1 @chapter Filtering Introduction
2 @c man begin FILTERING INTRODUCTION
4 Filtering in FFmpeg is enabled through the libavfilter library.
6 In libavfilter, a filter can have multiple inputs and multiple
8 To illustrate the sorts of things that are possible, we consider the
13 input --> split ---------------------> overlay --> output
16 +-----> crop --> vflip -------+
19 This filtergraph splits the input stream in two streams, then sends one
20 stream through the crop filter and the vflip filter, before merging it
21 back with the other stream by overlaying it on top. You can use the
22 following command to achieve this:
25 ffmpeg -i INPUT -vf "split [main][tmp]; [tmp] crop=iw:ih/2:0:0, vflip [flip]; [main][flip] overlay=0:H/2" OUTPUT
28 The result will be that the top half of the video is mirrored
29 onto the bottom half of the output video.
31 Filters in the same linear chain are separated by commas, and distinct
32 linear chains of filters are separated by semicolons. In our example,
33 @var{crop,vflip} are in one linear chain, @var{split} and
34 @var{overlay} are separately in another. The points where the linear
35 chains join are labelled by names enclosed in square brackets. In the
36 example, the split filter generates two outputs that are associated to
37 the labels @var{[main]} and @var{[tmp]}.
39 The stream sent to the second output of @var{split}, labelled as
40 @var{[tmp]}, is processed through the @var{crop} filter, which crops
41 away the lower half part of the video, and then vertically flipped. The
42 @var{overlay} filter takes in input the first unchanged output of the
43 split filter (which was labelled as @var{[main]}), and overlay on its
44 lower half the output generated by the @var{crop,vflip} filterchain.
46 Some filters take in input a list of parameters: they are specified
47 after the filter name and an equal sign, and are separated from each other
50 There exist so-called @var{source filters} that do not have an
51 audio/video input, and @var{sink filters} that will not have audio/video
54 @c man end FILTERING INTRODUCTION
57 @c man begin GRAPH2DOT
59 The @file{graph2dot} program included in the FFmpeg @file{tools}
60 directory can be used to parse a filtergraph description and issue a
61 corresponding textual representation in the dot language.
68 to see how to use @file{graph2dot}.
70 You can then pass the dot description to the @file{dot} program (from
71 the graphviz suite of programs) and obtain a graphical representation
74 For example the sequence of commands:
76 echo @var{GRAPH_DESCRIPTION} | \
77 tools/graph2dot -o graph.tmp && \
78 dot -Tpng graph.tmp -o graph.png && \
82 can be used to create and display an image representing the graph
83 described by the @var{GRAPH_DESCRIPTION} string. Note that this string must be
84 a complete self-contained graph, with its inputs and outputs explicitly defined.
85 For example if your command line is of the form:
87 ffmpeg -i infile -vf scale=640:360 outfile
89 your @var{GRAPH_DESCRIPTION} string will need to be of the form:
91 nullsrc,scale=640:360,nullsink
93 you may also need to set the @var{nullsrc} parameters and add a @var{format}
94 filter in order to simulate a specific input file.
98 @chapter Filtergraph description
99 @c man begin FILTERGRAPH DESCRIPTION
101 A filtergraph is a directed graph of connected filters. It can contain
102 cycles, and there can be multiple links between a pair of
103 filters. Each link has one input pad on one side connecting it to one
104 filter from which it takes its input, and one output pad on the other
105 side connecting it to one filter accepting its output.
107 Each filter in a filtergraph is an instance of a filter class
108 registered in the application, which defines the features and the
109 number of input and output pads of the filter.
111 A filter with no input pads is called a "source", and a filter with no
112 output pads is called a "sink".
114 @anchor{Filtergraph syntax}
115 @section Filtergraph syntax
117 A filtergraph has a textual representation, which is recognized by the
118 @option{-filter}/@option{-vf}/@option{-af} and
119 @option{-filter_complex} options in @command{ffmpeg} and
120 @option{-vf}/@option{-af} in @command{ffplay}, and by the
121 @code{avfilter_graph_parse_ptr()} function defined in
122 @file{libavfilter/avfilter.h}.
