
8023 This filter supports the all above options as @ref{commands}.
8025 @section colorchannelmixer
8027 Adjust video input frames by re-mixing color channels.
8029 This filter modifies a color channel by adding the values associated to
8030 the other channels of the same pixels. For example if the value to
8031 modify is red, the output value will be:
8033 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
8036 The filter accepts the following options:
8043 Adjust contribution of input red, green, blue and alpha channels for output red channel.
8044 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
8050 Adjust contribution of input red, green, blue and alpha channels for output green channel.
8051 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
8057 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
8058 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
8064 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
8065 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
8067 Allowed ranges for options are @code{[-2.0, 2.0]}.
8070 Preserve lightness when changing colors. Default is disabled.
8073 @subsection Examples
8077 Convert source to grayscale:
8079 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
8082 Simulate sepia tones:
8084 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
8088 @subsection Commands
8090 This filter supports the all above options as @ref{commands}.
8093 RGB colorspace color keying.
8095 The filter accepts the following options:
8099 The color which will be replaced with transparency.
8102 Similarity percentage with the key color.
8104 0.01 matches only the exact key color, while 1.0 matches everything.
8109 0.0 makes pixels either fully transparent, or not transparent at all.
8111 Higher values result in semi-transparent pixels, with a higher transparency
8112 the more similar the pixels color is to the key color.
8115 @subsection Examples
8119 Make every green pixel in the input image transparent:
8121 ffmpeg -i input.png -vf colorkey=green out.png
8125 Overlay a greenscreen-video on top of a static background image.
8127 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
8131 @subsection Commands
8132 This filter supports same @ref{commands} as options.
8133 The command accepts the same syntax of the corresponding option.
8135 If the specified expression is not valid, it is kept at its current
8139 Remove all color information for all RGB colors except for certain one.
8141 The filter accepts the following options:
8145 The color which will not be replaced with neutral gray.
8148 Similarity percentage with the above color.
8149 0.01 matches only the exact key color, while 1.0 matches everything.
8152 Blend percentage. 0.0 makes pixels fully gray.
8153 Higher values result in more preserved color.
8156 @subsection Commands
8157 This filter supports same @ref{commands} as options.
8158 The command accepts the same syntax of the corresponding option.
8160 If the specified expression is not valid, it is kept at its current
8163 @section colorlevels
8165 Adjust video input frames using levels.
8167 The filter accepts the following options:
8174 Adjust red, green, blue and alpha input black point.
8175 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
8181 Adjust red, green, blue and alpha input white point.
8182 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
8184 Input levels are used to lighten highlights (bright tones), darken shadows
8185 (dark tones), change the balance of bright and dark tones.
8191 Adjust red, green, blue and alpha output black point.
8192 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
8198 Adjust red, green, blue and alpha output white point.
8199 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
8201 Output levels allows manual selection of a constrained output level range.
8204 @subsection Examples
8208 Make video output darker:
8210 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
8216 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
8220 Make video output lighter:
8222 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
8226 Increase brightness:
8228 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
8232 @subsection Commands
8234 This filter supports the all above options as @ref{commands}.
8236 @section colormatrix
8238 Convert color matrix.
8240 The filter accepts the following options:
8245 Specify the source and destination color matrix. Both values must be
8248 The accepted values are:
8276 For example to convert from BT.601 to SMPTE-240M, use the command:
8278 colormatrix=bt601:smpte240m
8283 Convert colorspace, transfer characteristics or color primaries.
8284 Input video needs to have an even size.
8286 The filter accepts the following options:
8291 Specify all color properties at once.
8293 The accepted values are:
8323 Specify output colorspace.
8325 The accepted values are:
8334 BT.470BG or BT.601-6 625
8337 SMPTE-170M or BT.601-6 525
8346 BT.2020 with non-constant luminance
8352 Specify output transfer characteristics.
8354 The accepted values are:
8366 Constant gamma of 2.2
8369 Constant gamma of 2.8
8372 SMPTE-170M, BT.601-6 625 or BT.601-6 525
8390 BT.2020 for 10-bits content
8393 BT.2020 for 12-bits content
8399 Specify output color primaries.
8401 The accepted values are:
8410 BT.470BG or BT.601-6 625
8413 SMPTE-170M or BT.601-6 525
8437 Specify output color range.
8439 The accepted values are:
8442 TV (restricted) range
8445 MPEG (restricted) range
8456 Specify output color format.
8458 The accepted values are:
8461 YUV 4:2:0 planar 8-bits
8464 YUV 4:2:0 planar 10-bits
8467 YUV 4:2:0 planar 12-bits
8470 YUV 4:2:2 planar 8-bits
8473 YUV 4:2:2 planar 10-bits
8476 YUV 4:2:2 planar 12-bits
8479 YUV 4:4:4 planar 8-bits
8482 YUV 4:4:4 planar 10-bits
8485 YUV 4:4:4 planar 12-bits
8490 Do a fast conversion, which skips gamma/primary correction. This will take
8491 significantly less CPU, but will be mathematically incorrect. To get output
8492 compatible with that produced by the colormatrix filter, use fast=1.
8495 Specify dithering mode.
8497 The accepted values are:
8503 Floyd-Steinberg dithering
8507 Whitepoint adaptation mode.
8509 The accepted values are:
8512 Bradford whitepoint adaptation
8515 von Kries whitepoint adaptation
8518 identity whitepoint adaptation (i.e. no whitepoint adaptation)
8522 Override all input properties at once. Same accepted values as @ref{all}.
8525 Override input colorspace. Same accepted values as @ref{space}.
8528 Override input color primaries. Same accepted values as @ref{primaries}.
8531 Override input transfer characteristics. Same accepted values as @ref{trc}.
8534 Override input color range. Same accepted values as @ref{range}.
8538 The filter converts the transfer characteristics, color space and color
8539 primaries to the specified user values. The output value, if not specified,
8540 is set to a default value based on the "all" property. If that property is
8541 also not specified, the filter will log an error. The output color range and
8542 format default to the same value as the input color range and format. The
8543 input transfer characteristics, color space, color primaries and color range
8544 should be set on the input data. If any of these are missing, the filter will
8545 log an error and no conversion will take place.
8547 For example to convert the input to SMPTE-240M, use the command:
8549 colorspace=smpte240m
8552 @section convolution
8554 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
8556 The filter accepts the following options:
8563 Set matrix for each plane.
8564 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
8565 and from 1 to 49 odd number of signed integers in @var{row} mode.
8571 Set multiplier for calculated value for each plane.
8572 If unset or 0, it will be sum of all matrix elements.
8578 Set bias for each plane. This value is added to the result of the multiplication.
8579 Useful for making the overall image brighter or darker. Default is 0.0.
8585 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
8586 Default is @var{square}.
8589 @subsection Commands
8591 This filter supports the all above options as @ref{commands}.
8593 @subsection Examples
8599 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"
8605 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"
8611 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"
8617 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"
8621 Apply laplacian edge detector which includes diagonals:
8623 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"
8629 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"
8635 Apply 2D convolution of video stream in frequency domain using second stream
8638 The filter accepts the following options:
8642 Set which planes to process.
8645 Set which impulse video frames will be processed, can be @var{first}
8646 or @var{all}. Default is @var{all}.
8649 The @code{convolve} filter also supports the @ref{framesync} options.
8653 Copy the input video source unchanged to the output. This is mainly useful for
8658 Video filtering on GPU using Apple's CoreImage API on OSX.
8660 Hardware acceleration is based on an OpenGL context. Usually, this means it is
8661 processed by video hardware. However, software-based OpenGL implementations
8662 exist which means there is no guarantee for hardware processing. It depends on
8665 There are many filters and image generators provided by Apple that come with a
8666 large variety of options. The filter has to be referenced by its name along
8669 The coreimage filter accepts the following options:
8672 List all available filters and generators along with all their respective
8673 options as well as possible minimum and maximum values along with the default
8680 Specify all filters by their respective name and options.
8681 Use @var{list_filters} to determine all valid filter names and options.
8682 Numerical options are specified by a float value and are automatically clamped
8683 to their respective value range. Vector and color options have to be specified
8684 by a list of space separated float values. Character escaping has to be done.
8685 A special option name @code{default} is available to use default options for a
8688 It is required to specify either @code{default} or at least one of the filter options.
8689 All omitted options are used with their default values.
8690 The syntax of the filter string is as follows:
8692 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
8696 Specify a rectangle where the output of the filter chain is copied into the
8697 input image. It is given by a list of space separated float values:
8699 output_rect=x\ y\ width\ height
8701 If not given, the output rectangle equals the dimensions of the input image.
8702 The output rectangle is automatically cropped at the borders of the input
8703 image. Negative values are valid for each component.
8705 output_rect=25\ 25\ 100\ 100
8709 Several filters can be chained for successive processing without GPU-HOST
8710 transfers allowing for fast processing of complex filter chains.
8711 Currently, only filters with zero (generators) or exactly one (filters) input
8712 image and one output image are supported. Also, transition filters are not yet
8715 Some filters generate output images with additional padding depending on the
8716 respective filter kernel. The padding is automatically removed to ensure the
8717 filter output has the same size as the input image.
8719 For image generators, the size of the output image is determined by the
8720 previous output image of the filter chain or the input image of the whole
8721 filterchain, respectively. The generators do not use the pixel information of
8722 this image to generate their output. However, the generated output is
8723 blended onto this image, resulting in partial or complete coverage of the
8726 The @ref{coreimagesrc} video source can be used for generating input images
8727 which are directly fed into the filter chain. By using it, providing input
8728 images by another video source or an input video is not required.
8730 @subsection Examples
8735 List all filters available:
8737 coreimage=list_filters=true
8741 Use the CIBoxBlur filter with default options to blur an image:
8743 coreimage=filter=CIBoxBlur@@default
8747 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8748 its center at 100x100 and a radius of 50 pixels:
8750 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8754 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8755 given as complete and escaped command-line for Apple's standard bash shell:
8757 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8763 Cover a rectangular object
8765 It accepts the following options:
8769 Filepath of the optional cover image, needs to be in yuv420.
8774 It accepts the following values:
8777 cover it by the supplied image
8779 cover it by interpolating the surrounding pixels
8782 Default value is @var{blur}.
8785 @subsection Examples
8789 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8791 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8797 Crop the input video to given dimensions.
8799 It accepts the following parameters:
8803 The width of the output video. It defaults to @code{iw}.
8804 This expression is evaluated only once during the filter
8805 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8808 The height of the output video. It defaults to @code{ih}.
8809 This expression is evaluated only once during the filter
8810 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8813 The horizontal position, in the input video, of the left edge of the output
8814 video. It defaults to @code{(in_w-out_w)/2}.
8815 This expression is evaluated per-frame.
8818 The vertical position, in the input video, of the top edge of the output video.
8819 It defaults to @code{(in_h-out_h)/2}.
8820 This expression is evaluated per-frame.
8823 If set to 1 will force the output display aspect ratio
8824 to be the same of the input, by changing the output sample aspect
8825 ratio. It defaults to 0.
8828 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8829 width/height/x/y as specified and will not be rounded to nearest smaller value.
8833 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8834 expressions containing the following constants:
8839 The computed values for @var{x} and @var{y}. They are evaluated for
8844 The input width and height.
8848 These are the same as @var{in_w} and @var{in_h}.
8852 The output (cropped) width and height.
8856 These are the same as @var{out_w} and @var{out_h}.
8859 same as @var{iw} / @var{ih}
8862 input sample aspect ratio
8865 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8869 horizontal and vertical chroma subsample values. For example for the
8870 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8873 The number of the input frame, starting from 0.
8876 the position in the file of the input frame, NAN if unknown
8879 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8883 The expression for @var{out_w} may depend on the value of @var{out_h},
8884 and the expression for @var{out_h} may depend on @var{out_w}, but they
8885 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8886 evaluated after @var{out_w} and @var{out_h}.
8888 The @var{x} and @var{y} parameters specify the expressions for the
8889 position of the top-left corner of the output (non-cropped) area. They
8890 are evaluated for each frame. If the evaluated value is not valid, it
8891 is approximated to the nearest valid value.
8893 The expression for @var{x} may depend on @var{y}, and the expression
8894 for @var{y} may depend on @var{x}.
8896 @subsection Examples
8900 Crop area with size 100x100 at position (12,34).
8905 Using named options, the example above becomes:
8907 crop=w=100:h=100:x=12:y=34
8911 Crop the central input area with size 100x100:
8917 Crop the central input area with size 2/3 of the input video:
8919 crop=2/3*in_w:2/3*in_h
8923 Crop the input video central square:
8930 Delimit the rectangle with the top-left corner placed at position
8931 100:100 and the right-bottom corner corresponding to the right-bottom
8932 corner of the input image.
8934 crop=in_w-100:in_h-100:100:100
8938 Crop 10 pixels from the left and right borders, and 20 pixels from
8939 the top and bottom borders
8941 crop=in_w-2*10:in_h-2*20
8945 Keep only the bottom right quarter of the input image:
8947 crop=in_w/2:in_h/2:in_w/2:in_h/2
8951 Crop height for getting Greek harmony:
8953 crop=in_w:1/PHI*in_w
8957 Apply trembling effect:
8959 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)
8963 Apply erratic camera effect depending on timestamp:
8965 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)"
8969 Set x depending on the value of y:
8971 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8975 @subsection Commands
8977 This filter supports the following commands:
8983 Set width/height of the output video and the horizontal/vertical position
8985 The command accepts the same syntax of the corresponding option.
8987 If the specified expression is not valid, it is kept at its current
8993 Auto-detect the crop size.
8995 It calculates the necessary cropping parameters and prints the
8996 recommended parameters via the logging system. The detected dimensions
8997 correspond to the non-black area of the input video.
8999 It accepts the following parameters:
9004 Set higher black value threshold, which can be optionally specified
9005 from nothing (0) to everything (255 for 8-bit based formats). An intensity
9006 value greater to the set value is considered non-black. It defaults to 24.
9007 You can also specify a value between 0.0 and 1.0 which will be scaled depending
9008 on the bitdepth of the pixel format.
9011 The value which the width/height should be divisible by. It defaults to
9012 16. The offset is automatically adjusted to center the video. Use 2 to
9013 get only even dimensions (needed for 4:2:2 video). 16 is best when
9014 encoding to most video codecs.
9017 Set the number of initial frames for which evaluation is skipped.
9018 Default is 2. Range is 0 to INT_MAX.
9020 @item reset_count, reset
9021 Set the counter that determines after how many frames cropdetect will
9022 reset the previously detected largest video area and start over to
9023 detect the current optimal crop area. Default value is 0.
9025 This can be useful when channel logos distort the video area. 0
9026 indicates 'never reset', and returns the largest area encountered during
9033 Delay video filtering until a given wallclock timestamp. The filter first
9034 passes on @option{preroll} amount of frames, then it buffers at most
9035 @option{buffer} amount of frames and waits for the cue. After reaching the cue
9036 it forwards the buffered frames and also any subsequent frames coming in its
9039 The filter can be used synchronize the output of multiple ffmpeg processes for
9040 realtime output devices like decklink. By putting the delay in the filtering
9041 chain and pre-buffering frames the process can pass on data to output almost
9042 immediately after the target wallclock timestamp is reached.
9044 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
9050 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
9053 The duration of content to pass on as preroll expressed in seconds. Default is 0.
9056 The maximum duration of content to buffer before waiting for the cue expressed
9057 in seconds. Default is 0.
9064 Apply color adjustments using curves.
9066 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
9067 component (red, green and blue) has its values defined by @var{N} key points
9068 tied from each other using a smooth curve. The x-axis represents the pixel
9069 values from the input frame, and the y-axis the new pixel values to be set for
9072 By default, a component curve is defined by the two points @var{(0;0)} and
9073 @var{(1;1)}. This creates a straight line where each original pixel value is
9074 "adjusted" to its own value, which means no change to the image.
9076 The filter allows you to redefine these two points and add some more. A new
9077 curve (using a natural cubic spline interpolation) will be define to pass
9078 smoothly through all these new coordinates. The new defined points needs to be
9079 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
9080 be in the @var{[0;1]} interval. If the computed curves happened to go outside
9081 the vector spaces, the values will be clipped accordingly.
9083 The filter accepts the following options:
9087 Select one of the available color presets. This option can be used in addition
9088 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
9089 options takes priority on the preset values.
9090 Available presets are:
9093 @item color_negative
9096 @item increase_contrast
9098 @item linear_contrast
9099 @item medium_contrast
9101 @item strong_contrast
9104 Default is @code{none}.
9106 Set the master key points. These points will define a second pass mapping. It
9107 is sometimes called a "luminance" or "value" mapping. It can be used with
9108 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
9109 post-processing LUT.
9111 Set the key points for the red component.
9113 Set the key points for the green component.
9115 Set the key points for the blue component.
9117 Set the key points for all components (not including master).
9118 Can be used in addition to the other key points component
9119 options. In this case, the unset component(s) will fallback on this
9120 @option{all} setting.
9122 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
9124 Save Gnuplot script of the curves in specified file.
9127 To avoid some filtergraph syntax conflicts, each key points list need to be
9128 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
9130 @subsection Examples
9134 Increase slightly the middle level of blue:
9136 curves=blue='0/0 0.5/0.58 1/1'
9142 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'
9144 Here we obtain the following coordinates for each components:
9147 @code{(0;0.11) (0.42;0.51) (1;0.95)}
9149 @code{(0;0) (0.50;0.48) (1;1)}
9151 @code{(0;0.22) (0.49;0.44) (1;0.80)}
9155 The previous example can also be achieved with the associated built-in preset:
9157 curves=preset=vintage
9167 Use a Photoshop preset and redefine the points of the green component:
9169 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
9173 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
9174 and @command{gnuplot}:
9176 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
9177 gnuplot -p /tmp/curves.plt
9183 Video data analysis filter.
9185 This filter shows hexadecimal pixel values of part of video.
9187 The filter accepts the following options:
9191 Set output video size.
9194 Set x offset from where to pick pixels.
9197 Set y offset from where to pick pixels.
9200 Set scope mode, can be one of the following:
9203 Draw hexadecimal pixel values with white color on black background.
9206 Draw hexadecimal pixel values with input video pixel color on black
9210 Draw hexadecimal pixel values on color background picked from input video,
9211 the text color is picked in such way so its always visible.
9215 Draw rows and columns numbers on left and top of video.
9218 Set background opacity.
9221 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
9224 Set pixel components to display. By default all pixel components are displayed.
9228 Apply Directional blur filter.
9230 The filter accepts the following options:
9234 Set angle of directional blur. Default is @code{45}.
9237 Set radius of directional blur. Default is @code{5}.
9240 Set which planes to filter. By default all planes are filtered.
9243 @subsection Commands
9244 This filter supports same @ref{commands} as options.
9245 The command accepts the same syntax of the corresponding option.
9247 If the specified expression is not valid, it is kept at its current
9252 Denoise frames using 2D DCT (frequency domain filtering).
9254 This filter is not designed for real time.
9256 The filter accepts the following options:
9260 Set the noise sigma constant.
9262 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
9263 coefficient (absolute value) below this threshold with be dropped.
9265 If you need a more advanced filtering, see @option{expr}.
9267 Default is @code{0}.
9270 Set number overlapping pixels for each block. Since the filter can be slow, you
9271 may want to reduce this value, at the cost of a less effective filter and the
9272 risk of various artefacts.
9274 If the overlapping value doesn't permit processing the whole input width or
9275 height, a warning will be displayed and according borders won't be denoised.
9277 Default value is @var{blocksize}-1, which is the best possible setting.
9280 Set the coefficient factor expression.
9282 For each coefficient of a DCT block, this expression will be evaluated as a
9283 multiplier value for the coefficient.
9285 If this is option is set, the @option{sigma} option will be ignored.
9287 The absolute value of the coefficient can be accessed through the @var{c}
9291 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
9292 @var{blocksize}, which is the width and height of the processed blocks.
9294 The default value is @var{3} (8x8) and can be raised to @var{4} for a
9295 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
9296 on the speed processing. Also, a larger block size does not necessarily means a
9300 @subsection Examples
9302 Apply a denoise with a @option{sigma} of @code{4.5}:
9307 The same operation can be achieved using the expression system:
9309 dctdnoiz=e='gte(c, 4.5*3)'
9312 Violent denoise using a block size of @code{16x16}:
9319 Remove banding artifacts from input video.
9320 It works by replacing banded pixels with average value of referenced pixels.
9322 The filter accepts the following options:
9329 Set banding detection threshold for each plane. Default is 0.02.
9330 Valid range is 0.00003 to 0.5.
9331 If difference between current pixel and reference pixel is less than threshold,
9332 it will be considered as banded.
9335 Banding detection range in pixels. Default is 16. If positive, random number
9336 in range 0 to set value will be used. If negative, exact absolute value
9338 The range defines square of four pixels around current pixel.
9341 Set direction in radians from which four pixel will be compared. If positive,
9342 random direction from 0 to set direction will be picked. If negative, exact of
9343 absolute value will be picked. For example direction 0, -PI or -2*PI radians
9344 will pick only pixels on same row and -PI/2 will pick only pixels on same
9348 If enabled, current pixel is compared with average value of all four
9349 surrounding pixels. The default is enabled. If disabled current pixel is
9350 compared with all four surrounding pixels. The pixel is considered banded
9351 if only all four differences with surrounding pixels are less than threshold.
9354 If enabled, current pixel is changed if and only if all pixel components are banded,
9355 e.g. banding detection threshold is triggered for all color components.
9356 The default is disabled.
9361 Remove blocking artifacts from input video.
9363 The filter accepts the following options:
9367 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
9368 This controls what kind of deblocking is applied.
9371 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
9377 Set blocking detection thresholds. Allowed range is 0 to 1.
9378 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
9379 Using higher threshold gives more deblocking strength.
9380 Setting @var{alpha} controls threshold detection at exact edge of block.
9381 Remaining options controls threshold detection near the edge. Each one for
9382 below/above or left/right. Setting any of those to @var{0} disables
9386 Set planes to filter. Default is to filter all available planes.
9389 @subsection Examples
9393 Deblock using weak filter and block size of 4 pixels.
9395 deblock=filter=weak:block=4
9399 Deblock using strong filter, block size of 4 pixels and custom thresholds for
9400 deblocking more edges.
9402 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
9406 Similar as above, but filter only first plane.
9408 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
9412 Similar as above, but filter only second and third plane.
9414 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
9421 Drop duplicated frames at regular intervals.
9423 The filter accepts the following options:
9427 Set the number of frames from which one will be dropped. Setting this to
9428 @var{N} means one frame in every batch of @var{N} frames will be dropped.
9429 Default is @code{5}.
9432 Set the threshold for duplicate detection. If the difference metric for a frame
9433 is less than or equal to this value, then it is declared as duplicate. Default
9437 Set scene change threshold. Default is @code{15}.
9441 Set the size of the x and y-axis blocks used during metric calculations.
9442 Larger blocks give better noise suppression, but also give worse detection of
9443 small movements. Must be a power of two. Default is @code{32}.
9446 Mark main input as a pre-processed input and activate clean source input
9447 stream. This allows the input to be pre-processed with various filters to help
9448 the metrics calculation while keeping the frame selection lossless. When set to
9449 @code{1}, the first stream is for the pre-processed input, and the second
9450 stream is the clean source from where the kept frames are chosen. Default is
9454 Set whether or not chroma is considered in the metric calculations. Default is
9460 Apply 2D deconvolution of video stream in frequency domain using second stream
9463 The filter accepts the following options:
9467 Set which planes to process.
9470 Set which impulse video frames will be processed, can be @var{first}
9471 or @var{all}. Default is @var{all}.
9474 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
9475 and height are not same and not power of 2 or if stream prior to convolving
9479 The @code{deconvolve} filter also supports the @ref{framesync} options.
9483 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
9485 It accepts the following options:
9489 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
9490 @var{rainbows} for cross-color reduction.
9493 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
9496 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
9499 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
9502 Set temporal chroma threshold. Lower values increases reduction of cross-color.
9507 Apply deflate effect to the video.
9509 This filter replaces the pixel by the local(3x3) average by taking into account
9510 only values lower than the pixel.
9512 It accepts the following options:
9519 Limit the maximum change for each plane, default is 65535.
9520 If 0, plane will remain unchanged.
9523 @subsection Commands
9525 This filter supports the all above options as @ref{commands}.
9529 Remove temporal frame luminance variations.
9531 It accepts the following options:
9535 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
9538 Set averaging mode to smooth temporal luminance variations.
9540 Available values are:
9565 Do not actually modify frame. Useful when one only wants metadata.
9570 Remove judder produced by partially interlaced telecined content.
9572 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
9573 source was partially telecined content then the output of @code{pullup,dejudder}
9574 will have a variable frame rate. May change the recorded frame rate of the
9575 container. Aside from that change, this filter will not affect constant frame
9578 The option available in this filter is:
9582 Specify the length of the window over which the judder repeats.
9584 Accepts any integer greater than 1. Useful values are:
9588 If the original was telecined from 24 to 30 fps (Film to NTSC).
9591 If the original was telecined from 25 to 30 fps (PAL to NTSC).
9594 If a mixture of the two.
9597 The default is @samp{4}.
9602 Suppress a TV station logo by a simple interpolation of the surrounding
9603 pixels. Just set a rectangle covering the logo and watch it disappear
9604 (and sometimes something even uglier appear - your mileage may vary).
9606 It accepts the following parameters:
9611 Specify the top left corner coordinates of the logo. They must be
9616 Specify the width and height of the logo to clear. They must be
9620 Specify the thickness of the fuzzy edge of the rectangle (added to
9621 @var{w} and @var{h}). The default value is 1. This option is
9622 deprecated, setting higher values should no longer be necessary and
9626 When set to 1, a green rectangle is drawn on the screen to simplify
9627 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
9628 The default value is 0.
9630 The rectangle is drawn on the outermost pixels which will be (partly)
9631 replaced with interpolated values. The values of the next pixels
9632 immediately outside this rectangle in each direction will be used to
9633 compute the interpolated pixel values inside the rectangle.
9637 @subsection Examples
9641 Set a rectangle covering the area with top left corner coordinates 0,0
9642 and size 100x77, and a band of size 10:
9644 delogo=x=0:y=0:w=100:h=77:band=10
9652 Remove the rain in the input image/video by applying the derain methods based on
9653 convolutional neural networks. Supported models:
9657 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
9658 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
9661 Training as well as model generation scripts are provided in
9662 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
9664 Native model files (.model) can be generated from TensorFlow model
9665 files (.pb) by using tools/python/convert.py
9667 The filter accepts the following options:
9671 Specify which filter to use. This option accepts the following values:
9675 Derain filter. To conduct derain filter, you need to use a derain model.
9678 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
9680 Default value is @samp{derain}.
9683 Specify which DNN backend to use for model loading and execution. This option accepts
9684 the following values:
9688 Native implementation of DNN loading and execution.
9691 TensorFlow backend. To enable this backend you
9692 need to install the TensorFlow for C library (see
9693 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9694 @code{--enable-libtensorflow}
9696 Default value is @samp{native}.
9699 Set path to model file specifying network architecture and its parameters.
9700 Note that different backends use different file formats. TensorFlow and native
9701 backend can load files for only its format.
9704 It can also be finished with @ref{dnn_processing} filter.
9708 Attempt to fix small changes in horizontal and/or vertical shift. This
9709 filter helps remove camera shake from hand-holding a camera, bumping a
9710 tripod, moving on a vehicle, etc.
9712 The filter accepts the following options:
9720 Specify a rectangular area where to limit the search for motion
9722 If desired the search for motion vectors can be limited to a
9723 rectangular area of the frame defined by its top left corner, width
9724 and height. These parameters have the same meaning as the drawbox
9725 filter which can be used to visualise the position of the bounding
9728 This is useful when simultaneous movement of subjects within the frame
9729 might be confused for camera motion by the motion vector search.
9731 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
9732 then the full frame is used. This allows later options to be set
9733 without specifying the bounding box for the motion vector search.
9735 Default - search the whole frame.
9739 Specify the maximum extent of movement in x and y directions in the
9740 range 0-64 pixels. Default 16.
9743 Specify how to generate pixels to fill blanks at the edge of the
9744 frame. Available values are:
9747 Fill zeroes at blank locations
9749 Original image at blank locations
9751 Extruded edge value at blank locations
9753 Mirrored edge at blank locations
9755 Default value is @samp{mirror}.
9758 Specify the blocksize to use for motion search. Range 4-128 pixels,
9762 Specify the contrast threshold for blocks. Only blocks with more than
9763 the specified contrast (difference between darkest and lightest
9764 pixels) will be considered. Range 1-255, default 125.
9767 Specify the search strategy. Available values are:
9770 Set exhaustive search
9772 Set less exhaustive search.
9774 Default value is @samp{exhaustive}.
9777 If set then a detailed log of the motion search is written to the
9784 Remove unwanted contamination of foreground colors, caused by reflected color of
9785 greenscreen or bluescreen.
9787 This filter accepts the following options:
9791 Set what type of despill to use.
9794 Set how spillmap will be generated.
9797 Set how much to get rid of still remaining spill.
9800 Controls amount of red in spill area.
9803 Controls amount of green in spill area.
9804 Should be -1 for greenscreen.
9807 Controls amount of blue in spill area.
9808 Should be -1 for bluescreen.
9811 Controls brightness of spill area, preserving colors.
9814 Modify alpha from generated spillmap.
9817 @subsection Commands
9819 This filter supports the all above options as @ref{commands}.
9823 Apply an exact inverse of the telecine operation. It requires a predefined
9824 pattern specified using the pattern option which must be the same as that passed
9825 to the telecine filter.
9827 This filter accepts the following options:
9836 The default value is @code{top}.
9840 A string of numbers representing the pulldown pattern you wish to apply.
9841 The default value is @code{23}.
9844 A number representing position of the first frame with respect to the telecine
9845 pattern. This is to be used if the stream is cut. The default value is @code{0}.
9850 Apply dilation effect to the video.
9852 This filter replaces the pixel by the local(3x3) maximum.
9854 It accepts the following options:
9861 Limit the maximum change for each plane, default is 65535.
9862 If 0, plane will remain unchanged.
9865 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9868 Flags to local 3x3 coordinates maps like this:
9875 @subsection Commands
9877 This filter supports the all above options as @ref{commands}.
9881 Displace pixels as indicated by second and third input stream.
9883 It takes three input streams and outputs one stream, the first input is the
9884 source, and second and third input are displacement maps.
9886 The second input specifies how much to displace pixels along the
9887 x-axis, while the third input specifies how much to displace pixels
9889 If one of displacement map streams terminates, last frame from that
9890 displacement map will be used.
9892 Note that once generated, displacements maps can be reused over and over again.
9894 A description of the accepted options follows.
9898 Set displace behavior for pixels that are out of range.
9900 Available values are:
9903 Missing pixels are replaced by black pixels.
9906 Adjacent pixels will spread out to replace missing pixels.
9909 Out of range pixels are wrapped so they point to pixels of other side.
9912 Out of range pixels will be replaced with mirrored pixels.
9914 Default is @samp{smear}.
9918 @subsection Examples
9922 Add ripple effect to rgb input of video size hd720:
9924 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
9928 Add wave effect to rgb input of video size hd720:
9930 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
9934 @anchor{dnn_processing}
9935 @section dnn_processing
9937 Do image processing with deep neural networks. It works together with another filter
9938 which converts the pixel format of the Frame to what the dnn network requires.
9940 The filter accepts the following options:
9944 Specify which DNN backend to use for model loading and execution. This option accepts
9945 the following values:
9949 Native implementation of DNN loading and execution.
9952 TensorFlow backend. To enable this backend you
9953 need to install the TensorFlow for C library (see
9954 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9955 @code{--enable-libtensorflow}
9958 OpenVINO backend. To enable this backend you
9959 need to build and install the OpenVINO for C library (see
9960 @url{https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md}) and configure FFmpeg with
9961 @code{--enable-libopenvino} (--extra-cflags=-I... --extra-ldflags=-L... might
9962 be needed if the header files and libraries are not installed into system path)
9966 Default value is @samp{native}.
9969 Set path to model file specifying network architecture and its parameters.
9970 Note that different backends use different file formats. TensorFlow, OpenVINO and native
9971 backend can load files for only its format.
9973 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9976 Set the input name of the dnn network.
9979 Set the output name of the dnn network.
9982 use DNN async execution if set (default: set),
9983 roll back to sync execution if the backend does not support async.
9987 @subsection Examples
9991 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
9993 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
9997 Halve the pixel value of the frame with format gray32f:
9999 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
10003 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
10005 ./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
10009 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
10011 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
10018 Draw a colored box on the input image.
10020 It accepts the following parameters:
10025 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
10029 The expressions which specify the width and height of the box; if 0 they are interpreted as
10030 the input width and height. It defaults to 0.
10033 Specify the color of the box to write. For the general syntax of this option,
10034 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10035 value @code{invert} is used, the box edge color is the same as the
10036 video with inverted luma.
10039 The expression which sets the thickness of the box edge.
10040 A value of @code{fill} will create a filled box. Default value is @code{3}.
10042 See below for the list of accepted constants.
10045 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
10046 will overwrite the video's color and alpha pixels.
10047 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
10050 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10051 following constants:
10055 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10059 horizontal and vertical chroma subsample values. For example for the
10060 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10064 The input width and height.
10067 The input sample aspect ratio.
10071 The x and y offset coordinates where the box is drawn.
10075 The width and height of the drawn box.
10078 The thickness of the drawn box.
10080 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10081 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10085 @subsection Examples
10089 Draw a black box around the edge of the input image:
10095 Draw a box with color red and an opacity of 50%:
10097 drawbox=10:20:200:60:red@@0.5
10100 The previous example can be specified as:
10102 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
10106 Fill the box with pink color:
10108 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
10112 Draw a 2-pixel red 2.40:1 mask:
10114 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
10118 @subsection Commands
10119 This filter supports same commands as options.
10120 The command accepts the same syntax of the corresponding option.
10122 If the specified expression is not valid, it is kept at its current
10127 Draw a graph using input video metadata.
10129 It accepts the following parameters:
10133 Set 1st frame metadata key from which metadata values will be used to draw a graph.
10136 Set 1st foreground color expression.
10139 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
10142 Set 2nd foreground color expression.
10145 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
10148 Set 3rd foreground color expression.
10151 Set 4th frame metadata key from which metadata values will be used to draw a graph.
10154 Set 4th foreground color expression.
10157 Set minimal value of metadata value.
10160 Set maximal value of metadata value.
10163 Set graph background color. Default is white.
10168 Available values for mode is:
10175 Default is @code{line}.
10180 Available values for slide is:
10183 Draw new frame when right border is reached.
10186 Replace old columns with new ones.
10189 Scroll from right to left.
10192 Scroll from left to right.
10195 Draw single picture.
10198 Default is @code{frame}.
10201 Set size of graph video. For the syntax of this option, check the
10202 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
10203 The default value is @code{900x256}.
10206 Set the output frame rate. Default value is @code{25}.
10208 The foreground color expressions can use the following variables:
10211 Minimal value of metadata value.
10214 Maximal value of metadata value.
10217 Current metadata key value.
10220 The color is defined as 0xAABBGGRR.
10223 Example using metadata from @ref{signalstats} filter:
10225 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
10228 Example using metadata from @ref{ebur128} filter:
10230 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
10235 Draw a grid on the input image.
10237 It accepts the following parameters:
10242 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
10246 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
10247 input width and height, respectively, minus @code{thickness}, so image gets
10248 framed. Default to 0.
10251 Specify the color of the grid. For the general syntax of this option,
10252 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10253 value @code{invert} is used, the grid color is the same as the
10254 video with inverted luma.
10257 The expression which sets the thickness of the grid line. Default value is @code{1}.
10259 See below for the list of accepted constants.
10262 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
10263 will overwrite the video's color and alpha pixels.
10264 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
10267 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10268 following constants:
10272 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10276 horizontal and vertical chroma subsample values. For example for the
10277 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10281 The input grid cell width and height.
10284 The input sample aspect ratio.
10288 The x and y coordinates of some point of grid intersection (meant to configure offset).
10292 The width and height of the drawn cell.
10295 The thickness of the drawn cell.
10297 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10298 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10302 @subsection Examples
10306 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
10308 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
10312 Draw a white 3x3 grid with an opacity of 50%:
10314 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
10318 @subsection Commands
10319 This filter supports same commands as options.
10320 The command accepts the same syntax of the corresponding option.
10322 If the specified expression is not valid, it is kept at its current
10328 Draw a text string or text from a specified file on top of a video, using the
10329 libfreetype library.
10331 To enable compilation of this filter, you need to configure FFmpeg with
10332 @code{--enable-libfreetype}.
10333 To enable default font fallback and the @var{font} option you need to
10334 configure FFmpeg with @code{--enable-libfontconfig}.
10335 To enable the @var{text_shaping} option, you need to configure FFmpeg with
10336 @code{--enable-libfribidi}.
10340 It accepts the following parameters:
10345 Used to draw a box around text using the background color.
10346 The value must be either 1 (enable) or 0 (disable).
10347 The default value of @var{box} is 0.
10350 Set the width of the border to be drawn around the box using @var{boxcolor}.
10351 The default value of @var{boxborderw} is 0.
10354 The color to be used for drawing box around text. For the syntax of this
10355 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10357 The default value of @var{boxcolor} is "white".
10360 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
10361 The default value of @var{line_spacing} is 0.
10364 Set the width of the border to be drawn around the text using @var{bordercolor}.
10365 The default value of @var{borderw} is 0.
10368 Set the color to be used for drawing border around text. For the syntax of this
10369 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10371 The default value of @var{bordercolor} is "black".
10374 Select how the @var{text} is expanded. Can be either @code{none},
10375 @code{strftime} (deprecated) or
10376 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
10380 Set a start time for the count. Value is in microseconds. Only applied
10381 in the deprecated strftime expansion mode. To emulate in normal expansion
10382 mode use the @code{pts} function, supplying the start time (in seconds)
10383 as the second argument.
10386 If true, check and fix text coords to avoid clipping.
10389 The color to be used for drawing fonts. For the syntax of this option, check
10390 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10392 The default value of @var{fontcolor} is "black".
10394 @item fontcolor_expr
10395 String which is expanded the same way as @var{text} to obtain dynamic
10396 @var{fontcolor} value. By default this option has empty value and is not
10397 processed. When this option is set, it overrides @var{fontcolor} option.
10400 The font family to be used for drawing text. By default Sans.
10403 The font file to be used for drawing text. The path must be included.
10404 This parameter is mandatory if the fontconfig support is disabled.
10407 Draw the text applying alpha blending. The value can
10408 be a number between 0.0 and 1.0.
10409 The expression accepts the same variables @var{x, y} as well.
10410 The default value is 1.
10411 Please see @var{fontcolor_expr}.
10414 The font size to be used for drawing text.
10415 The default value of @var{fontsize} is 16.
10418 If set to 1, attempt to shape the text (for example, reverse the order of
10419 right-to-left text and join Arabic characters) before drawing it.
10420 Otherwise, just draw the text exactly as given.
10421 By default 1 (if supported).
10423 @item ft_load_flags
10424 The flags to be used for loading the fonts.
10426 The flags map the corresponding flags supported by libfreetype, and are
10427 a combination of the following values:
10434 @item vertical_layout
10435 @item force_autohint
10438 @item ignore_global_advance_width
10440 @item ignore_transform
10442 @item linear_design
10446 Default value is "default".
10448 For more information consult the documentation for the FT_LOAD_*
10452 The color to be used for drawing a shadow behind the drawn text. For the
10453 syntax of this option, check the @ref{color syntax,,"Color" section in the
10454 ffmpeg-utils manual,ffmpeg-utils}.
10456 The default value of @var{shadowcolor} is "black".
10460 The x and y offsets for the text shadow position with respect to the
10461 position of the text. They can be either positive or negative
10462 values. The default value for both is "0".
10465 The starting frame number for the n/frame_num variable. The default value
10469 The size in number of spaces to use for rendering the tab.
10470 Default value is 4.
10473 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
10474 format. It can be used with or without text parameter. @var{timecode_rate}
10475 option must be specified.
10477 @item timecode_rate, rate, r
10478 Set the timecode frame rate (timecode only). Value will be rounded to nearest
10479 integer. Minimum value is "1".
10480 Drop-frame timecode is supported for frame rates 30 & 60.
10483 If set to 1, the output of the timecode option will wrap around at 24 hours.
10484 Default is 0 (disabled).
10487 The text string to be drawn. The text must be a sequence of UTF-8
10488 encoded characters.
10489 This parameter is mandatory if no file is specified with the parameter
10493 A text file containing text to be drawn. The text must be a sequence
10494 of UTF-8 encoded characters.
10496 This parameter is mandatory if no text string is specified with the
10497 parameter @var{text}.
10499 If both @var{text} and @var{textfile} are specified, an error is thrown.
10502 If set to 1, the @var{textfile} will be reloaded before each frame.
10503 Be sure to update it atomically, or it may be read partially, or even fail.
10507 The expressions which specify the offsets where text will be drawn
10508 within the video frame. They are relative to the top/left border of the
10511 The default value of @var{x} and @var{y} is "0".
10513 See below for the list of accepted constants and functions.
10516 The parameters for @var{x} and @var{y} are expressions containing the
10517 following constants and functions:
10521 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
10525 horizontal and vertical chroma subsample values. For example for the
10526 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10529 the height of each text line
10537 @item max_glyph_a, ascent
10538 the maximum distance from the baseline to the highest/upper grid
10539 coordinate used to place a glyph outline point, for all the rendered
10541 It is a positive value, due to the grid's orientation with the Y axis
10544 @item max_glyph_d, descent
10545 the maximum distance from the baseline to the lowest grid coordinate
10546 used to place a glyph outline point, for all the rendered glyphs.
10547 This is a negative value, due to the grid's orientation, with the Y axis
10551 maximum glyph height, that is the maximum height for all the glyphs
10552 contained in the rendered text, it is equivalent to @var{ascent} -
10556 maximum glyph width, that is the maximum width for all the glyphs
10557 contained in the rendered text
10560 the number of input frame, starting from 0
10562 @item rand(min, max)
10563 return a random number included between @var{min} and @var{max}
10566 The input sample aspect ratio.
10569 timestamp expressed in seconds, NAN if the input timestamp is unknown
10572 the height of the rendered text
10575 the width of the rendered text
10579 the x and y offset coordinates where the text is drawn.
10581 These parameters allow the @var{x} and @var{y} expressions to refer
10582 to each other, so you can for example specify @code{y=x/dar}.
10585 A one character description of the current frame's picture type.
10588 The current packet's position in the input file or stream
10589 (in bytes, from the start of the input). A value of -1 indicates
10590 this info is not available.
10593 The current packet's duration, in seconds.
10596 The current packet's size (in bytes).
10599 @anchor{drawtext_expansion}
10600 @subsection Text expansion
10602 If @option{expansion} is set to @code{strftime},
10603 the filter recognizes strftime() sequences in the provided text and
10604 expands them accordingly. Check the documentation of strftime(). This
10605 feature is deprecated.
10607 If @option{expansion} is set to @code{none}, the text is printed verbatim.
10609 If @option{expansion} is set to @code{normal} (which is the default),
10610 the following expansion mechanism is used.
10612 The backslash character @samp{\}, followed by any character, always expands to
10613 the second character.
10615 Sequences of the form @code{%@{...@}} are expanded. The text between the
10616 braces is a function name, possibly followed by arguments separated by ':'.
10617 If the arguments contain special characters or delimiters (':' or '@}'),
10618 they should be escaped.
10620 Note that they probably must also be escaped as the value for the
10621 @option{text} option in the filter argument string and as the filter
10622 argument in the filtergraph description, and possibly also for the shell,
10623 that makes up to four levels of escaping; using a text file avoids these
10626 The following functions are available:
10631 The expression evaluation result.
10633 It must take one argument specifying the expression to be evaluated,
10634 which accepts the same constants and functions as the @var{x} and
10635 @var{y} values. Note that not all constants should be used, for
10636 example the text size is not known when evaluating the expression, so
10637 the constants @var{text_w} and @var{text_h} will have an undefined
10640 @item expr_int_format, eif
10641 Evaluate the expression's value and output as formatted integer.
10643 The first argument is the expression to be evaluated, just as for the @var{expr} function.
10644 The second argument specifies the output format. Allowed values are @samp{x},
10645 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
10646 @code{printf} function.
10647 The third parameter is optional and sets the number of positions taken by the output.
10648 It can be used to add padding with zeros from the left.
10651 The time at which the filter is running, expressed in UTC.
10652 It can accept an argument: a strftime() format string.
10655 The time at which the filter is running, expressed in the local time zone.
10656 It can accept an argument: a strftime() format string.
10659 Frame metadata. Takes one or two arguments.
10661 The first argument is mandatory and specifies the metadata key.
10663 The second argument is optional and specifies a default value, used when the
10664 metadata key is not found or empty.
10666 Available metadata can be identified by inspecting entries
10667 starting with TAG included within each frame section
10668 printed by running @code{ffprobe -show_frames}.
10670 String metadata generated in filters leading to
10671 the drawtext filter are also available.
10674 The frame number, starting from 0.
10677 A one character description of the current picture type.
10680 The timestamp of the current frame.
10681 It can take up to three arguments.
10683 The first argument is the format of the timestamp; it defaults to @code{flt}
10684 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
10685 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
10686 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
10687 @code{localtime} stands for the timestamp of the frame formatted as
10688 local time zone time.
10690 The second argument is an offset added to the timestamp.
10692 If the format is set to @code{hms}, a third argument @code{24HH} may be
10693 supplied to present the hour part of the formatted timestamp in 24h format
10696 If the format is set to @code{localtime} or @code{gmtime},
10697 a third argument may be supplied: a strftime() format string.
10698 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
10701 @subsection Commands
10703 This filter supports altering parameters via commands:
10706 Alter existing filter parameters.
10708 Syntax for the argument is the same as for filter invocation, e.g.
10711 fontsize=56:fontcolor=green:text='Hello World'
10714 Full filter invocation with sendcmd would look like this:
10717 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
10721 If the entire argument can't be parsed or applied as valid values then the filter will
10722 continue with its existing parameters.
10724 @subsection Examples
10728 Draw "Test Text" with font FreeSerif, using the default values for the
10729 optional parameters.
10732 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
10736 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
10737 and y=50 (counting from the top-left corner of the screen), text is
10738 yellow with a red box around it. Both the text and the box have an
10742 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
10743 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
10746 Note that the double quotes are not necessary if spaces are not used
10747 within the parameter list.
10750 Show the text at the center of the video frame:
10752 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
10756 Show the text at a random position, switching to a new position every 30 seconds:
10758 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)"
10762 Show a text line sliding from right to left in the last row of the video
10763 frame. The file @file{LONG_LINE} is assumed to contain a single line
10766 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
10770 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
10772 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
10776 Draw a single green letter "g", at the center of the input video.
10777 The glyph baseline is placed at half screen height.
10779 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
10783 Show text for 1 second every 3 seconds:
10785 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
10789 Use fontconfig to set the font. Note that the colons need to be escaped.
10791 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
10795 Draw "Test Text" with font size dependent on height of the video.
10797 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
10801 Print the date of a real-time encoding (see strftime(3)):
10803 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
10807 Show text fading in and out (appearing/disappearing):
10810 DS=1.0 # display start
10811 DE=10.0 # display end
10812 FID=1.5 # fade in duration
10813 FOD=5 # fade out duration
10814 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 @}"
10818 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
10819 and the @option{fontsize} value are included in the @option{y} offset.
10821 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
10822 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
10826 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
10827 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
10828 must have option @option{-export_path_metadata 1} for the special metadata fields
10829 to be available for filters.
10831 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
10836 For more information about libfreetype, check:
10837 @url{http://www.freetype.org/}.
10839 For more information about fontconfig, check:
10840 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
10842 For more information about libfribidi, check:
10843 @url{http://fribidi.org/}.
10845 @section edgedetect
10847 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
10849 The filter accepts the following options:
10854 Set low and high threshold values used by the Canny thresholding
10857 The high threshold selects the "strong" edge pixels, which are then
10858 connected through 8-connectivity with the "weak" edge pixels selected
10859 by the low threshold.
10861 @var{low} and @var{high} threshold values must be chosen in the range
10862 [0,1], and @var{low} should be lesser or equal to @var{high}.
10864 Default value for @var{low} is @code{20/255}, and default value for @var{high}
10868 Define the drawing mode.
10872 Draw white/gray wires on black background.
10875 Mix the colors to create a paint/cartoon effect.
10878 Apply Canny edge detector on all selected planes.
10880 Default value is @var{wires}.
10883 Select planes for filtering. By default all available planes are filtered.
10886 @subsection Examples
10890 Standard edge detection with custom values for the hysteresis thresholding:
10892 edgedetect=low=0.1:high=0.4
10896 Painting effect without thresholding:
10898 edgedetect=mode=colormix:high=0
10904 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10906 For each input image, the filter will compute the optimal mapping from
10907 the input to the output given the codebook length, that is the number
10908 of distinct output colors.
10910 This filter accepts the following options.
10913 @item codebook_length, l
10914 Set codebook length. The value must be a positive integer, and
10915 represents the number of distinct output colors. Default value is 256.
10918 Set the maximum number of iterations to apply for computing the optimal
10919 mapping. The higher the value the better the result and the higher the
10920 computation time. Default value is 1.
10923 Set a random seed, must be an integer included between 0 and
10924 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10925 will try to use a good random seed on a best effort basis.
10928 Set pal8 output pixel format. This option does not work with codebook
10929 length greater than 256.
10934 Measure graylevel entropy in histogram of color channels of video frames.
10936 It accepts the following parameters:
10940 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10942 @var{diff} mode measures entropy of histogram delta values, absolute differences
10943 between neighbour histogram values.
10947 Apply the EPX magnification filter which is designed for pixel art.
10949 It accepts the following option:
10953 Set the scaling dimension: @code{2} for @code{2xEPX}, @code{3} for
10955 Default is @code{3}.
10959 Set brightness, contrast, saturation and approximate gamma adjustment.
10961 The filter accepts the following options:
10965 Set the contrast expression. The value must be a float value in range
10966 @code{-1000.0} to @code{1000.0}. The default value is "1".
10969 Set the brightness expression. The value must be a float value in
10970 range @code{-1.0} to @code{1.0}. The default value is "0".
10973 Set the saturation expression. The value must be a float in
10974 range @code{0.0} to @code{3.0}. The default value is "1".
10977 Set the gamma expression. The value must be a float in range
10978 @code{0.1} to @code{10.0}. The default value is "1".
10981 Set the gamma expression for red. The value must be a float in
10982 range @code{0.1} to @code{10.0}. The default value is "1".
10985 Set the gamma expression for green. The value must be a float in range
10986 @code{0.1} to @code{10.0}. The default value is "1".
10989 Set the gamma expression for blue. The value must be a float in range
10990 @code{0.1} to @code{10.0}. The default value is "1".
10993 Set the gamma weight expression. It can be used to reduce the effect
10994 of a high gamma value on bright image areas, e.g. keep them from
10995 getting overamplified and just plain white. The value must be a float
10996 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
10997 gamma correction all the way down while @code{1.0} leaves it at its
10998 full strength. Default is "1".
11001 Set when the expressions for brightness, contrast, saturation and
11002 gamma expressions are evaluated.
11004 It accepts the following values:
11007 only evaluate expressions once during the filter initialization or
11008 when a command is processed
11011 evaluate expressions for each incoming frame
11014 Default value is @samp{init}.
11017 The expressions accept the following parameters:
11020 frame count of the input frame starting from 0
11023 byte position of the corresponding packet in the input file, NAN if
11027 frame rate of the input video, NAN if the input frame rate is unknown
11030 timestamp expressed in seconds, NAN if the input timestamp is unknown
11033 @subsection Commands
11034 The filter supports the following commands:
11038 Set the contrast expression.
11041 Set the brightness expression.
11044 Set the saturation expression.
11047 Set the gamma expression.
11050 Set the gamma_r expression.
11053 Set gamma_g expression.
11056 Set gamma_b expression.
11059 Set gamma_weight expression.
11061 The command accepts the same syntax of the corresponding option.
11063 If the specified expression is not valid, it is kept at its current
11070 Apply erosion effect to the video.
11072 This filter replaces the pixel by the local(3x3) minimum.
11074 It accepts the following options:
11081 Limit the maximum change for each plane, default is 65535.
11082 If 0, plane will remain unchanged.
11085 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
11088 Flags to local 3x3 coordinates maps like this:
11095 @subsection Commands
11097 This filter supports the all above options as @ref{commands}.
11101 Deinterlace the input video ("estdif" stands for "Edge Slope
11102 Tracing Deinterlacing Filter").
11104 Spatial only filter that uses edge slope tracing algorithm
11105 to interpolate missing lines.
11106 It accepts the following parameters:
11110 The interlacing mode to adopt. It accepts one of the following values:
11114 Output one frame for each frame.
11116 Output one frame for each field.
11119 The default value is @code{field}.
11122 The picture field parity assumed for the input interlaced video. It accepts one
11123 of the following values:
11127 Assume the top field is first.
11129 Assume the bottom field is first.
11131 Enable automatic detection of field parity.
11134 The default value is @code{auto}.
11135 If the interlacing is unknown or the decoder does not export this information,
11136 top field first will be assumed.
11139 Specify which frames to deinterlace. Accepts one of the following
11144 Deinterlace all frames.
11146 Only deinterlace frames marked as interlaced.
11149 The default value is @code{all}.
11152 Specify the search radius for edge slope tracing. Default value is 1.
11153 Allowed range is from 1 to 15.
11156 Specify the search radius for best edge matching. Default value is 2.
11157 Allowed range is from 0 to 15.
11160 Specify the interpolation used. Default is 4-point interpolation. It accepts one
11161 of the following values:
11165 Two-point interpolation.
11167 Four-point interpolation.
11169 Six-point interpolation.
11173 @subsection Commands
11174 This filter supports same @ref{commands} as options.
11176 @section extractplanes
11178 Extract color channel components from input video stream into
11179 separate grayscale video streams.
11181 The filter accepts the following option:
11185 Set plane(s) to extract.
11187 Available values for planes are:
11198 Choosing planes not available in the input will result in an error.
11199 That means you cannot select @code{r}, @code{g}, @code{b} planes
11200 with @code{y}, @code{u}, @code{v} planes at same time.
11203 @subsection Examples
11207 Extract luma, u and v color channel component from input video frame
11208 into 3 grayscale outputs:
11210 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
11216 Apply a fade-in/out effect to the input video.
11218 It accepts the following parameters:
11222 The effect type can be either "in" for a fade-in, or "out" for a fade-out
11224 Default is @code{in}.
11226 @item start_frame, s
11227 Specify the number of the frame to start applying the fade
11228 effect at. Default is 0.
11231 The number of frames that the fade effect lasts. At the end of the
11232 fade-in effect, the output video will have the same intensity as the input video.
11233 At the end of the fade-out transition, the output video will be filled with the
11234 selected @option{color}.
11238 If set to 1, fade only alpha channel, if one exists on the input.
11239 Default value is 0.
11241 @item start_time, st
11242 Specify the timestamp (in seconds) of the frame to start to apply the fade
11243 effect. If both start_frame and start_time are specified, the fade will start at
11244 whichever comes last. Default is 0.
11247 The number of seconds for which the fade effect has to last. At the end of the
11248 fade-in effect the output video will have the same intensity as the input video,
11249 at the end of the fade-out transition the output video will be filled with the
11250 selected @option{color}.
11251 If both duration and nb_frames are specified, duration is used. Default is 0
11252 (nb_frames is used by default).
11255 Specify the color of the fade. Default is "black".
11258 @subsection Examples
11262 Fade in the first 30 frames of video:
11267 The command above is equivalent to:
11273 Fade out the last 45 frames of a 200-frame video:
11276 fade=type=out:start_frame=155:nb_frames=45
11280 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
11282 fade=in:0:25, fade=out:975:25
11286 Make the first 5 frames yellow, then fade in from frame 5-24:
11288 fade=in:5:20:color=yellow
11292 Fade in alpha over first 25 frames of video:
11294 fade=in:0:25:alpha=1
11298 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
11300 fade=t=in:st=5.5:d=0.5
11306 Denoise frames using 3D FFT (frequency domain filtering).
11308 The filter accepts the following options:
11312 Set the noise sigma constant. This sets denoising strength.
11313 Default value is 1. Allowed range is from 0 to 30.
11314 Using very high sigma with low overlap may give blocking artifacts.
11317 Set amount of denoising. By default all detected noise is reduced.
11318 Default value is 1. Allowed range is from 0 to 1.
11321 Set size of block, Default is 4, can be 3, 4, 5 or 6.
11322 Actual size of block in pixels is 2 to power of @var{block}, so by default
11323 block size in pixels is 2^4 which is 16.
11326 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
11329 Set number of previous frames to use for denoising. By default is set to 0.
11332 Set number of next frames to to use for denoising. By default is set to 0.
11335 Set planes which will be filtered, by default are all available filtered
11340 Apply arbitrary expressions to samples in frequency domain
11344 Adjust the dc value (gain) of the luma plane of the image. The filter
11345 accepts an integer value in range @code{0} to @code{1000}. The default
11346 value is set to @code{0}.
11349 Adjust the dc value (gain) of the 1st chroma plane of the image. The
11350 filter accepts an integer value in range @code{0} to @code{1000}. The
11351 default value is set to @code{0}.
11354 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
11355 filter accepts an integer value in range @code{0} to @code{1000}. The
11356 default value is set to @code{0}.
11359 Set the frequency domain weight expression for the luma plane.
11362 Set the frequency domain weight expression for the 1st chroma plane.
11365 Set the frequency domain weight expression for the 2nd chroma plane.
11368 Set when the expressions are evaluated.
11370 It accepts the following values:
11373 Only evaluate expressions once during the filter initialization.
11376 Evaluate expressions for each incoming frame.
11379 Default value is @samp{init}.
11381 The filter accepts the following variables:
11384 The coordinates of the current sample.
11388 The width and height of the image.
11391 The number of input frame, starting from 0.
11394 @subsection Examples
11400 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
11406 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
11412 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
11418 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
11425 Extract a single field from an interlaced image using stride
11426 arithmetic to avoid wasting CPU time. The output frames are marked as
11429 The filter accepts the following options:
11433 Specify whether to extract the top (if the value is @code{0} or
11434 @code{top}) or the bottom field (if the value is @code{1} or
11440 Create new frames by copying the top and bottom fields from surrounding frames
11441 supplied as numbers by the hint file.
11445 Set file containing hints: absolute/relative frame numbers.
11447 There must be one line for each frame in a clip. Each line must contain two
11448 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
11449 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
11450 is current frame number for @code{absolute} mode or out of [-1, 1] range
11451 for @code{relative} mode. First number tells from which frame to pick up top
11452 field and second number tells from which frame to pick up bottom field.
11454 If optionally followed by @code{+} output frame will be marked as interlaced,
11455 else if followed by @code{-} output frame will be marked as progressive, else
11456 it will be marked same as input frame.
11457 If optionally followed by @code{t} output frame will use only top field, or in
11458 case of @code{b} it will use only bottom field.
11459 If line starts with @code{#} or @code{;} that line is skipped.
11462 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
11465 Example of first several lines of @code{hint} file for @code{relative} mode:
11467 0,0 - # first frame
11468 1,0 - # second frame, use third's frame top field and second's frame bottom field
11469 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
11484 @section fieldmatch
11486 Field matching filter for inverse telecine. It is meant to reconstruct the
11487 progressive frames from a telecined stream. The filter does not drop duplicated
11488 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
11489 followed by a decimation filter such as @ref{decimate} in the filtergraph.
11491 The separation of the field matching and the decimation is notably motivated by
11492 the possibility of inserting a de-interlacing filter fallback between the two.
11493 If the source has mixed telecined and real interlaced content,
11494 @code{fieldmatch} will not be able to match fields for the interlaced parts.
11495 But these remaining combed frames will be marked as interlaced, and thus can be
11496 de-interlaced by a later filter such as @ref{yadif} before decimation.
11498 In addition to the various configuration options, @code{fieldmatch} can take an
11499 optional second stream, activated through the @option{ppsrc} option. If
11500 enabled, the frames reconstruction will be based on the fields and frames from
11501 this second stream. This allows the first input to be pre-processed in order to
11502 help the various algorithms of the filter, while keeping the output lossless
11503 (assuming the fields are matched properly). Typically, a field-aware denoiser,
11504 or brightness/contrast adjustments can help.
11506 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
11507 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
11508 which @code{fieldmatch} is based on. While the semantic and usage are very
11509 close, some behaviour and options names can differ.
11511 The @ref{decimate} filter currently only works for constant frame rate input.
11512 If your input has mixed telecined (30fps) and progressive content with a lower
11513 framerate like 24fps use the following filterchain to produce the necessary cfr
11514 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
11516 The filter accepts the following options:
11520 Specify the assumed field order of the input stream. Available values are:
11524 Auto detect parity (use FFmpeg's internal parity value).
11526 Assume bottom field first.
11528 Assume top field first.
11531 Note that it is sometimes recommended not to trust the parity announced by the
11534 Default value is @var{auto}.
11537 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
11538 sense that it won't risk creating jerkiness due to duplicate frames when
11539 possible, but if there are bad edits or blended fields it will end up
11540 outputting combed frames when a good match might actually exist. On the other
11541 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
11542 but will almost always find a good frame if there is one. The other values are
11543 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
11544 jerkiness and creating duplicate frames versus finding good matches in sections
11545 with bad edits, orphaned fields, blended fields, etc.
11547 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
11549 Available values are:
11553 2-way matching (p/c)
11555 2-way matching, and trying 3rd match if still combed (p/c + n)
11557 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
11559 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
11560 still combed (p/c + n + u/b)
11562 3-way matching (p/c/n)
11564 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
11565 detected as combed (p/c/n + u/b)
11568 The parenthesis at the end indicate the matches that would be used for that
11569 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
11572 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
11575 Default value is @var{pc_n}.
11578 Mark the main input stream as a pre-processed input, and enable the secondary
11579 input stream as the clean source to pick the fields from. See the filter
11580 introduction for more details. It is similar to the @option{clip2} feature from
11583 Default value is @code{0} (disabled).
11586 Set the field to match from. It is recommended to set this to the same value as
11587 @option{order} unless you experience matching failures with that setting. In
11588 certain circumstances changing the field that is used to match from can have a
11589 large impact on matching performance. Available values are:
11593 Automatic (same value as @option{order}).
11595 Match from the bottom field.
11597 Match from the top field.
11600 Default value is @var{auto}.
11603 Set whether or not chroma is included during the match comparisons. In most
11604 cases it is recommended to leave this enabled. You should set this to @code{0}
11605 only if your clip has bad chroma problems such as heavy rainbowing or other
11606 artifacts. Setting this to @code{0} could also be used to speed things up at
11607 the cost of some accuracy.
11609 Default value is @code{1}.
11613 These define an exclusion band which excludes the lines between @option{y0} and
11614 @option{y1} from being included in the field matching decision. An exclusion
11615 band can be used to ignore subtitles, a logo, or other things that may
11616 interfere with the matching. @option{y0} sets the starting scan line and
11617 @option{y1} sets the ending line; all lines in between @option{y0} and
11618 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
11619 @option{y0} and @option{y1} to the same value will disable the feature.
11620 @option{y0} and @option{y1} defaults to @code{0}.
11623 Set the scene change detection threshold as a percentage of maximum change on
11624 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
11625 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
11626 @option{scthresh} is @code{[0.0, 100.0]}.
11628 Default value is @code{12.0}.
11631 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
11632 account the combed scores of matches when deciding what match to use as the
11633 final match. Available values are:
11637 No final matching based on combed scores.
11639 Combed scores are only used when a scene change is detected.
11641 Use combed scores all the time.
11644 Default is @var{sc}.
11647 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
11648 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
11649 Available values are:
11653 No forced calculation.
11655 Force p/c/n calculations.
11657 Force p/c/n/u/b calculations.
11660 Default value is @var{none}.
11663 This is the area combing threshold used for combed frame detection. This
11664 essentially controls how "strong" or "visible" combing must be to be detected.
11665 Larger values mean combing must be more visible and smaller values mean combing
11666 can be less visible or strong and still be detected. Valid settings are from
11667 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
11668 be detected as combed). This is basically a pixel difference value. A good
11669 range is @code{[8, 12]}.
11671 Default value is @code{9}.
11674 Sets whether or not chroma is considered in the combed frame decision. Only
11675 disable this if your source has chroma problems (rainbowing, etc.) that are
11676 causing problems for the combed frame detection with chroma enabled. Actually,
11677 using @option{chroma}=@var{0} is usually more reliable, except for the case
11678 where there is chroma only combing in the source.
11680 Default value is @code{0}.
11684 Respectively set the x-axis and y-axis size of the window used during combed
11685 frame detection. This has to do with the size of the area in which
11686 @option{combpel} pixels are required to be detected as combed for a frame to be
11687 declared combed. See the @option{combpel} parameter description for more info.
11688 Possible values are any number that is a power of 2 starting at 4 and going up
11691 Default value is @code{16}.
11694 The number of combed pixels inside any of the @option{blocky} by
11695 @option{blockx} size blocks on the frame for the frame to be detected as
11696 combed. While @option{cthresh} controls how "visible" the combing must be, this
11697 setting controls "how much" combing there must be in any localized area (a
11698 window defined by the @option{blockx} and @option{blocky} settings) on the
11699 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
11700 which point no frames will ever be detected as combed). This setting is known
11701 as @option{MI} in TFM/VFM vocabulary.
11703 Default value is @code{80}.
11706 @anchor{p/c/n/u/b meaning}
11707 @subsection p/c/n/u/b meaning
11709 @subsubsection p/c/n
11711 We assume the following telecined stream:
11714 Top fields: 1 2 2 3 4
11715 Bottom fields: 1 2 3 4 4
11718 The numbers correspond to the progressive frame the fields relate to. Here, the
11719 first two frames are progressive, the 3rd and 4th are combed, and so on.
11721 When @code{fieldmatch} is configured to run a matching from bottom
11722 (@option{field}=@var{bottom}) this is how this input stream get transformed:
11727 B 1 2 3 4 4 <-- matching reference
11736 As a result of the field matching, we can see that some frames get duplicated.
11737 To perform a complete inverse telecine, you need to rely on a decimation filter
11738 after this operation. See for instance the @ref{decimate} filter.
11740 The same operation now matching from top fields (@option{field}=@var{top})
11745 T 1 2 2 3 4 <-- matching reference
11755 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
11756 basically, they refer to the frame and field of the opposite parity:
11759 @item @var{p} matches the field of the opposite parity in the previous frame
11760 @item @var{c} matches the field of the opposite parity in the current frame
11761 @item @var{n} matches the field of the opposite parity in the next frame
11766 The @var{u} and @var{b} matching are a bit special in the sense that they match
11767 from the opposite parity flag. In the following examples, we assume that we are
11768 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
11769 'x' is placed above and below each matched fields.
11771 With bottom matching (@option{field}=@var{bottom}):
11776 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11777 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11785 With top matching (@option{field}=@var{top}):
11790 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11791 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11799 @subsection Examples
11801 Simple IVTC of a top field first telecined stream:
11803 fieldmatch=order=tff:combmatch=none, decimate
11806 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
11808 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
11811 @section fieldorder
11813 Transform the field order of the input video.
11815 It accepts the following parameters:
11820 The output field order. Valid values are @var{tff} for top field first or @var{bff}
11821 for bottom field first.
11824 The default value is @samp{tff}.
11826 The transformation is done by shifting the picture content up or down
11827 by one line, and filling the remaining line with appropriate picture content.
11828 This method is consistent with most broadcast field order converters.
11830 If the input video is not flagged as being interlaced, or it is already
11831 flagged as being of the required output field order, then this filter does
11832 not alter the incoming video.
11834 It is very useful when converting to or from PAL DV material,
11835 which is bottom field first.
11839 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
11842 @section fifo, afifo
11844 Buffer input images and send them when they are requested.
11846 It is mainly useful when auto-inserted by the libavfilter
11849 It does not take parameters.
11851 @section fillborders
11853 Fill borders of the input video, without changing video stream dimensions.
11854 Sometimes video can have garbage at the four edges and you may not want to
11855 crop video input to keep size multiple of some number.
11857 This filter accepts the following options:
11861 Number of pixels to fill from left border.
11864 Number of pixels to fill from right border.
11867 Number of pixels to fill from top border.
11870 Number of pixels to fill from bottom border.
11875 It accepts the following values:
11878 fill pixels using outermost pixels
11881 fill pixels using mirroring (half sample symmetric)
11884 fill pixels with constant value
11887 fill pixels using reflecting (whole sample symmetric)
11890 fill pixels using wrapping
11893 fade pixels to constant value
11896 Default is @var{smear}.
11899 Set color for pixels in fixed or fade mode. Default is @var{black}.
11902 @subsection Commands
11903 This filter supports same @ref{commands} as options.
11904 The command accepts the same syntax of the corresponding option.
11906 If the specified expression is not valid, it is kept at its current
11911 Find a rectangular object
11913 It accepts the following options:
11917 Filepath of the object image, needs to be in gray8.
11920 Detection threshold, default is 0.5.
11923 Number of mipmaps, default is 3.
11925 @item xmin, ymin, xmax, ymax
11926 Specifies the rectangle in which to search.
11929 @subsection Examples
11933 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
11935 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
11941 Flood area with values of same pixel components with another values.
11943 It accepts the following options:
11946 Set pixel x coordinate.
11949 Set pixel y coordinate.
11952 Set source #0 component value.
11955 Set source #1 component value.
11958 Set source #2 component value.
11961 Set source #3 component value.
11964 Set destination #0 component value.
11967 Set destination #1 component value.
11970 Set destination #2 component value.
11973 Set destination #3 component value.
11979 Convert the input video to one of the specified pixel formats.
11980 Libavfilter will try to pick one that is suitable as input to
11983 It accepts the following parameters:
11987 A '|'-separated list of pixel format names, such as
11988 "pix_fmts=yuv420p|monow|rgb24".
11992 @subsection Examples
11996 Convert the input video to the @var{yuv420p} format
11998 format=pix_fmts=yuv420p
12001 Convert the input video to any of the formats in the list
12003 format=pix_fmts=yuv420p|yuv444p|yuv410p
12010 Convert the video to specified constant frame rate by duplicating or dropping
12011 frames as necessary.
12013 It accepts the following parameters:
12017 The desired output frame rate. The default is @code{25}.
12020 Assume the first PTS should be the given value, in seconds. This allows for
12021 padding/trimming at the start of stream. By default, no assumption is made
12022 about the first frame's expected PTS, so no padding or trimming is done.
12023 For example, this could be set to 0 to pad the beginning with duplicates of
12024 the first frame if a video stream starts after the audio stream or to trim any
12025 frames with a negative PTS.
12028 Timestamp (PTS) rounding method.
12030 Possible values are:
12037 round towards -infinity
12039 round towards +infinity
12043 The default is @code{near}.
12046 Action performed when reading the last frame.
12048 Possible values are:
12051 Use same timestamp rounding method as used for other frames.
12053 Pass through last frame if input duration has not been reached yet.
12055 The default is @code{round}.
12059 Alternatively, the options can be specified as a flat string:
12060 @var{fps}[:@var{start_time}[:@var{round}]].
12062 See also the @ref{setpts} filter.
12064 @subsection Examples
12068 A typical usage in order to set the fps to 25:
12074 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
12076 fps=fps=film:round=near
12082 Pack two different video streams into a stereoscopic video, setting proper
12083 metadata on supported codecs. The two views should have the same size and
12084 framerate and processing will stop when the shorter video ends. Please note
12085 that you may conveniently adjust view properties with the @ref{scale} and
12088 It accepts the following parameters:
12092 The desired packing format. Supported values are:
12097 The views are next to each other (default).
12100 The views are on top of each other.
12103 The views are packed by line.
12106 The views are packed by column.
12109 The views are temporally interleaved.
12118 # Convert left and right views into a frame-sequential video
12119 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
12121 # Convert views into a side-by-side video with the same output resolution as the input
12122 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
12127 Change the frame rate by interpolating new video output frames from the source
12130 This filter is not designed to function correctly with interlaced media. If
12131 you wish to change the frame rate of interlaced media then you are required
12132 to deinterlace before this filter and re-interlace after this filter.
12134 A description of the accepted options follows.
12138 Specify the output frames per second. This option can also be specified
12139 as a value alone. The default is @code{50}.
12142 Specify the start of a range where the output frame will be created as a
12143 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12144 the default is @code{15}.
12147 Specify the end of a range where the output frame will be created as a
12148 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12149 the default is @code{240}.
12152 Specify the level at which a scene change is detected as a value between
12153 0 and 100 to indicate a new scene; a low value reflects a low
12154 probability for the current frame to introduce a new scene, while a higher
12155 value means the current frame is more likely to be one.
12156 The default is @code{8.2}.
12159 Specify flags influencing the filter process.
12161 Available value for @var{flags} is:
12164 @item scene_change_detect, scd
12165 Enable scene change detection using the value of the option @var{scene}.
12166 This flag is enabled by default.
12172 Select one frame every N-th frame.
12174 This filter accepts the following option:
12177 Select frame after every @code{step} frames.
12178 Allowed values are positive integers higher than 0. Default value is @code{1}.
12181 @section freezedetect
12183 Detect frozen video.
12185 This filter logs a message and sets frame metadata when it detects that the
12186 input video has no significant change in content during a specified duration.
12187 Video freeze detection calculates the mean average absolute difference of all
12188 the components of video frames and compares it to a noise floor.
12190 The printed times and duration are expressed in seconds. The
12191 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
12192 whose timestamp equals or exceeds the detection duration and it contains the
12193 timestamp of the first frame of the freeze. The
12194 @code{lavfi.freezedetect.freeze_duration} and
12195 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
12198 The filter accepts the following options:
12202 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
12203 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
12207 Set freeze duration until notification (default is 2 seconds).
12210 @section freezeframes
12212 Freeze video frames.
12214 This filter freezes video frames using frame from 2nd input.
12216 The filter accepts the following options:
12220 Set number of first frame from which to start freeze.
12223 Set number of last frame from which to end freeze.
12226 Set number of frame from 2nd input which will be used instead of replaced frames.
12232 Apply a frei0r effect to the input video.
12234 To enable the compilation of this filter, you need to install the frei0r
12235 header and configure FFmpeg with @code{--enable-frei0r}.
12237 It accepts the following parameters:
12242 The name of the frei0r effect to load. If the environment variable
12243 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
12244 directories specified by the colon-separated list in @env{FREI0R_PATH}.
12245 Otherwise, the standard frei0r paths are searched, in this order:
12246 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
12247 @file{/usr/lib/frei0r-1/}.
12249 @item filter_params
12250 A '|'-separated list of parameters to pass to the frei0r effect.
12254 A frei0r effect parameter can be a boolean (its value is either
12255 "y" or "n"), a double, a color (specified as
12256 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
12257 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
12258 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
12259 a position (specified as @var{X}/@var{Y}, where
12260 @var{X} and @var{Y} are floating point numbers) and/or a string.
12262 The number and types of parameters depend on the loaded effect. If an
12263 effect parameter is not specified, the default value is set.
12265 @subsection Examples
12269 Apply the distort0r effect, setting the first two double parameters:
12271 frei0r=filter_name=distort0r:filter_params=0.5|0.01
12275 Apply the colordistance effect, taking a color as the first parameter:
12277 frei0r=colordistance:0.2/0.3/0.4
12278 frei0r=colordistance:violet
12279 frei0r=colordistance:0x112233
12283 Apply the perspective effect, specifying the top left and top right image
12286 frei0r=perspective:0.2/0.2|0.8/0.2
12290 For more information, see
12291 @url{http://frei0r.dyne.org}
12293 @subsection Commands
12295 This filter supports the @option{filter_params} option as @ref{commands}.
12299 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
12301 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
12302 processing filter, one of them is performed once per block, not per pixel.
12303 This allows for much higher speed.
12305 The filter accepts the following options:
12309 Set quality. This option defines the number of levels for averaging. It accepts
12310 an integer in the range 4-5. Default value is @code{4}.
12313 Force a constant quantization parameter. It accepts an integer in range 0-63.
12314 If not set, the filter will use the QP from the video stream (if available).
12317 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
12318 more details but also more artifacts, while higher values make the image smoother
12319 but also blurrier. Default value is @code{0} − PSNR optimal.
12321 @item use_bframe_qp
12322 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
12323 option may cause flicker since the B-Frames have often larger QP. Default is
12324 @code{0} (not enabled).
12330 Apply Gaussian blur filter.
12332 The filter accepts the following options:
12336 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
12339 Set number of steps for Gaussian approximation. Default is @code{1}.
12342 Set which planes to filter. By default all planes are filtered.
12345 Set vertical sigma, if negative it will be same as @code{sigma}.
12346 Default is @code{-1}.
12349 @subsection Commands
12350 This filter supports same commands as options.
12351 The command accepts the same syntax of the corresponding option.
12353 If the specified expression is not valid, it is kept at its current
12358 Apply generic equation to each pixel.
12360 The filter accepts the following options:
12363 @item lum_expr, lum
12364 Set the luminance expression.
12366 Set the chrominance blue expression.
12368 Set the chrominance red expression.
12369 @item alpha_expr, a
12370 Set the alpha expression.
12372 Set the red expression.
12373 @item green_expr, g
12374 Set the green expression.
12376 Set the blue expression.
12379 The colorspace is selected according to the specified options. If one
12380 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
12381 options is specified, the filter will automatically select a YCbCr
12382 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
12383 @option{blue_expr} options is specified, it will select an RGB
12386 If one of the chrominance expression is not defined, it falls back on the other
12387 one. If no alpha expression is specified it will evaluate to opaque value.
12388 If none of chrominance expressions are specified, they will evaluate
12389 to the luminance expression.
12391 The expressions can use the following variables and functions:
12395 The sequential number of the filtered frame, starting from @code{0}.
12399 The coordinates of the current sample.
12403 The width and height of the image.
12407 Width and height scale depending on the currently filtered plane. It is the
12408 ratio between the corresponding luma plane number of pixels and the current
12409 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
12410 @code{0.5,0.5} for chroma planes.
12413 Time of the current frame, expressed in seconds.
12416 Return the value of the pixel at location (@var{x},@var{y}) of the current
12420 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
12424 Return the value of the pixel at location (@var{x},@var{y}) of the
12425 blue-difference chroma plane. Return 0 if there is no such plane.
12428 Return the value of the pixel at location (@var{x},@var{y}) of the
12429 red-difference chroma plane. Return 0 if there is no such plane.
12434 Return the value of the pixel at location (@var{x},@var{y}) of the
12435 red/green/blue component. Return 0 if there is no such component.
12438 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
12439 plane. Return 0 if there is no such plane.
12441 @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)
12442 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
12443 sums of samples within a rectangle. See the functions without the sum postfix.
12445 @item interpolation
12446 Set one of interpolation methods:
12451 Default is bilinear.
12454 For functions, if @var{x} and @var{y} are outside the area, the value will be
12455 automatically clipped to the closer edge.
12457 Please note that this filter can use multiple threads in which case each slice
12458 will have its own expression state. If you want to use only a single expression
12459 state because your expressions depend on previous state then you should limit
12460 the number of filter threads to 1.
12462 @subsection Examples
12466 Flip the image horizontally:
12472 Generate a bidimensional sine wave, with angle @code{PI/3} and a
12473 wavelength of 100 pixels:
12475 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
12479 Generate a fancy enigmatic moving light:
12481 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
12485 Generate a quick emboss effect:
12487 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
12491 Modify RGB components depending on pixel position:
12493 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
12497 Create a radial gradient that is the same size as the input (also see
12498 the @ref{vignette} filter):
12500 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
12506 Fix the banding artifacts that are sometimes introduced into nearly flat
12507 regions by truncation to 8-bit color depth.
12508 Interpolate the gradients that should go where the bands are, and
12511 It is designed for playback only. Do not use it prior to
12512 lossy compression, because compression tends to lose the dither and
12513 bring back the bands.
12515 It accepts the following parameters:
12520 The maximum amount by which the filter will change any one pixel. This is also
12521 the threshold for detecting nearly flat regions. Acceptable values range from
12522 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
12526 The neighborhood to fit the gradient to. A larger radius makes for smoother
12527 gradients, but also prevents the filter from modifying the pixels near detailed
12528 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
12529 values will be clipped to the valid range.
12533 Alternatively, the options can be specified as a flat string:
12534 @var{strength}[:@var{radius}]
12536 @subsection Examples
12540 Apply the filter with a @code{3.5} strength and radius of @code{8}:
12546 Specify radius, omitting the strength (which will fall-back to the default
12554 @anchor{graphmonitor}
12555 @section graphmonitor
12556 Show various filtergraph stats.
12558 With this filter one can debug complete filtergraph.
12559 Especially issues with links filling with queued frames.
12561 The filter accepts the following options:
12565 Set video output size. Default is @var{hd720}.
12568 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
12571 Set output mode, can be @var{fulll} or @var{compact}.
12572 In @var{compact} mode only filters with some queued frames have displayed stats.
12575 Set flags which enable which stats are shown in video.
12577 Available values for flags are:
12580 Display number of queued frames in each link.
12582 @item frame_count_in
12583 Display number of frames taken from filter.
12585 @item frame_count_out
12586 Display number of frames given out from filter.
12589 Display current filtered frame pts.
12592 Display current filtered frame time.
12595 Display time base for filter link.
12598 Display used format for filter link.
12601 Display video size or number of audio channels in case of audio used by filter link.
12604 Display video frame rate or sample rate in case of audio used by filter link.
12607 Display link output status.
12611 Set upper limit for video rate of output stream, Default value is @var{25}.
12612 This guarantee that output video frame rate will not be higher than this value.
12616 A color constancy variation filter which estimates scene illumination via grey edge algorithm
12617 and corrects the scene colors accordingly.
12619 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
12621 The filter accepts the following options:
12625 The order of differentiation to be applied on the scene. Must be chosen in the range
12626 [0,2] and default value is 1.
12629 The Minkowski parameter to be used for calculating the Minkowski distance. Must
12630 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
12631 max value instead of calculating Minkowski distance.
12634 The standard deviation of Gaussian blur to be applied on the scene. Must be
12635 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
12636 can't be equal to 0 if @var{difford} is greater than 0.
12639 @subsection Examples
12645 greyedge=difford=1:minknorm=5:sigma=2
12651 greyedge=difford=1:minknorm=0:sigma=2
12659 Apply a Hald CLUT to a video stream.
12661 First input is the video stream to process, and second one is the Hald CLUT.
12662 The Hald CLUT input can be a simple picture or a complete video stream.
12664 The filter accepts the following options:
12668 Force termination when the shortest input terminates. Default is @code{0}.
12670 Continue applying the last CLUT after the end of the stream. A value of
12671 @code{0} disable the filter after the last frame of the CLUT is reached.
12672 Default is @code{1}.
12675 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
12676 filters share the same internals).
12678 This filter also supports the @ref{framesync} options.
12680 More information about the Hald CLUT can be found on Eskil Steenberg's website
12681 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
12683 @subsection Workflow examples
12685 @subsubsection Hald CLUT video stream
12687 Generate an identity Hald CLUT stream altered with various effects:
12689 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
12692 Note: make sure you use a lossless codec.
12694 Then use it with @code{haldclut} to apply it on some random stream:
12696 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
12699 The Hald CLUT will be applied to the 10 first seconds (duration of
12700 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
12701 to the remaining frames of the @code{mandelbrot} stream.
12703 @subsubsection Hald CLUT with preview
12705 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
12706 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
12707 biggest possible square starting at the top left of the picture. The remaining
12708 padding pixels (bottom or right) will be ignored. This area can be used to add
12709 a preview of the Hald CLUT.
12711 Typically, the following generated Hald CLUT will be supported by the
12712 @code{haldclut} filter:
12715 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
12716 pad=iw+320 [padded_clut];
12717 smptebars=s=320x256, split [a][b];
12718 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
12719 [main][b] overlay=W-320" -frames:v 1 clut.png
12722 It contains the original and a preview of the effect of the CLUT: SMPTE color
12723 bars are displayed on the right-top, and below the same color bars processed by
12726 Then, the effect of this Hald CLUT can be visualized with:
12728 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
12733 Flip the input video horizontally.
12735 For example, to horizontally flip the input video with @command{ffmpeg}:
12737 ffmpeg -i in.avi -vf "hflip" out.avi
12741 This filter applies a global color histogram equalization on a
12744 It can be used to correct video that has a compressed range of pixel
12745 intensities. The filter redistributes the pixel intensities to
12746 equalize their distribution across the intensity range. It may be
12747 viewed as an "automatically adjusting contrast filter". This filter is
12748 useful only for correcting degraded or poorly captured source
12751 The filter accepts the following options:
12755 Determine the amount of equalization to be applied. As the strength
12756 is reduced, the distribution of pixel intensities more-and-more
12757 approaches that of the input frame. The value must be a float number
12758 in the range [0,1] and defaults to 0.200.
12761 Set the maximum intensity that can generated and scale the output
12762 values appropriately. The strength should be set as desired and then
12763 the intensity can be limited if needed to avoid washing-out. The value
12764 must be a float number in the range [0,1] and defaults to 0.210.