124 A filterchain consists of a sequence of connected filters, each one
125 connected to the previous one in the sequence. A filterchain is
126 represented by a list of ","-separated filter descriptions.
128 A filtergraph consists of a sequence of filterchains. A sequence of
129 filterchains is represented by a list of ";"-separated filterchain
132 A filter is represented by a string of the form:
133 [@var{in_link_1}]...[@var{in_link_N}]@var{filter_name}@@@var{id}=@var{arguments}[@var{out_link_1}]...[@var{out_link_M}]
135 @var{filter_name} is the name of the filter class of which the
136 described filter is an instance of, and has to be the name of one of
137 the filter classes registered in the program optionally followed by "@@@var{id}".
138 The name of the filter class is optionally followed by a string
141 @var{arguments} is a string which contains the parameters used to
142 initialize the filter instance. It may have one of two forms:
146 A ':'-separated list of @var{key=value} pairs.
149 A ':'-separated list of @var{value}. In this case, the keys are assumed to be
150 the option names in the order they are declared. E.g. the @code{fade} filter
151 declares three options in this order -- @option{type}, @option{start_frame} and
152 @option{nb_frames}. Then the parameter list @var{in:0:30} means that the value
153 @var{in} is assigned to the option @option{type}, @var{0} to
154 @option{start_frame} and @var{30} to @option{nb_frames}.
157 A ':'-separated list of mixed direct @var{value} and long @var{key=value}
158 pairs. The direct @var{value} must precede the @var{key=value} pairs, and
159 follow the same constraints order of the previous point. The following
160 @var{key=value} pairs can be set in any preferred order.
164 If the option value itself is a list of items (e.g. the @code{format} filter
165 takes a list of pixel formats), the items in the list are usually separated by
168 The list of arguments can be quoted using the character @samp{'} as initial
169 and ending mark, and the character @samp{\} for escaping the characters
170 within the quoted text; otherwise the argument string is considered
171 terminated when the next special character (belonging to the set
172 @samp{[]=;,}) is encountered.
174 The name and arguments of the filter are optionally preceded and
175 followed by a list of link labels.
176 A link label allows one to name a link and associate it to a filter output
177 or input pad. The preceding labels @var{in_link_1}
178 ... @var{in_link_N}, are associated to the filter input pads,
179 the following labels @var{out_link_1} ... @var{out_link_M}, are
180 associated to the output pads.
182 When two link labels with the same name are found in the
183 filtergraph, a link between the corresponding input and output pad is
186 If an output pad is not labelled, it is linked by default to the first
187 unlabelled input pad of the next filter in the filterchain.
188 For example in the filterchain
190 nullsrc, split[L1], [L2]overlay, nullsink
192 the split filter instance has two output pads, and the overlay filter
193 instance two input pads. The first output pad of split is labelled
194 "L1", the first input pad of overlay is labelled "L2", and the second
195 output pad of split is linked to the second input pad of overlay,
196 which are both unlabelled.
198 In a filter description, if the input label of the first filter is not
199 specified, "in" is assumed; if the output label of the last filter is not
200 specified, "out" is assumed.
202 In a complete filterchain all the unlabelled filter input and output
203 pads must be connected. A filtergraph is considered valid if all the
204 filter input and output pads of all the filterchains are connected.
206 Libavfilter will automatically insert @ref{scale} filters where format
207 conversion is required. It is possible to specify swscale flags
208 for those automatically inserted scalers by prepending
209 @code{sws_flags=@var{flags};}
210 to the filtergraph description.