12767 Set the antibanding level. If enabled the filter will randomly vary
12768 the luminance of output pixels by a small amount to avoid banding of
12769 the histogram. Possible values are @code{none}, @code{weak} or
12770 @code{strong}. It defaults to @code{none}.
12776 Compute and draw a color distribution histogram for the input video.
12778 The computed histogram is a representation of the color component
12779 distribution in an image.
12781 Standard histogram displays the color components distribution in an image.
12782 Displays color graph for each color component. Shows distribution of
12783 the Y, U, V, A or R, G, B components, depending on input format, in the
12784 current frame. Below each graph a color component scale meter is shown.
12786 The filter accepts the following options:
12790 Set height of level. Default value is @code{200}.
12791 Allowed range is [50, 2048].
12794 Set height of color scale. Default value is @code{12}.
12795 Allowed range is [0, 40].
12799 It accepts the following values:
12802 Per color component graphs are placed below each other.
12805 Per color component graphs are placed side by side.
12808 Presents information identical to that in the @code{parade}, except
12809 that the graphs representing color components are superimposed directly
12812 Default is @code{stack}.
12815 Set mode. Can be either @code{linear}, or @code{logarithmic}.
12816 Default is @code{linear}.
12819 Set what color components to display.
12820 Default is @code{7}.
12823 Set foreground opacity. Default is @code{0.7}.
12826 Set background opacity. Default is @code{0.5}.
12829 @subsection Examples
12834 Calculate and draw histogram:
12836 ffplay -i input -vf histogram
12844 This is a high precision/quality 3d denoise filter. It aims to reduce
12845 image noise, producing smooth images and making still images really
12846 still. It should enhance compressibility.
12848 It accepts the following optional parameters:
12852 A non-negative floating point number which specifies spatial luma strength.
12853 It defaults to 4.0.
12855 @item chroma_spatial
12856 A non-negative floating point number which specifies spatial chroma strength.
12857 It defaults to 3.0*@var{luma_spatial}/4.0.
12860 A floating point number which specifies luma temporal strength. It defaults to
12861 6.0*@var{luma_spatial}/4.0.
12864 A floating point number which specifies chroma temporal strength. It defaults to
12865 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
12868 @subsection Commands
12869 This filter supports same @ref{commands} as options.
12870 The command accepts the same syntax of the corresponding option.
12872 If the specified expression is not valid, it is kept at its current
12875 @anchor{hwdownload}
12876 @section hwdownload
12878 Download hardware frames to system memory.
12880 The input must be in hardware frames, and the output a non-hardware format.
12881 Not all formats will be supported on the output - it may be necessary to insert
12882 an additional @option{format} filter immediately following in the graph to get
12883 the output in a supported format.
12887 Map hardware frames to system memory or to another device.
12889 This filter has several different modes of operation; which one is used depends
12890 on the input and output formats:
12893 Hardware frame input, normal frame output
12895 Map the input frames to system memory and pass them to the output. If the
12896 original hardware frame is later required (for example, after overlaying
12897 something else on part of it), the @option{hwmap} filter can be used again
12898 in the next mode to retrieve it.
12900 Normal frame input, hardware frame output
12902 If the input is actually a software-mapped hardware frame, then unmap it -
12903 that is, return the original hardware frame.
12905 Otherwise, a device must be provided. Create new hardware surfaces on that
12906 device for the output, then map them back to the software format at the input
12907 and give those frames to the preceding filter. This will then act like the
12908 @option{hwupload} filter, but may be able to avoid an additional copy when
12909 the input is already in a compatible format.
12911 Hardware frame input and output
12913 A device must be supplied for the output, either directly or with the
12914 @option{derive_device} option. The input and output devices must be of
12915 different types and compatible - the exact meaning of this is
12916 system-dependent, but typically it means that they must refer to the same
12917 underlying hardware context (for example, refer to the same graphics card).
12919 If the input frames were originally created on the output device, then unmap
12920 to retrieve the original frames.
12922 Otherwise, map the frames to the output device - create new hardware frames
12923 on the output corresponding to the frames on the input.
12926 The following additional parameters are accepted:
12930 Set the frame mapping mode. Some combination of:
12933 The mapped frame should be readable.
12935 The mapped frame should be writeable.
12937 The mapping will always overwrite the entire frame.
12939 This may improve performance in some cases, as the original contents of the
12940 frame need not be loaded.
12942 The mapping must not involve any copying.
12944 Indirect mappings to copies of frames are created in some cases where either
12945 direct mapping is not possible or it would have unexpected properties.
12946 Setting this flag ensures that the mapping is direct and will fail if that is
12949 Defaults to @var{read+write} if not specified.
12951 @item derive_device @var{type}
12952 Rather than using the device supplied at initialisation, instead derive a new
12953 device of type @var{type} from the device the input frames exist on.
12956 In a hardware to hardware mapping, map in reverse - create frames in the sink
12957 and map them back to the source. This may be necessary in some cases where
12958 a mapping in one direction is required but only the opposite direction is
12959 supported by the devices being used.
12961 This option is dangerous - it may break the preceding filter in undefined
12962 ways if there are any additional constraints on that filter's output.
12963 Do not use it without fully understanding the implications of its use.
12969 Upload system memory frames to hardware surfaces.
12971 The device to upload to must be supplied when the filter is initialised. If
12972 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
12973 option or with the @option{derive_device} option. The input and output devices
12974 must be of different types and compatible - the exact meaning of this is
12975 system-dependent, but typically it means that they must refer to the same
12976 underlying hardware context (for example, refer to the same graphics card).
12978 The following additional parameters are accepted:
12981 @item derive_device @var{type}
12982 Rather than using the device supplied at initialisation, instead derive a new
12983 device of type @var{type} from the device the input frames exist on.
12986 @anchor{hwupload_cuda}
12987 @section hwupload_cuda
12989 Upload system memory frames to a CUDA device.
12991 It accepts the following optional parameters:
12995 The number of the CUDA device to use
13000 Apply a high-quality magnification filter designed for pixel art. This filter
13001 was originally created by Maxim Stepin.
13003 It accepts the following option:
13007 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
13008 @code{hq3x} and @code{4} for @code{hq4x}.
13009 Default is @code{3}.
13013 Stack input videos horizontally.
13015 All streams must be of same pixel format and of same height.
13017 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
13018 to create same output.
13020 The filter accepts the following option:
13024 Set number of input streams. Default is 2.
13027 If set to 1, force the output to terminate when the shortest input
13028 terminates. Default value is 0.
13033 Modify the hue and/or the saturation of the input.
13035 It accepts the following parameters:
13039 Specify the hue angle as a number of degrees. It accepts an expression,
13040 and defaults to "0".
13043 Specify the saturation in the [-10,10] range. It accepts an expression and
13047 Specify the hue angle as a number of radians. It accepts an
13048 expression, and defaults to "0".
13051 Specify the brightness in the [-10,10] range. It accepts an expression and
13055 @option{h} and @option{H} are mutually exclusive, and can't be
13056 specified at the same time.
13058 The @option{b}, @option{h}, @option{H} and @option{s} option values are
13059 expressions containing the following constants:
13063 frame count of the input frame starting from 0
13066 presentation timestamp of the input frame expressed in time base units
13069 frame rate of the input video, NAN if the input frame rate is unknown
13072 timestamp expressed in seconds, NAN if the input timestamp is unknown
13075 time base of the input video
13078 @subsection Examples
13082 Set the hue to 90 degrees and the saturation to 1.0:
13088 Same command but expressing the hue in radians:
13094 Rotate hue and make the saturation swing between 0
13095 and 2 over a period of 1 second:
13097 hue="H=2*PI*t: s=sin(2*PI*t)+1"
13101 Apply a 3 seconds saturation fade-in effect starting at 0:
13103 hue="s=min(t/3\,1)"
13106 The general fade-in expression can be written as:
13108 hue="s=min(0\, max((t-START)/DURATION\, 1))"
13112 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
13114 hue="s=max(0\, min(1\, (8-t)/3))"
13117 The general fade-out expression can be written as:
13119 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
13124 @subsection Commands
13126 This filter supports the following commands:
13132 Modify the hue and/or the saturation and/or brightness of the input video.
13133 The command accepts the same syntax of the corresponding option.
13135 If the specified expression is not valid, it is kept at its current
13139 @section hysteresis
13141 Grow first stream into second stream by connecting components.
13142 This makes it possible to build more robust edge masks.
13144 This filter accepts the following options:
13148 Set which planes will be processed as bitmap, unprocessed planes will be
13149 copied from first stream.
13150 By default value 0xf, all planes will be processed.
13153 Set threshold which is used in filtering. If pixel component value is higher than
13154 this value filter algorithm for connecting components is activated.
13155 By default value is 0.
13158 The @code{hysteresis} filter also supports the @ref{framesync} options.
13162 Detect video interlacing type.
13164 This filter tries to detect if the input frames are interlaced, progressive,
13165 top or bottom field first. It will also try to detect fields that are
13166 repeated between adjacent frames (a sign of telecine).
13168 Single frame detection considers only immediately adjacent frames when classifying each frame.
13169 Multiple frame detection incorporates the classification history of previous frames.
13171 The filter will log these metadata values:
13174 @item single.current_frame
13175 Detected type of current frame using single-frame detection. One of:
13176 ``tff'' (top field first), ``bff'' (bottom field first),
13177 ``progressive'', or ``undetermined''
13180 Cumulative number of frames detected as top field first using single-frame detection.
13183 Cumulative number of frames detected as top field first using multiple-frame detection.
13186 Cumulative number of frames detected as bottom field first using single-frame detection.
13188 @item multiple.current_frame
13189 Detected type of current frame using multiple-frame detection. One of:
13190 ``tff'' (top field first), ``bff'' (bottom field first),
13191 ``progressive'', or ``undetermined''
13194 Cumulative number of frames detected as bottom field first using multiple-frame detection.
13196 @item single.progressive
13197 Cumulative number of frames detected as progressive using single-frame detection.
13199 @item multiple.progressive
13200 Cumulative number of frames detected as progressive using multiple-frame detection.
13202 @item single.undetermined
13203 Cumulative number of frames that could not be classified using single-frame detection.
13205 @item multiple.undetermined
13206 Cumulative number of frames that could not be classified using multiple-frame detection.
13208 @item repeated.current_frame
13209 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
13211 @item repeated.neither
13212 Cumulative number of frames with no repeated field.
13215 Cumulative number of frames with the top field repeated from the previous frame's top field.
13217 @item repeated.bottom
13218 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
13221 The filter accepts the following options:
13225 Set interlacing threshold.
13227 Set progressive threshold.
13229 Threshold for repeated field detection.
13231 Number of frames after which a given frame's contribution to the
13232 statistics is halved (i.e., it contributes only 0.5 to its
13233 classification). The default of 0 means that all frames seen are given
13234 full weight of 1.0 forever.
13235 @item analyze_interlaced_flag
13236 When this is not 0 then idet will use the specified number of frames to determine
13237 if the interlaced flag is accurate, it will not count undetermined frames.
13238 If the flag is found to be accurate it will be used without any further
13239 computations, if it is found to be inaccurate it will be cleared without any
13240 further computations. This allows inserting the idet filter as a low computational
13241 method to clean up the interlaced flag
13246 Deinterleave or interleave fields.
13248 This filter allows one to process interlaced images fields without
13249 deinterlacing them. Deinterleaving splits the input frame into 2
13250 fields (so called half pictures). Odd lines are moved to the top
13251 half of the output image, even lines to the bottom half.
13252 You can process (filter) them independently and then re-interleave them.
13254 The filter accepts the following options:
13258 @item chroma_mode, c
13259 @item alpha_mode, a
13260 Available values for @var{luma_mode}, @var{chroma_mode} and
13261 @var{alpha_mode} are:
13267 @item deinterleave, d
13268 Deinterleave fields, placing one above the other.
13270 @item interleave, i
13271 Interleave fields. Reverse the effect of deinterleaving.
13273 Default value is @code{none}.
13275 @item luma_swap, ls
13276 @item chroma_swap, cs
13277 @item alpha_swap, as
13278 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
13281 @subsection Commands
13283 This filter supports the all above options as @ref{commands}.
13287 Apply inflate effect to the video.
13289 This filter replaces the pixel by the local(3x3) average by taking into account
13290 only values higher than the pixel.
13292 It accepts the following options:
13299 Limit the maximum change for each plane, default is 65535.
13300 If 0, plane will remain unchanged.
13303 @subsection Commands
13305 This filter supports the all above options as @ref{commands}.
13309 Simple interlacing filter from progressive contents. This interleaves upper (or
13310 lower) lines from odd frames with lower (or upper) lines from even frames,
13311 halving the frame rate and preserving image height.
13314 Original Original New Frame
13315 Frame 'j' Frame 'j+1' (tff)
13316 ========== =========== ==================
13317 Line 0 --------------------> Frame 'j' Line 0
13318 Line 1 Line 1 ----> Frame 'j+1' Line 1
13319 Line 2 ---------------------> Frame 'j' Line 2
13320 Line 3 Line 3 ----> Frame 'j+1' Line 3
13322 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
13325 It accepts the following optional parameters:
13329 This determines whether the interlaced frame is taken from the even
13330 (tff - default) or odd (bff) lines of the progressive frame.
13333 Vertical lowpass filter to avoid twitter interlacing and
13334 reduce moire patterns.
13338 Disable vertical lowpass filter
13341 Enable linear filter (default)
13344 Enable complex filter. This will slightly less reduce twitter and moire
13345 but better retain detail and subjective sharpness impression.
13352 Deinterlace input video by applying Donald Graft's adaptive kernel
13353 deinterling. Work on interlaced parts of a video to produce
13354 progressive frames.
13356 The description of the accepted parameters follows.
13360 Set the threshold which affects the filter's tolerance when
13361 determining if a pixel line must be processed. It must be an integer
13362 in the range [0,255] and defaults to 10. A value of 0 will result in
13363 applying the process on every pixels.
13366 Paint pixels exceeding the threshold value to white if set to 1.
13370 Set the fields order. Swap fields if set to 1, leave fields alone if
13374 Enable additional sharpening if set to 1. Default is 0.
13377 Enable twoway sharpening if set to 1. Default is 0.
13380 @subsection Examples
13384 Apply default values:
13386 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
13390 Enable additional sharpening:
13396 Paint processed pixels in white:
13404 Slowly update darker pixels.
13406 This filter makes short flashes of light appear longer.
13407 This filter accepts the following options:
13411 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
13414 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
13417 @section lenscorrection
13419 Correct radial lens distortion
13421 This filter can be used to correct for radial distortion as can result from the use
13422 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
13423 one can use tools available for example as part of opencv or simply trial-and-error.
13424 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
13425 and extract the k1 and k2 coefficients from the resulting matrix.
13427 Note that effectively the same filter is available in the open-source tools Krita and
13428 Digikam from the KDE project.
13430 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
13431 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
13432 brightness distribution, so you may want to use both filters together in certain
13433 cases, though you will have to take care of ordering, i.e. whether vignetting should
13434 be applied before or after lens correction.
13436 @subsection Options
13438 The filter accepts the following options:
13442 Relative x-coordinate of the focal point of the image, and thereby the center of the
13443 distortion. This value has a range [0,1] and is expressed as fractions of the image
13444 width. Default is 0.5.
13446 Relative y-coordinate of the focal point of the image, and thereby the center of the
13447 distortion. This value has a range [0,1] and is expressed as fractions of the image
13448 height. Default is 0.5.
13450 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
13451 no correction. Default is 0.
13453 Coefficient of the double quadratic correction term. This value has a range [-1,1].
13454 0 means no correction. Default is 0.
13456 Set interpolation type. Can be @code{nearest} or @code{bilinear}.
13457 Default is @code{nearest}.
13459 Specify the color of the unmapped pixels. For the syntax of this option,
13460 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
13461 manual,ffmpeg-utils}. Default color is @code{black@@0}.
13464 The formula that generates the correction is:
13466 @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)
13468 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
13469 distances from the focal point in the source and target images, respectively.
13473 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
13475 The @code{lensfun} filter requires the camera make, camera model, and lens model
13476 to apply the lens correction. The filter will load the lensfun database and
13477 query it to find the corresponding camera and lens entries in the database. As
13478 long as these entries can be found with the given options, the filter can
13479 perform corrections on frames. Note that incomplete strings will result in the
13480 filter choosing the best match with the given options, and the filter will
13481 output the chosen camera and lens models (logged with level "info"). You must
13482 provide the make, camera model, and lens model as they are required.
13484 The filter accepts the following options:
13488 The make of the camera (for example, "Canon"). This option is required.
13491 The model of the camera (for example, "Canon EOS 100D"). This option is
13495 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
13496 option is required.
13499 The type of correction to apply. The following values are valid options:
13503 Enables fixing lens vignetting.
13506 Enables fixing lens geometry. This is the default.
13509 Enables fixing chromatic aberrations.
13512 Enables fixing lens vignetting and lens geometry.
13515 Enables fixing lens vignetting and chromatic aberrations.
13518 Enables fixing both lens geometry and chromatic aberrations.
13521 Enables all possible corrections.
13525 The focal length of the image/video (zoom; expected constant for video). For
13526 example, a 18--55mm lens has focal length range of [18--55], so a value in that
13527 range should be chosen when using that lens. Default 18.
13530 The aperture of the image/video (expected constant for video). Note that
13531 aperture is only used for vignetting correction. Default 3.5.
13533 @item focus_distance
13534 The focus distance of the image/video (expected constant for video). Note that
13535 focus distance is only used for vignetting and only slightly affects the
13536 vignetting correction process. If unknown, leave it at the default value (which
13540 The scale factor which is applied after transformation. After correction the
13541 video is no longer necessarily rectangular. This parameter controls how much of
13542 the resulting image is visible. The value 0 means that a value will be chosen
13543 automatically such that there is little or no unmapped area in the output
13544 image. 1.0 means that no additional scaling is done. Lower values may result
13545 in more of the corrected image being visible, while higher values may avoid
13546 unmapped areas in the output.
13548 @item target_geometry
13549 The target geometry of the output image/video. The following values are valid
13553 @item rectilinear (default)
13556 @item equirectangular
13557 @item fisheye_orthographic
13558 @item fisheye_stereographic
13559 @item fisheye_equisolid
13560 @item fisheye_thoby
13563 Apply the reverse of image correction (instead of correcting distortion, apply
13566 @item interpolation
13567 The type of interpolation used when correcting distortion. The following values
13572 @item linear (default)
13577 @subsection Examples
13581 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
13582 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
13586 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
13590 Apply the same as before, but only for the first 5 seconds of video.
13593 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
13600 Obtain the VMAF (Video Multi-Method Assessment Fusion)
13601 score between two input videos.
13603 The obtained VMAF score is printed through the logging system.
13605 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
13606 After installing the library it can be enabled using:
13607 @code{./configure --enable-libvmaf}.
13608 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
13610 The filter has following options:
13614 Set the model path which is to be used for SVM.
13615 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
13618 Set the file path to be used to store logs.
13621 Set the format of the log file (csv, json or xml).
13623 @item enable_transform
13624 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
13625 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
13626 Default value: @code{false}
13629 Invokes the phone model which will generate VMAF scores higher than in the
13630 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
13631 Default value: @code{false}
13634 Enables computing psnr along with vmaf.
13635 Default value: @code{false}
13638 Enables computing ssim along with vmaf.
13639 Default value: @code{false}
13642 Enables computing ms_ssim along with vmaf.
13643 Default value: @code{false}
13646 Set the pool method to be used for computing vmaf.
13647 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
13650 Set number of threads to be used when computing vmaf.
13651 Default value: @code{0}, which makes use of all available logical processors.
13654 Set interval for frame subsampling used when computing vmaf.
13655 Default value: @code{1}
13657 @item enable_conf_interval
13658 Enables confidence interval.
13659 Default value: @code{false}
13662 This filter also supports the @ref{framesync} options.
13664 @subsection Examples
13667 On the below examples the input file @file{main.mpg} being processed is
13668 compared with the reference file @file{ref.mpg}.
13671 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
13675 Example with options:
13677 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
13681 Example with options and different containers:
13683 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 -
13689 Limits the pixel components values to the specified range [min, max].
13691 The filter accepts the following options:
13695 Lower bound. Defaults to the lowest allowed value for the input.
13698 Upper bound. Defaults to the highest allowed value for the input.
13701 Specify which planes will be processed. Defaults to all available.
13704 @subsection Commands
13706 This filter supports the all above options as @ref{commands}.
13712 The filter accepts the following options:
13716 Set the number of loops. Setting this value to -1 will result in infinite loops.
13720 Set maximal size in number of frames. Default is 0.
13723 Set first frame of loop. Default is 0.
13726 @subsection Examples
13730 Loop single first frame infinitely:
13732 loop=loop=-1:size=1:start=0
13736 Loop single first frame 10 times:
13738 loop=loop=10:size=1:start=0
13742 Loop 10 first frames 5 times:
13744 loop=loop=5:size=10:start=0
13750 Apply a 1D LUT to an input video.
13752 The filter accepts the following options:
13756 Set the 1D LUT file name.
13758 Currently supported formats:
13767 Select interpolation mode.
13769 Available values are:
13773 Use values from the nearest defined point.
13775 Interpolate values using the linear interpolation.
13777 Interpolate values using the cosine interpolation.
13779 Interpolate values using the cubic interpolation.
13781 Interpolate values using the spline interpolation.
13788 Apply a 3D LUT to an input video.
13790 The filter accepts the following options:
13794 Set the 3D LUT file name.
13796 Currently supported formats:
13810 Select interpolation mode.
13812 Available values are:
13816 Use values from the nearest defined point.
13818 Interpolate values using the 8 points defining a cube.
13820 Interpolate values using a tetrahedron.
13826 Turn certain luma values into transparency.
13828 The filter accepts the following options:
13832 Set the luma which will be used as base for transparency.
13833 Default value is @code{0}.
13836 Set the range of luma values to be keyed out.
13837 Default value is @code{0.01}.
13840 Set the range of softness. Default value is @code{0}.
13841 Use this to control gradual transition from zero to full transparency.
13844 @subsection Commands
13845 This filter supports same @ref{commands} as options.
13846 The command accepts the same syntax of the corresponding option.
13848 If the specified expression is not valid, it is kept at its current
13851 @section lut, lutrgb, lutyuv
13853 Compute a look-up table for binding each pixel component input value
13854 to an output value, and apply it to the input video.
13856 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
13857 to an RGB input video.
13859 These filters accept the following parameters:
13862 set first pixel component expression
13864 set second pixel component expression
13866 set third pixel component expression
13868 set fourth pixel component expression, corresponds to the alpha component
13871 set red component expression
13873 set green component expression
13875 set blue component expression
13877 alpha component expression
13880 set Y/luminance component expression
13882 set U/Cb component expression
13884 set V/Cr component expression
13887 Each of them specifies the expression to use for computing the lookup table for
13888 the corresponding pixel component values.
13890 The exact component associated to each of the @var{c*} options depends on the
13893 The @var{lut} filter requires either YUV or RGB pixel formats in input,
13894 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
13896 The expressions can contain the following constants and functions:
13901 The input width and height.
13904 The input value for the pixel component.
13907 The input value, clipped to the @var{minval}-@var{maxval} range.
13910 The maximum value for the pixel component.
13913 The minimum value for the pixel component.
13916 The negated value for the pixel component value, clipped to the
13917 @var{minval}-@var{maxval} range; it corresponds to the expression
13918 "maxval-clipval+minval".
13921 The computed value in @var{val}, clipped to the
13922 @var{minval}-@var{maxval} range.
13924 @item gammaval(gamma)
13925 The computed gamma correction value of the pixel component value,
13926 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
13928 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
13932 All expressions default to "val".
13934 @subsection Examples
13938 Negate input video:
13940 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
13941 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
13944 The above is the same as:
13946 lutrgb="r=negval:g=negval:b=negval"
13947 lutyuv="y=negval:u=negval:v=negval"
13957 Remove chroma components, turning the video into a graytone image:
13959 lutyuv="u=128:v=128"
13963 Apply a luma burning effect:
13969 Remove green and blue components:
13975 Set a constant alpha channel value on input:
13977 format=rgba,lutrgb=a="maxval-minval/2"
13981 Correct luminance gamma by a factor of 0.5:
13983 lutyuv=y=gammaval(0.5)
13987 Discard least significant bits of luma:
13989 lutyuv=y='bitand(val, 128+64+32)'
13993 Technicolor like effect:
13995 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
13999 @section lut2, tlut2
14001 The @code{lut2} filter takes two input streams and outputs one
14004 The @code{tlut2} (time lut2) filter takes two consecutive frames
14005 from one single stream.
14007 This filter accepts the following parameters:
14010 set first pixel component expression
14012 set second pixel component expression
14014 set third pixel component expression
14016 set fourth pixel component expression, corresponds to the alpha component
14019 set output bit depth, only available for @code{lut2} filter. By default is 0,
14020 which means bit depth is automatically picked from first input format.
14023 The @code{lut2} filter also supports the @ref{framesync} options.
14025 Each of them specifies the expression to use for computing the lookup table for
14026 the corresponding pixel component values.
14028 The exact component associated to each of the @var{c*} options depends on the
14031 The expressions can contain the following constants:
14036 The input width and height.
14039 The first input value for the pixel component.
14042 The second input value for the pixel component.
14045 The first input video bit depth.
14048 The second input video bit depth.
14051 All expressions default to "x".
14053 @subsection Examples
14057 Highlight differences between two RGB video streams:
14059 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)'
14063 Highlight differences between two YUV video streams:
14065 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)'
14069 Show max difference between two video streams:
14071 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)))'
14075 @section maskedclamp
14077 Clamp the first input stream with the second input and third input stream.
14079 Returns the value of first stream to be between second input
14080 stream - @code{undershoot} and third input stream + @code{overshoot}.
14082 This filter accepts the following options:
14085 Default value is @code{0}.
14088 Default value is @code{0}.
14091 Set which planes will be processed as bitmap, unprocessed planes will be
14092 copied from first stream.
14093 By default value 0xf, all planes will be processed.
14096 @subsection Commands
14098 This filter supports the all above options as @ref{commands}.
14102 Merge the second and third input stream into output stream using absolute differences
14103 between second input stream and first input stream and absolute difference between
14104 third input stream and first input stream. The picked value will be from second input
14105 stream if second absolute difference is greater than first one or from third input stream
14108 This filter accepts the following options:
14111 Set which planes will be processed as bitmap, unprocessed planes will be
14112 copied from first stream.
14113 By default value 0xf, all planes will be processed.
14116 @subsection Commands
14118 This filter supports the all above options as @ref{commands}.
14120 @section maskedmerge
14122 Merge the first input stream with the second input stream using per pixel
14123 weights in the third input stream.
14125 A value of 0 in the third stream pixel component means that pixel component
14126 from first stream is returned unchanged, while maximum value (eg. 255 for
14127 8-bit videos) means that pixel component from second stream is returned
14128 unchanged. Intermediate values define the amount of merging between both
14129 input stream's pixel components.
14131 This filter accepts the following options:
14134 Set which planes will be processed as bitmap, unprocessed planes will be
14135 copied from first stream.
14136 By default value 0xf, all planes will be processed.
14139 @subsection Commands
14141 This filter supports the all above options as @ref{commands}.
14145 Merge the second and third input stream into output stream using absolute differences
14146 between second input stream and first input stream and absolute difference between
14147 third input stream and first input stream. The picked value will be from second input
14148 stream if second absolute difference is less than first one or from third input stream
14151 This filter accepts the following options:
14154 Set which planes will be processed as bitmap, unprocessed planes will be
14155 copied from first stream.
14156 By default value 0xf, all planes will be processed.
14159 @subsection Commands
14161 This filter supports the all above options as @ref{commands}.
14163 @section maskedthreshold
14164 Pick pixels comparing absolute difference of two video streams with fixed
14167 If absolute difference between pixel component of first and second video
14168 stream is equal or lower than user supplied threshold than pixel component
14169 from first video stream is picked, otherwise pixel component from second
14170 video stream is picked.
14172 This filter accepts the following options:
14175 Set threshold used when picking pixels from absolute difference from two input
14179 Set which planes will be processed as bitmap, unprocessed planes will be
14180 copied from second stream.
14181 By default value 0xf, all planes will be processed.
14184 @subsection Commands
14186 This filter supports the all above options as @ref{commands}.
14189 Create mask from input video.
14191 For example it is useful to create motion masks after @code{tblend} filter.
14193 This filter accepts the following options:
14197 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
14200 Set high threshold. Any pixel component higher than this value will be set to max value
14201 allowed for current pixel format.
14204 Set planes to filter, by default all available planes are filtered.
14207 Fill all frame pixels with this value.
14210 Set max average pixel value for frame. If sum of all pixel components is higher that this
14211 average, output frame will be completely filled with value set by @var{fill} option.
14212 Typically useful for scene changes when used in combination with @code{tblend} filter.
14217 Apply motion-compensation deinterlacing.
14219 It needs one field per frame as input and must thus be used together
14220 with yadif=1/3 or equivalent.
14222 This filter accepts the following options:
14225 Set the deinterlacing mode.
14227 It accepts one of the following values:
14232 use iterative motion estimation
14234 like @samp{slow}, but use multiple reference frames.
14236 Default value is @samp{fast}.
14239 Set the picture field parity assumed for the input video. It must be
14240 one of the following values:
14244 assume top field first
14246 assume bottom field first
14249 Default value is @samp{bff}.
14252 Set per-block quantization parameter (QP) used by the internal
14255 Higher values should result in a smoother motion vector field but less
14256 optimal individual vectors. Default value is 1.
14261 Pick median pixel from certain rectangle defined by radius.
14263 This filter accepts the following options:
14267 Set horizontal radius size. Default value is @code{1}.
14268 Allowed range is integer from 1 to 127.
14271 Set which planes to process. Default is @code{15}, which is all available planes.
14274 Set vertical radius size. Default value is @code{0}.
14275 Allowed range is integer from 0 to 127.
14276 If it is 0, value will be picked from horizontal @code{radius} option.
14279 Set median percentile. Default value is @code{0.5}.
14280 Default value of @code{0.5} will pick always median values, while @code{0} will pick
14281 minimum values, and @code{1} maximum values.
14284 @subsection Commands
14285 This filter supports same @ref{commands} as options.
14286 The command accepts the same syntax of the corresponding option.
14288 If the specified expression is not valid, it is kept at its current
14291 @section mergeplanes
14293 Merge color channel components from several video streams.
14295 The filter accepts up to 4 input streams, and merge selected input
14296 planes to the output video.
14298 This filter accepts the following options:
14301 Set input to output plane mapping. Default is @code{0}.
14303 The mappings is specified as a bitmap. It should be specified as a
14304 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
14305 mapping for the first plane of the output stream. 'A' sets the number of
14306 the input stream to use (from 0 to 3), and 'a' the plane number of the
14307 corresponding input to use (from 0 to 3). The rest of the mappings is
14308 similar, 'Bb' describes the mapping for the output stream second
14309 plane, 'Cc' describes the mapping for the output stream third plane and
14310 'Dd' describes the mapping for the output stream fourth plane.
14313 Set output pixel format. Default is @code{yuva444p}.
14316 @subsection Examples
14320 Merge three gray video streams of same width and height into single video stream:
14322 [a0][a1][a2]mergeplanes=0x001020:yuv444p
14326 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
14328 [a0][a1]mergeplanes=0x00010210:yuva444p
14332 Swap Y and A plane in yuva444p stream:
14334 format=yuva444p,mergeplanes=0x03010200:yuva444p
14338 Swap U and V plane in yuv420p stream:
14340 format=yuv420p,mergeplanes=0x000201:yuv420p
14344 Cast a rgb24 clip to yuv444p:
14346 format=rgb24,mergeplanes=0x000102:yuv444p
14352 Estimate and export motion vectors using block matching algorithms.
14353 Motion vectors are stored in frame side data to be used by other filters.
14355 This filter accepts the following options:
14358 Specify the motion estimation method. Accepts one of the following values:
14362 Exhaustive search algorithm.
14364 Three step search algorithm.
14366 Two dimensional logarithmic search algorithm.
14368 New three step search algorithm.
14370 Four step search algorithm.
14372 Diamond search algorithm.
14374 Hexagon-based search algorithm.
14376 Enhanced predictive zonal search algorithm.
14378 Uneven multi-hexagon search algorithm.
14380 Default value is @samp{esa}.
14383 Macroblock size. Default @code{16}.
14386 Search parameter. Default @code{7}.
14389 @section midequalizer
14391 Apply Midway Image Equalization effect using two video streams.
14393 Midway Image Equalization adjusts a pair of images to have the same
14394 histogram, while maintaining their dynamics as much as possible. It's
14395 useful for e.g. matching exposures from a pair of stereo cameras.
14397 This filter has two inputs and one output, which must be of same pixel format, but
14398 may be of different sizes. The output of filter is first input adjusted with
14399 midway histogram of both inputs.
14401 This filter accepts the following option:
14405 Set which planes to process. Default is @code{15}, which is all available planes.
14408 @section minterpolate
14410 Convert the video to specified frame rate using motion interpolation.
14412 This filter accepts the following options:
14415 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}.
14418 Motion interpolation mode. Following values are accepted:
14421 Duplicate previous or next frame for interpolating new ones.
14423 Blend source frames. Interpolated frame is mean of previous and next frames.
14425 Motion compensated interpolation. Following options are effective when this mode is selected:
14429 Motion compensation mode. Following values are accepted:
14432 Overlapped block motion compensation.
14434 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
14436 Default mode is @samp{obmc}.
14439 Motion estimation mode. Following values are accepted:
14442 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
14444 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
14446 Default mode is @samp{bilat}.
14449 The algorithm to be used for motion estimation. Following values are accepted:
14452 Exhaustive search algorithm.
14454 Three step search algorithm.
14456 Two dimensional logarithmic search algorithm.
14458 New three step search algorithm.
14460 Four step search algorithm.
14462 Diamond search algorithm.
14464 Hexagon-based search algorithm.
14466 Enhanced predictive zonal search algorithm.
14468 Uneven multi-hexagon search algorithm.
14470 Default algorithm is @samp{epzs}.
14473 Macroblock size. Default @code{16}.
14476 Motion estimation search parameter. Default @code{32}.
14479 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).
14484 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:
14487 Disable scene change detection.
14489 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
14491 Default method is @samp{fdiff}.
14493 @item scd_threshold
14494 Scene change detection threshold. Default is @code{10.}.
14499 Mix several video input streams into one video stream.
14501 A description of the accepted options follows.
14505 The number of inputs. If unspecified, it defaults to 2.
14508 Specify weight of each input video stream as sequence.
14509 Each weight is separated by space. If number of weights
14510 is smaller than number of @var{frames} last specified
14511 weight will be used for all remaining unset weights.
14514 Specify scale, if it is set it will be multiplied with sum
14515 of each weight multiplied with pixel values to give final destination
14516 pixel value. By default @var{scale} is auto scaled to sum of weights.
14519 Specify how end of stream is determined.
14522 The duration of the longest input. (default)
14525 The duration of the shortest input.
14528 The duration of the first input.
14532 @section mpdecimate
14534 Drop frames that do not differ greatly from the previous frame in
14535 order to reduce frame rate.
14537 The main use of this filter is for very-low-bitrate encoding
14538 (e.g. streaming over dialup modem), but it could in theory be used for
14539 fixing movies that were inverse-telecined incorrectly.
14541 A description of the accepted options follows.
14545 Set the maximum number of consecutive frames which can be dropped (if
14546 positive), or the minimum interval between dropped frames (if
14547 negative). If the value is 0, the frame is dropped disregarding the
14548 number of previous sequentially dropped frames.
14550 Default value is 0.
14555 Set the dropping threshold values.
14557 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
14558 represent actual pixel value differences, so a threshold of 64
14559 corresponds to 1 unit of difference for each pixel, or the same spread
14560 out differently over the block.
14562 A frame is a candidate for dropping if no 8x8 blocks differ by more
14563 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
14564 meaning the whole image) differ by more than a threshold of @option{lo}.
14566 Default value for @option{hi} is 64*12, default value for @option{lo} is
14567 64*5, and default value for @option{frac} is 0.33.
14573 Negate (invert) the input video.
14575 It accepts the following option:
14580 With value 1, it negates the alpha component, if present. Default value is 0.
14586 Denoise frames using Non-Local Means algorithm.
14588 Each pixel is adjusted by looking for other pixels with similar contexts. This
14589 context similarity is defined by comparing their surrounding patches of size
14590 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
14593 Note that the research area defines centers for patches, which means some
14594 patches will be made of pixels outside that research area.
14596 The filter accepts the following options.
14600 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
14603 Set patch size. Default is 7. Must be odd number in range [0, 99].
14606 Same as @option{p} but for chroma planes.
14608 The default value is @var{0} and means automatic.
14611 Set research size. Default is 15. Must be odd number in range [0, 99].
14614 Same as @option{r} but for chroma planes.
14616 The default value is @var{0} and means automatic.
14621 Deinterlace video using neural network edge directed interpolation.
14623 This filter accepts the following options:
14627 Mandatory option, without binary file filter can not work.
14628 Currently file can be found here:
14629 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
14632 Set which frames to deinterlace, by default it is @code{all}.
14633 Can be @code{all} or @code{interlaced}.
14636 Set mode of operation.
14638 Can be one of the following:
14642 Use frame flags, both fields.
14644 Use frame flags, single field.
14646 Use top field only.
14648 Use bottom field only.
14650 Use both fields, top first.
14652 Use both fields, bottom first.
14656 Set which planes to process, by default filter process all frames.
14659 Set size of local neighborhood around each pixel, used by the predictor neural
14662 Can be one of the following:
14675 Set the number of neurons in predictor neural network.
14676 Can be one of the following:
14687 Controls the number of different neural network predictions that are blended
14688 together to compute the final output value. Can be @code{fast}, default or
14692 Set which set of weights to use in the predictor.
14693 Can be one of the following:
14697 weights trained to minimize absolute error
14699 weights trained to minimize squared error
14703 Controls whether or not the prescreener neural network is used to decide
14704 which pixels should be processed by the predictor neural network and which
14705 can be handled by simple cubic interpolation.
14706 The prescreener is trained to know whether cubic interpolation will be
14707 sufficient for a pixel or whether it should be predicted by the predictor nn.
14708 The computational complexity of the prescreener nn is much less than that of
14709 the predictor nn. Since most pixels can be handled by cubic interpolation,
14710 using the prescreener generally results in much faster processing.
14711 The prescreener is pretty accurate, so the difference between using it and not
14712 using it is almost always unnoticeable.
14714 Can be one of the following:
14724 Default is @code{new}.
14727 @subsection Commands
14728 This filter supports same @ref{commands} as options, excluding @var{weights} option.
14732 Force libavfilter not to use any of the specified pixel formats for the
14733 input to the next filter.
14735 It accepts the following parameters:
14739 A '|'-separated list of pixel format names, such as
14740 pix_fmts=yuv420p|monow|rgb24".
14744 @subsection Examples
14748 Force libavfilter to use a format different from @var{yuv420p} for the
14749 input to the vflip filter:
14751 noformat=pix_fmts=yuv420p,vflip
14755 Convert the input video to any of the formats not contained in the list:
14757 noformat=yuv420p|yuv444p|yuv410p
14763 Add noise on video input frame.
14765 The filter accepts the following options:
14773 Set noise seed for specific pixel component or all pixel components in case
14774 of @var{all_seed}. Default value is @code{123457}.
14776 @item all_strength, alls
14777 @item c0_strength, c0s
14778 @item c1_strength, c1s
14779 @item c2_strength, c2s
14780 @item c3_strength, c3s
14781 Set noise strength for specific pixel component or all pixel components in case
14782 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
14784 @item all_flags, allf
14785 @item c0_flags, c0f
14786 @item c1_flags, c1f
14787 @item c2_flags, c2f
14788 @item c3_flags, c3f
14789 Set pixel component flags or set flags for all components if @var{all_flags}.
14790 Available values for component flags are:
14793 averaged temporal noise (smoother)
14795 mix random noise with a (semi)regular pattern
14797 temporal noise (noise pattern changes between frames)
14799 uniform noise (gaussian otherwise)
14803 @subsection Examples
14805 Add temporal and uniform noise to input video:
14807 noise=alls=20:allf=t+u
14812 Normalize RGB video (aka histogram stretching, contrast stretching).
14813 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
14815 For each channel of each frame, the filter computes the input range and maps
14816 it linearly to the user-specified output range. The output range defaults
14817 to the full dynamic range from pure black to pure white.
14819 Temporal smoothing can be used on the input range to reduce flickering (rapid
14820 changes in brightness) caused when small dark or bright objects enter or leave
14821 the scene. This is similar to the auto-exposure (automatic gain control) on a
14822 video camera, and, like a video camera, it may cause a period of over- or
14823 under-exposure of the video.
14825 The R,G,B channels can be normalized independently, which may cause some
14826 color shifting, or linked together as a single channel, which prevents
14827 color shifting. Linked normalization preserves hue. Independent normalization
14828 does not, so it can be used to remove some color casts. Independent and linked
14829 normalization can be combined in any ratio.
14831 The normalize filter accepts the following options:
14836 Colors which define the output range. The minimum input value is mapped to
14837 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
14838 The defaults are black and white respectively. Specifying white for
14839 @var{blackpt} and black for @var{whitept} will give color-inverted,
14840 normalized video. Shades of grey can be used to reduce the dynamic range
14841 (contrast). Specifying saturated colors here can create some interesting
14845 The number of previous frames to use for temporal smoothing. The input range
14846 of each channel is smoothed using a rolling average over the current frame
14847 and the @var{smoothing} previous frames. The default is 0 (no temporal
14851 Controls the ratio of independent (color shifting) channel normalization to
14852 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
14853 independent. Defaults to 1.0 (fully independent).
14856 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
14857 expensive no-op. Defaults to 1.0 (full strength).
14861 @subsection Commands
14862 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
14863 The command accepts the same syntax of the corresponding option.
14865 If the specified expression is not valid, it is kept at its current
14868 @subsection Examples
14870 Stretch video contrast to use the full dynamic range, with no temporal
14871 smoothing; may flicker depending on the source content:
14873 normalize=blackpt=black:whitept=white:smoothing=0
14876 As above, but with 50 frames of temporal smoothing; flicker should be
14877 reduced, depending on the source content:
14879 normalize=blackpt=black:whitept=white:smoothing=50
14882 As above, but with hue-preserving linked channel normalization:
14884 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
14887 As above, but with half strength:
14889 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
14892 Map the darkest input color to red, the brightest input color to cyan:
14894 normalize=blackpt=red:whitept=cyan
14899 Pass the video source unchanged to the output.
14902 Optical Character Recognition
14904 This filter uses Tesseract for optical character recognition. To enable
14905 compilation of this filter, you need to configure FFmpeg with
14906 @code{--enable-libtesseract}.
14908 It accepts the following options:
14912 Set datapath to tesseract data. Default is to use whatever was
14913 set at installation.
14916 Set language, default is "eng".
14919 Set character whitelist.
14922 Set character blacklist.
14925 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
14926 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
14930 Apply a video transform using libopencv.
14932 To enable this filter, install the libopencv library and headers and
14933 configure FFmpeg with @code{--enable-libopencv}.
14935 It accepts the following parameters:
14940 The name of the libopencv filter to apply.
14942 @item filter_params
14943 The parameters to pass to the libopencv filter. If not specified, the default
14944 values are assumed.
14948 Refer to the official libopencv documentation for more precise
14950 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
14952 Several libopencv filters are supported; see the following subsections.
14957 Dilate an image by using a specific structuring element.
14958 It corresponds to the libopencv function @code{cvDilate}.
14960 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
14962 @var{struct_el} represents a structuring element, and has the syntax:
14963 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
14965 @var{cols} and @var{rows} represent the number of columns and rows of
14966 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
14967 point, and @var{shape} the shape for the structuring element. @var{shape}
14968 must be "rect", "cross", "ellipse", or "custom".
14970 If the value for @var{shape} is "custom", it must be followed by a
14971 string of the form "=@var{filename}". The file with name
14972 @var{filename} is assumed to represent a binary image, with each
14973 printable character corresponding to a bright pixel. When a custom
14974 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
14975 or columns and rows of the read file are assumed instead.
14977 The default value for @var{struct_el} is "3x3+0x0/rect".
14979 @var{nb_iterations} specifies the number of times the transform is
14980 applied to the image, and defaults to 1.
14984 # Use the default values
14987 # Dilate using a structuring element with a 5x5 cross, iterating two times
14988 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
14990 # Read the shape from the file diamond.shape, iterating two times.
14991 # The file diamond.shape may contain a pattern of characters like this
14997 # The specified columns and rows are ignored
14998 # but the anchor point coordinates are not
14999 ocv=dilate:0x0+2x2/custom=diamond.shape|2
15004 Erode an image by using a specific structuring element.
15005 It corresponds to the libopencv function @code{cvErode}.
15007 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
15008 with the same syntax and semantics as the @ref{dilate} filter.
15012 Smooth the input video.
15014 The filter takes the following parameters:
15015 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
15017 @var{type} is the type of smooth filter to apply, and must be one of
15018 the following values: "blur", "blur_no_scale", "median", "gaussian",
15019 or "bilateral". The default value is "gaussian".
15021 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
15022 depends on the smooth type. @var{param1} and
15023 @var{param2} accept integer positive values or 0. @var{param3} and
15024 @var{param4} accept floating point values.
15026 The default value for @var{param1} is 3. The default value for the
15027 other parameters is 0.
15029 These parameters correspond to the parameters assigned to the
15030 libopencv function @code{cvSmooth}.
15032 @section oscilloscope
15034 2D Video Oscilloscope.
15036 Useful to measure spatial impulse, step responses, chroma delays, etc.
15038 It accepts the following parameters:
15042 Set scope center x position.
15045 Set scope center y position.
15048 Set scope size, relative to frame diagonal.
15051 Set scope tilt/rotation.
15057 Set trace center x position.
15060 Set trace center y position.
15063 Set trace width, relative to width of frame.
15066 Set trace height, relative to height of frame.
15069 Set which components to trace. By default it traces first three components.
15072 Draw trace grid. By default is enabled.
15075 Draw some statistics. By default is enabled.
15078 Draw scope. By default is enabled.
15081 @subsection Commands
15082 This filter supports same @ref{commands} as options.
15083 The command accepts the same syntax of the corresponding option.
15085 If the specified expression is not valid, it is kept at its current
15088 @subsection Examples
15092 Inspect full first row of video frame.
15094 oscilloscope=x=0.5:y=0:s=1
15098 Inspect full last row of video frame.
15100 oscilloscope=x=0.5:y=1:s=1
15104 Inspect full 5th line of video frame of height 1080.
15106 oscilloscope=x=0.5:y=5/1080:s=1
15110 Inspect full last column of video frame.
15112 oscilloscope=x=1:y=0.5:s=1:t=1
15120 Overlay one video on top of another.
15122 It takes two inputs and has one output. The first input is the "main"
15123 video on which the second input is overlaid.
15125 It accepts the following parameters:
15127 A description of the accepted options follows.
15132 Set the expression for the x and y coordinates of the overlaid video
15133 on the main video. Default value is "0" for both expressions. In case
15134 the expression is invalid, it is set to a huge value (meaning that the
15135 overlay will not be displayed within the output visible area).
15138 See @ref{framesync}.
15141 Set when the expressions for @option{x}, and @option{y} are evaluated.
15143 It accepts the following values:
15146 only evaluate expressions once during the filter initialization or
15147 when a command is processed
15150 evaluate expressions for each incoming frame
15153 Default value is @samp{frame}.
15156 See @ref{framesync}.
15159 Set the format for the output video.
15161 It accepts the following values:
15164 force YUV420 output
15167 force YUV420p10 output
15170 force YUV422 output
15173 force YUV422p10 output
15176 force YUV444 output
15179 force packed RGB output
15182 force planar RGB output
15185 automatically pick format
15188 Default value is @samp{yuv420}.
15191 See @ref{framesync}.
15194 Set format of alpha of the overlaid video, it can be @var{straight} or
15195 @var{premultiplied}. Default is @var{straight}.
15198 The @option{x}, and @option{y} expressions can contain the following
15204 The main input width and height.
15208 The overlay input width and height.
15212 The computed values for @var{x} and @var{y}. They are evaluated for
15217 horizontal and vertical chroma subsample values of the output
15218 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
15222 the number of input frame, starting from 0
15225 the position in the file of the input frame, NAN if unknown
15228 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
15232 This filter also supports the @ref{framesync} options.
15234 Note that the @var{n}, @var{pos}, @var{t} variables are available only
15235 when evaluation is done @emph{per frame}, and will evaluate to NAN
15236 when @option{eval} is set to @samp{init}.
15238 Be aware that frames are taken from each input video in timestamp
15239 order, hence, if their initial timestamps differ, it is a good idea
15240 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
15241 have them begin in the same zero timestamp, as the example for
15242 the @var{movie} filter does.
15244 You can chain together more overlays but you should test the
15245 efficiency of such approach.
15247 @subsection Commands
15249 This filter supports the following commands:
15253 Modify the x and y of the overlay input.
15254 The command accepts the same syntax of the corresponding option.
15256 If the specified expression is not valid, it is kept at its current
15260 @subsection Examples
15264 Draw the overlay at 10 pixels from the bottom right corner of the main
15267 overlay=main_w-overlay_w-10:main_h-overlay_h-10
15270 Using named options the example above becomes:
15272 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
15276 Insert a transparent PNG logo in the bottom left corner of the input,
15277 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
15279 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
15283 Insert 2 different transparent PNG logos (second logo on bottom
15284 right corner) using the @command{ffmpeg} tool:
15286 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
15290 Add a transparent color layer on top of the main video; @code{WxH}
15291 must specify the size of the main input to the overlay filter:
15293 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
15297 Play an original video and a filtered version (here with the deshake
15298 filter) side by side using the @command{ffplay} tool:
15300 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
15303 The above command is the same as:
15305 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
15309 Make a sliding overlay appearing from the left to the right top part of the
15310 screen starting since time 2:
15312 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
15316 Compose output by putting two input videos side to side:
15318 ffmpeg -i left.avi -i right.avi -filter_complex "
15319 nullsrc=size=200x100 [background];
15320 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
15321 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
15322 [background][left] overlay=shortest=1 [background+left];
15323 [background+left][right] overlay=shortest=1:x=100 [left+right]
15328 Mask 10-20 seconds of a video by applying the delogo filter to a section
15330 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
15331 -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]'
15336 Chain several overlays in cascade:
15338 nullsrc=s=200x200 [bg];
15339 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
15340 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
15341 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
15342 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
15343 [in3] null, [mid2] overlay=100:100 [out0]
15348 @anchor{overlay_cuda}
15349 @section overlay_cuda
15351 Overlay one video on top of another.
15353 This is the CUDA variant of the @ref{overlay} filter.
15354 It only accepts CUDA frames. The underlying input pixel formats have to match.
15356 It takes two inputs and has one output. The first input is the "main"
15357 video on which the second input is overlaid.
15359 It accepts the following parameters:
15364 Set the x and y coordinates of the overlaid video on the main video.
15365 Default value is "0" for both expressions.
15368 See @ref{framesync}.
15371 See @ref{framesync}.
15374 See @ref{framesync}.
15378 This filter also supports the @ref{framesync} options.
15382 Apply Overcomplete Wavelet denoiser.
15384 The filter accepts the following options:
15390 Larger depth values will denoise lower frequency components more, but
15391 slow down filtering.
15393 Must be an int in the range 8-16, default is @code{8}.
15395 @item luma_strength, ls
15398 Must be a double value in the range 0-1000, default is @code{1.0}.
15400 @item chroma_strength, cs
15401 Set chroma strength.
15403 Must be a double value in the range 0-1000, default is @code{1.0}.
15409 Add paddings to the input image, and place the original input at the
15410 provided @var{x}, @var{y} coordinates.
15412 It accepts the following parameters:
15417 Specify an expression for the size of the output image with the
15418 paddings added. If the value for @var{width} or @var{height} is 0, the
15419 corresponding input size is used for the output.
15421 The @var{width} expression can reference the value set by the
15422 @var{height} expression, and vice versa.
15424 The default value of @var{width} and @var{height} is 0.
15428 Specify the offsets to place the input image at within the padded area,
15429 with respect to the top/left border of the output image.
15431 The @var{x} expression can reference the value set by the @var{y}
15432 expression, and vice versa.
15434 The default value of @var{x} and @var{y} is 0.
15436 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
15437 so the input image is centered on the padded area.
15440 Specify the color of the padded area. For the syntax of this option,
15441 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
15442 manual,ffmpeg-utils}.
15444 The default value of @var{color} is "black".
15447 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
15449 It accepts the following values:
15453 Only evaluate expressions once during the filter initialization or when
15454 a command is processed.
15457 Evaluate expressions for each incoming frame.
15461 Default value is @samp{init}.
15464 Pad to aspect instead to a resolution.
15468 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
15469 options are expressions containing the following constants:
15474 The input video width and height.
15478 These are the same as @var{in_w} and @var{in_h}.
15482 The output width and height (the size of the padded area), as
15483 specified by the @var{width} and @var{height} expressions.
15487 These are the same as @var{out_w} and @var{out_h}.
15491 The x and y offsets as specified by the @var{x} and @var{y}
15492 expressions, or NAN if not yet specified.
15495 same as @var{iw} / @var{ih}
15498 input sample aspect ratio
15501 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
15505 The horizontal and vertical chroma subsample values. For example for the
15506 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15509 @subsection Examples
15513 Add paddings with the color "violet" to the input video. The output video
15514 size is 640x480, and the top-left corner of the input video is placed at
15517 pad=640:480:0:40:violet
15520 The example above is equivalent to the following command:
15522 pad=width=640:height=480:x=0:y=40:color=violet
15526 Pad the input to get an output with dimensions increased by 3/2,
15527 and put the input video at the center of the padded area:
15529 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
15533 Pad the input to get a squared output with size equal to the maximum
15534 value between the input width and height, and put the input video at
15535 the center of the padded area:
15537 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
15541 Pad the input to get a final w/h ratio of 16:9:
15543 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
15547 In case of anamorphic video, in order to set the output display aspect
15548 correctly, it is necessary to use @var{sar} in the expression,
15549 according to the relation:
15551 (ih * X / ih) * sar = output_dar
15552 X = output_dar / sar
15555 Thus the previous example needs to be modified to:
15557 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
15561 Double the output size and put the input video in the bottom-right
15562 corner of the output padded area:
15564 pad="2*iw:2*ih:ow-iw:oh-ih"
15568 @anchor{palettegen}
15569 @section palettegen
15571 Generate one palette for a whole video stream.
15573 It accepts the following options:
15577 Set the maximum number of colors to quantize in the palette.
15578 Note: the palette will still contain 256 colors; the unused palette entries
15581 @item reserve_transparent
15582 Create a palette of 255 colors maximum and reserve the last one for
15583 transparency. Reserving the transparency color is useful for GIF optimization.
15584 If not set, the maximum of colors in the palette will be 256. You probably want
15585 to disable this option for a standalone image.
15588 @item transparency_color
15589 Set the color that will be used as background for transparency.
15592 Set statistics mode.
15594 It accepts the following values:
15597 Compute full frame histograms.
15599 Compute histograms only for the part that differs from previous frame. This
15600 might be relevant to give more importance to the moving part of your input if
15601 the background is static.
15603 Compute new histogram for each frame.
15606 Default value is @var{full}.
15609 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
15610 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
15611 color quantization of the palette. This information is also visible at
15612 @var{info} logging level.
15614 @subsection Examples
15618 Generate a representative palette of a given video using @command{ffmpeg}:
15620 ffmpeg -i input.mkv -vf palettegen palette.png
15624 @section paletteuse
15626 Use a palette to downsample an input video stream.
15628 The filter takes two inputs: one video stream and a palette. The palette must
15629 be a 256 pixels image.
15631 It accepts the following options:
15635 Select dithering mode. Available algorithms are:
15638 Ordered 8x8 bayer dithering (deterministic)
15640 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
15641 Note: this dithering is sometimes considered "wrong" and is included as a
15643 @item floyd_steinberg
15644 Floyd and Steingberg dithering (error diffusion)
15646 Frankie Sierra dithering v2 (error diffusion)
15648 Frankie Sierra dithering v2 "Lite" (error diffusion)
15651 Default is @var{sierra2_4a}.
15654 When @var{bayer} dithering is selected, this option defines the scale of the
15655 pattern (how much the crosshatch pattern is visible). A low value means more
15656 visible pattern for less banding, and higher value means less visible pattern
15657 at the cost of more banding.
15659 The option must be an integer value in the range [0,5]. Default is @var{2}.
15662 If set, define the zone to process
15666 Only the changing rectangle will be reprocessed. This is similar to GIF
15667 cropping/offsetting compression mechanism. This option can be useful for speed
15668 if only a part of the image is changing, and has use cases such as limiting the
15669 scope of the error diffusal @option{dither} to the rectangle that bounds the
15670 moving scene (it leads to more deterministic output if the scene doesn't change
15671 much, and as a result less moving noise and better GIF compression).
15674 Default is @var{none}.
15677 Take new palette for each output frame.
15679 @item alpha_threshold
15680 Sets the alpha threshold for transparency. Alpha values above this threshold
15681 will be treated as completely opaque, and values below this threshold will be
15682 treated as completely transparent.
15684 The option must be an integer value in the range [0,255]. Default is @var{128}.
15687 @subsection Examples
15691 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
15692 using @command{ffmpeg}:
15694 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
15698 @section perspective
15700 Correct perspective of video not recorded perpendicular to the screen.
15702 A description of the accepted parameters follows.
15713 Set coordinates expression for top left, top right, bottom left and bottom right corners.
15714 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
15715 If the @code{sense} option is set to @code{source}, then the specified points will be sent
15716 to the corners of the destination. If the @code{sense} option is set to @code{destination},
15717 then the corners of the source will be sent to the specified coordinates.
15719 The expressions can use the following variables:
15724 the width and height of video frame.
15728 Output frame count.
15731 @item interpolation
15732 Set interpolation for perspective correction.
15734 It accepts the following values:
15740 Default value is @samp{linear}.
15743 Set interpretation of coordinate options.
15745 It accepts the following values:
15749 Send point in the source specified by the given coordinates to
15750 the corners of the destination.
15752 @item 1, destination
15754 Send the corners of the source to the point in the destination specified
15755 by the given coordinates.
15757 Default value is @samp{source}.
15761 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
15763 It accepts the following values:
15766 only evaluate expressions once during the filter initialization or
15767 when a command is processed
15770 evaluate expressions for each incoming frame
15773 Default value is @samp{init}.
15778 Delay interlaced video by one field time so that the field order changes.
15780 The intended use is to fix PAL movies that have been captured with the
15781 opposite field order to the film-to-video transfer.
15783 A description of the accepted parameters follows.
15789 It accepts the following values:
15792 Capture field order top-first, transfer bottom-first.
15793 Filter will delay the bottom field.
15796 Capture field order bottom-first, transfer top-first.
15797 Filter will delay the top field.
15800 Capture and transfer with the same field order. This mode only exists
15801 for the documentation of the other options to refer to, but if you
15802 actually select it, the filter will faithfully do nothing.
15805 Capture field order determined automatically by field flags, transfer
15807 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
15808 basis using field flags. If no field information is available,
15809 then this works just like @samp{u}.
15812 Capture unknown or varying, transfer opposite.
15813 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
15814 analyzing the images and selecting the alternative that produces best
15815 match between the fields.
15818 Capture top-first, transfer unknown or varying.
15819 Filter selects among @samp{t} and @samp{p} using image analysis.
15822 Capture bottom-first, transfer unknown or varying.
15823 Filter selects among @samp{b} and @samp{p} using image analysis.
15826 Capture determined by field flags, transfer unknown or varying.
15827 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
15828 image analysis. If no field information is available, then this works just
15829 like @samp{U}. This is the default mode.
15832 Both capture and transfer unknown or varying.
15833 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
15837 @subsection Commands
15839 This filter supports the all above options as @ref{commands}.
15841 @section photosensitivity
15842 Reduce various flashes in video, so to help users with epilepsy.
15844 It accepts the following options:
15847 Set how many frames to use when filtering. Default is 30.
15850 Set detection threshold factor. Default is 1.
15854 Set how many pixels to skip when sampling frames. Default is 1.
15855 Allowed range is from 1 to 1024.
15858 Leave frames unchanged. Default is disabled.
15861 @section pixdesctest
15863 Pixel format descriptor test filter, mainly useful for internal
15864 testing. The output video should be equal to the input video.
15868 format=monow, pixdesctest
15871 can be used to test the monowhite pixel format descriptor definition.
15875 Display sample values of color channels. Mainly useful for checking color
15876 and levels. Minimum supported resolution is 640x480.
15878 The filters accept the following options:
15882 Set scope X position, relative offset on X axis.
15885 Set scope Y position, relative offset on Y axis.
15894 Set window opacity. This window also holds statistics about pixel area.
15897 Set window X position, relative offset on X axis.
15900 Set window Y position, relative offset on Y axis.
15905 Enable the specified chain of postprocessing subfilters using libpostproc. This
15906 library should be automatically selected with a GPL build (@code{--enable-gpl}).
15907 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
15908 Each subfilter and some options have a short and a long name that can be used
15909 interchangeably, i.e. dr/dering are the same.
15911 The filters accept the following options:
15915 Set postprocessing subfilters string.
15918 All subfilters share common options to determine their scope:
15922 Honor the quality commands for this subfilter.
15925 Do chrominance filtering, too (default).
15928 Do luminance filtering only (no chrominance).
15931 Do chrominance filtering only (no luminance).
15934 These options can be appended after the subfilter name, separated by a '|'.
15936 Available subfilters are:
15939 @item hb/hdeblock[|difference[|flatness]]
15940 Horizontal deblocking filter
15943 Difference factor where higher values mean more deblocking (default: @code{32}).
15945 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15948 @item vb/vdeblock[|difference[|flatness]]
15949 Vertical deblocking filter
15952 Difference factor where higher values mean more deblocking (default: @code{32}).
15954 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15957 @item ha/hadeblock[|difference[|flatness]]
15958 Accurate horizontal deblocking filter
15961 Difference factor where higher values mean more deblocking (default: @code{32}).
15963 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15966 @item va/vadeblock[|difference[|flatness]]
15967 Accurate vertical deblocking filter
15970 Difference factor where higher values mean more deblocking (default: @code{32}).
15972 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15976 The horizontal and vertical deblocking filters share the difference and
15977 flatness values so you cannot set different horizontal and vertical
15981 @item h1/x1hdeblock
15982 Experimental horizontal deblocking filter
15984 @item v1/x1vdeblock
15985 Experimental vertical deblocking filter
15990 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
15993 larger -> stronger filtering
15995 larger -> stronger filtering
15997 larger -> stronger filtering
16000 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
16003 Stretch luminance to @code{0-255}.
16006 @item lb/linblenddeint
16007 Linear blend deinterlacing filter that deinterlaces the given block by
16008 filtering all lines with a @code{(1 2 1)} filter.
16010 @item li/linipoldeint
16011 Linear interpolating deinterlacing filter that deinterlaces the given block by
16012 linearly interpolating every second line.
16014 @item ci/cubicipoldeint
16015 Cubic interpolating deinterlacing filter deinterlaces the given block by
16016 cubically interpolating every second line.
16018 @item md/mediandeint
16019 Median deinterlacing filter that deinterlaces the given block by applying a
16020 median filter to every second line.
16022 @item fd/ffmpegdeint
16023 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
16024 second line with a @code{(-1 4 2 4 -1)} filter.
16027 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
16028 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
16030 @item fq/forceQuant[|quantizer]
16031 Overrides the quantizer table from the input with the constant quantizer you
16039 Default pp filter combination (@code{hb|a,vb|a,dr|a})
16042 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
16045 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
16048 @subsection Examples
16052 Apply horizontal and vertical deblocking, deringing and automatic
16053 brightness/contrast:
16059 Apply default filters without brightness/contrast correction:
16065 Apply default filters and temporal denoiser:
16067 pp=default/tmpnoise|1|2|3
16071 Apply deblocking on luminance only, and switch vertical deblocking on or off
16072 automatically depending on available CPU time:
16079 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
16080 similar to spp = 6 with 7 point DCT, where only the center sample is
16083 The filter accepts the following options:
16087 Force a constant quantization parameter. It accepts an integer in range
16088 0 to 63. If not set, the filter will use the QP from the video stream
16092 Set thresholding mode. Available modes are:
16096 Set hard thresholding.
16098 Set soft thresholding (better de-ringing effect, but likely blurrier).
16100 Set medium thresholding (good results, default).
16104 @section premultiply
16105 Apply alpha premultiply effect to input video stream using first plane
16106 of second stream as alpha.
16108 Both streams must have same dimensions and same pixel format.
16110 The filter accepts the following option:
16114 Set which planes will be processed, unprocessed planes will be copied.
16115 By default value 0xf, all planes will be processed.
16118 Do not require 2nd input for processing, instead use alpha plane from input stream.
16122 Apply prewitt operator to input video stream.
16124 The filter accepts the following option:
16128 Set which planes will be processed, unprocessed planes will be copied.
16129 By default value 0xf, all planes will be processed.
16132 Set value which will be multiplied with filtered result.
16135 Set value which will be added to filtered result.
16138 @subsection Commands
16140 This filter supports the all above options as @ref{commands}.
16142 @section pseudocolor
16144 Alter frame colors in video with pseudocolors.
16146 This filter accepts the following options:
16150 set pixel first component expression
16153 set pixel second component expression
16156 set pixel third component expression
16159 set pixel fourth component expression, corresponds to the alpha component
16162 set component to use as base for altering colors
16165 Each of them specifies the expression to use for computing the lookup table for
16166 the corresponding pixel component values.
16168 The expressions can contain the following constants and functions:
16173 The input width and height.
16176 The input value for the pixel component.
16178 @item ymin, umin, vmin, amin
16179 The minimum allowed component value.
16181 @item ymax, umax, vmax, amax
16182 The maximum allowed component value.
16185 All expressions default to "val".
16187 @subsection Examples
16191 Change too high luma values to gradient:
16193 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'"
16199 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
16200 Ratio) between two input videos.
16202 This filter takes in input two input videos, the first input is
16203 considered the "main" source and is passed unchanged to the
16204 output. The second input is used as a "reference" video for computing
16207 Both video inputs must have the same resolution and pixel format for
16208 this filter to work correctly. Also it assumes that both inputs
16209 have the same number of frames, which are compared one by one.
16211 The obtained average PSNR is printed through the logging system.
16213 The filter stores the accumulated MSE (mean squared error) of each
16214 frame, and at the end of the processing it is averaged across all frames
16215 equally, and the following formula is applied to obtain the PSNR:
16218 PSNR = 10*log10(MAX^2/MSE)
16221 Where MAX is the average of the maximum values of each component of the
16224 The description of the accepted parameters follows.
16227 @item stats_file, f
16228 If specified the filter will use the named file to save the PSNR of
16229 each individual frame. When filename equals "-" the data is sent to
16232 @item stats_version
16233 Specifies which version of the stats file format to use. Details of
16234 each format are written below.
16235 Default value is 1.
16237 @item stats_add_max
16238 Determines whether the max value is output to the stats log.
16239 Default value is 0.
16240 Requires stats_version >= 2. If this is set and stats_version < 2,
16241 the filter will return an error.
16244 This filter also supports the @ref{framesync} options.
16246 The file printed if @var{stats_file} is selected, contains a sequence of
16247 key/value pairs of the form @var{key}:@var{value} for each compared
16250 If a @var{stats_version} greater than 1 is specified, a header line precedes
16251 the list of per-frame-pair stats, with key value pairs following the frame
16252 format with the following parameters:
16255 @item psnr_log_version
16256 The version of the log file format. Will match @var{stats_version}.
16259 A comma separated list of the per-frame-pair parameters included in
16263 A description of each shown per-frame-pair parameter follows:
16267 sequential number of the input frame, starting from 1
16270 Mean Square Error pixel-by-pixel average difference of the compared
16271 frames, averaged over all the image components.
16273 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
16274 Mean Square Error pixel-by-pixel average difference of the compared
16275 frames for the component specified by the suffix.
16277 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
16278 Peak Signal to Noise ratio of the compared frames for the component
16279 specified by the suffix.
16281 @item max_avg, max_y, max_u, max_v
16282 Maximum allowed value for each channel, and average over all
16286 @subsection Examples
16291 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
16292 [main][ref] psnr="stats_file=stats.log" [out]
16295 On this example the input file being processed is compared with the
16296 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
16297 is stored in @file{stats.log}.
16300 Another example with different containers:
16302 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 -
16309 Pulldown reversal (inverse telecine) filter, capable of handling mixed
16310 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
16313 The pullup filter is designed to take advantage of future context in making
16314 its decisions. This filter is stateless in the sense that it does not lock
16315 onto a pattern to follow, but it instead looks forward to the following
16316 fields in order to identify matches and rebuild progressive frames.
16318 To produce content with an even framerate, insert the fps filter after
16319 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
16320 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
16322 The filter accepts the following options:
16329 These options set the amount of "junk" to ignore at the left, right, top, and
16330 bottom of the image, respectively. Left and right are in units of 8 pixels,
16331 while top and bottom are in units of 2 lines.
16332 The default is 8 pixels on each side.
16335 Set the strict breaks. Setting this option to 1 will reduce the chances of
16336 filter generating an occasional mismatched frame, but it may also cause an
16337 excessive number of frames to be dropped during high motion sequences.
16338 Conversely, setting it to -1 will make filter match fields more easily.
16339 This may help processing of video where there is slight blurring between
16340 the fields, but may also cause there to be interlaced frames in the output.
16341 Default value is @code{0}.
16344 Set the metric plane to use. It accepts the following values:
16350 Use chroma blue plane.
16353 Use chroma red plane.
16356 This option may be set to use chroma plane instead of the default luma plane
16357 for doing filter's computations. This may improve accuracy on very clean
16358 source material, but more likely will decrease accuracy, especially if there
16359 is chroma noise (rainbow effect) or any grayscale video.
16360 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
16361 load and make pullup usable in realtime on slow machines.
16364 For best results (without duplicated frames in the output file) it is
16365 necessary to change the output frame rate. For example, to inverse
16366 telecine NTSC input:
16368 ffmpeg -i input -vf pullup -r 24000/1001 ...
16373 Change video quantization parameters (QP).
16375 The filter accepts the following option:
16379 Set expression for quantization parameter.
16382 The expression is evaluated through the eval API and can contain, among others,
16383 the following constants:
16387 1 if index is not 129, 0 otherwise.
16390 Sequential index starting from -129 to 128.
16393 @subsection Examples
16397 Some equation like:
16405 Flush video frames from internal cache of frames into a random order.
16406 No frame is discarded.
16407 Inspired by @ref{frei0r} nervous filter.
16411 Set size in number of frames of internal cache, in range from @code{2} to
16412 @code{512}. Default is @code{30}.
16415 Set seed for random number generator, must be an integer included between
16416 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
16417 less than @code{0}, the filter will try to use a good random seed on a
16421 @section readeia608
16423 Read closed captioning (EIA-608) information from the top lines of a video frame.
16425 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
16426 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
16427 with EIA-608 data (starting from 0). A description of each metadata value follows:
16430 @item lavfi.readeia608.X.cc
16431 The two bytes stored as EIA-608 data (printed in hexadecimal).
16433 @item lavfi.readeia608.X.line
16434 The number of the line on which the EIA-608 data was identified and read.
16437 This filter accepts the following options:
16441 Set the line to start scanning for EIA-608 data. Default is @code{0}.
16444 Set the line to end scanning for EIA-608 data. Default is @code{29}.
16447 Set the ratio of width reserved for sync code detection.
16448 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
16451 Enable checking the parity bit. In the event of a parity error, the filter will output
16452 @code{0x00} for that character. Default is false.
16455 Lowpass lines prior to further processing. Default is enabled.
16458 @subsection Commands
16460 This filter supports the all above options as @ref{commands}.
16462 @subsection Examples
16466 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
16468 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
16474 Read vertical interval timecode (VITC) information from the top lines of a
16477 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
16478 timecode value, if a valid timecode has been detected. Further metadata key
16479 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
16480 timecode data has been found or not.
16482 This filter accepts the following options:
16486 Set the maximum number of lines to scan for VITC data. If the value is set to
16487 @code{-1} the full video frame is scanned. Default is @code{45}.
16490 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
16491 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
16494 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
16495 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
16498 @subsection Examples
16502 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
16503 draw @code{--:--:--:--} as a placeholder:
16505 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
16511 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
16513 Destination pixel at position (X, Y) will be picked from source (x, y) position
16514 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
16515 value for pixel will be used for destination pixel.
16517 Xmap and Ymap input video streams must be of same dimensions. Output video stream
16518 will have Xmap/Ymap video stream dimensions.
16519 Xmap and Ymap input video streams are 16bit depth, single channel.
16523 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
16524 Default is @code{color}.
16527 Specify the color of the unmapped pixels. For the syntax of this option,
16528 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
16529 manual,ffmpeg-utils}. Default color is @code{black}.
16532 @section removegrain
16534 The removegrain filter is a spatial denoiser for progressive video.
16538 Set mode for the first plane.
16541 Set mode for the second plane.
16544 Set mode for the third plane.
16547 Set mode for the fourth plane.
16550 Range of mode is from 0 to 24. Description of each mode follows:
16554 Leave input plane unchanged. Default.
16557 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
16560 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
16563 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
16566 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
16567 This is equivalent to a median filter.
16570 Line-sensitive clipping giving the minimal change.
16573 Line-sensitive clipping, intermediate.
16576 Line-sensitive clipping, intermediate.
16579 Line-sensitive clipping, intermediate.
16582 Line-sensitive clipping on a line where the neighbours pixels are the closest.
16585 Replaces the target pixel with the closest neighbour.
16588 [1 2 1] horizontal and vertical kernel blur.
16594 Bob mode, interpolates top field from the line where the neighbours
16595 pixels are the closest.
16598 Bob mode, interpolates bottom field from the line where the neighbours
16599 pixels are the closest.
16602 Bob mode, interpolates top field. Same as 13 but with a more complicated
16603 interpolation formula.
16606 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
16607 interpolation formula.
16610 Clips the pixel with the minimum and maximum of respectively the maximum and
16611 minimum of each pair of opposite neighbour pixels.
16614 Line-sensitive clipping using opposite neighbours whose greatest distance from
16615 the current pixel is minimal.
16618 Replaces the pixel with the average of its 8 neighbours.
16621 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
16624 Clips pixels using the averages of opposite neighbour.
16627 Same as mode 21 but simpler and faster.
16630 Small edge and halo removal, but reputed useless.
16636 @section removelogo
16638 Suppress a TV station logo, using an image file to determine which
16639 pixels comprise the logo. It works by filling in the pixels that
16640 comprise the logo with neighboring pixels.
16642 The filter accepts the following options:
16646 Set the filter bitmap file, which can be any image format supported by
16647 libavformat. The width and height of the image file must match those of the
16648 video stream being processed.
16651 Pixels in the provided bitmap image with a value of zero are not
16652 considered part of the logo, non-zero pixels are considered part of
16653 the logo. If you use white (255) for the logo and black (0) for the
16654 rest, you will be safe. For making the filter bitmap, it is
16655 recommended to take a screen capture of a black frame with the logo
16656 visible, and then using a threshold filter followed by the erode
16657 filter once or twice.
16659 If needed, little splotches can be fixed manually. Remember that if
16660 logo pixels are not covered, the filter quality will be much
16661 reduced. Marking too many pixels as part of the logo does not hurt as
16662 much, but it will increase the amount of blurring needed to cover over
16663 the image and will destroy more information than necessary, and extra
16664 pixels will slow things down on a large logo.
16666 @section repeatfields
16668 This filter uses the repeat_field flag from the Video ES headers and hard repeats
16669 fields based on its value.
16673 Reverse a video clip.
16675 Warning: This filter requires memory to buffer the entire clip, so trimming
16678 @subsection Examples
16682 Take the first 5 seconds of a clip, and reverse it.
16689 Shift R/G/B/A pixels horizontally and/or vertically.
16691 The filter accepts the following options:
16694 Set amount to shift red horizontally.
16696 Set amount to shift red vertically.
16698 Set amount to shift green horizontally.
16700 Set amount to shift green vertically.
16702 Set amount to shift blue horizontally.
16704 Set amount to shift blue vertically.
16706 Set amount to shift alpha horizontally.
16708 Set amount to shift alpha vertically.
16710 Set edge mode, can be @var{smear}, default, or @var{warp}.
16713 @subsection Commands
16715 This filter supports the all above options as @ref{commands}.
16718 Apply roberts cross operator to input video stream.
16720 The filter accepts the following option:
16724 Set which planes will be processed, unprocessed planes will be copied.
16725 By default value 0xf, all planes will be processed.
16728 Set value which will be multiplied with filtered result.
16731 Set value which will be added to filtered result.
16734 @subsection Commands
16736 This filter supports the all above options as @ref{commands}.
16740 Rotate video by an arbitrary angle expressed in radians.
16742 The filter accepts the following options:
16744 A description of the optional parameters follows.
16747 Set an expression for the angle by which to rotate the input video
16748 clockwise, expressed as a number of radians. A negative value will
16749 result in a counter-clockwise rotation. By default it is set to "0".
16751 This expression is evaluated for each frame.
16754 Set the output width expression, default value is "iw".
16755 This expression is evaluated just once during configuration.
16758 Set the output height expression, default value is "ih".
16759 This expression is evaluated just once during configuration.
16762 Enable bilinear interpolation if set to 1, a value of 0 disables
16763 it. Default value is 1.
16766 Set the color used to fill the output area not covered by the rotated
16767 image. For the general syntax of this option, check the
16768 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
16769 If the special value "none" is selected then no
16770 background is printed (useful for example if the background is never shown).
16772 Default value is "black".
16775 The expressions for the angle and the output size can contain the
16776 following constants and functions:
16780 sequential number of the input frame, starting from 0. It is always NAN
16781 before the first frame is filtered.
16784 time in seconds of the input frame, it is set to 0 when the filter is
16785 configured. It is always NAN before the first frame is filtered.
16789 horizontal and vertical chroma subsample values. For example for the
16790 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16794 the input video width and height
16798 the output width and height, that is the size of the padded area as
16799 specified by the @var{width} and @var{height} expressions
16803 the minimal width/height required for completely containing the input
16804 video rotated by @var{a} radians.
16806 These are only available when computing the @option{out_w} and
16807 @option{out_h} expressions.
16810 @subsection Examples
16814 Rotate the input by PI/6 radians clockwise:
16820 Rotate the input by PI/6 radians counter-clockwise:
16826 Rotate the input by 45 degrees clockwise:
16832 Apply a constant rotation with period T, starting from an angle of PI/3:
16834 rotate=PI/3+2*PI*t/T
16838 Make the input video rotation oscillating with a period of T
16839 seconds and an amplitude of A radians:
16841 rotate=A*sin(2*PI/T*t)
16845 Rotate the video, output size is chosen so that the whole rotating
16846 input video is always completely contained in the output:
16848 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
16852 Rotate the video, reduce the output size so that no background is ever
16855 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
16859 @subsection Commands
16861 The filter supports the following commands:
16865 Set the angle expression.
16866 The command accepts the same syntax of the corresponding option.
16868 If the specified expression is not valid, it is kept at its current
16874 Apply Shape Adaptive Blur.
16876 The filter accepts the following options:
16879 @item luma_radius, lr
16880 Set luma blur filter strength, must be a value in range 0.1-4.0, default
16881 value is 1.0. A greater value will result in a more blurred image, and
16882 in slower processing.
16884 @item luma_pre_filter_radius, lpfr
16885 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
16888 @item luma_strength, ls
16889 Set luma maximum difference between pixels to still be considered, must
16890 be a value in the 0.1-100.0 range, default value is 1.0.
16892 @item chroma_radius, cr
16893 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
16894 greater value will result in a more blurred image, and in slower
16897 @item chroma_pre_filter_radius, cpfr
16898 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
16900 @item chroma_strength, cs
16901 Set chroma maximum difference between pixels to still be considered,
16902 must be a value in the -0.9-100.0 range.
16905 Each chroma option value, if not explicitly specified, is set to the
16906 corresponding luma option value.
16911 Scale (resize) the input video, using the libswscale library.
16913 The scale filter forces the output display aspect ratio to be the same
16914 of the input, by changing the output sample aspect ratio.
16916 If the input image format is different from the format requested by
16917 the next filter, the scale filter will convert the input to the
16920 @subsection Options
16921 The filter accepts the following options, or any of the options
16922 supported by the libswscale scaler.
16924 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
16925 the complete list of scaler options.
16930 Set the output video dimension expression. Default value is the input
16933 If the @var{width} or @var{w} value is 0, the input width is used for
16934 the output. If the @var{height} or @var{h} value is 0, the input height
16935 is used for the output.
16937 If one and only one of the values is -n with n >= 1, the scale filter
16938 will use a value that maintains the aspect ratio of the input image,
16939 calculated from the other specified dimension. After that it will,
16940 however, make sure that the calculated dimension is divisible by n and
16941 adjust the value if necessary.
16943 If both values are -n with n >= 1, the behavior will be identical to
16944 both values being set to 0 as previously detailed.
16946 See below for the list of accepted constants for use in the dimension
16950 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
16954 Only evaluate expressions once during the filter initialization or when a command is processed.
16957 Evaluate expressions for each incoming frame.
16961 Default value is @samp{init}.
16965 Set the interlacing mode. It accepts the following values:
16969 Force interlaced aware scaling.
16972 Do not apply interlaced scaling.
16975 Select interlaced aware scaling depending on whether the source frames
16976 are flagged as interlaced or not.
16979 Default value is @samp{0}.
16982 Set libswscale scaling flags. See
16983 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16984 complete list of values. If not explicitly specified the filter applies
16988 @item param0, param1
16989 Set libswscale input parameters for scaling algorithms that need them. See
16990 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16991 complete documentation. If not explicitly specified the filter applies
16997 Set the video size. For the syntax of this option, check the
16998 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17000 @item in_color_matrix
17001 @item out_color_matrix
17002 Set in/output YCbCr color space type.
17004 This allows the autodetected value to be overridden as well as allows forcing
17005 a specific value used for the output and encoder.
17007 If not specified, the color space type depends on the pixel format.
17013 Choose automatically.
17016 Format conforming to International Telecommunication Union (ITU)
17017 Recommendation BT.709.
17020 Set color space conforming to the United States Federal Communications
17021 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
17026 Set color space conforming to:
17030 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
17033 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
17036 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
17041 Set color space conforming to SMPTE ST 240:1999.
17044 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
17049 Set in/output YCbCr sample range.
17051 This allows the autodetected value to be overridden as well as allows forcing
17052 a specific value used for the output and encoder. If not specified, the
17053 range depends on the pixel format. Possible values:
17057 Choose automatically.
17060 Set full range (0-255 in case of 8-bit luma).
17062 @item mpeg/limited/tv
17063 Set "MPEG" range (16-235 in case of 8-bit luma).
17066 @item force_original_aspect_ratio
17067 Enable decreasing or increasing output video width or height if necessary to
17068 keep the original aspect ratio. Possible values:
17072 Scale the video as specified and disable this feature.
17075 The output video dimensions will automatically be decreased if needed.
17078 The output video dimensions will automatically be increased if needed.
17082 One useful instance of this option is that when you know a specific device's
17083 maximum allowed resolution, you can use this to limit the output video to
17084 that, while retaining the aspect ratio. For example, device A allows
17085 1280x720 playback, and your video is 1920x800. Using this option (set it to
17086 decrease) and specifying 1280x720 to the command line makes the output
17089 Please note that this is a different thing than specifying -1 for @option{w}
17090 or @option{h}, you still need to specify the output resolution for this option
17093 @item force_divisible_by
17094 Ensures that both the output dimensions, width and height, are divisible by the
17095 given integer when used together with @option{force_original_aspect_ratio}. This
17096 works similar to using @code{-n} in the @option{w} and @option{h} options.
17098 This option respects the value set for @option{force_original_aspect_ratio},
17099 increasing or decreasing the resolution accordingly. The video's aspect ratio
17100 may be slightly modified.
17102 This option can be handy if you need to have a video fit within or exceed
17103 a defined resolution using @option{force_original_aspect_ratio} but also have
17104 encoder restrictions on width or height divisibility.
17108 The values of the @option{w} and @option{h} options are expressions
17109 containing the following constants:
17114 The input width and height
17118 These are the same as @var{in_w} and @var{in_h}.
17122 The output (scaled) width and height
17126 These are the same as @var{out_w} and @var{out_h}
17129 The same as @var{iw} / @var{ih}
17132 input sample aspect ratio
17135 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
17139 horizontal and vertical input chroma subsample values. For example for the
17140 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17144 horizontal and vertical output chroma subsample values. For example for the
17145 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17148 The (sequential) number of the input frame, starting from 0.
17149 Only available with @code{eval=frame}.
17152 The presentation timestamp of the input frame, expressed as a number of
17153 seconds. Only available with @code{eval=frame}.
17156 The position (byte offset) of the frame in the input stream, or NaN if
17157 this information is unavailable and/or meaningless (for example in case of synthetic video).