212 Here is a BNF description of the filtergraph syntax:
214 @var{NAME} ::= sequence of alphanumeric characters and '_'
215 @var{FILTER_NAME} ::= @var{NAME}["@@"@var{NAME}]
216 @var{LINKLABEL} ::= "[" @var{NAME} "]"
217 @var{LINKLABELS} ::= @var{LINKLABEL} [@var{LINKLABELS}]
218 @var{FILTER_ARGUMENTS} ::= sequence of chars (possibly quoted)
219 @var{FILTER} ::= [@var{LINKLABELS}] @var{FILTER_NAME} ["=" @var{FILTER_ARGUMENTS}] [@var{LINKLABELS}]
220 @var{FILTERCHAIN} ::= @var{FILTER} [,@var{FILTERCHAIN}]
221 @var{FILTERGRAPH} ::= [sws_flags=@var{flags};] @var{FILTERCHAIN} [;@var{FILTERGRAPH}]
224 @anchor{filtergraph escaping}
225 @section Notes on filtergraph escaping
227 Filtergraph description composition entails several levels of
228 escaping. See @ref{quoting_and_escaping,,the "Quoting and escaping"
229 section in the ffmpeg-utils(1) manual,ffmpeg-utils} for more
230 information about the employed escaping procedure.
232 A first level escaping affects the content of each filter option
233 value, which may contain the special character @code{:} used to
234 separate values, or one of the escaping characters @code{\'}.
236 A second level escaping affects the whole filter description, which
237 may contain the escaping characters @code{\'} or the special
238 characters @code{[],;} used by the filtergraph description.
240 Finally, when you specify a filtergraph on a shell commandline, you
241 need to perform a third level escaping for the shell special
242 characters contained within it.
244 For example, consider the following string to be embedded in
245 the @ref{drawtext} filter description @option{text} value:
247 this is a 'string': may contain one, or more, special characters
250 This string contains the @code{'} special escaping character, and the
251 @code{:} special character, so it needs to be escaped in this way:
253 text=this is a \'string\'\: may contain one, or more, special characters
256 A second level of escaping is required when embedding the filter
257 description in a filtergraph description, in order to escape all the
258 filtergraph special characters. Thus the example above becomes:
260 drawtext=text=this is a \\\'string\\\'\\: may contain one\, or more\, special characters
262 (note that in addition to the @code{\'} escaping special characters,
263 also @code{,} needs to be escaped).
265 Finally an additional level of escaping is needed when writing the
266 filtergraph description in a shell command, which depends on the
267 escaping rules of the adopted shell. For example, assuming that
268 @code{\} is special and needs to be escaped with another @code{\}, the
269 previous string will finally result in:
271 -vf "drawtext=text=this is a \\\\\\'string\\\\\\'\\\\: may contain one\\, or more\\, special characters"
274 @chapter Timeline editing
276 Some filters support a generic @option{enable} option. For the filters
277 supporting timeline editing, this option can be set to an expression which is
278 evaluated before sending a frame to the filter. If the evaluation is non-zero,
279 the filter will be enabled, otherwise the frame will be sent unchanged to the
280 next filter in the filtergraph.
282 The expression accepts the following values:
285 timestamp expressed in seconds, NAN if the input timestamp is unknown
288 sequential number of the input frame, starting from 0
291 the position in the file of the input frame, NAN if unknown
295 width and height of the input frame if video
298 Additionally, these filters support an @option{enable} command that can be used
299 to re-define the expression.
301 Like any other filtering option, the @option{enable} option follows the same
304 For example, to enable a blur filter (@ref{smartblur}) from 10 seconds to 3
305 minutes, and a @ref{curves} filter starting at 3 seconds:
307 smartblur = enable='between(t,10,3*60)',
308 curves = enable='gte(t,3)' : preset=cross_process
311 See @code{ffmpeg -filters} to view which filters have timeline support.
313 @c man end FILTERGRAPH DESCRIPTION
316 @chapter Changing options at runtime with a command
318 Some options can be changed during the operation of the filter using
319 a command. These options are marked 'T' on the output of
320 @command{ffmpeg} @option{-h filter=<name of filter>}.
321 The name of the command is the name of the option and the argument is
325 @chapter Options for filters with several inputs (framesync)
326 @c man begin OPTIONS FOR FILTERS WITH SEVERAL INPUTS
328 Some filters with several inputs support a common set of options.
329 These options can only be set by name, not with the short notation.
333 The action to take when EOF is encountered on the secondary input; it accepts
334 one of the following values:
338 Repeat the last frame (the default).
342 Pass the main input through.
346 If set to 1, force the output to terminate when the shortest input
347 terminates. Default value is 0.
350 If set to 1, force the filter to extend the last frame of secondary streams
351 until the end of the primary stream. A value of 0 disables this behavior.
355 @c man end OPTIONS FOR FILTERS WITH SEVERAL INPUTS
357 @chapter Audio Filters
358 @c man begin AUDIO FILTERS
360 When you configure your FFmpeg build, you can disable any of the
361 existing filters using @code{--disable-filters}.
362 The configure output will show the audio filters included in your
365 Below is a description of the currently available audio filters.
369 A compressor is mainly used to reduce the dynamic range of a signal.
370 Especially modern music is mostly compressed at a high ratio to
371 improve the overall loudness. It's done to get the highest attention
372 of a listener, "fatten" the sound and bring more "power" to the track.
373 If a signal is compressed too much it may sound dull or "dead"
374 afterwards or it may start to "pump" (which could be a powerful effect
375 but can also destroy a track completely).
376 The right compression is the key to reach a professional sound and is
377 the high art of mixing and mastering. Because of its complex settings
378 it may take a long time to get the right feeling for this kind of effect.
380 Compression is done by detecting the volume above a chosen level
381 @code{threshold} and dividing it by the factor set with @code{ratio}.
382 So if you set the threshold to -12dB and your signal reaches -6dB a ratio
383 of 2:1 will result in a signal at -9dB. Because an exact manipulation of
384 the signal would cause distortion of the waveform the reduction can be
385 levelled over the time. This is done by setting "Attack" and "Release".
386 @code{attack} determines how long the signal has to rise above the threshold
387 before any reduction will occur and @code{release} sets the time the signal
388 has to fall below the threshold to reduce the reduction again. Shorter signals
389 than the chosen attack time will be left untouched.
390 The overall reduction of the signal can be made up afterwards with the
391 @code{makeup} setting. So compressing the peaks of a signal about 6dB and
392 raising the makeup to this level results in a signal twice as loud than the
393 source. To gain a softer entry in the compression the @code{knee} flattens the
394 hard edge at the threshold in the range of the chosen decibels.
396 The filter accepts the following options:
400 Set input gain. Default is 1. Range is between 0.015625 and 64.
403 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
404 Default is @code{downward}.
407 If a signal of stream rises above this level it will affect the gain
409 By default it is 0.125. Range is between 0.00097563 and 1.
412 Set a ratio by which the signal is reduced. 1:2 means that if the level
413 rose 4dB above the threshold, it will be only 2dB above after the reduction.
414 Default is 2. Range is between 1 and 20.
417 Amount of milliseconds the signal has to rise above the threshold before gain
418 reduction starts. Default is 20. Range is between 0.01 and 2000.
421 Amount of milliseconds the signal has to fall below the threshold before
422 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
425 Set the amount by how much signal will be amplified after processing.
426 Default is 1. Range is from 1 to 64.
429 Curve the sharp knee around the threshold to enter gain reduction more softly.
430 Default is 2.82843. Range is between 1 and 8.
433 Choose if the @code{average} level between all channels of input stream
434 or the louder(@code{maximum}) channel of input stream affects the
435 reduction. Default is @code{average}.
438 Should the exact signal be taken in case of @code{peak} or an RMS one in case
439 of @code{rms}. Default is @code{rms} which is mostly smoother.
442 How much to use compressed signal in output. Default is 1.
443 Range is between 0 and 1.
448 This filter supports the all above options as @ref{commands}.
451 Simple audio dynamic range compression/expansion filter.
453 The filter accepts the following options:
457 Set contrast. Default is 33. Allowed range is between 0 and 100.
462 Copy the input audio source unchanged to the output. This is mainly useful for
467 Apply cross fade from one input audio stream to another input audio stream.
468 The cross fade is applied for specified duration near the end of first stream.
470 The filter accepts the following options:
474 Specify the number of samples for which the cross fade effect has to last.
475 At the end of the cross fade effect the first input audio will be completely
476 silent. Default is 44100.