17158 Only available with @code{eval=frame}.
17161 @subsection Examples
17165 Scale the input video to a size of 200x100
17170 This is equivalent to:
17181 Specify a size abbreviation for the output size:
17186 which can also be written as:
17192 Scale the input to 2x:
17194 scale=w=2*iw:h=2*ih
17198 The above is the same as:
17200 scale=2*in_w:2*in_h
17204 Scale the input to 2x with forced interlaced scaling:
17206 scale=2*iw:2*ih:interl=1
17210 Scale the input to half size:
17212 scale=w=iw/2:h=ih/2
17216 Increase the width, and set the height to the same size:
17222 Seek Greek harmony:
17229 Increase the height, and set the width to 3/2 of the height:
17231 scale=w=3/2*oh:h=3/5*ih
17235 Increase the size, making the size a multiple of the chroma
17238 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
17242 Increase the width to a maximum of 500 pixels,
17243 keeping the same aspect ratio as the input:
17245 scale=w='min(500\, iw*3/2):h=-1'
17249 Make pixels square by combining scale and setsar:
17251 scale='trunc(ih*dar):ih',setsar=1/1
17255 Make pixels square by combining scale and setsar,
17256 making sure the resulting resolution is even (required by some codecs):
17258 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
17262 @subsection Commands
17264 This filter supports the following commands:
17268 Set the output video dimension expression.
17269 The command accepts the same syntax of the corresponding option.
17271 If the specified expression is not valid, it is kept at its current
17277 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
17278 format conversion on CUDA video frames. Setting the output width and height
17279 works in the same way as for the @var{scale} filter.
17281 The following additional options are accepted:
17284 The pixel format of the output CUDA frames. If set to the string "same" (the
17285 default), the input format will be kept. Note that automatic format negotiation
17286 and conversion is not yet supported for hardware frames
17289 The interpolation algorithm used for resizing. One of the following:
17296 @item cubic2p_bspline
17297 2-parameter cubic (B=1, C=0)
17299 @item cubic2p_catmullrom
17300 2-parameter cubic (B=0, C=1/2)
17302 @item cubic2p_b05c03
17303 2-parameter cubic (B=1/2, C=3/10)
17311 @item force_original_aspect_ratio
17312 Enable decreasing or increasing output video width or height if necessary to
17313 keep the original aspect ratio. Possible values:
17317 Scale the video as specified and disable this feature.
17320 The output video dimensions will automatically be decreased if needed.
17323 The output video dimensions will automatically be increased if needed.
17327 One useful instance of this option is that when you know a specific device's
17328 maximum allowed resolution, you can use this to limit the output video to
17329 that, while retaining the aspect ratio. For example, device A allows
17330 1280x720 playback, and your video is 1920x800. Using this option (set it to
17331 decrease) and specifying 1280x720 to the command line makes the output
17334 Please note that this is a different thing than specifying -1 for @option{w}
17335 or @option{h}, you still need to specify the output resolution for this option
17338 @item force_divisible_by
17339 Ensures that both the output dimensions, width and height, are divisible by the
17340 given integer when used together with @option{force_original_aspect_ratio}. This
17341 works similar to using @code{-n} in the @option{w} and @option{h} options.
17343 This option respects the value set for @option{force_original_aspect_ratio},
17344 increasing or decreasing the resolution accordingly. The video's aspect ratio
17345 may be slightly modified.
17347 This option can be handy if you need to have a video fit within or exceed
17348 a defined resolution using @option{force_original_aspect_ratio} but also have
17349 encoder restrictions on width or height divisibility.
17355 Scale (resize) the input video, based on a reference video.
17357 See the scale filter for available options, scale2ref supports the same but
17358 uses the reference video instead of the main input as basis. scale2ref also
17359 supports the following additional constants for the @option{w} and
17360 @option{h} options:
17365 The main input video's width and height
17368 The same as @var{main_w} / @var{main_h}
17371 The main input video's sample aspect ratio
17373 @item main_dar, mdar
17374 The main input video's display aspect ratio. Calculated from
17375 @code{(main_w / main_h) * main_sar}.
17379 The main input video's horizontal and vertical chroma subsample values.
17380 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
17384 The (sequential) number of the main input frame, starting from 0.
17385 Only available with @code{eval=frame}.
17388 The presentation timestamp of the main input frame, expressed as a number of
17389 seconds. Only available with @code{eval=frame}.
17392 The position (byte offset) of the frame in the main input stream, or NaN if
17393 this information is unavailable and/or meaningless (for example in case of synthetic video).
17394 Only available with @code{eval=frame}.
17397 @subsection Examples
17401 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
17403 'scale2ref[b][a];[a][b]overlay'
17407 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
17409 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
17413 @subsection Commands
17415 This filter supports the following commands:
17419 Set the output video dimension expression.
17420 The command accepts the same syntax of the corresponding option.
17422 If the specified expression is not valid, it is kept at its current
17427 Scroll input video horizontally and/or vertically by constant speed.
17429 The filter accepts the following options:
17431 @item horizontal, h
17432 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
17433 Negative values changes scrolling direction.
17436 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
17437 Negative values changes scrolling direction.
17440 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
17443 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
17446 @subsection Commands
17448 This filter supports the following @ref{commands}:
17450 @item horizontal, h
17451 Set the horizontal scrolling speed.
17453 Set the vertical scrolling speed.
17459 Detect video scene change.
17461 This filter sets frame metadata with mafd between frame, the scene score, and
17462 forward the frame to the next filter, so they can use these metadata to detect
17463 scene change or others.
17465 In addition, this filter logs a message and sets frame metadata when it detects
17466 a scene change by @option{threshold}.
17468 @code{lavfi.scd.mafd} metadata keys are set with mafd for every frame.
17470 @code{lavfi.scd.score} metadata keys are set with scene change score for every frame
17471 to detect scene change.
17473 @code{lavfi.scd.time} metadata keys are set with current filtered frame time which
17474 detect scene change with @option{threshold}.
17476 The filter accepts the following options:
17480 Set the scene change detection threshold as a percentage of maximum change. Good
17481 values are in the @code{[8.0, 14.0]} range. The range for @option{threshold} is
17484 Default value is @code{10.}.
17487 Set the flag to pass scene change frames to the next filter. Default value is @code{0}
17488 You can enable it if you want to get snapshot of scene change frames only.
17491 @anchor{selectivecolor}
17492 @section selectivecolor
17494 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
17495 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
17496 by the "purity" of the color (that is, how saturated it already is).
17498 This filter is similar to the Adobe Photoshop Selective Color tool.
17500 The filter accepts the following options:
17503 @item correction_method
17504 Select color correction method.
17506 Available values are:
17509 Specified adjustments are applied "as-is" (added/subtracted to original pixel
17512 Specified adjustments are relative to the original component value.
17514 Default is @code{absolute}.
17516 Adjustments for red pixels (pixels where the red component is the maximum)
17518 Adjustments for yellow pixels (pixels where the blue component is the minimum)
17520 Adjustments for green pixels (pixels where the green component is the maximum)
17522 Adjustments for cyan pixels (pixels where the red component is the minimum)
17524 Adjustments for blue pixels (pixels where the blue component is the maximum)
17526 Adjustments for magenta pixels (pixels where the green component is the minimum)
17528 Adjustments for white pixels (pixels where all components are greater than 128)
17530 Adjustments for all pixels except pure black and pure white
17532 Adjustments for black pixels (pixels where all components are lesser than 128)
17534 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
17537 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
17538 4 space separated floating point adjustment values in the [-1,1] range,
17539 respectively to adjust the amount of cyan, magenta, yellow and black for the
17540 pixels of its range.
17542 @subsection Examples
17546 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
17547 increase magenta by 27% in blue areas:
17549 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
17553 Use a Photoshop selective color preset:
17555 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
17559 @anchor{separatefields}
17560 @section separatefields
17562 The @code{separatefields} takes a frame-based video input and splits
17563 each frame into its components fields, producing a new half height clip
17564 with twice the frame rate and twice the frame count.
17566 This filter use field-dominance information in frame to decide which
17567 of each pair of fields to place first in the output.
17568 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
17570 @section setdar, setsar
17572 The @code{setdar} filter sets the Display Aspect Ratio for the filter
17575 This is done by changing the specified Sample (aka Pixel) Aspect
17576 Ratio, according to the following equation:
17578 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
17581 Keep in mind that the @code{setdar} filter does not modify the pixel
17582 dimensions of the video frame. Also, the display aspect ratio set by
17583 this filter may be changed by later filters in the filterchain,
17584 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
17587 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
17588 the filter output video.
17590 Note that as a consequence of the application of this filter, the
17591 output display aspect ratio will change according to the equation
17594 Keep in mind that the sample aspect ratio set by the @code{setsar}
17595 filter may be changed by later filters in the filterchain, e.g. if
17596 another "setsar" or a "setdar" filter is applied.
17598 It accepts the following parameters:
17601 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
17602 Set the aspect ratio used by the filter.
17604 The parameter can be a floating point number string, an expression, or
17605 a string of the form @var{num}:@var{den}, where @var{num} and
17606 @var{den} are the numerator and denominator of the aspect ratio. If
17607 the parameter is not specified, it is assumed the value "0".
17608 In case the form "@var{num}:@var{den}" is used, the @code{:} character
17612 Set the maximum integer value to use for expressing numerator and
17613 denominator when reducing the expressed aspect ratio to a rational.
17614 Default value is @code{100}.
17618 The parameter @var{sar} is an expression containing
17619 the following constants:
17623 These are approximated values for the mathematical constants e
17624 (Euler's number), pi (Greek pi), and phi (the golden ratio).
17627 The input width and height.
17630 These are the same as @var{w} / @var{h}.
17633 The input sample aspect ratio.
17636 The input display aspect ratio. It is the same as
17637 (@var{w} / @var{h}) * @var{sar}.
17640 Horizontal and vertical chroma subsample values. For example, for the
17641 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17644 @subsection Examples
17649 To change the display aspect ratio to 16:9, specify one of the following:
17656 To change the sample aspect ratio to 10:11, specify:
17662 To set a display aspect ratio of 16:9, and specify a maximum integer value of
17663 1000 in the aspect ratio reduction, use the command:
17665 setdar=ratio=16/9:max=1000
17673 Force field for the output video frame.
17675 The @code{setfield} filter marks the interlace type field for the
17676 output frames. It does not change the input frame, but only sets the
17677 corresponding property, which affects how the frame is treated by
17678 following filters (e.g. @code{fieldorder} or @code{yadif}).
17680 The filter accepts the following options:
17685 Available values are:
17689 Keep the same field property.
17692 Mark the frame as bottom-field-first.
17695 Mark the frame as top-field-first.
17698 Mark the frame as progressive.
17705 Force frame parameter for the output video frame.
17707 The @code{setparams} filter marks interlace and color range for the
17708 output frames. It does not change the input frame, but only sets the
17709 corresponding property, which affects how the frame is treated by
17714 Available values are:
17718 Keep the same field property (default).
17721 Mark the frame as bottom-field-first.
17724 Mark the frame as top-field-first.
17727 Mark the frame as progressive.
17731 Available values are:
17735 Keep the same color range property (default).
17737 @item unspecified, unknown
17738 Mark the frame as unspecified color range.
17740 @item limited, tv, mpeg
17741 Mark the frame as limited range.
17743 @item full, pc, jpeg
17744 Mark the frame as full range.
17747 @item color_primaries
17748 Set the color primaries.
17749 Available values are:
17753 Keep the same color primaries property (default).
17770 Set the color transfer.
17771 Available values are:
17775 Keep the same color trc property (default).
17797 Set the colorspace.
17798 Available values are:
17802 Keep the same colorspace property (default).
17815 @item chroma-derived-nc
17816 @item chroma-derived-c
17823 Show a line containing various information for each input video frame.
17824 The input video is not modified.
17826 This filter supports the following options:
17830 Calculate checksums of each plane. By default enabled.
17833 The shown line contains a sequence of key/value pairs of the form
17834 @var{key}:@var{value}.
17836 The following values are shown in the output:
17840 The (sequential) number of the input frame, starting from 0.
17843 The Presentation TimeStamp of the input frame, expressed as a number of
17844 time base units. The time base unit depends on the filter input pad.
17847 The Presentation TimeStamp of the input frame, expressed as a number of
17851 The position of the frame in the input stream, or -1 if this information is
17852 unavailable and/or meaningless (for example in case of synthetic video).
17855 The pixel format name.
17858 The sample aspect ratio of the input frame, expressed in the form
17859 @var{num}/@var{den}.
17862 The size of the input frame. For the syntax of this option, check the
17863 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17866 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
17867 for bottom field first).
17870 This is 1 if the frame is a key frame, 0 otherwise.
17873 The picture type of the input frame ("I" for an I-frame, "P" for a
17874 P-frame, "B" for a B-frame, or "?" for an unknown type).
17875 Also refer to the documentation of the @code{AVPictureType} enum and of
17876 the @code{av_get_picture_type_char} function defined in
17877 @file{libavutil/avutil.h}.
17880 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
17882 @item plane_checksum
17883 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
17884 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
17887 The mean value of pixels in each plane of the input frame, expressed in the form
17888 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
17891 The standard deviation of pixel values in each plane of the input frame, expressed
17892 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
17896 @section showpalette
17898 Displays the 256 colors palette of each frame. This filter is only relevant for
17899 @var{pal8} pixel format frames.
17901 It accepts the following option:
17905 Set the size of the box used to represent one palette color entry. Default is
17906 @code{30} (for a @code{30x30} pixel box).
17909 @section shuffleframes
17911 Reorder and/or duplicate and/or drop video frames.
17913 It accepts the following parameters:
17917 Set the destination indexes of input frames.
17918 This is space or '|' separated list of indexes that maps input frames to output
17919 frames. Number of indexes also sets maximal value that each index may have.
17920 '-1' index have special meaning and that is to drop frame.
17923 The first frame has the index 0. The default is to keep the input unchanged.
17925 @subsection Examples
17929 Swap second and third frame of every three frames of the input:
17931 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
17935 Swap 10th and 1st frame of every ten frames of the input:
17937 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
17941 @section shufflepixels
17943 Reorder pixels in video frames.
17945 This filter accepts the following options:
17949 Set shuffle direction. Can be forward or inverse direction.
17950 Default direction is forward.
17953 Set shuffle mode. Can be horizontal, vertical or block mode.
17957 Set shuffle block_size. In case of horizontal shuffle mode only width
17958 part of size is used, and in case of vertical shuffle mode only height
17959 part of size is used.
17962 Set random seed used with shuffling pixels. Mainly useful to set to be able
17963 to reverse filtering process to get original input.
17964 For example, to reverse forward shuffle you need to use same parameters
17965 and exact same seed and to set direction to inverse.
17968 @section shuffleplanes
17970 Reorder and/or duplicate video planes.
17972 It accepts the following parameters:
17977 The index of the input plane to be used as the first output plane.
17980 The index of the input plane to be used as the second output plane.
17983 The index of the input plane to be used as the third output plane.
17986 The index of the input plane to be used as the fourth output plane.
17990 The first plane has the index 0. The default is to keep the input unchanged.
17992 @subsection Examples
17996 Swap the second and third planes of the input:
17998 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
18002 @anchor{signalstats}
18003 @section signalstats
18004 Evaluate various visual metrics that assist in determining issues associated
18005 with the digitization of analog video media.
18007 By default the filter will log these metadata values:
18011 Display the minimal Y value contained within the input frame. Expressed in
18015 Display the Y value at the 10% percentile within the input frame. Expressed in
18019 Display the average Y value within the input frame. Expressed in range of
18023 Display the Y value at the 90% percentile within the input frame. Expressed in
18027 Display the maximum Y value contained within the input frame. Expressed in
18031 Display the minimal U value contained within the input frame. Expressed in
18035 Display the U value at the 10% percentile within the input frame. Expressed in
18039 Display the average U value within the input frame. Expressed in range of
18043 Display the U value at the 90% percentile within the input frame. Expressed in
18047 Display the maximum U value contained within the input frame. Expressed in
18051 Display the minimal V value contained within the input frame. Expressed in
18055 Display the V value at the 10% percentile within the input frame. Expressed in
18059 Display the average V value within the input frame. Expressed in range of
18063 Display the V value at the 90% percentile within the input frame. Expressed in
18067 Display the maximum V value contained within the input frame. Expressed in
18071 Display the minimal saturation value contained within the input frame.
18072 Expressed in range of [0-~181.02].
18075 Display the saturation value at the 10% percentile within the input frame.
18076 Expressed in range of [0-~181.02].
18079 Display the average saturation value within the input frame. Expressed in range
18083 Display the saturation value at the 90% percentile within the input frame.
18084 Expressed in range of [0-~181.02].
18087 Display the maximum saturation value contained within the input frame.
18088 Expressed in range of [0-~181.02].
18091 Display the median value for hue within the input frame. Expressed in range of
18095 Display the average value for hue within the input frame. Expressed in range of
18099 Display the average of sample value difference between all values of the Y
18100 plane in the current frame and corresponding values of the previous input frame.
18101 Expressed in range of [0-255].
18104 Display the average of sample value difference between all values of the U
18105 plane in the current frame and corresponding values of the previous input frame.
18106 Expressed in range of [0-255].
18109 Display the average of sample value difference between all values of the V
18110 plane in the current frame and corresponding values of the previous input frame.
18111 Expressed in range of [0-255].
18114 Display bit depth of Y plane in current frame.
18115 Expressed in range of [0-16].
18118 Display bit depth of U plane in current frame.
18119 Expressed in range of [0-16].
18122 Display bit depth of V plane in current frame.
18123 Expressed in range of [0-16].
18126 The filter accepts the following options:
18132 @option{stat} specify an additional form of image analysis.
18133 @option{out} output video with the specified type of pixel highlighted.
18135 Both options accept the following values:
18139 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
18140 unlike the neighboring pixels of the same field. Examples of temporal outliers
18141 include the results of video dropouts, head clogs, or tape tracking issues.
18144 Identify @var{vertical line repetition}. Vertical line repetition includes
18145 similar rows of pixels within a frame. In born-digital video vertical line
18146 repetition is common, but this pattern is uncommon in video digitized from an
18147 analog source. When it occurs in video that results from the digitization of an
18148 analog source it can indicate concealment from a dropout compensator.
18151 Identify pixels that fall outside of legal broadcast range.
18155 Set the highlight color for the @option{out} option. The default color is
18159 @subsection Examples
18163 Output data of various video metrics:
18165 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
18169 Output specific data about the minimum and maximum values of the Y plane per frame:
18171 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
18175 Playback video while highlighting pixels that are outside of broadcast range in red.
18177 ffplay example.mov -vf signalstats="out=brng:color=red"
18181 Playback video with signalstats metadata drawn over the frame.
18183 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
18186 The contents of signalstat_drawtext.txt used in the command are:
18189 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
18190 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
18191 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
18192 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
18200 Calculates the MPEG-7 Video Signature. The filter can handle more than one
18201 input. In this case the matching between the inputs can be calculated additionally.
18202 The filter always passes through the first input. The signature of each stream can
18203 be written into a file.
18205 It accepts the following options:
18209 Enable or disable the matching process.
18211 Available values are:
18215 Disable the calculation of a matching (default).
18217 Calculate the matching for the whole video and output whether the whole video
18218 matches or only parts.
18220 Calculate only until a matching is found or the video ends. Should be faster in
18225 Set the number of inputs. The option value must be a non negative integer.
18226 Default value is 1.
18229 Set the path to which the output is written. If there is more than one input,
18230 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
18231 integer), that will be replaced with the input number. If no filename is
18232 specified, no output will be written. This is the default.
18235 Choose the output format.
18237 Available values are:
18241 Use the specified binary representation (default).
18243 Use the specified xml representation.
18247 Set threshold to detect one word as similar. The option value must be an integer
18248 greater than zero. The default value is 9000.
18251 Set threshold to detect all words as similar. The option value must be an integer
18252 greater than zero. The default value is 60000.
18255 Set threshold to detect frames as similar. The option value must be an integer
18256 greater than zero. The default value is 116.
18259 Set the minimum length of a sequence in frames to recognize it as matching
18260 sequence. The option value must be a non negative integer value.
18261 The default value is 0.
18264 Set the minimum relation, that matching frames to all frames must have.
18265 The option value must be a double value between 0 and 1. The default value is 0.5.
18268 @subsection Examples
18272 To calculate the signature of an input video and store it in signature.bin:
18274 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
18278 To detect whether two videos match and store the signatures in XML format in
18279 signature0.xml and signature1.xml:
18281 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 -
18289 Blur the input video without impacting the outlines.
18291 It accepts the following options:
18294 @item luma_radius, lr
18295 Set the luma radius. The option value must be a float number in
18296 the range [0.1,5.0] that specifies the variance of the gaussian filter
18297 used to blur the image (slower if larger). Default value is 1.0.
18299 @item luma_strength, ls
18300 Set the luma strength. The option value must be a float number
18301 in the range [-1.0,1.0] that configures the blurring. A value included
18302 in [0.0,1.0] will blur the image whereas a value included in
18303 [-1.0,0.0] will sharpen the image. Default value is 1.0.
18305 @item luma_threshold, lt
18306 Set the luma threshold used as a coefficient to determine
18307 whether a pixel should be blurred or not. The option value must be an
18308 integer in the range [-30,30]. A value of 0 will filter all the image,
18309 a value included in [0,30] will filter flat areas and a value included
18310 in [-30,0] will filter edges. Default value is 0.
18312 @item chroma_radius, cr
18313 Set the chroma radius. The option value must be a float number in
18314 the range [0.1,5.0] that specifies the variance of the gaussian filter
18315 used to blur the image (slower if larger). Default value is @option{luma_radius}.
18317 @item chroma_strength, cs
18318 Set the chroma strength. The option value must be a float number
18319 in the range [-1.0,1.0] that configures the blurring. A value included
18320 in [0.0,1.0] will blur the image whereas a value included in
18321 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
18323 @item chroma_threshold, ct
18324 Set the chroma threshold used as a coefficient to determine
18325 whether a pixel should be blurred or not. The option value must be an
18326 integer in the range [-30,30]. A value of 0 will filter all the image,
18327 a value included in [0,30] will filter flat areas and a value included
18328 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
18331 If a chroma option is not explicitly set, the corresponding luma value
18335 Apply sobel operator to input video stream.
18337 The filter accepts the following option:
18341 Set which planes will be processed, unprocessed planes will be copied.
18342 By default value 0xf, all planes will be processed.
18345 Set value which will be multiplied with filtered result.
18348 Set value which will be added to filtered result.
18351 @subsection Commands
18353 This filter supports the all above options as @ref{commands}.
18358 Apply a simple postprocessing filter that compresses and decompresses the image
18359 at several (or - in the case of @option{quality} level @code{6} - all) shifts
18360 and average the results.
18362 The filter accepts the following options:
18366 Set quality. This option defines the number of levels for averaging. It accepts
18367 an integer in the range 0-6. If set to @code{0}, the filter will have no
18368 effect. A value of @code{6} means the higher quality. For each increment of
18369 that value the speed drops by a factor of approximately 2. Default value is
18373 Force a constant quantization parameter. If not set, the filter will use the QP
18374 from the video stream (if available).
18377 Set thresholding mode. Available modes are:
18381 Set hard thresholding (default).
18383 Set soft thresholding (better de-ringing effect, but likely blurrier).
18386 @item use_bframe_qp
18387 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
18388 option may cause flicker since the B-Frames have often larger QP. Default is
18389 @code{0} (not enabled).
18392 @subsection Commands
18394 This filter supports the following commands:
18396 @item quality, level
18397 Set quality level. The value @code{max} can be used to set the maximum level,
18398 currently @code{6}.
18404 Scale the input by applying one of the super-resolution methods based on
18405 convolutional neural networks. Supported models:
18409 Super-Resolution Convolutional Neural Network model (SRCNN).
18410 See @url{https://arxiv.org/abs/1501.00092}.
18413 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
18414 See @url{https://arxiv.org/abs/1609.05158}.
18417 Training scripts as well as scripts for model file (.pb) saving can be found at
18418 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
18419 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
18421 Native model files (.model) can be generated from TensorFlow model
18422 files (.pb) by using tools/python/convert.py
18424 The filter accepts the following options:
18428 Specify which DNN backend to use for model loading and execution. This option accepts
18429 the following values:
18433 Native implementation of DNN loading and execution.
18436 TensorFlow backend. To enable this backend you
18437 need to install the TensorFlow for C library (see
18438 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
18439 @code{--enable-libtensorflow}
18442 Default value is @samp{native}.
18445 Set path to model file specifying network architecture and its parameters.
18446 Note that different backends use different file formats. TensorFlow backend
18447 can load files for both formats, while native backend can load files for only
18451 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
18452 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
18453 input upscaled using bicubic upscaling with proper scale factor.
18456 This feature can also be finished with @ref{dnn_processing} filter.
18460 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
18462 This filter takes in input two input videos, the first input is
18463 considered the "main" source and is passed unchanged to the
18464 output. The second input is used as a "reference" video for computing
18467 Both video inputs must have the same resolution and pixel format for
18468 this filter to work correctly. Also it assumes that both inputs
18469 have the same number of frames, which are compared one by one.
18471 The filter stores the calculated SSIM of each frame.
18473 The description of the accepted parameters follows.
18476 @item stats_file, f
18477 If specified the filter will use the named file to save the SSIM of
18478 each individual frame. When filename equals "-" the data is sent to
18482 The file printed if @var{stats_file} is selected, contains a sequence of
18483 key/value pairs of the form @var{key}:@var{value} for each compared
18486 A description of each shown parameter follows:
18490 sequential number of the input frame, starting from 1
18492 @item Y, U, V, R, G, B
18493 SSIM of the compared frames for the component specified by the suffix.
18496 SSIM of the compared frames for the whole frame.
18499 Same as above but in dB representation.
18502 This filter also supports the @ref{framesync} options.
18504 @subsection Examples
18509 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
18510 [main][ref] ssim="stats_file=stats.log" [out]
18513 On this example the input file being processed is compared with the
18514 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
18515 is stored in @file{stats.log}.
18518 Another example with both psnr and ssim at same time:
18520 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
18524 Another example with different containers:
18526 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 -
18532 Convert between different stereoscopic image formats.
18534 The filters accept the following options:
18538 Set stereoscopic image format of input.
18540 Available values for input image formats are:
18543 side by side parallel (left eye left, right eye right)
18546 side by side crosseye (right eye left, left eye right)
18549 side by side parallel with half width resolution
18550 (left eye left, right eye right)
18553 side by side crosseye with half width resolution
18554 (right eye left, left eye right)
18558 above-below (left eye above, right eye below)
18562 above-below (right eye above, left eye below)
18566 above-below with half height resolution
18567 (left eye above, right eye below)
18571 above-below with half height resolution
18572 (right eye above, left eye below)
18575 alternating frames (left eye first, right eye second)
18578 alternating frames (right eye first, left eye second)
18581 interleaved rows (left eye has top row, right eye starts on next row)
18584 interleaved rows (right eye has top row, left eye starts on next row)
18587 interleaved columns, left eye first
18590 interleaved columns, right eye first
18592 Default value is @samp{sbsl}.
18596 Set stereoscopic image format of output.
18600 side by side parallel (left eye left, right eye right)
18603 side by side crosseye (right eye left, left eye right)
18606 side by side parallel with half width resolution
18607 (left eye left, right eye right)
18610 side by side crosseye with half width resolution
18611 (right eye left, left eye right)
18615 above-below (left eye above, right eye below)
18619 above-below (right eye above, left eye below)
18623 above-below with half height resolution
18624 (left eye above, right eye below)
18628 above-below with half height resolution
18629 (right eye above, left eye below)
18632 alternating frames (left eye first, right eye second)
18635 alternating frames (right eye first, left eye second)
18638 interleaved rows (left eye has top row, right eye starts on next row)
18641 interleaved rows (right eye has top row, left eye starts on next row)
18644 anaglyph red/blue gray
18645 (red filter on left eye, blue filter on right eye)
18648 anaglyph red/green gray
18649 (red filter on left eye, green filter on right eye)
18652 anaglyph red/cyan gray
18653 (red filter on left eye, cyan filter on right eye)
18656 anaglyph red/cyan half colored
18657 (red filter on left eye, cyan filter on right eye)
18660 anaglyph red/cyan color
18661 (red filter on left eye, cyan filter on right eye)
18664 anaglyph red/cyan color optimized with the least squares projection of dubois
18665 (red filter on left eye, cyan filter on right eye)
18668 anaglyph green/magenta gray
18669 (green filter on left eye, magenta filter on right eye)
18672 anaglyph green/magenta half colored
18673 (green filter on left eye, magenta filter on right eye)
18676 anaglyph green/magenta colored
18677 (green filter on left eye, magenta filter on right eye)
18680 anaglyph green/magenta color optimized with the least squares projection of dubois
18681 (green filter on left eye, magenta filter on right eye)
18684 anaglyph yellow/blue gray
18685 (yellow filter on left eye, blue filter on right eye)
18688 anaglyph yellow/blue half colored
18689 (yellow filter on left eye, blue filter on right eye)
18692 anaglyph yellow/blue colored
18693 (yellow filter on left eye, blue filter on right eye)
18696 anaglyph yellow/blue color optimized with the least squares projection of dubois
18697 (yellow filter on left eye, blue filter on right eye)
18700 mono output (left eye only)
18703 mono output (right eye only)
18706 checkerboard, left eye first
18709 checkerboard, right eye first
18712 interleaved columns, left eye first
18715 interleaved columns, right eye first
18721 Default value is @samp{arcd}.
18724 @subsection Examples
18728 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
18734 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
18740 @section streamselect, astreamselect
18741 Select video or audio streams.
18743 The filter accepts the following options:
18747 Set number of inputs. Default is 2.
18750 Set input indexes to remap to outputs.
18753 @subsection Commands
18755 The @code{streamselect} and @code{astreamselect} filter supports the following
18760 Set input indexes to remap to outputs.
18763 @subsection Examples
18767 Select first 5 seconds 1st stream and rest of time 2nd stream:
18769 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
18773 Same as above, but for audio:
18775 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
18782 Draw subtitles on top of input video using the libass library.
18784 To enable compilation of this filter you need to configure FFmpeg with
18785 @code{--enable-libass}. This filter also requires a build with libavcodec and
18786 libavformat to convert the passed subtitles file to ASS (Advanced Substation
18787 Alpha) subtitles format.
18789 The filter accepts the following options:
18793 Set the filename of the subtitle file to read. It must be specified.
18795 @item original_size
18796 Specify the size of the original video, the video for which the ASS file
18797 was composed. For the syntax of this option, check the
18798 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18799 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
18800 correctly scale the fonts if the aspect ratio has been changed.
18803 Set a directory path containing fonts that can be used by the filter.
18804 These fonts will be used in addition to whatever the font provider uses.
18807 Process alpha channel, by default alpha channel is untouched.
18810 Set subtitles input character encoding. @code{subtitles} filter only. Only
18811 useful if not UTF-8.
18813 @item stream_index, si
18814 Set subtitles stream index. @code{subtitles} filter only.
18817 Override default style or script info parameters of the subtitles. It accepts a
18818 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
18821 If the first key is not specified, it is assumed that the first value
18822 specifies the @option{filename}.
18824 For example, to render the file @file{sub.srt} on top of the input
18825 video, use the command:
18830 which is equivalent to:
18832 subtitles=filename=sub.srt
18835 To render the default subtitles stream from file @file{video.mkv}, use:
18837 subtitles=video.mkv
18840 To render the second subtitles stream from that file, use:
18842 subtitles=video.mkv:si=1
18845 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
18846 @code{DejaVu Serif}, use:
18848 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
18851 @section super2xsai
18853 Scale the input by 2x and smooth using the Super2xSaI (Scale and
18854 Interpolate) pixel art scaling algorithm.
18856 Useful for enlarging pixel art images without reducing sharpness.
18860 Swap two rectangular objects in video.
18862 This filter accepts the following options:
18872 Set 1st rect x coordinate.
18875 Set 1st rect y coordinate.
18878 Set 2nd rect x coordinate.
18881 Set 2nd rect y coordinate.
18883 All expressions are evaluated once for each frame.
18886 The all options are expressions containing the following constants:
18891 The input width and height.
18894 same as @var{w} / @var{h}
18897 input sample aspect ratio
18900 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
18903 The number of the input frame, starting from 0.
18906 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
18909 the position in the file of the input frame, NAN if unknown
18916 Blend successive video frames.
18922 Apply telecine process to the video.
18924 This filter accepts the following options:
18933 The default value is @code{top}.
18937 A string of numbers representing the pulldown pattern you wish to apply.
18938 The default value is @code{23}.
18942 Some typical patterns:
18947 24p: 2332 (preferred)
18954 24p: 222222222223 ("Euro pulldown")
18959 @section thistogram
18961 Compute and draw a color distribution histogram for the input video across time.
18963 Unlike @ref{histogram} video filter which only shows histogram of single input frame
18964 at certain time, this filter shows also past histograms of number of frames defined
18965 by @code{width} option.
18967 The computed histogram is a representation of the color component
18968 distribution in an image.
18970 The filter accepts the following options:
18974 Set width of single color component output. Default value is @code{0}.
18975 Value of @code{0} means width will be picked from input video.
18976 This also set number of passed histograms to keep.
18977 Allowed range is [0, 8192].
18979 @item display_mode, d
18981 It accepts the following values:
18984 Per color component graphs are placed below each other.
18987 Per color component graphs are placed side by side.
18990 Presents information identical to that in the @code{parade}, except
18991 that the graphs representing color components are superimposed directly
18994 Default is @code{stack}.
18996 @item levels_mode, m
18997 Set mode. Can be either @code{linear}, or @code{logarithmic}.
18998 Default is @code{linear}.
19000 @item components, c
19001 Set what color components to display.
19002 Default is @code{7}.
19005 Set background opacity. Default is @code{0.9}.
19008 Show envelope. Default is disabled.
19011 Set envelope color. Default is @code{gold}.
19016 Available values for slide is:
19019 Draw new frame when right border is reached.
19022 Replace old columns with new ones.
19025 Scroll from right to left.
19028 Scroll from left to right.
19031 Draw single picture.
19034 Default is @code{replace}.
19039 Apply threshold effect to video stream.
19041 This filter needs four video streams to perform thresholding.
19042 First stream is stream we are filtering.
19043 Second stream is holding threshold values, third stream is holding min values,
19044 and last, fourth stream is holding max values.
19046 The filter accepts the following option:
19050 Set which planes will be processed, unprocessed planes will be copied.
19051 By default value 0xf, all planes will be processed.
19054 For example if first stream pixel's component value is less then threshold value
19055 of pixel component from 2nd threshold stream, third stream value will picked,
19056 otherwise fourth stream pixel component value will be picked.
19058 Using color source filter one can perform various types of thresholding:
19060 @subsection Examples
19064 Binary threshold, using gray color as threshold:
19066 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
19070 Inverted binary threshold, using gray color as threshold:
19072 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
19076 Truncate binary threshold, using gray color as threshold:
19078 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
19082 Threshold to zero, using gray color as threshold:
19084 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
19088 Inverted threshold to zero, using gray color as threshold:
19090 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
19095 Select the most representative frame in a given sequence of consecutive frames.
19097 The filter accepts the following options:
19101 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
19102 will pick one of them, and then handle the next batch of @var{n} frames until
19103 the end. Default is @code{100}.
19106 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
19107 value will result in a higher memory usage, so a high value is not recommended.
19109 @subsection Examples
19113 Extract one picture each 50 frames:
19119 Complete example of a thumbnail creation with @command{ffmpeg}:
19121 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
19128 Tile several successive frames together.
19130 The @ref{untile} filter can do the reverse.
19132 The filter accepts the following options:
19137 Set the grid size (i.e. the number of lines and columns). For the syntax of
19138 this option, check the
19139 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19142 Set the maximum number of frames to render in the given area. It must be less
19143 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
19144 the area will be used.
19147 Set the outer border margin in pixels.
19150 Set the inner border thickness (i.e. the number of pixels between frames). For
19151 more advanced padding options (such as having different values for the edges),
19152 refer to the pad video filter.
19155 Specify the color of the unused area. For the syntax of this option, check the
19156 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
19157 The default value of @var{color} is "black".
19160 Set the number of frames to overlap when tiling several successive frames together.
19161 The value must be between @code{0} and @var{nb_frames - 1}.
19164 Set the number of frames to initially be empty before displaying first output frame.
19165 This controls how soon will one get first output frame.
19166 The value must be between @code{0} and @var{nb_frames - 1}.
19169 @subsection Examples
19173 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
19175 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
19177 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
19178 duplicating each output frame to accommodate the originally detected frame
19182 Display @code{5} pictures in an area of @code{3x2} frames,
19183 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
19184 mixed flat and named options:
19186 tile=3x2:nb_frames=5:padding=7:margin=2
19190 @section tinterlace
19192 Perform various types of temporal field interlacing.
19194 Frames are counted starting from 1, so the first input frame is
19197 The filter accepts the following options:
19202 Specify the mode of the interlacing. This option can also be specified
19203 as a value alone. See below for a list of values for this option.
19205 Available values are:
19209 Move odd frames into the upper field, even into the lower field,
19210 generating a double height frame at half frame rate.
19214 Frame 1 Frame 2 Frame 3 Frame 4
19216 11111 22222 33333 44444
19217 11111 22222 33333 44444
19218 11111 22222 33333 44444
19219 11111 22222 33333 44444
19233 Only output odd frames, even frames are dropped, generating a frame with
19234 unchanged height at half frame rate.
19239 Frame 1 Frame 2 Frame 3 Frame 4
19241 11111 22222 33333 44444
19242 11111 22222 33333 44444
19243 11111 22222 33333 44444
19244 11111 22222 33333 44444
19254 Only output even frames, odd frames are dropped, generating a frame with
19255 unchanged height at half frame rate.
19260 Frame 1 Frame 2 Frame 3 Frame 4
19262 11111 22222 33333 44444
19263 11111 22222 33333 44444
19264 11111 22222 33333 44444
19265 11111 22222 33333 44444
19275 Expand each frame to full height, but pad alternate lines with black,
19276 generating a frame with double height at the same input frame rate.
19281 Frame 1 Frame 2 Frame 3 Frame 4
19283 11111 22222 33333 44444
19284 11111 22222 33333 44444
19285 11111 22222 33333 44444
19286 11111 22222 33333 44444
19289 11111 ..... 33333 .....
19290 ..... 22222 ..... 44444
19291 11111 ..... 33333 .....
19292 ..... 22222 ..... 44444
19293 11111 ..... 33333 .....
19294 ..... 22222 ..... 44444
19295 11111 ..... 33333 .....
19296 ..... 22222 ..... 44444
19300 @item interleave_top, 4
19301 Interleave the upper field from odd frames with the lower field from
19302 even frames, generating a frame with unchanged height at half frame rate.
19307 Frame 1 Frame 2 Frame 3 Frame 4
19309 11111<- 22222 33333<- 44444
19310 11111 22222<- 33333 44444<-
19311 11111<- 22222 33333<- 44444
19312 11111 22222<- 33333 44444<-
19322 @item interleave_bottom, 5
19323 Interleave the lower field from odd frames with the upper field from
19324 even frames, generating a frame with unchanged height at half frame rate.
19329 Frame 1 Frame 2 Frame 3 Frame 4
19331 11111 22222<- 33333 44444<-
19332 11111<- 22222 33333<- 44444
19333 11111 22222<- 33333 44444<-
19334 11111<- 22222 33333<- 44444
19344 @item interlacex2, 6
19345 Double frame rate with unchanged height. Frames are inserted each
19346 containing the second temporal field from the previous input frame and
19347 the first temporal field from the next input frame. This mode relies on
19348 the top_field_first flag. Useful for interlaced video displays with no
19349 field synchronisation.
19354 Frame 1 Frame 2 Frame 3 Frame 4
19356 11111 22222 33333 44444
19357 11111 22222 33333 44444
19358 11111 22222 33333 44444
19359 11111 22222 33333 44444
19362 11111 22222 22222 33333 33333 44444 44444
19363 11111 11111 22222 22222 33333 33333 44444
19364 11111 22222 22222 33333 33333 44444 44444
19365 11111 11111 22222 22222 33333 33333 44444
19370 Move odd frames into the upper field, even into the lower field,
19371 generating a double height frame at same frame rate.
19376 Frame 1 Frame 2 Frame 3 Frame 4
19378 11111 22222 33333 44444
19379 11111 22222 33333 44444
19380 11111 22222 33333 44444
19381 11111 22222 33333 44444
19384 11111 33333 33333 55555
19385 22222 22222 44444 44444
19386 11111 33333 33333 55555
19387 22222 22222 44444 44444
19388 11111 33333 33333 55555
19389 22222 22222 44444 44444
19390 11111 33333 33333 55555
19391 22222 22222 44444 44444
19396 Numeric values are deprecated but are accepted for backward
19397 compatibility reasons.
19399 Default mode is @code{merge}.
19402 Specify flags influencing the filter process.
19404 Available value for @var{flags} is:
19407 @item low_pass_filter, vlpf
19408 Enable linear vertical low-pass filtering in the filter.
19409 Vertical low-pass filtering is required when creating an interlaced
19410 destination from a progressive source which contains high-frequency
19411 vertical detail. Filtering will reduce interlace 'twitter' and Moire
19414 @item complex_filter, cvlpf
19415 Enable complex vertical low-pass filtering.
19416 This will slightly less reduce interlace 'twitter' and Moire
19417 patterning but better retain detail and subjective sharpness impression.
19420 Bypass already interlaced frames, only adjust the frame rate.
19423 Vertical low-pass filtering and bypassing already interlaced frames can only be
19424 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
19429 Pick median pixels from several successive input video frames.
19431 The filter accepts the following options:
19435 Set radius of median filter.
19436 Default is 1. Allowed range is from 1 to 127.
19439 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
19442 Set median percentile. Default value is @code{0.5}.
19443 Default value of @code{0.5} will pick always median values, while @code{0} will pick
19444 minimum values, and @code{1} maximum values.
19447 @subsection Commands
19449 This filter supports all above options as @ref{commands}, excluding option @code{radius}.
19451 @section tmidequalizer
19453 Apply Temporal Midway Video Equalization effect.
19455 Midway Video Equalization adjusts a sequence of video frames to have the same
19456 histograms, while maintaining their dynamics as much as possible. It's
19457 useful for e.g. matching exposures from a video frames sequence.