479 Specify the duration of the cross fade effect. See
480 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
481 for the accepted syntax.
482 By default the duration is determined by @var{nb_samples}.
483 If set this option is used instead of @var{nb_samples}.
486 Should first stream end overlap with second stream start. Default is enabled.
489 Set curve for cross fade transition for first stream.
492 Set curve for cross fade transition for second stream.
494 For description of available curve types see @ref{afade} filter description.
501 Cross fade from one input to another:
503 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
507 Cross fade from one input to another but without overlapping:
509 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
514 Split audio stream into several bands.
516 This filter splits audio stream into two or more frequency ranges.
517 Summing all streams back will give flat output.
519 The filter accepts the following options:
523 Set split frequencies. Those must be positive and increasing.
526 Set filter order for each band split. This controls filter roll-off or steepness
527 of filter transfer function.
528 Available values are:
553 Default is @var{4th}.
556 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
559 Set output gain for each band. Default value is 1 for all bands.
566 Split input audio stream into two bands (low and high) with split frequency of 1500 Hz,
567 each band will be in separate stream:
569 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
573 Same as above, but with higher filter order:
575 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500:order=8th[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
579 Same as above, but also with additional middle band (frequencies between 1500 and 8000):
581 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500 8000:order=8th[LOW][MID][HIGH]' -map '[LOW]' low.wav -map '[MID]' mid.wav -map '[HIGH]' high.wav
587 Reduce audio bit resolution.
589 This filter is bit crusher with enhanced functionality. A bit crusher
590 is used to audibly reduce number of bits an audio signal is sampled
591 with. This doesn't change the bit depth at all, it just produces the
592 effect. Material reduced in bit depth sounds more harsh and "digital".
593 This filter is able to even round to continuous values instead of discrete
595 Additionally it has a D/C offset which results in different crushing of
596 the lower and the upper half of the signal.
597 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
599 Another feature of this filter is the logarithmic mode.
600 This setting switches from linear distances between bits to logarithmic ones.
601 The result is a much more "natural" sounding crusher which doesn't gate low
602 signals for example. The human ear has a logarithmic perception,
603 so this kind of crushing is much more pleasant.
604 Logarithmic crushing is also able to get anti-aliased.
606 The filter accepts the following options:
622 Can be linear: @code{lin} or logarithmic: @code{log}.
631 Set sample reduction.
634 Enable LFO. By default disabled.
645 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
649 Remove impulsive noise from input audio.
651 Samples detected as impulsive noise are replaced by interpolated samples using
652 autoregressive modelling.
656 Set window size, in milliseconds. Allowed range is from @code{10} to
657 @code{100}. Default value is @code{55} milliseconds.
658 This sets size of window which will be processed at once.
661 Set window overlap, in percentage of window size. Allowed range is from
662 @code{50} to @code{95}. Default value is @code{75} percent.
663 Setting this to a very high value increases impulsive noise removal but makes
664 whole process much slower.
667 Set autoregression order, in percentage of window size. Allowed range is from
668 @code{0} to @code{25}. Default value is @code{2} percent. This option also
669 controls quality of interpolated samples using neighbour good samples.
672 Set threshold value. Allowed range is from @code{1} to @code{100}.
673 Default value is @code{2}.
674 This controls the strength of impulsive noise which is going to be removed.
675 The lower value, the more samples will be detected as impulsive noise.
678 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
679 @code{10}. Default value is @code{2}.
680 If any two samples detected as noise are spaced less than this value then any
681 sample between those two samples will be also detected as noise.
686 It accepts the following values:
689 Select overlap-add method. Even not interpolated samples are slightly
690 changed with this method.
693 Select overlap-save method. Not interpolated samples remain unchanged.
696 Default value is @code{a}.
700 Remove clipped samples from input audio.
702 Samples detected as clipped are replaced by interpolated samples using
703 autoregressive modelling.
707 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
708 Default value is @code{55} milliseconds.
709 This sets size of window which will be processed at once.
712 Set window overlap, in percentage of window size. Allowed range is from @code{50}
713 to @code{95}. Default value is @code{75} percent.
716 Set autoregression order, in percentage of window size. Allowed range is from
717 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
718 quality of interpolated samples using neighbour good samples.
721 Set threshold value. Allowed range is from @code{1} to @code{100}.
722 Default value is @code{10}. Higher values make clip detection less aggressive.
725 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
726 Default value is @code{1000}. Higher values make clip detection less aggressive.
731 It accepts the following values:
734 Select overlap-add method. Even not interpolated samples are slightly changed
738 Select overlap-save method. Not interpolated samples remain unchanged.
741 Default value is @code{a}.
746 Delay one or more audio channels.
748 Samples in delayed channel are filled with silence.
750 The filter accepts the following option:
754 Set list of delays in milliseconds for each channel separated by '|'.
755 Unused delays will be silently ignored. If number of given delays is
756 smaller than number of channels all remaining channels will not be delayed.
757 If you want to delay exact number of samples, append 'S' to number.
758 If you want instead to delay in seconds, append 's' to number.
761 Use last set delay for all remaining channels. By default is disabled.
762 This option if enabled changes how option @code{delays} is interpreted.
769 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
770 the second channel (and any other channels that may be present) unchanged.
776 Delay second channel by 500 samples, the third channel by 700 samples and leave
777 the first channel (and any other channels that may be present) unchanged.
783 Delay all channels by same number of samples:
785 adelay=delays=64S:all=1
790 Remedy denormals in audio by adding extremely low-level noise.
792 This filter shall be placed before any filter that can produce denormals.
794 A description of the accepted parameters follows.
798 Set level of added noise in dB. Default is @code{-351}.
799 Allowed range is from -451 to -90.
802 Set type of added noise.
815 Default is @code{dc}.
820 This filter supports the all above options as @ref{commands}.
822 @section aderivative, aintegral
824 Compute derivative/integral of audio stream.
826 Applying both filters one after another produces original audio.
830 Apply echoing to the input audio.
832 Echoes are reflected sound and can occur naturally amongst mountains
833 (and sometimes large buildings) when talking or shouting; digital echo
834 effects emulate this behaviour and are often used to help fill out the
835 sound of a single instrument or vocal. The time difference between the
836 original signal and the reflection is the @code{delay}, and the
837 loudness of the reflected signal is the @code{decay}.
838 Multiple echoes can have different delays and decays.
840 A description of the accepted parameters follows.
844 Set input gain of reflected signal. Default is @code{0.6}.
847 Set output gain of reflected signal. Default is @code{0.3}.
850 Set list of time intervals in milliseconds between original signal and reflections
851 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
852 Default is @code{1000}.
855 Set list of loudness of reflected signals separated by '|'.
856 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
857 Default is @code{0.5}.
864 Make it sound as if there are twice as many instruments as are actually playing:
866 aecho=0.8:0.88:60:0.4
870 If delay is very short, then it sounds like a (metallic) robot playing music:
876 A longer delay will sound like an open air concert in the mountains:
878 aecho=0.8:0.9:1000:0.3
882 Same as above but with one more mountain:
884 aecho=0.8:0.9:1000|1800:0.3|0.25
889 Audio emphasis filter creates or restores material directly taken from LPs or
890 emphased CDs with different filter curves. E.g. to store music on vinyl the
891 signal has to be altered by a filter first to even out the disadvantages of
892 this recording medium.
893 Once the material is played back the inverse filter has to be applied to
894 restore the distortion of the frequency response.
896 The filter accepts the following options:
906 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
907 use @code{production} mode. Default is @code{reproduction} mode.
910 Set filter type. Selects medium. Can be one of the following:
922 select Compact Disc (CD).
928 select 50µs (FM-KF).
930 select 75µs (FM-KF).
936 This filter supports the all above options as @ref{commands}.
940 Modify an audio signal according to the specified expressions.
942 This filter accepts one or more expressions (one for each channel),
943 which are evaluated and used to modify a corresponding audio signal.