19459 This filter accepts the following option:
19463 Set filtering radius. Default is @code{5}. Allowed range is from 1 to 127.
19466 Set filtering sigma. Default is @code{0.5}. This controls strength of filtering.
19467 Setting this option to 0 effectively does nothing.
19470 Set which planes to process. Default is @code{15}, which is all available planes.
19475 Mix successive video frames.
19477 A description of the accepted options follows.
19481 The number of successive frames to mix. If unspecified, it defaults to 3.
19484 Specify weight of each input video frame.
19485 Each weight is separated by space. If number of weights is smaller than
19486 number of @var{frames} last specified weight will be used for all remaining
19490 Specify scale, if it is set it will be multiplied with sum
19491 of each weight multiplied with pixel values to give final destination
19492 pixel value. By default @var{scale} is auto scaled to sum of weights.
19495 @subsection Examples
19499 Average 7 successive frames:
19501 tmix=frames=7:weights="1 1 1 1 1 1 1"
19505 Apply simple temporal convolution:
19507 tmix=frames=3:weights="-1 3 -1"
19511 Similar as above but only showing temporal differences:
19513 tmix=frames=3:weights="-1 2 -1":scale=1
19519 Tone map colors from different dynamic ranges.
19521 This filter expects data in single precision floating point, as it needs to
19522 operate on (and can output) out-of-range values. Another filter, such as
19523 @ref{zscale}, is needed to convert the resulting frame to a usable format.
19525 The tonemapping algorithms implemented only work on linear light, so input
19526 data should be linearized beforehand (and possibly correctly tagged).
19529 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
19532 @subsection Options
19533 The filter accepts the following options.
19537 Set the tone map algorithm to use.
19539 Possible values are:
19542 Do not apply any tone map, only desaturate overbright pixels.
19545 Hard-clip any out-of-range values. Use it for perfect color accuracy for
19546 in-range values, while distorting out-of-range values.
19549 Stretch the entire reference gamut to a linear multiple of the display.
19552 Fit a logarithmic transfer between the tone curves.
19555 Preserve overall image brightness with a simple curve, using nonlinear
19556 contrast, which results in flattening details and degrading color accuracy.
19559 Preserve both dark and bright details better than @var{reinhard}, at the cost
19560 of slightly darkening everything. Use it when detail preservation is more
19561 important than color and brightness accuracy.
19564 Smoothly map out-of-range values, while retaining contrast and colors for
19565 in-range material as much as possible. Use it when color accuracy is more
19566 important than detail preservation.
19572 Tune the tone mapping algorithm.
19574 This affects the following algorithms:
19580 Specifies the scale factor to use while stretching.
19584 Specifies the exponent of the function.
19588 Specify an extra linear coefficient to multiply into the signal before clipping.
19592 Specify the local contrast coefficient at the display peak.
19593 Default to 0.5, which means that in-gamut values will be about half as bright
19600 Specify the transition point from linear to mobius transform. Every value
19601 below this point is guaranteed to be mapped 1:1. The higher the value, the
19602 more accurate the result will be, at the cost of losing bright details.
19603 Default to 0.3, which due to the steep initial slope still preserves in-range
19604 colors fairly accurately.
19608 Apply desaturation for highlights that exceed this level of brightness. The
19609 higher the parameter, the more color information will be preserved. This
19610 setting helps prevent unnaturally blown-out colors for super-highlights, by
19611 (smoothly) turning into white instead. This makes images feel more natural,
19612 at the cost of reducing information about out-of-range colors.
19614 The default of 2.0 is somewhat conservative and will mostly just apply to
19615 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
19617 This option works only if the input frame has a supported color tag.
19620 Override signal/nominal/reference peak with this value. Useful when the
19621 embedded peak information in display metadata is not reliable or when tone
19622 mapping from a lower range to a higher range.
19627 Temporarily pad video frames.
19629 The filter accepts the following options:
19633 Specify number of delay frames before input video stream. Default is 0.
19636 Specify number of padding frames after input video stream.
19637 Set to -1 to pad indefinitely. Default is 0.
19640 Set kind of frames added to beginning of stream.
19641 Can be either @var{add} or @var{clone}.
19642 With @var{add} frames of solid-color are added.
19643 With @var{clone} frames are clones of first frame.
19644 Default is @var{add}.
19647 Set kind of frames added to end of stream.
19648 Can be either @var{add} or @var{clone}.
19649 With @var{add} frames of solid-color are added.
19650 With @var{clone} frames are clones of last frame.
19651 Default is @var{add}.
19653 @item start_duration, stop_duration
19654 Specify the duration of the start/stop delay. See
19655 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19656 for the accepted syntax.
19657 These options override @var{start} and @var{stop}. Default is 0.
19660 Specify the color of the padded area. For the syntax of this option,
19661 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
19662 manual,ffmpeg-utils}.
19664 The default value of @var{color} is "black".
19670 Transpose rows with columns in the input video and optionally flip it.
19672 It accepts the following parameters:
19677 Specify the transposition direction.
19679 Can assume the following values:
19681 @item 0, 4, cclock_flip
19682 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
19690 Rotate by 90 degrees clockwise, that is:
19698 Rotate by 90 degrees counterclockwise, that is:
19705 @item 3, 7, clock_flip
19706 Rotate by 90 degrees clockwise and vertically flip, that is:
19714 For values between 4-7, the transposition is only done if the input
19715 video geometry is portrait and not landscape. These values are
19716 deprecated, the @code{passthrough} option should be used instead.
19718 Numerical values are deprecated, and should be dropped in favor of
19719 symbolic constants.
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.
19728 Preserve portrait geometry (when @var{height} >= @var{width}).
19730 Preserve landscape geometry (when @var{width} >= @var{height}).
19733 Default value is @code{none}.
19736 For example to rotate by 90 degrees clockwise and preserve portrait
19739 transpose=dir=1:passthrough=portrait
19742 The command above can also be specified as:
19744 transpose=1:portrait
19747 @section transpose_npp
19749 Transpose rows with columns in the input video and optionally flip it.
19750 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
19752 It accepts the following parameters:
19757 Specify the transposition direction.
19759 Can assume the following values:
19762 Rotate by 90 degrees counterclockwise and vertically flip. (default)
19765 Rotate by 90 degrees clockwise.
19768 Rotate by 90 degrees counterclockwise.
19771 Rotate by 90 degrees clockwise and vertically flip.
19775 Do not apply the transposition if the input geometry matches the one
19776 specified by the specified value. It accepts the following values:
19779 Always apply transposition. (default)
19781 Preserve portrait geometry (when @var{height} >= @var{width}).
19783 Preserve landscape geometry (when @var{width} >= @var{height}).
19789 Trim the input so that the output contains one continuous subpart of the input.
19791 It accepts the following parameters:
19794 Specify the time of the start of the kept section, i.e. the frame with the
19795 timestamp @var{start} will be the first frame in the output.
19798 Specify the time of the first frame that will be dropped, i.e. the frame
19799 immediately preceding the one with the timestamp @var{end} will be the last
19800 frame in the output.
19803 This is the same as @var{start}, except this option sets the start timestamp
19804 in timebase units instead of seconds.
19807 This is the same as @var{end}, except this option sets the end timestamp
19808 in timebase units instead of seconds.
19811 The maximum duration of the output in seconds.
19814 The number of the first frame that should be passed to the output.
19817 The number of the first frame that should be dropped.
19820 @option{start}, @option{end}, and @option{duration} are expressed as time
19821 duration specifications; see
19822 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19823 for the accepted syntax.
19825 Note that the first two sets of the start/end options and the @option{duration}
19826 option look at the frame timestamp, while the _frame variants simply count the
19827 frames that pass through the filter. Also note that this filter does not modify
19828 the timestamps. If you wish for the output timestamps to start at zero, insert a
19829 setpts filter after the trim filter.
19831 If multiple start or end options are set, this filter tries to be greedy and
19832 keep all the frames that match at least one of the specified constraints. To keep
19833 only the part that matches all the constraints at once, chain multiple trim
19836 The defaults are such that all the input is kept. So it is possible to set e.g.
19837 just the end values to keep everything before the specified time.
19842 Drop everything except the second minute of input:
19844 ffmpeg -i INPUT -vf trim=60:120
19848 Keep only the first second:
19850 ffmpeg -i INPUT -vf trim=duration=1
19855 @section unpremultiply
19856 Apply alpha unpremultiply effect to input video stream using first plane
19857 of second stream as alpha.
19859 Both streams must have same dimensions and same pixel format.
19861 The filter accepts the following option:
19865 Set which planes will be processed, unprocessed planes will be copied.
19866 By default value 0xf, all planes will be processed.
19868 If the format has 1 or 2 components, then luma is bit 0.
19869 If the format has 3 or 4 components:
19870 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
19871 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
19872 If present, the alpha channel is always the last bit.
19875 Do not require 2nd input for processing, instead use alpha plane from input stream.
19881 Sharpen or blur the input video.
19883 It accepts the following parameters:
19886 @item luma_msize_x, lx
19887 Set the luma matrix horizontal size. It must be an odd integer between
19888 3 and 23. The default value is 5.
19890 @item luma_msize_y, ly
19891 Set the luma matrix vertical size. It must be an odd integer between 3
19892 and 23. The default value is 5.
19894 @item luma_amount, la
19895 Set the luma effect strength. It must be a floating point number, reasonable
19896 values lay between -1.5 and 1.5.
19898 Negative values will blur the input video, while positive values will
19899 sharpen it, a value of zero will disable the effect.
19901 Default value is 1.0.
19903 @item chroma_msize_x, cx
19904 Set the chroma matrix horizontal size. It must be an odd integer
19905 between 3 and 23. The default value is 5.
19907 @item chroma_msize_y, cy
19908 Set the chroma matrix vertical size. It must be an odd integer
19909 between 3 and 23. The default value is 5.
19911 @item chroma_amount, ca
19912 Set the chroma effect strength. It must be a floating point number, reasonable
19913 values lay between -1.5 and 1.5.
19915 Negative values will blur the input video, while positive values will
19916 sharpen it, a value of zero will disable the effect.
19918 Default value is 0.0.
19922 All parameters are optional and default to the equivalent of the
19923 string '5:5:1.0:5:5:0.0'.
19925 @subsection Examples
19929 Apply strong luma sharpen effect:
19931 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
19935 Apply a strong blur of both luma and chroma parameters:
19937 unsharp=7:7:-2:7:7:-2
19944 Decompose a video made of tiled images into the individual images.
19946 The frame rate of the output video is the frame rate of the input video
19947 multiplied by the number of tiles.
19949 This filter does the reverse of @ref{tile}.
19951 The filter accepts the following options:
19956 Set the grid size (i.e. the number of lines and columns). For the syntax of
19957 this option, check the
19958 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19961 @subsection Examples
19965 Produce a 1-second video from a still image file made of 25 frames stacked
19966 vertically, like an analogic film reel:
19968 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
19974 Apply ultra slow/simple postprocessing filter that compresses and decompresses
19975 the image at several (or - in the case of @option{quality} level @code{8} - all)
19976 shifts and average the results.
19978 The way this differs from the behavior of spp is that uspp actually encodes &
19979 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
19980 DCT similar to MJPEG.
19982 The filter accepts the following options:
19986 Set quality. This option defines the number of levels for averaging. It accepts
19987 an integer in the range 0-8. If set to @code{0}, the filter will have no
19988 effect. A value of @code{8} means the higher quality. For each increment of
19989 that value the speed drops by a factor of approximately 2. Default value is
19993 Force a constant quantization parameter. If not set, the filter will use the QP
19994 from the video stream (if available).
19999 Convert 360 videos between various formats.
20001 The filter accepts the following options:
20007 Set format of the input/output video.
20015 Equirectangular projection.
20020 Cubemap with 3x2/6x1/1x6 layout.
20022 Format specific options:
20027 Set padding proportion for the input/output cubemap. Values in decimals.
20034 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)
20037 Default value is @b{@samp{0}}.
20038 Maximum value is @b{@samp{0.1}}.
20042 Set fixed padding for the input/output cubemap. Values in pixels.
20044 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
20048 Set order of faces for the input/output cubemap. Choose one direction for each position.
20050 Designation of directions:
20066 Default value is @b{@samp{rludfb}}.
20070 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
20072 Designation of angles:
20075 0 degrees clockwise
20077 90 degrees clockwise
20079 180 degrees clockwise
20081 270 degrees clockwise
20084 Default value is @b{@samp{000000}}.
20088 Equi-Angular Cubemap.
20095 Format specific options:
20100 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20102 If diagonal field of view is set it overrides horizontal and vertical field of view.
20107 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20109 If diagonal field of view is set it overrides horizontal and vertical field of view.
20115 Format specific options:
20120 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20122 If diagonal field of view is set it overrides horizontal and vertical field of view.
20127 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20129 If diagonal field of view is set it overrides horizontal and vertical field of view.
20135 Facebook's 360 formats.
20138 Stereographic format.
20140 Format specific options:
20145 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20147 If diagonal field of view is set it overrides horizontal and vertical field of view.
20152 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20154 If diagonal field of view is set it overrides horizontal and vertical field of view.
20161 Ball format, gives significant distortion toward the back.
20164 Hammer-Aitoff map projection format.
20167 Sinusoidal map projection format.
20170 Fisheye projection.
20172 Format specific options:
20177 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20179 If diagonal field of view is set it overrides horizontal and vertical field of view.
20184 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20186 If diagonal field of view is set it overrides horizontal and vertical field of view.
20190 Pannini projection.
20192 Format specific options:
20195 Set output pannini parameter.
20198 Set input pannini parameter.
20202 Cylindrical projection.
20204 Format specific options:
20209 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20211 If diagonal field of view is set it overrides horizontal and vertical field of view.
20216 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20218 If diagonal field of view is set it overrides horizontal and vertical field of view.
20222 Perspective projection. @i{(output only)}
20224 Format specific options:
20227 Set perspective parameter.
20231 Tetrahedron projection.
20234 Truncated square pyramid projection.
20238 Half equirectangular projection.
20243 Format specific options:
20248 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20250 If diagonal field of view is set it overrides horizontal and vertical field of view.
20255 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20257 If diagonal field of view is set it overrides horizontal and vertical field of view.
20261 Orthographic format.
20263 Format specific options:
20268 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20270 If diagonal field of view is set it overrides horizontal and vertical field of view.
20275 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20277 If diagonal field of view is set it overrides horizontal and vertical field of view.
20281 Octahedron projection.
20285 Set interpolation method.@*
20286 @i{Note: more complex interpolation methods require much more memory to run.}
20296 Bilinear interpolation.
20298 Lagrange9 interpolation.
20301 Bicubic interpolation.
20304 Lanczos interpolation.
20307 Spline16 interpolation.
20310 Gaussian interpolation.
20312 Mitchell interpolation.
20315 Default value is @b{@samp{line}}.
20319 Set the output video resolution.
20321 Default resolution depends on formats.
20325 Set the input/output stereo format.
20336 Default value is @b{@samp{2d}} for input and output format.
20341 Set rotation for the output video. Values in degrees.
20344 Set rotation order for the output video. Choose one item for each position.
20355 Default value is @b{@samp{ypr}}.
20360 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
20364 Set if input video is flipped horizontally/vertically. Boolean values.
20367 Set if input video is transposed. Boolean value, by default disabled.
20370 Set if output video needs to be transposed. Boolean value, by default disabled.
20373 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
20376 @subsection Examples
20380 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
20382 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
20385 Extract back view of Equi-Angular Cubemap:
20387 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
20390 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
20392 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
20396 @subsection Commands
20398 This filter supports subset of above options as @ref{commands}.
20400 @section vaguedenoiser
20402 Apply a wavelet based denoiser.
20404 It transforms each frame from the video input into the wavelet domain,
20405 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
20406 the obtained coefficients. It does an inverse wavelet transform after.
20407 Due to wavelet properties, it should give a nice smoothed result, and
20408 reduced noise, without blurring picture features.
20410 This filter accepts the following options:
20414 The filtering strength. The higher, the more filtered the video will be.
20415 Hard thresholding can use a higher threshold than soft thresholding
20416 before the video looks overfiltered. Default value is 2.
20419 The filtering method the filter will use.
20421 It accepts the following values:
20424 All values under the threshold will be zeroed.
20427 All values under the threshold will be zeroed. All values above will be
20428 reduced by the threshold.
20431 Scales or nullifies coefficients - intermediary between (more) soft and
20432 (less) hard thresholding.
20435 Default is garrote.
20438 Number of times, the wavelet will decompose the picture. Picture can't
20439 be decomposed beyond a particular point (typically, 8 for a 640x480
20440 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
20443 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
20446 A list of the planes to process. By default all planes are processed.
20449 The threshold type the filter will use.
20451 It accepts the following values:
20454 Threshold used is same for all decompositions.
20457 Threshold used depends also on each decomposition coefficients.
20460 Default is universal.
20463 @section vectorscope
20465 Display 2 color component values in the two dimensional graph (which is called
20468 This filter accepts the following options:
20472 Set vectorscope mode.
20474 It accepts the following values:
20478 Gray values are displayed on graph, higher brightness means more pixels have
20479 same component color value on location in graph. This is the default mode.
20482 Gray values are displayed on graph. Surrounding pixels values which are not
20483 present in video frame are drawn in gradient of 2 color components which are
20484 set by option @code{x} and @code{y}. The 3rd color component is static.
20487 Actual color components values present in video frame are displayed on graph.
20490 Similar as color2 but higher frequency of same values @code{x} and @code{y}
20491 on graph increases value of another color component, which is luminance by
20492 default values of @code{x} and @code{y}.
20495 Actual colors present in video frame are displayed on graph. If two different
20496 colors map to same position on graph then color with higher value of component
20497 not present in graph is picked.
20500 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
20501 component picked from radial gradient.
20505 Set which color component will be represented on X-axis. Default is @code{1}.
20508 Set which color component will be represented on Y-axis. Default is @code{2}.
20511 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
20512 of color component which represents frequency of (X, Y) location in graph.
20517 No envelope, this is default.
20520 Instant envelope, even darkest single pixel will be clearly highlighted.
20523 Hold maximum and minimum values presented in graph over time. This way you
20524 can still spot out of range values without constantly looking at vectorscope.
20527 Peak and instant envelope combined together.
20531 Set what kind of graticule to draw.
20540 Set graticule opacity.
20543 Set graticule flags.
20547 Draw graticule for white point.
20550 Draw graticule for black point.
20553 Draw color points short names.
20557 Set background opacity.
20559 @item lthreshold, l
20560 Set low threshold for color component not represented on X or Y axis.
20561 Values lower than this value will be ignored. Default is 0.
20562 Note this value is multiplied with actual max possible value one pixel component
20563 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
20566 @item hthreshold, h
20567 Set high threshold for color component not represented on X or Y axis.
20568 Values higher than this value will be ignored. Default is 1.
20569 Note this value is multiplied with actual max possible value one pixel component
20570 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
20571 is 0.9 * 255 = 230.
20573 @item colorspace, c
20574 Set what kind of colorspace to use when drawing graticule.
20584 Set color tint for gray/tint vectorscope mode. By default both options are zero.
20585 This means no tint, and output will remain gray.
20588 @anchor{vidstabdetect}
20589 @section vidstabdetect
20591 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
20592 @ref{vidstabtransform} for pass 2.
20594 This filter generates a file with relative translation and rotation
20595 transform information about subsequent frames, which is then used by
20596 the @ref{vidstabtransform} filter.
20598 To enable compilation of this filter you need to configure FFmpeg with
20599 @code{--enable-libvidstab}.
20601 This filter accepts the following options:
20605 Set the path to the file used to write the transforms information.
20606 Default value is @file{transforms.trf}.
20609 Set how shaky the video is and how quick the camera is. It accepts an
20610 integer in the range 1-10, a value of 1 means little shakiness, a
20611 value of 10 means strong shakiness. Default value is 5.
20614 Set the accuracy of the detection process. It must be a value in the
20615 range 1-15. A value of 1 means low accuracy, a value of 15 means high
20616 accuracy. Default value is 15.
20619 Set stepsize of the search process. The region around minimum is
20620 scanned with 1 pixel resolution. Default value is 6.
20623 Set minimum contrast. Below this value a local measurement field is
20624 discarded. Must be a floating point value in the range 0-1. Default
20628 Set reference frame number for tripod mode.
20630 If enabled, the motion of the frames is compared to a reference frame
20631 in the filtered stream, identified by the specified number. The idea
20632 is to compensate all movements in a more-or-less static scene and keep
20633 the camera view absolutely still.
20635 If set to 0, it is disabled. The frames are counted starting from 1.
20638 Show fields and transforms in the resulting frames. It accepts an
20639 integer in the range 0-2. Default value is 0, which disables any
20643 @subsection Examples
20647 Use default values:
20653 Analyze strongly shaky movie and put the results in file
20654 @file{mytransforms.trf}:
20656 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
20660 Visualize the result of internal transformations in the resulting
20663 vidstabdetect=show=1
20667 Analyze a video with medium shakiness using @command{ffmpeg}:
20669 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
20673 @anchor{vidstabtransform}
20674 @section vidstabtransform
20676 Video stabilization/deshaking: pass 2 of 2,
20677 see @ref{vidstabdetect} for pass 1.
20679 Read a file with transform information for each frame and
20680 apply/compensate them. Together with the @ref{vidstabdetect}
20681 filter this can be used to deshake videos. See also
20682 @url{http://public.hronopik.de/vid.stab}. It is important to also use
20683 the @ref{unsharp} filter, see below.
20685 To enable compilation of this filter you need to configure FFmpeg with
20686 @code{--enable-libvidstab}.
20688 @subsection Options
20692 Set path to the file used to read the transforms. Default value is
20693 @file{transforms.trf}.
20696 Set the number of frames (value*2 + 1) used for lowpass filtering the
20697 camera movements. Default value is 10.
20699 For example a number of 10 means that 21 frames are used (10 in the
20700 past and 10 in the future) to smoothen the motion in the video. A
20701 larger value leads to a smoother video, but limits the acceleration of
20702 the camera (pan/tilt movements). 0 is a special case where a static
20703 camera is simulated.
20706 Set the camera path optimization algorithm.
20708 Accepted values are:
20711 gaussian kernel low-pass filter on camera motion (default)
20713 averaging on transformations
20717 Set maximal number of pixels to translate frames. Default value is -1,
20721 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
20722 value is -1, meaning no limit.
20725 Specify how to deal with borders that may be visible due to movement
20728 Available values are:
20731 keep image information from previous frame (default)
20733 fill the border black
20737 Invert transforms if set to 1. Default value is 0.
20740 Consider transforms as relative to previous frame if set to 1,
20741 absolute if set to 0. Default value is 0.
20744 Set percentage to zoom. A positive value will result in a zoom-in
20745 effect, a negative value in a zoom-out effect. Default value is 0 (no
20749 Set optimal zooming to avoid borders.
20751 Accepted values are:
20756 optimal static zoom value is determined (only very strong movements
20757 will lead to visible borders) (default)
20759 optimal adaptive zoom value is determined (no borders will be
20760 visible), see @option{zoomspeed}
20763 Note that the value given at zoom is added to the one calculated here.
20766 Set percent to zoom maximally each frame (enabled when
20767 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
20771 Specify type of interpolation.
20773 Available values are:
20778 linear only horizontal
20780 linear in both directions (default)
20782 cubic in both directions (slow)
20786 Enable virtual tripod mode if set to 1, which is equivalent to
20787 @code{relative=0:smoothing=0}. Default value is 0.
20789 Use also @code{tripod} option of @ref{vidstabdetect}.
20792 Increase log verbosity if set to 1. Also the detected global motions
20793 are written to the temporary file @file{global_motions.trf}. Default
20797 @subsection Examples
20801 Use @command{ffmpeg} for a typical stabilization with default values:
20803 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
20806 Note the use of the @ref{unsharp} filter which is always recommended.
20809 Zoom in a bit more and load transform data from a given file:
20811 vidstabtransform=zoom=5:input="mytransforms.trf"
20815 Smoothen the video even more:
20817 vidstabtransform=smoothing=30
20823 Flip the input video vertically.
20825 For example, to vertically flip a video with @command{ffmpeg}:
20827 ffmpeg -i in.avi -vf "vflip" out.avi
20832 Detect variable frame rate video.
20834 This filter tries to detect if the input is variable or constant frame rate.
20836 At end it will output number of frames detected as having variable delta pts,
20837 and ones with constant delta pts.
20838 If there was frames with variable delta, than it will also show min, max and
20839 average delta encountered.
20843 Boost or alter saturation.
20845 The filter accepts the following options:
20848 Set strength of boost if positive value or strength of alter if negative value.
20849 Default is 0. Allowed range is from -2 to 2.
20852 Set the red balance. Default is 1. Allowed range is from -10 to 10.
20855 Set the green balance. Default is 1. Allowed range is from -10 to 10.
20858 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
20861 Set the red luma coefficient.
20864 Set the green luma coefficient.
20867 Set the blue luma coefficient.
20870 If @code{intensity} is negative and this is set to 1, colors will change,
20871 otherwise colors will be less saturated, more towards gray.
20874 @subsection Commands
20876 This filter supports the all above options as @ref{commands}.
20881 Make or reverse a natural vignetting effect.
20883 The filter accepts the following options:
20887 Set lens angle expression as a number of radians.
20889 The value is clipped in the @code{[0,PI/2]} range.
20891 Default value: @code{"PI/5"}
20895 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
20899 Set forward/backward mode.
20901 Available modes are:
20904 The larger the distance from the central point, the darker the image becomes.
20907 The larger the distance from the central point, the brighter the image becomes.
20908 This can be used to reverse a vignette effect, though there is no automatic
20909 detection to extract the lens @option{angle} and other settings (yet). It can
20910 also be used to create a burning effect.
20913 Default value is @samp{forward}.
20916 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
20918 It accepts the following values:
20921 Evaluate expressions only once during the filter initialization.
20924 Evaluate expressions for each incoming frame. This is way slower than the
20925 @samp{init} mode since it requires all the scalers to be re-computed, but it
20926 allows advanced dynamic expressions.
20929 Default value is @samp{init}.
20932 Set dithering to reduce the circular banding effects. Default is @code{1}
20936 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
20937 Setting this value to the SAR of the input will make a rectangular vignetting
20938 following the dimensions of the video.
20940 Default is @code{1/1}.
20943 @subsection Expressions
20945 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
20946 following parameters.
20951 input width and height
20954 the number of input frame, starting from 0
20957 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
20958 @var{TB} units, NAN if undefined
20961 frame rate of the input video, NAN if the input frame rate is unknown
20964 the PTS (Presentation TimeStamp) of the filtered video frame,
20965 expressed in seconds, NAN if undefined
20968 time base of the input video
20972 @subsection Examples
20976 Apply simple strong vignetting effect:
20982 Make a flickering vignetting:
20984 vignette='PI/4+random(1)*PI/50':eval=frame
20989 @section vmafmotion
20991 Obtain the average VMAF motion score of a video.
20992 It is one of the component metrics of VMAF.
20994 The obtained average motion score is printed through the logging system.
20996 The filter accepts the following options:
21000 If specified, the filter will use the named file to save the motion score of
21001 each frame with respect to the previous frame.
21002 When filename equals "-" the data is sent to standard output.
21007 ffmpeg -i ref.mpg -vf vmafmotion -f null -
21011 Stack input videos vertically.
21013 All streams must be of same pixel format and of same width.
21015 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
21016 to create same output.
21018 The filter accepts the following options:
21022 Set number of input streams. Default is 2.
21025 If set to 1, force the output to terminate when the shortest input
21026 terminates. Default value is 0.
21031 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
21032 Deinterlacing Filter").
21034 Based on the process described by Martin Weston for BBC R&D, and
21035 implemented based on the de-interlace algorithm written by Jim
21036 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
21037 uses filter coefficients calculated by BBC R&D.
21039 This filter uses field-dominance information in frame to decide which
21040 of each pair of fields to place first in the output.
21041 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
21043 There are two sets of filter coefficients, so called "simple"
21044 and "complex". Which set of filter coefficients is used can
21045 be set by passing an optional parameter:
21049 Set the interlacing filter coefficients. Accepts one of the following values:
21053 Simple filter coefficient set.
21055 More-complex filter coefficient set.
21057 Default value is @samp{complex}.
21060 The interlacing mode to adopt. It accepts one of the following values:
21064 Output one frame for each frame.
21066 Output one frame for each field.
21069 The default value is @code{field}.
21072 The picture field parity assumed for the input interlaced video. It accepts one
21073 of the following values:
21077 Assume the top field is first.
21079 Assume the bottom field is first.
21081 Enable automatic detection of field parity.
21084 The default value is @code{auto}.
21085 If the interlacing is unknown or the decoder does not export this information,
21086 top field first will be assumed.
21089 Specify which frames to deinterlace. Accepts one of the following values:
21093 Deinterlace all frames,
21095 Only deinterlace frames marked as interlaced.
21098 Default value is @samp{all}.
21101 @subsection Commands
21102 This filter supports same @ref{commands} as options.
21105 Video waveform monitor.
21107 The waveform monitor plots color component intensity. By default luminance
21108 only. Each column of the waveform corresponds to a column of pixels in the
21111 It accepts the following options:
21115 Can be either @code{row}, or @code{column}. Default is @code{column}.
21116 In row mode, the graph on the left side represents color component value 0 and
21117 the right side represents value = 255. In column mode, the top side represents
21118 color component value = 0 and bottom side represents value = 255.
21121 Set intensity. Smaller values are useful to find out how many values of the same
21122 luminance are distributed across input rows/columns.
21123 Default value is @code{0.04}. Allowed range is [0, 1].
21126 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
21127 In mirrored mode, higher values will be represented on the left
21128 side for @code{row} mode and at the top for @code{column} mode. Default is
21129 @code{1} (mirrored).
21133 It accepts the following values:
21136 Presents information identical to that in the @code{parade}, except
21137 that the graphs representing color components are superimposed directly
21140 This display mode makes it easier to spot relative differences or similarities
21141 in overlapping areas of the color components that are supposed to be identical,
21142 such as neutral whites, grays, or blacks.
21145 Display separate graph for the color components side by side in
21146 @code{row} mode or one below the other in @code{column} mode.
21149 Display separate graph for the color components side by side in
21150 @code{column} mode or one below the other in @code{row} mode.
21152 Using this display mode makes it easy to spot color casts in the highlights
21153 and shadows of an image, by comparing the contours of the top and the bottom
21154 graphs of each waveform. Since whites, grays, and blacks are characterized
21155 by exactly equal amounts of red, green, and blue, neutral areas of the picture
21156 should display three waveforms of roughly equal width/height. If not, the
21157 correction is easy to perform by making level adjustments the three waveforms.
21159 Default is @code{stack}.
21161 @item components, c
21162 Set which color components to display. Default is 1, which means only luminance
21163 or red color component if input is in RGB colorspace. If is set for example to
21164 7 it will display all 3 (if) available color components.
21169 No envelope, this is default.
21172 Instant envelope, minimum and maximum values presented in graph will be easily
21173 visible even with small @code{step} value.
21176 Hold minimum and maximum values presented in graph across time. This way you
21177 can still spot out of range values without constantly looking at waveforms.
21180 Peak and instant envelope combined together.
21186 No filtering, this is default.
21189 Luma and chroma combined together.
21192 Similar as above, but shows difference between blue and red chroma.
21195 Similar as above, but use different colors.
21198 Similar as above, but again with different colors.
21201 Displays only chroma.
21204 Displays actual color value on waveform.
21207 Similar as above, but with luma showing frequency of chroma values.
21211 Set which graticule to display.
21215 Do not display graticule.
21218 Display green graticule showing legal broadcast ranges.
21221 Display orange graticule showing legal broadcast ranges.
21224 Display invert graticule showing legal broadcast ranges.
21228 Set graticule opacity.
21231 Set graticule flags.
21235 Draw numbers above lines. By default enabled.
21238 Draw dots instead of lines.
21242 Set scale used for displaying graticule.
21249 Default is digital.
21252 Set background opacity.
21256 Set tint for output.
21257 Only used with lowpass filter and when display is not overlay and input
21258 pixel formats are not RGB.
21261 @section weave, doubleweave
21263 The @code{weave} takes a field-based video input and join
21264 each two sequential fields into single frame, producing a new double
21265 height clip with half the frame rate and half the frame count.
21267 The @code{doubleweave} works same as @code{weave} but without
21268 halving frame rate and frame count.
21270 It accepts the following option:
21274 Set first field. Available values are:
21278 Set the frame as top-field-first.
21281 Set the frame as bottom-field-first.
21285 @subsection Examples
21289 Interlace video using @ref{select} and @ref{separatefields} filter:
21291 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
21296 Apply the xBR high-quality magnification filter which is designed for pixel
21297 art. It follows a set of edge-detection rules, see
21298 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
21300 It accepts the following option:
21304 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
21305 @code{3xBR} and @code{4} for @code{4xBR}.
21306 Default is @code{3}.
21311 Apply cross fade from one input video stream to another input video stream.
21312 The cross fade is applied for specified duration.
21314 The filter accepts the following options:
21318 Set one of available transition effects:
21366 Default transition effect is fade.
21369 Set cross fade duration in seconds.
21370 Default duration is 1 second.
21373 Set cross fade start relative to first input stream in seconds.
21374 Default offset is 0.
21377 Set expression for custom transition effect.
21379 The expressions can use the following variables and functions:
21384 The coordinates of the current sample.
21388 The width and height of the image.
21391 Progress of transition effect.
21394 Currently processed plane.
21397 Return value of first input at current location and plane.
21400 Return value of second input at current location and plane.
21406 Return the value of the pixel at location (@var{x},@var{y}) of the
21407 first/second/third/fourth component of first input.
21413 Return the value of the pixel at location (@var{x},@var{y}) of the
21414 first/second/third/fourth component of second input.
21418 @subsection Examples
21422 Cross fade from one input video to another input video, with fade transition and duration of transition
21423 of 2 seconds starting at offset of 5 seconds:
21425 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
21430 Pick median pixels from several input videos.
21432 The filter accepts the following options:
21436 Set number of inputs.
21437 Default is 3. Allowed range is from 3 to 255.
21438 If number of inputs is even number, than result will be mean value between two median values.
21441 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
21444 Set median percentile. Default value is @code{0.5}.
21445 Default value of @code{0.5} will pick always median values, while @code{0} will pick
21446 minimum values, and @code{1} maximum values.
21449 @subsection Commands
21451 This filter supports all above options as @ref{commands}, excluding option @code{inputs}.
21454 Stack video inputs into custom layout.
21456 All streams must be of same pixel format.
21458 The filter accepts the following options:
21462 Set number of input streams. Default is 2.
21465 Specify layout of inputs.
21466 This option requires the desired layout configuration to be explicitly set by the user.
21467 This sets position of each video input in output. Each input
21468 is separated by '|'.
21469 The first number represents the column, and the second number represents the row.
21470 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
21471 where X is video input from which to take width or height.
21472 Multiple values can be used when separated by '+'. In such
21473 case values are summed together.
21475 Note that if inputs are of different sizes gaps may appear, as not all of
21476 the output video frame will be filled. Similarly, videos can overlap each
21477 other if their position doesn't leave enough space for the full frame of
21480 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
21481 a layout must be set by the user.
21484 If set to 1, force the output to terminate when the shortest input
21485 terminates. Default value is 0.
21488 If set to valid color, all unused pixels will be filled with that color.
21489 By default fill is set to none, so it is disabled.
21492 @subsection Examples
21496 Display 4 inputs into 2x2 grid.
21500 input1(0, 0) | input3(w0, 0)
21501 input2(0, h0) | input4(w0, h0)
21505 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
21508 Note that if inputs are of different sizes, gaps or overlaps may occur.
21511 Display 4 inputs into 1x4 grid.
21518 input4(0, h0+h1+h2)
21522 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
21525 Note that if inputs are of different widths, unused space will appear.
21528 Display 9 inputs into 3x3 grid.
21532 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
21533 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
21534 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
21538 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
21541 Note that if inputs are of different sizes, gaps or overlaps may occur.
21544 Display 16 inputs into 4x4 grid.
21548 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
21549 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
21550 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
21551 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
21555 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|
21556 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
21559 Note that if inputs are of different sizes, gaps or overlaps may occur.
21566 Deinterlace the input video ("yadif" means "yet another deinterlacing
21569 It accepts the following parameters:
21575 The interlacing mode to adopt. It accepts one of the following values:
21578 @item 0, send_frame
21579 Output one frame for each frame.
21580 @item 1, send_field
21581 Output one frame for each field.
21582 @item 2, send_frame_nospatial
21583 Like @code{send_frame}, but it skips the spatial interlacing check.
21584 @item 3, send_field_nospatial
21585 Like @code{send_field}, but it skips the spatial interlacing check.
21588 The default value is @code{send_frame}.
21591 The picture field parity assumed for the input interlaced video. It accepts one
21592 of the following values:
21596 Assume the top field is first.
21598 Assume the bottom field is first.
21600 Enable automatic detection of field parity.
21603 The default value is @code{auto}.
21604 If the interlacing is unknown or the decoder does not export this information,
21605 top field first will be assumed.
21608 Specify which frames to deinterlace. Accepts one of the following
21613 Deinterlace all frames.
21614 @item 1, interlaced
21615 Only deinterlace frames marked as interlaced.
21618 The default value is @code{all}.
21621 @section yadif_cuda
21623 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
21624 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
21627 It accepts the following parameters:
21633 The interlacing mode to adopt. It accepts one of the following values:
21636 @item 0, send_frame
21637 Output one frame for each frame.
21638 @item 1, send_field
21639 Output one frame for each field.
21640 @item 2, send_frame_nospatial
21641 Like @code{send_frame}, but it skips the spatial interlacing check.
21642 @item 3, send_field_nospatial
21643 Like @code{send_field}, but it skips the spatial interlacing check.
21646 The default value is @code{send_frame}.
21649 The picture field parity assumed for the input interlaced video. It accepts one
21650 of the following values:
21654 Assume the top field is first.
21656 Assume the bottom field is first.
21658 Enable automatic detection of field parity.
21661 The default value is @code{auto}.
21662 If the interlacing is unknown or the decoder does not export this information,
21663 top field first will be assumed.
21666 Specify which frames to deinterlace. Accepts one of the following
21671 Deinterlace all frames.
21672 @item 1, interlaced
21673 Only deinterlace frames marked as interlaced.
21676 The default value is @code{all}.
21681 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
21682 The algorithm is described in
21683 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
21685 It accepts the following parameters:
21689 Set the window radius. Default value is 3.
21692 Set which planes to filter. Default is only the first plane.
21695 Set blur strength. Default value is 128.
21698 @subsection Commands
21699 This filter supports same @ref{commands} as options.
21703 Apply Zoom & Pan effect.
21705 This filter accepts the following options:
21709 Set the zoom expression. Range is 1-10. Default is 1.
21713 Set the x and y expression. Default is 0.
21716 Set the duration expression in number of frames.
21717 This sets for how many number of frames effect will last for
21718 single input image.
21721 Set the output image size, default is 'hd720'.
21724 Set the output frame rate, default is '25'.
21727 Each expression can contain the following constants:
21746 Output frame count.
21749 The input timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
21751 @item out_time, time, ot
21752 The output timestamp expressed in seconds.
21756 Last calculated 'x' and 'y' position from 'x' and 'y' expression
21757 for current input frame.
21761 'x' and 'y' of last output frame of previous input frame or 0 when there was
21762 not yet such frame (first input frame).
21765 Last calculated zoom from 'z' expression for current input frame.
21768 Last calculated zoom of last output frame of previous input frame.
21771 Number of output frames for current input frame. Calculated from 'd' expression
21772 for each input frame.
21775 number of output frames created for previous input frame
21778 Rational number: input width / input height
21781 sample aspect ratio
21784 display aspect ratio
21788 @subsection Examples
21792 Zoom in up to 1.5x and pan at same time to some spot near center of picture:
21794 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
21798 Zoom in up to 1.5x and pan always at center of picture:
21800 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21804 Same as above but without pausing:
21806 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21810 Zoom in 2x into center of picture only for the first second of the input video:
21812 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21819 Scale (resize) the input video, using the z.lib library:
21820 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
21821 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
21823 The zscale filter forces the output display aspect ratio to be the same
21824 as the input, by changing the output sample aspect ratio.
21826 If the input image format is different from the format requested by
21827 the next filter, the zscale filter will convert the input to the
21830 @subsection Options
21831 The filter accepts the following options.
21836 Set the output video dimension expression. Default value is the input
21839 If the @var{width} or @var{w} value is 0, the input width is used for
21840 the output. If the @var{height} or @var{h} value is 0, the input height
21841 is used for the output.
21843 If one and only one of the values is -n with n >= 1, the zscale filter
21844 will use a value that maintains the aspect ratio of the input image,
21845 calculated from the other specified dimension. After that it will,
21846 however, make sure that the calculated dimension is divisible by n and
21847 adjust the value if necessary.
21849 If both values are -n with n >= 1, the behavior will be identical to
21850 both values being set to 0 as previously detailed.
21852 See below for the list of accepted constants for use in the dimension
21856 Set the video size. For the syntax of this option, check the
21857 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21860 Set the dither type.
21862 Possible values are:
21867 @item error_diffusion
21873 Set the resize filter type.
21875 Possible values are:
21885 Default is bilinear.
21888 Set the color range.
21890 Possible values are:
21897 Default is same as input.
21900 Set the color primaries.
21902 Possible values are:
21912 Default is same as input.
21915 Set the transfer characteristics.
21917 Possible values are:
21931 Default is same as input.
21934 Set the colorspace matrix.
21936 Possible value are:
21947 Default is same as input.
21950 Set the input color range.
21952 Possible values are:
21959 Default is same as input.
21961 @item primariesin, pin
21962 Set the input color primaries.
21964 Possible values are:
21974 Default is same as input.
21976 @item transferin, tin
21977 Set the input transfer characteristics.
21979 Possible values are:
21990 Default is same as input.
21992 @item matrixin, min
21993 Set the input colorspace matrix.
21995 Possible value are:
22007 Set the output chroma location.
22009 Possible values are:
22020 @item chromalin, cin
22021 Set the input chroma location.
22023 Possible values are:
22035 Set the nominal peak luminance.
22038 The values of the @option{w} and @option{h} options are expressions
22039 containing the following constants:
22044 The input width and height
22048 These are the same as @var{in_w} and @var{in_h}.
22052 The output (scaled) width and height
22056 These are the same as @var{out_w} and @var{out_h}
22059 The same as @var{iw} / @var{ih}
22062 input sample aspect ratio
22065 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
22069 horizontal and vertical input chroma subsample values. For example for the
22070 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
22074 horizontal and vertical output chroma subsample values. For example for the
22075 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
22078 @subsection Commands
22080 This filter supports the following commands:
22084 Set the output video dimension expression.
22085 The command accepts the same syntax of the corresponding option.
22087 If the specified expression is not valid, it is kept at its current
22091 @c man end VIDEO FILTERS
22093 @chapter OpenCL Video Filters
22094 @c man begin OPENCL VIDEO FILTERS
22096 Below is a description of the currently available OpenCL video filters.
22098 To enable compilation of these filters you need to configure FFmpeg with
22099 @code{--enable-opencl}.
22101 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
22104 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
22105 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
22106 given device parameters.
22108 @item -filter_hw_device @var{name}
22109 Pass the hardware device called @var{name} to all filters in any filter graph.
22113 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
22117 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
22119 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
22123 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.
22125 @section avgblur_opencl
22127 Apply average blur filter.
22129 The filter accepts the following options:
22133 Set horizontal radius size.
22134 Range is @code{[1, 1024]} and default value is @code{1}.
22137 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22140 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
22143 @subsection Example
22147 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.
22149 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
22153 @section boxblur_opencl
22155 Apply a boxblur algorithm to the input video.
22157 It accepts the following parameters:
22161 @item luma_radius, lr
22162 @item luma_power, lp
22163 @item chroma_radius, cr
22164 @item chroma_power, cp
22165 @item alpha_radius, ar
22166 @item alpha_power, ap
22170 A description of the accepted options follows.
22173 @item luma_radius, lr
22174 @item chroma_radius, cr
22175 @item alpha_radius, ar
22176 Set an expression for the box radius in pixels used for blurring the
22177 corresponding input plane.
22179 The radius value must be a non-negative number, and must not be
22180 greater than the value of the expression @code{min(w,h)/2} for the
22181 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
22184 Default value for @option{luma_radius} is "2". If not specified,
22185 @option{chroma_radius} and @option{alpha_radius} default to the
22186 corresponding value set for @option{luma_radius}.
22188 The expressions can contain the following constants:
22192 The input width and height in pixels.
22196 The input chroma image width and height in pixels.
22200 The horizontal and vertical chroma subsample values. For example, for the
22201 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
22204 @item luma_power, lp
22205 @item chroma_power, cp
22206 @item alpha_power, ap
22207 Specify how many times the boxblur filter is applied to the
22208 corresponding plane.
22210 Default value for @option{luma_power} is 2. If not specified,
22211 @option{chroma_power} and @option{alpha_power} default to the
22212 corresponding value set for @option{luma_power}.
22214 A value of 0 will disable the effect.
22217 @subsection Examples
22219 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.
22223 Apply a boxblur filter with the luma, chroma, and alpha radius
22224 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.
22226 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
22227 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
22231 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.
22233 For the luma plane, a 2x2 box radius will be run once.
22235 For the chroma plane, a 4x4 box radius will be run 5 times.
22237 For the alpha plane, a 3x3 box radius will be run 7 times.
22239 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
22243 @section colorkey_opencl
22244 RGB colorspace color keying.
22246 The filter accepts the following options:
22250 The color which will be replaced with transparency.
22253 Similarity percentage with the key color.
22255 0.01 matches only the exact key color, while 1.0 matches everything.
22260 0.0 makes pixels either fully transparent, or not transparent at all.
22262 Higher values result in semi-transparent pixels, with a higher transparency
22263 the more similar the pixels color is to the key color.
22266 @subsection Examples
22270 Make every semi-green pixel in the input transparent with some slight blending:
22272 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
22276 @section convolution_opencl
22278 Apply convolution of 3x3, 5x5, 7x7 matrix.
22280 The filter accepts the following options:
22287 Set matrix for each plane.
22288 Matrix is sequence of 9, 25 or 49 signed numbers.
22289 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
22295 Set multiplier for calculated value for each plane.
22296 If unset or 0, it will be sum of all matrix elements.
22297 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
22303 Set bias for each plane. This value is added to the result of the multiplication.
22304 Useful for making the overall image brighter or darker.
22305 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
22309 @subsection Examples
22315 -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
22321 -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
22325 Apply edge enhance:
22327 -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
22333 -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
22337 Apply laplacian edge detector which includes diagonals:
22339 -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
22345 -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
22349 @section erosion_opencl
22351 Apply erosion effect to the video.
22353 This filter replaces the pixel by the local(3x3) minimum.
22355 It accepts the following options:
22362 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22363 If @code{0}, plane will remain unchanged.
22366 Flag which specifies the pixel to refer to.
22367 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22369 Flags to local 3x3 coordinates region centered on @code{x}:
22378 @subsection Example
22382 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.
22384 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22388 @section deshake_opencl
22389 Feature-point based video stabilization filter.
22391 The filter accepts the following options:
22395 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
22398 Whether or not additional debug info should be displayed, both in the processed output and in the console.
22400 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
22402 Viewing point matches in the output video is only supported for RGB input.
22404 Defaults to @code{0}.
22406 @item adaptive_crop
22407 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
22409 Defaults to @code{1}.
22411 @item refine_features
22412 Whether or not feature points should be refined at a sub-pixel level.
22414 This can be turned off for a slight performance gain at the cost of precision.
22416 Defaults to @code{1}.
22418 @item smooth_strength
22419 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
22421 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
22423 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
22425 Defaults to @code{0.0}.
22427 @item smooth_window_multiplier
22428 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
22430 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
22432 Acceptable values range from @code{0.1} to @code{10.0}.
22434 Larger values increase the amount of motion data available for determining how to smooth the camera path,
22435 potentially improving smoothness, but also increase latency and memory usage.
22437 Defaults to @code{2.0}.
22441 @subsection Examples
22445 Stabilize a video with a fixed, medium smoothing strength:
22447 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
22451 Stabilize a video with debugging (both in console and in rendered video):
22453 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
22457 @section dilation_opencl
22459 Apply dilation effect to the video.
22461 This filter replaces the pixel by the local(3x3) maximum.
22463 It accepts the following options:
22470 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22471 If @code{0}, plane will remain unchanged.
22474 Flag which specifies the pixel to refer to.
22475 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22477 Flags to local 3x3 coordinates region centered on @code{x}:
22486 @subsection Example
22490 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.
22492 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22496 @section nlmeans_opencl
22498 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
22500 @section overlay_opencl
22502 Overlay one video on top of another.
22504 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
22505 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
22507 The filter accepts the following options:
22512 Set the x coordinate of the overlaid video on the main video.
22513 Default value is @code{0}.
22516 Set the y coordinate of the overlaid video on the main video.
22517 Default value is @code{0}.
22521 @subsection Examples
22525 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
22527 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22530 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
22532 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22537 @section pad_opencl
22539 Add paddings to the input image, and place the original input at the
22540 provided @var{x}, @var{y} coordinates.
22542 It accepts the following options:
22547 Specify an expression for the size of the output image with the
22548 paddings added. If the value for @var{width} or @var{height} is 0, the
22549 corresponding input size is used for the output.
22551 The @var{width} expression can reference the value set by the
22552 @var{height} expression, and vice versa.
22554 The default value of @var{width} and @var{height} is 0.
22558 Specify the offsets to place the input image at within the padded area,
22559 with respect to the top/left border of the output image.
22561 The @var{x} expression can reference the value set by the @var{y}
22562 expression, and vice versa.
22564 The default value of @var{x} and @var{y} is 0.
22566 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
22567 so the input image is centered on the padded area.
22570 Specify the color of the padded area. For the syntax of this option,
22571 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
22572 manual,ffmpeg-utils}.
22575 Pad to an aspect instead to a resolution.
22578 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
22579 options are expressions containing the following constants:
22584 The input video width and height.
22588 These are the same as @var{in_w} and @var{in_h}.
22592 The output width and height (the size of the padded area), as
22593 specified by the @var{width} and @var{height} expressions.
22597 These are the same as @var{out_w} and @var{out_h}.
22601 The x and y offsets as specified by the @var{x} and @var{y}
22602 expressions, or NAN if not yet specified.
22605 same as @var{iw} / @var{ih}
22608 input sample aspect ratio
22611 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
22614 @section prewitt_opencl
22616 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
22618 The filter accepts the following option:
22622 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22625 Set value which will be multiplied with filtered result.
22626 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22629 Set value which will be added to filtered result.
22630 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22633 @subsection Example
22637 Apply the Prewitt operator with scale set to 2 and delta set to 10.
22639 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
22643 @anchor{program_opencl}
22644 @section program_opencl
22646 Filter video using an OpenCL program.
22651 OpenCL program source file.
22654 Kernel name in program.
22657 Number of inputs to the filter. Defaults to 1.
22660 Size of output frames. Defaults to the same as the first input.
22664 The @code{program_opencl} filter also supports the @ref{framesync} options.
22666 The program source file must contain a kernel function with the given name,
22667 which will be run once for each plane of the output. Each run on a plane
22668 gets enqueued as a separate 2D global NDRange with one work-item for each
22669 pixel to be generated. The global ID offset for each work-item is therefore
22670 the coordinates of a pixel in the destination image.
22672 The kernel function needs to take the following arguments:
22675 Destination image, @var{__write_only image2d_t}.
22677 This image will become the output; the kernel should write all of it.
22679 Frame index, @var{unsigned int}.
22681 This is a counter starting from zero and increasing by one for each frame.
22683 Source images, @var{__read_only image2d_t}.
22685 These are the most recent images on each input. The kernel may read from
22686 them to generate the output, but they can't be written to.
22693 Copy the input to the output (output must be the same size as the input).
22695 __kernel void copy(__write_only image2d_t destination,
22696 unsigned int index,
22697 __read_only image2d_t source)
22699 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
22701 int2 location = (int2)(get_global_id(0), get_global_id(1));
22703 float4 value = read_imagef(source, sampler, location);
22705 write_imagef(destination, location, value);
22710 Apply a simple transformation, rotating the input by an amount increasing
22711 with the index counter. Pixel values are linearly interpolated by the
22712 sampler, and the output need not have the same dimensions as the input.
22714 __kernel void rotate_image(__write_only image2d_t dst,
22715 unsigned int index,
22716 __read_only image2d_t src)
22718 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22719 CLK_FILTER_LINEAR);
22721 float angle = (float)index / 100.0f;
22723 float2 dst_dim = convert_float2(get_image_dim(dst));
22724 float2 src_dim = convert_float2(get_image_dim(src));
22726 float2 dst_cen = dst_dim / 2.0f;
22727 float2 src_cen = src_dim / 2.0f;
22729 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22731 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
22733 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
22734 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
22736 src_pos = src_pos * src_dim / dst_dim;
22738 float2 src_loc = src_pos + src_cen;
22740 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
22741 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
22742 write_imagef(dst, dst_loc, 0.5f);
22744 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
22749 Blend two inputs together, with the amount of each input used varying
22750 with the index counter.
22752 __kernel void blend_images(__write_only image2d_t dst,
22753 unsigned int index,
22754 __read_only image2d_t src1,
22755 __read_only image2d_t src2)
22757 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22758 CLK_FILTER_LINEAR);
22760 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
22762 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22763 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
22764 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
22766 float4 val1 = read_imagef(src1, sampler, src1_loc);
22767 float4 val2 = read_imagef(src2, sampler, src2_loc);
22769 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
22775 @section roberts_opencl
22776 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
22778 The filter accepts the following option:
22782 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22785 Set value which will be multiplied with filtered result.
22786 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22789 Set value which will be added to filtered result.
22790 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22793 @subsection Example
22797 Apply the Roberts cross operator with scale set to 2 and delta set to 10
22799 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
22803 @section sobel_opencl
22805 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
22807 The filter accepts the following option:
22811 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22814 Set value which will be multiplied with filtered result.
22815 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22818 Set value which will be added to filtered result.
22819 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22822 @subsection Example
22826 Apply sobel operator with scale set to 2 and delta set to 10
22828 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
22832 @section tonemap_opencl
22834 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
22836 It accepts the following parameters:
22840 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
22843 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
22846 Apply desaturation for highlights that exceed this level of brightness. The
22847 higher the parameter, the more color information will be preserved. This
22848 setting helps prevent unnaturally blown-out colors for super-highlights, by
22849 (smoothly) turning into white instead. This makes images feel more natural,
22850 at the cost of reducing information about out-of-range colors.
22852 The default value is 0.5, and the algorithm here is a little different from
22853 the cpu version tonemap currently. A setting of 0.0 disables this option.
22856 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
22857 is used to detect whether the scene has changed or not. If the distance between
22858 the current frame average brightness and the current running average exceeds
22859 a threshold value, we would re-calculate scene average and peak brightness.
22860 The default value is 0.2.
22863 Specify the output pixel format.
22865 Currently supported formats are:
22872 Set the output color range.
22874 Possible values are:
22880 Default is same as input.
22883 Set the output color primaries.
22885 Possible values are:
22891 Default is same as input.
22894 Set the output transfer characteristics.
22896 Possible values are:
22905 Set the output colorspace matrix.
22907 Possible value are:
22913 Default is same as input.
22917 @subsection Example
22921 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
22923 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
22927 @section unsharp_opencl
22929 Sharpen or blur the input video.
22931 It accepts the following parameters:
22934 @item luma_msize_x, lx
22935 Set the luma matrix horizontal size.
22936 Range is @code{[1, 23]} and default value is @code{5}.
22938 @item luma_msize_y, ly
22939 Set the luma matrix vertical size.
22940 Range is @code{[1, 23]} and default value is @code{5}.
22942 @item luma_amount, la
22943 Set the luma effect strength.
22944 Range is @code{[-10, 10]} and default value is @code{1.0}.
22946 Negative values will blur the input video, while positive values will
22947 sharpen it, a value of zero will disable the effect.
22949 @item chroma_msize_x, cx
22950 Set the chroma matrix horizontal size.
22951 Range is @code{[1, 23]} and default value is @code{5}.
22953 @item chroma_msize_y, cy
22954 Set the chroma matrix vertical size.
22955 Range is @code{[1, 23]} and default value is @code{5}.
22957 @item chroma_amount, ca
22958 Set the chroma effect strength.
22959 Range is @code{[-10, 10]} and default value is @code{0.0}.
22961 Negative values will blur the input video, while positive values will
22962 sharpen it, a value of zero will disable the effect.
22966 All parameters are optional and default to the equivalent of the
22967 string '5:5:1.0:5:5:0.0'.
22969 @subsection Examples
22973 Apply strong luma sharpen effect:
22975 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
22979 Apply a strong blur of both luma and chroma parameters:
22981 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
22985 @section xfade_opencl
22987 Cross fade two videos with custom transition effect by using OpenCL.
22989 It accepts the following options:
22993 Set one of possible transition effects.
22997 Select custom transition effect, the actual transition description
22998 will be picked from source and kernel options.
23010 Default transition is fade.
23014 OpenCL program source file for custom transition.
23017 Set name of kernel to use for custom transition from program source file.
23020 Set duration of video transition.
23023 Set time of start of transition relative to first video.
23026 The program source file must contain a kernel function with the given name,
23027 which will be run once for each plane of the output. Each run on a plane
23028 gets enqueued as a separate 2D global NDRange with one work-item for each
23029 pixel to be generated. The global ID offset for each work-item is therefore
23030 the coordinates of a pixel in the destination image.
23032 The kernel function needs to take the following arguments:
23035 Destination image, @var{__write_only image2d_t}.
23037 This image will become the output; the kernel should write all of it.
23040 First Source image, @var{__read_only image2d_t}.
23041 Second Source image, @var{__read_only image2d_t}.
23043 These are the most recent images on each input. The kernel may read from
23044 them to generate the output, but they can't be written to.
23047 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
23054 Apply dots curtain transition effect:
23056 __kernel void blend_images(__write_only image2d_t dst,
23057 __read_only image2d_t src1,
23058 __read_only image2d_t src2,
23061 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
23062 CLK_FILTER_LINEAR);
23063 int2 p = (int2)(get_global_id(0), get_global_id(1));
23064 float2 rp = (float2)(get_global_id(0), get_global_id(1));
23065 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
23068 float2 dots = (float2)(20.0, 20.0);
23069 float2 center = (float2)(0,0);
23072 float4 val1 = read_imagef(src1, sampler, p);
23073 float4 val2 = read_imagef(src2, sampler, p);
23074 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
23076 write_imagef(dst, p, next ? val1 : val2);
23082 @c man end OPENCL VIDEO FILTERS
23084 @chapter VAAPI Video Filters
23085 @c man begin VAAPI VIDEO FILTERS
23087 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
23089 To enable compilation of these filters you need to configure FFmpeg with
23090 @code{--enable-vaapi}.
23092 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}
23094 @section tonemap_vaapi
23096 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
23097 It maps the dynamic range of HDR10 content to the SDR content.
23098 It currently only accepts HDR10 as input.
23100 It accepts the following parameters:
23104 Specify the output pixel format.
23106 Currently supported formats are:
23115 Set the output color primaries.
23117 Default is same as input.
23120 Set the output transfer characteristics.
23125 Set the output colorspace matrix.
23127 Default is same as input.
23131 @subsection Example
23135 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
23137 tonemap_vaapi=format=p010:t=bt2020-10
23141 @c man end VAAPI VIDEO FILTERS
23143 @chapter Video Sources
23144 @c man begin VIDEO SOURCES
23146 Below is a description of the currently available video sources.
23150 Buffer video frames, and make them available to the filter chain.
23152 This source is mainly intended for a programmatic use, in particular
23153 through the interface defined in @file{libavfilter/buffersrc.h}.
23155 It accepts the following parameters:
23160 Specify the size (width and height) of the buffered video frames. For the
23161 syntax of this option, check the
23162 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23165 The input video width.
23168 The input video height.
23171 A string representing the pixel format of the buffered video frames.
23172 It may be a number corresponding to a pixel format, or a pixel format
23176 Specify the timebase assumed by the timestamps of the buffered frames.
23179 Specify the frame rate expected for the video stream.
23181 @item pixel_aspect, sar
23182 The sample (pixel) aspect ratio of the input video.
23185 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
23186 to the filtergraph description to specify swscale flags for automatically
23187 inserted scalers. See @ref{Filtergraph syntax}.
23189 @item hw_frames_ctx
23190 When using a hardware pixel format, this should be a reference to an
23191 AVHWFramesContext describing input frames.
23196 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
23199 will instruct the source to accept video frames with size 320x240 and
23200 with format "yuv410p", assuming 1/24 as the timestamps timebase and
23201 square pixels (1:1 sample aspect ratio).
23202 Since the pixel format with name "yuv410p" corresponds to the number 6
23203 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
23204 this example corresponds to:
23206 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
23209 Alternatively, the options can be specified as a flat string, but this
23210 syntax is deprecated:
23212 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
23216 Create a pattern generated by an elementary cellular automaton.
23218 The initial state of the cellular automaton can be defined through the
23219 @option{filename} and @option{pattern} options. If such options are
23220 not specified an initial state is created randomly.
23222 At each new frame a new row in the video is filled with the result of
23223 the cellular automaton next generation. The behavior when the whole
23224 frame is filled is defined by the @option{scroll} option.
23226 This source accepts the following options:
23230 Read the initial cellular automaton state, i.e. the starting row, from
23231 the specified file.
23232 In the file, each non-whitespace character is considered an alive
23233 cell, a newline will terminate the row, and further characters in the
23234 file will be ignored.
23237 Read the initial cellular automaton state, i.e. the starting row, from
23238 the specified string.
23240 Each non-whitespace character in the string is considered an alive
23241 cell, a newline will terminate the row, and further characters in the
23242 string will be ignored.
23245 Set the video rate, that is the number of frames generated per second.
23248 @item random_fill_ratio, ratio
23249 Set the random fill ratio for the initial cellular automaton row. It
23250 is a floating point number value ranging from 0 to 1, defaults to
23253 This option is ignored when a file or a pattern is specified.
23255 @item random_seed, seed
23256 Set the seed for filling randomly the initial row, must be an integer
23257 included between 0 and UINT32_MAX. If not specified, or if explicitly
23258 set to -1, the filter will try to use a good random seed on a best
23262 Set the cellular automaton rule, it is a number ranging from 0 to 255.
23263 Default value is 110.
23266 Set the size of the output video. For the syntax of this option, check the
23267 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23269 If @option{filename} or @option{pattern} is specified, the size is set
23270 by default to the width of the specified initial state row, and the
23271 height is set to @var{width} * PHI.
23273 If @option{size} is set, it must contain the width of the specified
23274 pattern string, and the specified pattern will be centered in the
23277 If a filename or a pattern string is not specified, the size value
23278 defaults to "320x518" (used for a randomly generated initial state).
23281 If set to 1, scroll the output upward when all the rows in the output
23282 have been already filled. If set to 0, the new generated row will be
23283 written over the top row just after the bottom row is filled.
23286 @item start_full, full
23287 If set to 1, completely fill the output with generated rows before
23288 outputting the first frame.
23289 This is the default behavior, for disabling set the value to 0.
23292 If set to 1, stitch the left and right row edges together.
23293 This is the default behavior, for disabling set the value to 0.
23296 @subsection Examples
23300 Read the initial state from @file{pattern}, and specify an output of
23303 cellauto=f=pattern:s=200x400
23307 Generate a random initial row with a width of 200 cells, with a fill
23310 cellauto=ratio=2/3:s=200x200
23314 Create a pattern generated by rule 18 starting by a single alive cell
23315 centered on an initial row with width 100:
23317 cellauto=p=@@:s=100x400:full=0:rule=18
23321 Specify a more elaborated initial pattern:
23323 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
23328 @anchor{coreimagesrc}
23329 @section coreimagesrc
23330 Video source generated on GPU using Apple's CoreImage API on OSX.
23332 This video source is a specialized version of the @ref{coreimage} video filter.
23333 Use a core image generator at the beginning of the applied filterchain to
23334 generate the content.
23336 The coreimagesrc video source accepts the following options:
23338 @item list_generators
23339 List all available generators along with all their respective options as well as
23340 possible minimum and maximum values along with the default values.
23342 list_generators=true
23346 Specify the size of the sourced video. For the syntax of this option, check the
23347 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23348 The default value is @code{320x240}.
23351 Specify the frame rate of the sourced video, as the number of frames
23352 generated per second. It has to be a string in the format
23353 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23354 number or a valid video frame rate abbreviation. The default value is
23358 Set the sample aspect ratio of the sourced video.
23361 Set the duration of the sourced video. See
23362 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23363 for the accepted syntax.
23365 If not specified, or the expressed duration is negative, the video is
23366 supposed to be generated forever.
23369 Additionally, all options of the @ref{coreimage} video filter are accepted.
23370 A complete filterchain can be used for further processing of the
23371 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
23372 and examples for details.
23374 @subsection Examples
23379 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
23380 given as complete and escaped command-line for Apple's standard bash shell:
23382 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
23384 This example is equivalent to the QRCode example of @ref{coreimage} without the
23385 need for a nullsrc video source.
23390 Generate several gradients.
23394 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23395 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23398 Set frame rate, expressed as number of frames per second. Default
23401 @item c0, c1, c2, c3, c4, c5, c6, c7
23402 Set 8 colors. Default values for colors is to pick random one.
23404 @item x0, y0, y0, y1
23405 Set gradient line source and destination points. If negative or out of range, random ones
23409 Set number of colors to use at once. Allowed range is from 2 to 8. Default value is 2.
23412 Set seed for picking gradient line points.
23415 Set the duration of the sourced video. See
23416 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23417 for the accepted syntax.
23419 If not specified, or the expressed duration is negative, the video is
23420 supposed to be generated forever.
23423 Set speed of gradients rotation.
23427 @section mandelbrot
23429 Generate a Mandelbrot set fractal, and progressively zoom towards the
23430 point specified with @var{start_x} and @var{start_y}.
23432 This source accepts the following options:
23437 Set the terminal pts value. Default value is 400.
23440 Set the terminal scale value.
23441 Must be a floating point value. Default value is 0.3.
23444 Set the inner coloring mode, that is the algorithm used to draw the
23445 Mandelbrot fractal internal region.
23447 It shall assume one of the following values:
23452 Show time until convergence.
23454 Set color based on point closest to the origin of the iterations.
23459 Default value is @var{mincol}.
23462 Set the bailout value. Default value is 10.0.
23465 Set the maximum of iterations performed by the rendering
23466 algorithm. Default value is 7189.
23469 Set outer coloring mode.
23470 It shall assume one of following values:
23472 @item iteration_count
23473 Set iteration count mode.
23474 @item normalized_iteration_count
23475 set normalized iteration count mode.
23477 Default value is @var{normalized_iteration_count}.
23480 Set frame rate, expressed as number of frames per second. Default
23484 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23485 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23488 Set the initial scale value. Default value is 3.0.
23491 Set the initial x position. Must be a floating point value between
23492 -100 and 100. Default value is -0.743643887037158704752191506114774.
23495 Set the initial y position. Must be a floating point value between
23496 -100 and 100. Default value is -0.131825904205311970493132056385139.
23501 Generate various test patterns, as generated by the MPlayer test filter.
23503 The size of the generated video is fixed, and is 256x256.
23504 This source is useful in particular for testing encoding features.
23506 This source accepts the following options:
23511 Specify the frame rate of the sourced video, as the number of frames
23512 generated per second. It has to be a string in the format
23513 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23514 number or a valid video frame rate abbreviation. The default value is
23518 Set the duration of the sourced video. See
23519 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23520 for the accepted syntax.
23522 If not specified, or the expressed duration is negative, the video is
23523 supposed to be generated forever.
23527 Set the number or the name of the test to perform. Supported tests are:
23541 @item max_frames, m
23542 Set the maximum number of frames generated for each test, default value is 30.
23546 Default value is "all", which will cycle through the list of all tests.
23551 mptestsrc=t=dc_luma
23554 will generate a "dc_luma" test pattern.
23556 @section frei0r_src
23558 Provide a frei0r source.
23560 To enable compilation of this filter you need to install the frei0r
23561 header and configure FFmpeg with @code{--enable-frei0r}.
23563 This source accepts the following parameters:
23568 The size of the video to generate. For the syntax of this option, check the
23569 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23572 The framerate of the generated video. It may be a string of the form
23573 @var{num}/@var{den} or a frame rate abbreviation.
23576 The name to the frei0r source to load. For more information regarding frei0r and
23577 how to set the parameters, read the @ref{frei0r} section in the video filters
23580 @item filter_params
23581 A '|'-separated list of parameters to pass to the frei0r source.
23585 For example, to generate a frei0r partik0l source with size 200x200
23586 and frame rate 10 which is overlaid on the overlay filter main input:
23588 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
23593 Generate a life pattern.
23595 This source is based on a generalization of John Conway's life game.
23597 The sourced input represents a life grid, each pixel represents a cell
23598 which can be in one of two possible states, alive or dead. Every cell
23599 interacts with its eight neighbours, which are the cells that are
23600 horizontally, vertically, or diagonally adjacent.
23602 At each interaction the grid evolves according to the adopted rule,
23603 which specifies the number of neighbor alive cells which will make a
23604 cell stay alive or born. The @option{rule} option allows one to specify
23607 This source accepts the following options:
23611 Set the file from which to read the initial grid state. In the file,
23612 each non-whitespace character is considered an alive cell, and newline
23613 is used to delimit the end of each row.
23615 If this option is not specified, the initial grid is generated
23619 Set the video rate, that is the number of frames generated per second.
23622 @item random_fill_ratio, ratio
23623 Set the random fill ratio for the initial random grid. It is a
23624 floating point number value ranging from 0 to 1, defaults to 1/PHI.
23625 It is ignored when a file is specified.
23627 @item random_seed, seed
23628 Set the seed for filling the initial random grid, must be an integer
23629 included between 0 and UINT32_MAX. If not specified, or if explicitly
23630 set to -1, the filter will try to use a good random seed on a best
23636 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
23637 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
23638 @var{NS} specifies the number of alive neighbor cells which make a
23639 live cell stay alive, and @var{NB} the number of alive neighbor cells
23640 which make a dead cell to become alive (i.e. to "born").
23641 "s" and "b" can be used in place of "S" and "B", respectively.
23643 Alternatively a rule can be specified by an 18-bits integer. The 9
23644 high order bits are used to encode the next cell state if it is alive
23645 for each number of neighbor alive cells, the low order bits specify
23646 the rule for "borning" new cells. Higher order bits encode for an
23647 higher number of neighbor cells.
23648 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
23649 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
23651 Default value is "S23/B3", which is the original Conway's game of life
23652 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
23653 cells, and will born a new cell if there are three alive cells around
23657 Set the size of the output video. For the syntax of this option, check the
23658 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23660 If @option{filename} is specified, the size is set by default to the
23661 same size of the input file. If @option{size} is set, it must contain
23662 the size specified in the input file, and the initial grid defined in
23663 that file is centered in the larger resulting area.
23665 If a filename is not specified, the size value defaults to "320x240"
23666 (used for a randomly generated initial grid).
23669 If set to 1, stitch the left and right grid edges together, and the
23670 top and bottom edges also. Defaults to 1.
23673 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
23674 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
23675 value from 0 to 255.
23678 Set the color of living (or new born) cells.
23681 Set the color of dead cells. If @option{mold} is set, this is the first color
23682 used to represent a dead cell.
23685 Set mold color, for definitely dead and moldy cells.
23687 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
23688 ffmpeg-utils manual,ffmpeg-utils}.
23691 @subsection Examples
23695 Read a grid from @file{pattern}, and center it on a grid of size
23698 life=f=pattern:s=300x300
23702 Generate a random grid of size 200x200, with a fill ratio of 2/3:
23704 life=ratio=2/3:s=200x200
23708 Specify a custom rule for evolving a randomly generated grid:
23714 Full example with slow death effect (mold) using @command{ffplay}:
23716 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
23723 @anchor{haldclutsrc}
23726 @anchor{pal100bars}
23727 @anchor{rgbtestsrc}
23729 @anchor{smptehdbars}
23732 @anchor{yuvtestsrc}
23733 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
23735 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
23737 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
23739 The @code{color} source provides an uniformly colored input.
23741 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
23742 @ref{haldclut} filter.
23744 The @code{nullsrc} source returns unprocessed video frames. It is
23745 mainly useful to be employed in analysis / debugging tools, or as the
23746 source for filters which ignore the input data.
23748 The @code{pal75bars} source generates a color bars pattern, based on
23749 EBU PAL recommendations with 75% color levels.
23751 The @code{pal100bars} source generates a color bars pattern, based on
23752 EBU PAL recommendations with 100% color levels.
23754 The @code{rgbtestsrc} source generates an RGB test pattern useful for
23755 detecting RGB vs BGR issues. You should see a red, green and blue
23756 stripe from top to bottom.
23758 The @code{smptebars} source generates a color bars pattern, based on
23759 the SMPTE Engineering Guideline EG 1-1990.
23761 The @code{smptehdbars} source generates a color bars pattern, based on
23762 the SMPTE RP 219-2002.
23764 The @code{testsrc} source generates a test video pattern, showing a
23765 color pattern, a scrolling gradient and a timestamp. This is mainly
23766 intended for testing purposes.
23768 The @code{testsrc2} source is similar to testsrc, but supports more
23769 pixel formats instead of just @code{rgb24}. This allows using it as an
23770 input for other tests without requiring a format conversion.
23772 The @code{yuvtestsrc} source generates an YUV test pattern. You should
23773 see a y, cb and cr stripe from top to bottom.
23775 The sources accept the following parameters:
23780 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
23781 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
23782 pixels to be used as identity matrix for 3D lookup tables. Each component is
23783 coded on a @code{1/(N*N)} scale.
23786 Specify the color of the source, only available in the @code{color}
23787 source. For the syntax of this option, check the
23788 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
23791 Specify the size of the sourced video. For the syntax of this option, check the
23792 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23793 The default value is @code{320x240}.
23795 This option is not available with the @code{allrgb}, @code{allyuv}, and
23796 @code{haldclutsrc} filters.
23799 Specify the frame rate of the sourced video, as the number of frames
23800 generated per second. It has to be a string in the format
23801 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23802 number or a valid video frame rate abbreviation. The default value is
23806 Set the duration of the sourced video. See
23807 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23808 for the accepted syntax.
23810 If not specified, or the expressed duration is negative, the video is
23811 supposed to be generated forever.
23813 Since the frame rate is used as time base, all frames including the last one
23814 will have their full duration. If the specified duration is not a multiple
23815 of the frame duration, it will be rounded up.
23818 Set the sample aspect ratio of the sourced video.
23821 Specify the alpha (opacity) of the background, only available in the
23822 @code{testsrc2} source. The value must be between 0 (fully transparent) and
23823 255 (fully opaque, the default).
23826 Set the number of decimals to show in the timestamp, only available in the
23827 @code{testsrc} source.
23829 The displayed timestamp value will correspond to the original
23830 timestamp value multiplied by the power of 10 of the specified
23831 value. Default value is 0.
23834 @subsection Examples
23838 Generate a video with a duration of 5.3 seconds, with size
23839 176x144 and a frame rate of 10 frames per second:
23841 testsrc=duration=5.3:size=qcif:rate=10
23845 The following graph description will generate a red source
23846 with an opacity of 0.2, with size "qcif" and a frame rate of 10
23849 color=c=red@@0.2:s=qcif:r=10
23853 If the input content is to be ignored, @code{nullsrc} can be used. The
23854 following command generates noise in the luminance plane by employing
23855 the @code{geq} filter:
23857 nullsrc=s=256x256, geq=random(1)*255:128:128
23861 @subsection Commands
23863 The @code{color} source supports the following commands:
23867 Set the color of the created image. Accepts the same syntax of the
23868 corresponding @option{color} option.
23873 Generate video using an OpenCL program.
23878 OpenCL program source file.
23881 Kernel name in program.
23884 Size of frames to generate. This must be set.
23887 Pixel format to use for the generated frames. This must be set.
23890 Number of frames generated every second. Default value is '25'.
23894 For details of how the program loading works, see the @ref{program_opencl}
23901 Generate a colour ramp by setting pixel values from the position of the pixel
23902 in the output image. (Note that this will work with all pixel formats, but
23903 the generated output will not be the same.)
23905 __kernel void ramp(__write_only image2d_t dst,
23906 unsigned int index)
23908 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23911 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
23913 write_imagef(dst, loc, val);
23918 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
23920 __kernel void sierpinski_carpet(__write_only image2d_t dst,
23921 unsigned int index)
23923 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23925 float4 value = 0.0f;
23926 int x = loc.x + index;
23927 int y = loc.y + index;
23928 while (x > 0 || y > 0) {
23929 if (x % 3 == 1 && y % 3 == 1) {
23937 write_imagef(dst, loc, value);
23943 @section sierpinski
23945 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
23947 This source accepts the following options:
23951 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23952 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23955 Set frame rate, expressed as number of frames per second. Default
23959 Set seed which is used for random panning.
23962 Set max jump for single pan destination. Allowed range is from 1 to 10000.
23965 Set fractal type, can be default @code{carpet} or @code{triangle}.
23968 @c man end VIDEO SOURCES
23970 @chapter Video Sinks
23971 @c man begin VIDEO SINKS
23973 Below is a description of the currently available video sinks.
23975 @section buffersink
23977 Buffer video frames, and make them available to the end of the filter
23980 This sink is mainly intended for programmatic use, in particular
23981 through the interface defined in @file{libavfilter/buffersink.h}
23982 or the options system.
23984 It accepts a pointer to an AVBufferSinkContext structure, which
23985 defines the incoming buffers' formats, to be passed as the opaque
23986 parameter to @code{avfilter_init_filter} for initialization.
23990 Null video sink: do absolutely nothing with the input video. It is
23991 mainly useful as a template and for use in analysis / debugging
23994 @c man end VIDEO SINKS
23996 @chapter Multimedia Filters
23997 @c man begin MULTIMEDIA FILTERS
23999 Below is a description of the currently available multimedia filters.
24003 Convert input audio to a video output, displaying the audio bit scope.
24005 The filter accepts the following options:
24009 Set frame rate, expressed as number of frames per second. Default
24013 Specify the video size for the output. For the syntax of this option, check the
24014 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24015 Default value is @code{1024x256}.
24018 Specify list of colors separated by space or by '|' which will be used to
24019 draw channels. Unrecognized or missing colors will be replaced
24023 @section adrawgraph
24024 Draw a graph using input audio metadata.
24026 See @ref{drawgraph}
24028 @section agraphmonitor
24030 See @ref{graphmonitor}.
24032 @section ahistogram
24034 Convert input audio to a video output, displaying the volume histogram.
24036 The filter accepts the following options:
24040 Specify how histogram is calculated.
24042 It accepts the following values:
24045 Use single histogram for all channels.
24047 Use separate histogram for each channel.
24049 Default is @code{single}.
24052 Set frame rate, expressed as number of frames per second. Default
24056 Specify the video size for the output. For the syntax of this option, check the
24057 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24058 Default value is @code{hd720}.
24063 It accepts the following values:
24074 reverse logarithmic
24076 Default is @code{log}.
24079 Set amplitude scale.
24081 It accepts the following values:
24088 Default is @code{log}.
24091 Set how much frames to accumulate in histogram.
24092 Default is 1. Setting this to -1 accumulates all frames.
24095 Set histogram ratio of window height.
24098 Set sonogram sliding.
24100 It accepts the following values:
24103 replace old rows with new ones.
24105 scroll from top to bottom.
24107 Default is @code{replace}.
24110 @section aphasemeter
24112 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
24113 representing mean phase of current audio frame. A video output can also be produced and is
24114 enabled by default. The audio is passed through as first output.
24116 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
24117 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
24118 and @code{1} means channels are in phase.
24120 The filter accepts the following options, all related to its video output:
24124 Set the output frame rate. Default value is @code{25}.
24127 Set the video size for the output. For the syntax of this option, check the
24128 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24129 Default value is @code{800x400}.
24134 Specify the red, green, blue contrast. Default values are @code{2},
24135 @code{7} and @code{1}.
24136 Allowed range is @code{[0, 255]}.
24139 Set color which will be used for drawing median phase. If color is
24140 @code{none} which is default, no median phase value will be drawn.
24143 Enable video output. Default is enabled.
24146 @subsection phasing detection
24148 The filter also detects out of phase and mono sequences in stereo streams.
24149 It logs the sequence start, end and duration when it lasts longer or as long as the minimum set.
24151 The filter accepts the following options for this detection:
24155 Enable mono and out of phase detection. Default is disabled.
24158 Set phase tolerance for mono detection, in amplitude ratio. Default is @code{0}.
24159 Allowed range is @code{[0, 1]}.
24162 Set angle threshold for out of phase detection, in degree. Default is @code{170}.
24163 Allowed range is @code{[90, 180]}.
24166 Set mono or out of phase duration until notification, expressed in seconds. Default is @code{2}.
24169 @subsection Examples
24173 Complete example with @command{ffmpeg} to detect 1 second of mono with 0.001 phase tolerance:
24175 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
24179 @section avectorscope
24181 Convert input audio to a video output, representing the audio vector
24184 The filter is used to measure the difference between channels of stereo
24185 audio stream. A monaural signal, consisting of identical left and right
24186 signal, results in straight vertical line. Any stereo separation is visible
24187 as a deviation from this line, creating a Lissajous figure.
24188 If the straight (or deviation from it) but horizontal line appears this
24189 indicates that the left and right channels are out of phase.
24191 The filter accepts the following options:
24195 Set the vectorscope mode.
24197 Available values are:
24200 Lissajous rotated by 45 degrees.
24203 Same as above but not rotated.
24206 Shape resembling half of circle.
24209 Default value is @samp{lissajous}.
24212 Set the video size for the output. For the syntax of this option, check the
24213 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24214 Default value is @code{400x400}.
24217 Set the output frame rate. Default value is @code{25}.
24223 Specify the red, green, blue and alpha contrast. Default values are @code{40},
24224 @code{160}, @code{80} and @code{255}.
24225 Allowed range is @code{[0, 255]}.
24231 Specify the red, green, blue and alpha fade. Default values are @code{15},
24232 @code{10}, @code{5} and @code{5}.
24233 Allowed range is @code{[0, 255]}.
24236 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
24237 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
24240 Set the vectorscope drawing mode.
24242 Available values are:
24245 Draw dot for each sample.
24248 Draw line between previous and current sample.
24251 Default value is @samp{dot}.
24254 Specify amplitude scale of audio samples.
24256 Available values are:
24272 Swap left channel axis with right channel axis.
24282 Mirror only x axis.
24285 Mirror only y axis.
24293 @subsection Examples
24297 Complete example using @command{ffplay}:
24299 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
24300 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
24304 @section bench, abench
24306 Benchmark part of a filtergraph.
24308 The filter accepts the following options:
24312 Start or stop a timer.
24314 Available values are:
24317 Get the current time, set it as frame metadata (using the key
24318 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
24321 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
24322 the input frame metadata to get the time difference. Time difference, average,
24323 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
24324 @code{min}) are then printed. The timestamps are expressed in seconds.
24328 @subsection Examples
24332 Benchmark @ref{selectivecolor} filter:
24334 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
24340 Concatenate audio and video streams, joining them together one after the
24343 The filter works on segments of synchronized video and audio streams. All
24344 segments must have the same number of streams of each type, and that will
24345 also be the number of streams at output.
24347 The filter accepts the following options:
24352 Set the number of segments. Default is 2.
24355 Set the number of output video streams, that is also the number of video
24356 streams in each segment. Default is 1.
24359 Set the number of output audio streams, that is also the number of audio
24360 streams in each segment. Default is 0.
24363 Activate unsafe mode: do not fail if segments have a different format.
24367 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
24368 @var{a} audio outputs.
24370 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
24371 segment, in the same order as the outputs, then the inputs for the second
24374 Related streams do not always have exactly the same duration, for various
24375 reasons including codec frame size or sloppy authoring. For that reason,
24376 related synchronized streams (e.g. a video and its audio track) should be
24377 concatenated at once. The concat filter will use the duration of the longest
24378 stream in each segment (except the last one), and if necessary pad shorter
24379 audio streams with silence.
24381 For this filter to work correctly, all segments must start at timestamp 0.
24383 All corresponding streams must have the same parameters in all segments; the
24384 filtering system will automatically select a common pixel format for video
24385 streams, and a common sample format, sample rate and channel layout for
24386 audio streams, but other settings, such as resolution, must be converted
24387 explicitly by the user.
24389 Different frame rates are acceptable but will result in variable frame rate
24390 at output; be sure to configure the output file to handle it.
24392 @subsection Examples
24396 Concatenate an opening, an episode and an ending, all in bilingual version
24397 (video in stream 0, audio in streams 1 and 2):
24399 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
24400 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
24401 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
24402 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
24406 Concatenate two parts, handling audio and video separately, using the
24407 (a)movie sources, and adjusting the resolution:
24409 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
24410 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
24411 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
24413 Note that a desync will happen at the stitch if the audio and video streams
24414 do not have exactly the same duration in the first file.
24418 @subsection Commands
24420 This filter supports the following commands:
24423 Close the current segment and step to the next one
24429 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
24430 level. By default, it logs a message at a frequency of 10Hz with the
24431 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
24432 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
24434 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
24435 sample format is double-precision floating point. The input stream will be converted to
24436 this specification, if needed. Users may need to insert aformat and/or aresample filters
24437 after this filter to obtain the original parameters.
24439 The filter also has a video output (see the @var{video} option) with a real
24440 time graph to observe the loudness evolution. The graphic contains the logged
24441 message mentioned above, so it is not printed anymore when this option is set,
24442 unless the verbose logging is set. The main graphing area contains the
24443 short-term loudness (3 seconds of analysis), and the gauge on the right is for
24444 the momentary loudness (400 milliseconds), but can optionally be configured
24445 to instead display short-term loudness (see @var{gauge}).
24447 The green area marks a +/- 1LU target range around the target loudness
24448 (-23LUFS by default, unless modified through @var{target}).
24450 More information about the Loudness Recommendation EBU R128 on
24451 @url{http://tech.ebu.ch/loudness}.
24453 The filter accepts the following options:
24458 Activate the video output. The audio stream is passed unchanged whether this
24459 option is set or no. The video stream will be the first output stream if
24460 activated. Default is @code{0}.
24463 Set the video size. This option is for video only. For the syntax of this
24465 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24466 Default and minimum resolution is @code{640x480}.
24469 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
24470 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
24471 other integer value between this range is allowed.
24474 Set metadata injection. If set to @code{1}, the audio input will be segmented
24475 into 100ms output frames, each of them containing various loudness information
24476 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
24478 Default is @code{0}.
24481 Force the frame logging level.
24483 Available values are:
24486 information logging level
24488 verbose logging level
24491 By default, the logging level is set to @var{info}. If the @option{video} or
24492 the @option{metadata} options are set, it switches to @var{verbose}.
24497 Available modes can be cumulated (the option is a @code{flag} type). Possible
24501 Disable any peak mode (default).
24503 Enable sample-peak mode.
24505 Simple peak mode looking for the higher sample value. It logs a message
24506 for sample-peak (identified by @code{SPK}).
24508 Enable true-peak mode.
24510 If enabled, the peak lookup is done on an over-sampled version of the input
24511 stream for better peak accuracy. It logs a message for true-peak.
24512 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
24513 This mode requires a build with @code{libswresample}.
24517 Treat mono input files as "dual mono". If a mono file is intended for playback
24518 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
24519 If set to @code{true}, this option will compensate for this effect.
24520 Multi-channel input files are not affected by this option.
24523 Set a specific pan law to be used for the measurement of dual mono files.
24524 This parameter is optional, and has a default value of -3.01dB.
24527 Set a specific target level (in LUFS) used as relative zero in the visualization.
24528 This parameter is optional and has a default value of -23LUFS as specified
24529 by EBU R128. However, material published online may prefer a level of -16LUFS
24530 (e.g. for use with podcasts or video platforms).
24533 Set the value displayed by the gauge. Valid values are @code{momentary} and s
24534 @code{shortterm}. By default the momentary value will be used, but in certain
24535 scenarios it may be more useful to observe the short term value instead (e.g.
24539 Sets the display scale for the loudness. Valid parameters are @code{absolute}
24540 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
24541 video output, not the summary or continuous log output.
24544 @subsection Examples
24548 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
24550 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
24554 Run an analysis with @command{ffmpeg}:
24556 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
24560 @section interleave, ainterleave
24562 Temporally interleave frames from several inputs.
24564 @code{interleave} works with video inputs, @code{ainterleave} with audio.
24566 These filters read frames from several inputs and send the oldest
24567 queued frame to the output.
24569 Input streams must have well defined, monotonically increasing frame
24572 In order to submit one frame to output, these filters need to enqueue
24573 at least one frame for each input, so they cannot work in case one
24574 input is not yet terminated and will not receive incoming frames.
24576 For example consider the case when one input is a @code{select} filter
24577 which always drops input frames. The @code{interleave} filter will keep
24578 reading from that input, but it will never be able to send new frames
24579 to output until the input sends an end-of-stream signal.
24581 Also, depending on inputs synchronization, the filters will drop
24582 frames in case one input receives more frames than the other ones, and
24583 the queue is already filled.
24585 These filters accept the following options:
24589 Set the number of different inputs, it is 2 by default.
24592 How to determine the end-of-stream.
24596 The duration of the longest input. (default)
24599 The duration of the shortest input.
24602 The duration of the first input.
24607 @subsection Examples
24611 Interleave frames belonging to different streams using @command{ffmpeg}:
24613 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
24617 Add flickering blur effect:
24619 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
24623 @section metadata, ametadata
24625 Manipulate frame metadata.
24627 This filter accepts the following options:
24631 Set mode of operation of the filter.
24633 Can be one of the following:
24637 If both @code{value} and @code{key} is set, select frames
24638 which have such metadata. If only @code{key} is set, select
24639 every frame that has such key in metadata.
24642 Add new metadata @code{key} and @code{value}. If key is already available
24646 Modify value of already present key.
24649 If @code{value} is set, delete only keys that have such value.
24650 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
24654 Print key and its value if metadata was found. If @code{key} is not set print all
24655 metadata values available in frame.
24659 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
24662 Set metadata value which will be used. This option is mandatory for
24663 @code{modify} and @code{add} mode.
24666 Which function to use when comparing metadata value and @code{value}.
24668 Can be one of following:
24672 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
24675 Values are interpreted as strings, returns true if metadata value starts with
24676 the @code{value} option string.
24679 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
24682 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
24685 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
24688 Values are interpreted as floats, returns true if expression from option @code{expr}
24692 Values are interpreted as strings, returns true if metadata value ends with
24693 the @code{value} option string.
24697 Set expression which is used when @code{function} is set to @code{expr}.
24698 The expression is evaluated through the eval API and can contain the following
24703 Float representation of @code{value} from metadata key.
24706 Float representation of @code{value} as supplied by user in @code{value} option.
24710 If specified in @code{print} mode, output is written to the named file. Instead of
24711 plain filename any writable url can be specified. Filename ``-'' is a shorthand
24712 for standard output. If @code{file} option is not set, output is written to the log
24713 with AV_LOG_INFO loglevel.
24716 Reduces buffering in print mode when output is written to a URL set using @var{file}.
24720 @subsection Examples
24724 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
24727 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
24730 Print silencedetect output to file @file{metadata.txt}.
24732 silencedetect,ametadata=mode=print:file=metadata.txt
24735 Direct all metadata to a pipe with file descriptor 4.
24737 metadata=mode=print:file='pipe\:4'
24741 @section perms, aperms
24743 Set read/write permissions for the output frames.
24745 These filters are mainly aimed at developers to test direct path in the
24746 following filter in the filtergraph.
24748 The filters accept the following options:
24752 Select the permissions mode.
24754 It accepts the following values:
24757 Do nothing. This is the default.
24759 Set all the output frames read-only.
24761 Set all the output frames directly writable.
24763 Make the frame read-only if writable, and writable if read-only.
24765 Set each output frame read-only or writable randomly.
24769 Set the seed for the @var{random} mode, must be an integer included between
24770 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
24771 @code{-1}, the filter will try to use a good random seed on a best effort
24775 Note: in case of auto-inserted filter between the permission filter and the
24776 following one, the permission might not be received as expected in that
24777 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
24778 perms/aperms filter can avoid this problem.
24780 @section realtime, arealtime
24782 Slow down filtering to match real time approximately.
24784 These filters will pause the filtering for a variable amount of time to
24785 match the output rate with the input timestamps.
24786 They are similar to the @option{re} option to @code{ffmpeg}.
24788 They accept the following options:
24792 Time limit for the pauses. Any pause longer than that will be considered
24793 a timestamp discontinuity and reset the timer. Default is 2 seconds.
24795 Speed factor for processing. The value must be a float larger than zero.
24796 Values larger than 1.0 will result in faster than realtime processing,
24797 smaller will slow processing down. The @var{limit} is automatically adapted
24798 accordingly. Default is 1.0.
24800 A processing speed faster than what is possible without these filters cannot
24805 @section select, aselect
24807 Select frames to pass in output.
24809 This filter accepts the following options:
24814 Set expression, which is evaluated for each input frame.
24816 If the expression is evaluated to zero, the frame is discarded.
24818 If the evaluation result is negative or NaN, the frame is sent to the
24819 first output; otherwise it is sent to the output with index
24820 @code{ceil(val)-1}, assuming that the input index starts from 0.
24822 For example a value of @code{1.2} corresponds to the output with index
24823 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
24826 Set the number of outputs. The output to which to send the selected
24827 frame is based on the result of the evaluation. Default value is 1.
24830 The expression can contain the following constants:
24834 The (sequential) number of the filtered frame, starting from 0.
24837 The (sequential) number of the selected frame, starting from 0.
24839 @item prev_selected_n
24840 The sequential number of the last selected frame. It's NAN if undefined.
24843 The timebase of the input timestamps.
24846 The PTS (Presentation TimeStamp) of the filtered video frame,
24847 expressed in @var{TB} units. It's NAN if undefined.
24850 The PTS of the filtered video frame,
24851 expressed in seconds. It's NAN if undefined.
24854 The PTS of the previously filtered video frame. It's NAN if undefined.
24856 @item prev_selected_pts
24857 The PTS of the last previously filtered video frame. It's NAN if undefined.
24859 @item prev_selected_t
24860 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
24863 The PTS of the first video frame in the video. It's NAN if undefined.
24866 The time of the first video frame in the video. It's NAN if undefined.
24868 @item pict_type @emph{(video only)}
24869 The type of the filtered frame. It can assume one of the following
24881 @item interlace_type @emph{(video only)}
24882 The frame interlace type. It can assume one of the following values:
24885 The frame is progressive (not interlaced).
24887 The frame is top-field-first.
24889 The frame is bottom-field-first.
24892 @item consumed_sample_n @emph{(audio only)}
24893 the number of selected samples before the current frame
24895 @item samples_n @emph{(audio only)}
24896 the number of samples in the current frame
24898 @item sample_rate @emph{(audio only)}
24899 the input sample rate
24902 This is 1 if the filtered frame is a key-frame, 0 otherwise.
24905 the position in the file of the filtered frame, -1 if the information
24906 is not available (e.g. for synthetic video)
24908 @item scene @emph{(video only)}
24909 value between 0 and 1 to indicate a new scene; a low value reflects a low
24910 probability for the current frame to introduce a new scene, while a higher
24911 value means the current frame is more likely to be one (see the example below)
24913 @item concatdec_select
24914 The concat demuxer can select only part of a concat input file by setting an
24915 inpoint and an outpoint, but the output packets may not be entirely contained
24916 in the selected interval. By using this variable, it is possible to skip frames
24917 generated by the concat demuxer which are not exactly contained in the selected
24920 This works by comparing the frame pts against the @var{lavf.concat.start_time}
24921 and the @var{lavf.concat.duration} packet metadata values which are also
24922 present in the decoded frames.
24924 The @var{concatdec_select} variable is -1 if the frame pts is at least
24925 start_time and either the duration metadata is missing or the frame pts is less
24926 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
24929 That basically means that an input frame is selected if its pts is within the
24930 interval set by the concat demuxer.
24934 The default value of the select expression is "1".
24936 @subsection Examples
24940 Select all frames in input:
24945 The example above is the same as:
24957 Select only I-frames:
24959 select='eq(pict_type\,I)'
24963 Select one frame every 100:
24965 select='not(mod(n\,100))'
24969 Select only frames contained in the 10-20 time interval:
24971 select=between(t\,10\,20)
24975 Select only I-frames contained in the 10-20 time interval:
24977 select=between(t\,10\,20)*eq(pict_type\,I)
24981 Select frames with a minimum distance of 10 seconds:
24983 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
24987 Use aselect to select only audio frames with samples number > 100:
24989 aselect='gt(samples_n\,100)'
24993 Create a mosaic of the first scenes:
24995 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
24998 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
25002 Send even and odd frames to separate outputs, and compose them:
25004 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
25008 Select useful frames from an ffconcat file which is using inpoints and
25009 outpoints but where the source files are not intra frame only.
25011 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
25015 @section sendcmd, asendcmd
25017 Send commands to filters in the filtergraph.
25019 These filters read commands to be sent to other filters in the
25022 @code{sendcmd} must be inserted between two video filters,
25023 @code{asendcmd} must be inserted between two audio filters, but apart
25024 from that they act the same way.
25026 The specification of commands can be provided in the filter arguments
25027 with the @var{commands} option, or in a file specified by the
25028 @var{filename} option.
25030 These filters accept the following options:
25033 Set the commands to be read and sent to the other filters.
25035 Set the filename of the commands to be read and sent to the other
25039 @subsection Commands syntax
25041 A commands description consists of a sequence of interval
25042 specifications, comprising a list of commands to be executed when a
25043 particular event related to that interval occurs. The occurring event
25044 is typically the current frame time entering or leaving a given time
25047 An interval is specified by the following syntax:
25049 @var{START}[-@var{END}] @var{COMMANDS};
25052 The time interval is specified by the @var{START} and @var{END} times.
25053 @var{END} is optional and defaults to the maximum time.
25055 The current frame time is considered within the specified interval if
25056 it is included in the interval [@var{START}, @var{END}), that is when
25057 the time is greater or equal to @var{START} and is lesser than
25060 @var{COMMANDS} consists of a sequence of one or more command
25061 specifications, separated by ",", relating to that interval. The
25062 syntax of a command specification is given by:
25064 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
25067 @var{FLAGS} is optional and specifies the type of events relating to
25068 the time interval which enable sending the specified command, and must
25069 be a non-null sequence of identifier flags separated by "+" or "|" and
25070 enclosed between "[" and "]".
25072 The following flags are recognized:
25075 The command is sent when the current frame timestamp enters the
25076 specified interval. In other words, the command is sent when the
25077 previous frame timestamp was not in the given interval, and the
25081 The command is sent when the current frame timestamp leaves the
25082 specified interval. In other words, the command is sent when the
25083 previous frame timestamp was in the given interval, and the
25087 The command @var{ARG} is interpreted as expression and result of
25088 expression is passed as @var{ARG}.
25090 The expression is evaluated through the eval API and can contain the following
25095 Original position in the file of the frame, or undefined if undefined
25096 for the current frame.
25099 The presentation timestamp in input.
25102 The count of the input frame for video or audio, starting from 0.
25105 The time in seconds of the current frame.
25108 The start time in seconds of the current command interval.
25111 The end time in seconds of the current command interval.
25114 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
25119 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
25122 @var{TARGET} specifies the target of the command, usually the name of
25123 the filter class or a specific filter instance name.
25125 @var{COMMAND} specifies the name of the command for the target filter.
25127 @var{ARG} is optional and specifies the optional list of argument for
25128 the given @var{COMMAND}.
25130 Between one interval specification and another, whitespaces, or
25131 sequences of characters starting with @code{#} until the end of line,
25132 are ignored and can be used to annotate comments.
25134 A simplified BNF description of the commands specification syntax
25137 @var{COMMAND_FLAG} ::= "enter" | "leave"
25138 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
25139 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
25140 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
25141 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
25142 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
25145 @subsection Examples
25149 Specify audio tempo change at second 4:
25151 asendcmd=c='4.0 atempo tempo 1.5',atempo
25155 Target a specific filter instance:
25157 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
25161 Specify a list of drawtext and hue commands in a file.
25163 # show text in the interval 5-10
25164 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
25165 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
25167 # desaturate the image in the interval 15-20
25168 15.0-20.0 [enter] hue s 0,
25169 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
25171 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
25173 # apply an exponential saturation fade-out effect, starting from time 25
25174 25 [enter] hue s exp(25-t)
25177 A filtergraph allowing to read and process the above command list
25178 stored in a file @file{test.cmd}, can be specified with:
25180 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
25185 @section setpts, asetpts
25187 Change the PTS (presentation timestamp) of the input frames.
25189 @code{setpts} works on video frames, @code{asetpts} on audio frames.
25191 This filter accepts the following options:
25196 The expression which is evaluated for each frame to construct its timestamp.
25200 The expression is evaluated through the eval API and can contain the following
25204 @item FRAME_RATE, FR
25205 frame rate, only defined for constant frame-rate video
25208 The presentation timestamp in input
25211 The count of the input frame for video or the number of consumed samples,
25212 not including the current frame for audio, starting from 0.
25214 @item NB_CONSUMED_SAMPLES
25215 The number of consumed samples, not including the current frame (only
25218 @item NB_SAMPLES, S
25219 The number of samples in the current frame (only audio)
25221 @item SAMPLE_RATE, SR
25222 The audio sample rate.
25225 The PTS of the first frame.
25228 the time in seconds of the first frame
25231 State whether the current frame is interlaced.
25234 the time in seconds of the current frame
25237 original position in the file of the frame, or undefined if undefined
25238 for the current frame
25241 The previous input PTS.
25244 previous input time in seconds
25247 The previous output PTS.
25250 previous output time in seconds
25253 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
25257 The wallclock (RTC) time at the start of the movie in microseconds.
25260 The timebase of the input timestamps.
25264 @subsection Examples
25268 Start counting PTS from zero
25270 setpts=PTS-STARTPTS
25274 Apply fast motion effect:
25280 Apply slow motion effect:
25286 Set fixed rate of 25 frames per second:
25292 Set fixed rate 25 fps with some jitter:
25294 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
25298 Apply an offset of 10 seconds to the input PTS:
25304 Generate timestamps from a "live source" and rebase onto the current timebase:
25306 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
25310 Generate timestamps by counting samples:
25319 Force color range for the output video frame.
25321 The @code{setrange} filter marks the color range property for the
25322 output frames. It does not change the input frame, but only sets the
25323 corresponding property, which affects how the frame is treated by
25326 The filter accepts the following options:
25331 Available values are:
25335 Keep the same color range property.
25337 @item unspecified, unknown
25338 Set the color range as unspecified.
25340 @item limited, tv, mpeg
25341 Set the color range as limited.
25343 @item full, pc, jpeg
25344 Set the color range as full.
25348 @section settb, asettb
25350 Set the timebase to use for the output frames timestamps.
25351 It is mainly useful for testing timebase configuration.
25353 It accepts the following parameters:
25358 The expression which is evaluated into the output timebase.
25362 The value for @option{tb} is an arithmetic expression representing a
25363 rational. The expression can contain the constants "AVTB" (the default
25364 timebase), "intb" (the input timebase) and "sr" (the sample rate,
25365 audio only). Default value is "intb".
25367 @subsection Examples
25371 Set the timebase to 1/25:
25377 Set the timebase to 1/10:
25383 Set the timebase to 1001/1000:
25389 Set the timebase to 2*intb:
25395 Set the default timebase value:
25402 Convert input audio to a video output representing frequency spectrum
25403 logarithmically using Brown-Puckette constant Q transform algorithm with
25404 direct frequency domain coefficient calculation (but the transform itself
25405 is not really constant Q, instead the Q factor is actually variable/clamped),
25406 with musical tone scale, from E0 to D#10.
25408 The filter accepts the following options:
25412 Specify the video size for the output. It must be even. For the syntax of this option,
25413 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25414 Default value is @code{1920x1080}.
25417 Set the output frame rate. Default value is @code{25}.
25420 Set the bargraph height. It must be even. Default value is @code{-1} which
25421 computes the bargraph height automatically.
25424 Set the axis height. It must be even. Default value is @code{-1} which computes
25425 the axis height automatically.
25428 Set the sonogram height. It must be even. Default value is @code{-1} which
25429 computes the sonogram height automatically.
25432 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
25433 instead. Default value is @code{1}.
25435 @item sono_v, volume
25436 Specify the sonogram volume expression. It can contain variables:
25439 the @var{bar_v} evaluated expression
25440 @item frequency, freq, f
25441 the frequency where it is evaluated
25442 @item timeclamp, tc
25443 the value of @var{timeclamp} option
25447 @item a_weighting(f)
25448 A-weighting of equal loudness
25449 @item b_weighting(f)
25450 B-weighting of equal loudness
25451 @item c_weighting(f)
25452 C-weighting of equal loudness.
25454 Default value is @code{16}.
25456 @item bar_v, volume2
25457 Specify the bargraph volume expression. It can contain variables:
25460 the @var{sono_v} evaluated expression
25461 @item frequency, freq, f
25462 the frequency where it is evaluated
25463 @item timeclamp, tc
25464 the value of @var{timeclamp} option
25468 @item a_weighting(f)
25469 A-weighting of equal loudness
25470 @item b_weighting(f)
25471 B-weighting of equal loudness
25472 @item c_weighting(f)
25473 C-weighting of equal loudness.
25475 Default value is @code{sono_v}.
25477 @item sono_g, gamma
25478 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
25479 higher gamma makes the spectrum having more range. Default value is @code{3}.
25480 Acceptable range is @code{[1, 7]}.
25482 @item bar_g, gamma2
25483 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
25487 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
25488 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
25490 @item timeclamp, tc
25491 Specify the transform timeclamp. At low frequency, there is trade-off between
25492 accuracy in time domain and frequency domain. If timeclamp is lower,
25493 event in time domain is represented more accurately (such as fast bass drum),
25494 otherwise event in frequency domain is represented more accurately
25495 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
25498 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
25499 limits future samples by applying asymmetric windowing in time domain, useful
25500 when low latency is required. Accepted range is @code{[0, 1]}.
25503 Specify the transform base frequency. Default value is @code{20.01523126408007475},
25504 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
25507 Specify the transform end frequency. Default value is @code{20495.59681441799654},
25508 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
25511 This option is deprecated and ignored.
25514 Specify the transform length in time domain. Use this option to control accuracy
25515 trade-off between time domain and frequency domain at every frequency sample.
25516 It can contain variables:
25518 @item frequency, freq, f
25519 the frequency where it is evaluated
25520 @item timeclamp, tc
25521 the value of @var{timeclamp} option.
25523 Default value is @code{384*tc/(384+tc*f)}.
25526 Specify the transform count for every video frame. Default value is @code{6}.
25527 Acceptable range is @code{[1, 30]}.
25530 Specify the transform count for every single pixel. Default value is @code{0},
25531 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
25534 Specify font file for use with freetype to draw the axis. If not specified,
25535 use embedded font. Note that drawing with font file or embedded font is not
25536 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
25540 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
25541 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
25545 Specify font color expression. This is arithmetic expression that should return
25546 integer value 0xRRGGBB. It can contain variables:
25548 @item frequency, freq, f
25549 the frequency where it is evaluated
25550 @item timeclamp, tc
25551 the value of @var{timeclamp} option
25556 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
25557 @item r(x), g(x), b(x)
25558 red, green, and blue value of intensity x.
25560 Default value is @code{st(0, (midi(f)-59.5)/12);
25561 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
25562 r(1-ld(1)) + b(ld(1))}.
25565 Specify image file to draw the axis. This option override @var{fontfile} and
25566 @var{fontcolor} option.
25569 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
25570 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
25571 Default value is @code{1}.
25574 Set colorspace. The accepted values are:
25577 Unspecified (default)
25586 BT.470BG or BT.601-6 625
25589 SMPTE-170M or BT.601-6 525
25595 BT.2020 with non-constant luminance
25600 Set spectrogram color scheme. This is list of floating point values with format
25601 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
25602 The default is @code{1|0.5|0|0|0.5|1}.
25606 @subsection Examples
25610 Playing audio while showing the spectrum:
25612 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
25616 Same as above, but with frame rate 30 fps:
25618 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
25622 Playing at 1280x720:
25624 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
25628 Disable sonogram display:
25634 A1 and its harmonics: A1, A2, (near)E3, A3:
25636 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),
25637 asplit[a][out1]; [a] showcqt [out0]'
25641 Same as above, but with more accuracy in frequency domain:
25643 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),
25644 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
25650 bar_v=10:sono_v=bar_v*a_weighting(f)
25654 Custom gamma, now spectrum is linear to the amplitude.
25660 Custom tlength equation:
25662 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)))'
25666 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
25668 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
25672 Custom font using fontconfig:
25674 font='Courier New,Monospace,mono|bold'
25678 Custom frequency range with custom axis using image file:
25680 axisfile=myaxis.png:basefreq=40:endfreq=10000
25686 Convert input audio to video output representing the audio power spectrum.
25687 Audio amplitude is on Y-axis while frequency is on X-axis.
25689 The filter accepts the following options:
25693 Specify size of video. For the syntax of this option, check the
25694 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25695 Default is @code{1024x512}.
25699 This set how each frequency bin will be represented.
25701 It accepts the following values:
25707 Default is @code{bar}.
25710 Set amplitude scale.
25712 It accepts the following values:
25726 Default is @code{log}.
25729 Set frequency scale.
25731 It accepts the following values:
25740 Reverse logarithmic scale.
25742 Default is @code{lin}.
25745 Set window size. Allowed range is from 16 to 65536.
25747 Default is @code{2048}
25750 Set windowing function.
25752 It accepts the following values:
25775 Default is @code{hanning}.
25778 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25779 which means optimal overlap for selected window function will be picked.
25782 Set time averaging. Setting this to 0 will display current maximal peaks.
25783 Default is @code{1}, which means time averaging is disabled.
25786 Specify list of colors separated by space or by '|' which will be used to
25787 draw channel frequencies. Unrecognized or missing colors will be replaced
25791 Set channel display mode.
25793 It accepts the following values:
25798 Default is @code{combined}.
25801 Set minimum amplitude used in @code{log} amplitude scaler.
25804 Set data display mode.
25806 It accepts the following values:
25812 Default is @code{magnitude}.
25815 @section showspatial
25817 Convert stereo input audio to a video output, representing the spatial relationship
25818 between two channels.
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{512x512}.
25829 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
25832 Set window function.
25834 It accepts the following values:
25859 Default value is @code{hann}.
25862 Set ratio of overlap window. Default value is @code{0.5}.
25863 When value is @code{1} overlap is set to recommended size for specific
25864 window function currently used.
25867 @anchor{showspectrum}
25868 @section showspectrum
25870 Convert input audio to a video output, representing the audio frequency
25873 The filter accepts the following options:
25877 Specify the video size for the output. For the syntax of this option, check the
25878 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25879 Default value is @code{640x512}.
25882 Specify how the spectrum should slide along the window.
25884 It accepts the following values:
25887 the samples start again on the left when they reach the right
25889 the samples scroll from right to left
25891 frames are only produced when the samples reach the right
25893 the samples scroll from left to right
25896 Default value is @code{replace}.
25899 Specify display mode.
25901 It accepts the following values:
25904 all channels are displayed in the same row
25906 all channels are displayed in separate rows
25909 Default value is @samp{combined}.
25912 Specify display color mode.
25914 It accepts the following values:
25917 each channel is displayed in a separate color
25919 each channel is displayed using the same color scheme
25921 each channel is displayed using the rainbow color scheme
25923 each channel is displayed using the moreland color scheme
25925 each channel is displayed using the nebulae color scheme
25927 each channel is displayed using the fire color scheme
25929 each channel is displayed using the fiery color scheme
25931 each channel is displayed using the fruit color scheme
25933 each channel is displayed using the cool color scheme
25935 each channel is displayed using the magma color scheme
25937 each channel is displayed using the green color scheme
25939 each channel is displayed using the viridis color scheme
25941 each channel is displayed using the plasma color scheme
25943 each channel is displayed using the cividis color scheme
25945 each channel is displayed using the terrain color scheme
25948 Default value is @samp{channel}.
25951 Specify scale used for calculating intensity color values.
25953 It accepts the following values:
25958 square root, default
25969 Default value is @samp{sqrt}.
25972 Specify frequency scale.
25974 It accepts the following values:
25982 Default value is @samp{lin}.
25985 Set saturation modifier for displayed colors. Negative values provide
25986 alternative color scheme. @code{0} is no saturation at all.
25987 Saturation must be in [-10.0, 10.0] range.
25988 Default value is @code{1}.
25991 Set window function.
25993 It accepts the following values:
26018 Default value is @code{hann}.
26021 Set orientation of time vs frequency axis. Can be @code{vertical} or
26022 @code{horizontal}. Default is @code{vertical}.
26025 Set ratio of overlap window. Default value is @code{0}.
26026 When value is @code{1} overlap is set to recommended size for specific
26027 window function currently used.
26030 Set scale gain for calculating intensity color values.
26031 Default value is @code{1}.
26034 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
26037 Set color rotation, must be in [-1.0, 1.0] range.
26038 Default value is @code{0}.
26041 Set start frequency from which to display spectrogram. Default is @code{0}.
26044 Set stop frequency to which to display spectrogram. Default is @code{0}.
26047 Set upper frame rate limit. Default is @code{auto}, unlimited.
26050 Draw time and frequency axes and legends. Default is disabled.
26053 The usage is very similar to the showwaves filter; see the examples in that
26056 @subsection Examples
26060 Large window with logarithmic color scaling:
26062 showspectrum=s=1280x480:scale=log
26066 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
26068 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
26069 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
26073 @section showspectrumpic
26075 Convert input audio to a single video frame, representing the audio frequency
26078 The filter accepts the following options:
26082 Specify the video size for the output. For the syntax of this option, check the
26083 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26084 Default value is @code{4096x2048}.
26087 Specify display mode.
26089 It accepts the following values:
26092 all channels are displayed in the same row
26094 all channels are displayed in separate rows
26096 Default value is @samp{combined}.
26099 Specify display color mode.
26101 It accepts the following values:
26104 each channel is displayed in a separate color
26106 each channel is displayed using the same color scheme
26108 each channel is displayed using the rainbow color scheme
26110 each channel is displayed using the moreland color scheme
26112 each channel is displayed using the nebulae color scheme
26114 each channel is displayed using the fire color scheme
26116 each channel is displayed using the fiery color scheme
26118 each channel is displayed using the fruit color scheme
26120 each channel is displayed using the cool color scheme
26122 each channel is displayed using the magma color scheme
26124 each channel is displayed using the green color scheme
26126 each channel is displayed using the viridis color scheme
26128 each channel is displayed using the plasma color scheme
26130 each channel is displayed using the cividis color scheme
26132 each channel is displayed using the terrain color scheme
26134 Default value is @samp{intensity}.
26137 Specify scale used for calculating intensity color values.
26139 It accepts the following values:
26144 square root, default
26154 Default value is @samp{log}.
26157 Specify frequency scale.
26159 It accepts the following values:
26167 Default value is @samp{lin}.
26170 Set saturation modifier for displayed colors. Negative values provide
26171 alternative color scheme. @code{0} is no saturation at all.
26172 Saturation must be in [-10.0, 10.0] range.
26173 Default value is @code{1}.
26176 Set window function.
26178 It accepts the following values:
26202 Default value is @code{hann}.
26205 Set orientation of time vs frequency axis. Can be @code{vertical} or
26206 @code{horizontal}. Default is @code{vertical}.
26209 Set scale gain for calculating intensity color values.
26210 Default value is @code{1}.
26213 Draw time and frequency axes and legends. Default is enabled.
26216 Set color rotation, must be in [-1.0, 1.0] range.
26217 Default value is @code{0}.
26220 Set start frequency from which to display spectrogram. Default is @code{0}.
26223 Set stop frequency to which to display spectrogram. Default is @code{0}.
26226 @subsection Examples
26230 Extract an audio spectrogram of a whole audio track
26231 in a 1024x1024 picture using @command{ffmpeg}:
26233 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
26237 @section showvolume
26239 Convert input audio volume to a video output.
26241 The filter accepts the following options:
26248 Set border width, allowed range is [0, 5]. Default is 1.
26251 Set channel width, allowed range is [80, 8192]. Default is 400.
26254 Set channel height, allowed range is [1, 900]. Default is 20.
26257 Set fade, allowed range is [0, 1]. Default is 0.95.
26260 Set volume color expression.
26262 The expression can use the following variables:
26266 Current max volume of channel in dB.
26272 Current channel number, starting from 0.
26276 If set, displays channel names. Default is enabled.
26279 If set, displays volume values. Default is enabled.
26282 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
26283 default is @code{h}.
26286 Set step size, allowed range is [0, 5]. Default is 0, which means
26290 Set background opacity, allowed range is [0, 1]. Default is 0.
26293 Set metering mode, can be peak: @code{p} or rms: @code{r},
26294 default is @code{p}.
26297 Set display scale, can be linear: @code{lin} or log: @code{log},
26298 default is @code{lin}.
26302 If set to > 0., display a line for the max level
26303 in the previous seconds.
26304 default is disabled: @code{0.}
26307 The color of the max line. Use when @code{dm} option is set to > 0.
26308 default is: @code{orange}
26313 Convert input audio to a video output, representing the samples waves.
26315 The filter accepts the following options:
26319 Specify the video size for the output. For the syntax of this option, check the
26320 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26321 Default value is @code{600x240}.
26326 Available values are:
26329 Draw a point for each sample.
26332 Draw a vertical line for each sample.
26335 Draw a point for each sample and a line between them.
26338 Draw a centered vertical line for each sample.
26341 Default value is @code{point}.
26344 Set the number of samples which are printed on the same column. A
26345 larger value will decrease the frame rate. Must be a positive
26346 integer. This option can be set only if the value for @var{rate}
26347 is not explicitly specified.
26350 Set the (approximate) output frame rate. This is done by setting the
26351 option @var{n}. Default value is "25".
26353 @item split_channels
26354 Set if channels should be drawn separately or overlap. Default value is 0.
26357 Set colors separated by '|' which are going to be used for drawing of each channel.
26360 Set amplitude scale.
26362 Available values are:
26380 Set the draw mode. This is mostly useful to set for high @var{n}.
26382 Available values are:
26385 Scale pixel values for each drawn sample.
26388 Draw every sample directly.
26391 Default value is @code{scale}.
26394 @subsection Examples
26398 Output the input file audio and the corresponding video representation
26401 amovie=a.mp3,asplit[out0],showwaves[out1]
26405 Create a synthetic signal and show it with showwaves, forcing a
26406 frame rate of 30 frames per second:
26408 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
26412 @section showwavespic
26414 Convert input audio to a single video frame, representing the samples waves.
26416 The filter accepts the following options:
26420 Specify the video size for the output. For the syntax of this option, check the
26421 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26422 Default value is @code{600x240}.
26424 @item split_channels
26425 Set if channels should be drawn separately or overlap. Default value is 0.
26428 Set colors separated by '|' which are going to be used for drawing of each channel.
26431 Set amplitude scale.
26433 Available values are:
26453 Available values are:
26456 Scale pixel values for each drawn sample.
26459 Draw every sample directly.
26462 Default value is @code{scale}.
26465 Set the filter mode.
26467 Available values are:
26470 Use average samples values for each drawn sample.
26473 Use peak samples values for each drawn sample.
26476 Default value is @code{average}.
26479 @subsection Examples
26483 Extract a channel split representation of the wave form of a whole audio track
26484 in a 1024x800 picture using @command{ffmpeg}:
26486 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
26490 @section sidedata, asidedata
26492 Delete frame side data, or select frames based on it.
26494 This filter accepts the following options:
26498 Set mode of operation of the filter.
26500 Can be one of the following:
26504 Select every frame with side data of @code{type}.
26507 Delete side data of @code{type}. If @code{type} is not set, delete all side
26513 Set side data type used with all modes. Must be set for @code{select} mode. For
26514 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
26515 in @file{libavutil/frame.h}. For example, to choose
26516 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
26520 @section spectrumsynth
26522 Synthesize audio from 2 input video spectrums, first input stream represents
26523 magnitude across time and second represents phase across time.
26524 The filter will transform from frequency domain as displayed in videos back
26525 to time domain as presented in audio output.
26527 This filter is primarily created for reversing processed @ref{showspectrum}
26528 filter outputs, but can synthesize sound from other spectrograms too.
26529 But in such case results are going to be poor if the phase data is not
26530 available, because in such cases phase data need to be recreated, usually
26531 it's just recreated from random noise.
26532 For best results use gray only output (@code{channel} color mode in
26533 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
26534 @code{lin} scale for phase video. To produce phase, for 2nd video, use
26535 @code{data} option. Inputs videos should generally use @code{fullframe}
26536 slide mode as that saves resources needed for decoding video.
26538 The filter accepts the following options:
26542 Specify sample rate of output audio, the sample rate of audio from which
26543 spectrum was generated may differ.
26546 Set number of channels represented in input video spectrums.
26549 Set scale which was used when generating magnitude input spectrum.
26550 Can be @code{lin} or @code{log}. Default is @code{log}.
26553 Set slide which was used when generating inputs spectrums.
26554 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
26555 Default is @code{fullframe}.
26558 Set window function used for resynthesis.
26561 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
26562 which means optimal overlap for selected window function will be picked.
26565 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
26566 Default is @code{vertical}.
26569 @subsection Examples
26573 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
26574 then resynthesize videos back to audio with spectrumsynth:
26576 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
26577 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
26578 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
26582 @section split, asplit
26584 Split input into several identical outputs.
26586 @code{asplit} works with audio input, @code{split} with video.
26588 The filter accepts a single parameter which specifies the number of outputs. If
26589 unspecified, it defaults to 2.
26591 @subsection Examples
26595 Create two separate outputs from the same input:
26597 [in] split [out0][out1]
26601 To create 3 or more outputs, you need to specify the number of
26604 [in] asplit=3 [out0][out1][out2]
26608 Create two separate outputs from the same input, one cropped and
26611 [in] split [splitout1][splitout2];
26612 [splitout1] crop=100:100:0:0 [cropout];
26613 [splitout2] pad=200:200:100:100 [padout];
26617 Create 5 copies of the input audio with @command{ffmpeg}:
26619 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
26625 Receive commands sent through a libzmq client, and forward them to
26626 filters in the filtergraph.
26628 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
26629 must be inserted between two video filters, @code{azmq} between two
26630 audio filters. Both are capable to send messages to any filter type.
26632 To enable these filters you need to install the libzmq library and
26633 headers and configure FFmpeg with @code{--enable-libzmq}.
26635 For more information about libzmq see:
26636 @url{http://www.zeromq.org/}
26638 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
26639 receives messages sent through a network interface defined by the
26640 @option{bind_address} (or the abbreviation "@option{b}") option.
26641 Default value of this option is @file{tcp://localhost:5555}. You may
26642 want to alter this value to your needs, but do not forget to escape any
26643 ':' signs (see @ref{filtergraph escaping}).
26645 The received message must be in the form:
26647 @var{TARGET} @var{COMMAND} [@var{ARG}]
26650 @var{TARGET} specifies the target of the command, usually the name of
26651 the filter class or a specific filter instance name. The default
26652 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
26653 but you can override this by using the @samp{filter_name@@id} syntax
26654 (see @ref{Filtergraph syntax}).
26656 @var{COMMAND} specifies the name of the command for the target filter.
26658 @var{ARG} is optional and specifies the optional argument list for the
26659 given @var{COMMAND}.
26661 Upon reception, the message is processed and the corresponding command
26662 is injected into the filtergraph. Depending on the result, the filter
26663 will send a reply to the client, adopting the format:
26665 @var{ERROR_CODE} @var{ERROR_REASON}
26669 @var{MESSAGE} is optional.
26671 @subsection Examples
26673 Look at @file{tools/zmqsend} for an example of a zmq client which can
26674 be used to send commands processed by these filters.
26676 Consider the following filtergraph generated by @command{ffplay}.
26677 In this example the last overlay filter has an instance name. All other
26678 filters will have default instance names.
26681 ffplay -dumpgraph 1 -f lavfi "
26682 color=s=100x100:c=red [l];
26683 color=s=100x100:c=blue [r];
26684 nullsrc=s=200x100, zmq [bg];
26685 [bg][l] overlay [bg+l];
26686 [bg+l][r] overlay@@my=x=100 "
26689 To change the color of the left side of the video, the following
26690 command can be used:
26692 echo Parsed_color_0 c yellow | tools/zmqsend
26695 To change the right side:
26697 echo Parsed_color_1 c pink | tools/zmqsend
26700 To change the position of the right side:
26702 echo overlay@@my x 150 | tools/zmqsend
26706 @c man end MULTIMEDIA FILTERS
26708 @chapter Multimedia Sources
26709 @c man begin MULTIMEDIA SOURCES
26711 Below is a description of the currently available multimedia sources.
26715 This is the same as @ref{movie} source, except it selects an audio
26721 Read audio and/or video stream(s) from a movie container.
26723 It accepts the following parameters:
26727 The name of the resource to read (not necessarily a file; it can also be a
26728 device or a stream accessed through some protocol).
26730 @item format_name, f
26731 Specifies the format assumed for the movie to read, and can be either
26732 the name of a container or an input device. If not specified, the
26733 format is guessed from @var{movie_name} or by probing.
26735 @item seek_point, sp
26736 Specifies the seek point in seconds. The frames will be output
26737 starting from this seek point. The parameter is evaluated with
26738 @code{av_strtod}, so the numerical value may be suffixed by an IS
26739 postfix. The default value is "0".
26742 Specifies the streams to read. Several streams can be specified,
26743 separated by "+". The source will then have as many outputs, in the
26744 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
26745 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
26746 respectively the default (best suited) video and audio stream. Default
26747 is "dv", or "da" if the filter is called as "amovie".
26749 @item stream_index, si
26750 Specifies the index of the video stream to read. If the value is -1,
26751 the most suitable video stream will be automatically selected. The default
26752 value is "-1". Deprecated. If the filter is called "amovie", it will select
26753 audio instead of video.
26756 Specifies how many times to read the stream in sequence.
26757 If the value is 0, the stream will be looped infinitely.
26758 Default value is "1".
26760 Note that when the movie is looped the source timestamps are not
26761 changed, so it will generate non monotonically increasing timestamps.
26763 @item discontinuity
26764 Specifies the time difference between frames above which the point is
26765 considered a timestamp discontinuity which is removed by adjusting the later
26769 It allows overlaying a second video on top of the main input of
26770 a filtergraph, as shown in this graph:
26772 input -----------> deltapts0 --> overlay --> output
26775 movie --> scale--> deltapts1 -------+
26777 @subsection Examples
26781 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
26782 on top of the input labelled "in":
26784 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
26785 [in] setpts=PTS-STARTPTS [main];
26786 [main][over] overlay=16:16 [out]
26790 Read from a video4linux2 device, and overlay it on top of the input
26793 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
26794 [in] setpts=PTS-STARTPTS [main];
26795 [main][over] overlay=16:16 [out]
26799 Read the first video stream and the audio stream with id 0x81 from
26800 dvd.vob; the video is connected to the pad named "video" and the audio is
26801 connected to the pad named "audio":
26803 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
26807 @subsection Commands
26809 Both movie and amovie support the following commands:
26812 Perform seek using "av_seek_frame".
26813 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
26816 @var{stream_index}: If stream_index is -1, a default
26817 stream is selected, and @var{timestamp} is automatically converted
26818 from AV_TIME_BASE units to the stream specific time_base.
26820 @var{timestamp}: Timestamp in AVStream.time_base units
26821 or, if no stream is specified, in AV_TIME_BASE units.
26823 @var{flags}: Flags which select direction and seeking mode.
26827 Get movie duration in AV_TIME_BASE units.
26831 @c man end MULTIMEDIA SOURCES