945 It accepts the following parameters:
949 Set the '|'-separated expressions list for each separate channel. If
950 the number of input channels is greater than the number of
951 expressions, the last specified expression is used for the remaining
954 @item channel_layout, c
955 Set output channel layout. If not specified, the channel layout is
956 specified by the number of expressions. If set to @samp{same}, it will
957 use by default the same input channel layout.
960 Each expression in @var{exprs} can contain the following constants and functions:
964 channel number of the current expression
967 number of the evaluated sample, starting from 0
973 time of the evaluated sample expressed in seconds
976 @item nb_out_channels
977 input and output number of channels
980 the value of input channel with number @var{CH}
983 Note: this filter is slow. For faster processing you should use a
992 aeval=val(ch)/2:c=same
996 Invert phase of the second channel:
1005 Apply fade-in/out effect to input audio.
1007 A description of the accepted parameters follows.
1011 Specify the effect type, can be either @code{in} for fade-in, or
1012 @code{out} for a fade-out effect. Default is @code{in}.
1014 @item start_sample, ss
1015 Specify the number of the start sample for starting to apply the fade
1016 effect. Default is 0.
1018 @item nb_samples, ns
1019 Specify the number of samples for which the fade effect has to last. At
1020 the end of the fade-in effect the output audio will have the same
1021 volume as the input audio, at the end of the fade-out transition
1022 the output audio will be silence. Default is 44100.
1024 @item start_time, st
1025 Specify the start time of the fade effect. Default is 0.
1026 The value must be specified as a time duration; see
1027 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1028 for the accepted syntax.
1029 If set this option is used instead of @var{start_sample}.
1032 Specify the duration of the fade effect. See
1033 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1034 for the accepted syntax.
1035 At the end of the fade-in effect the output audio will have the same
1036 volume as the input audio, at the end of the fade-out transition
1037 the output audio will be silence.
1038 By default the duration is determined by @var{nb_samples}.
1039 If set this option is used instead of @var{nb_samples}.
1042 Set curve for fade transition.
1044 It accepts the following values:
1047 select triangular, linear slope (default)
1049 select quarter of sine wave
1051 select half of sine wave
1053 select exponential sine wave
1057 select inverted parabola
1071 select inverted quarter of sine wave
1073 select inverted half of sine wave
1075 select double-exponential seat
1077 select double-exponential sigmoid
1079 select logistic sigmoid
1081 select sine cardinal function
1083 select inverted sine cardinal function
1089 @subsection Commands
1091 This filter supports the all above options as @ref{commands}.
1093 @subsection Examples
1097 Fade in first 15 seconds of audio:
1099 afade=t=in:ss=0:d=15
1103 Fade out last 25 seconds of a 900 seconds audio:
1105 afade=t=out:st=875:d=25
1110 Denoise audio samples with FFT.
1112 A description of the accepted parameters follows.
1116 Set the noise reduction in dB, allowed range is 0.01 to 97.
1117 Default value is 12 dB.
1120 Set the noise floor in dB, allowed range is -80 to -20.
1121 Default value is -50 dB.
1126 It accepts the following values:
1135 Select shellac noise.
1138 Select custom noise, defined in @code{bn} option.
1140 Default value is white noise.
1144 Set custom band noise for every one of 15 bands.
1145 Bands are separated by ' ' or '|'.
1148 Set the residual floor in dB, allowed range is -80 to -20.
1149 Default value is -38 dB.
1152 Enable noise tracking. By default is disabled.
1153 With this enabled, noise floor is automatically adjusted.
1156 Enable residual tracking. By default is disabled.
1159 Set the output mode.
1161 It accepts the following values:
1164 Pass input unchanged.
1167 Pass noise filtered out.
1172 Default value is @var{o}.
1176 @subsection Commands
1178 This filter supports the following commands:
1180 @item sample_noise, sn
1181 Start or stop measuring noise profile.
1182 Syntax for the command is : "start" or "stop" string.
1183 After measuring noise profile is stopped it will be
1184 automatically applied in filtering.
1186 @item noise_reduction, nr
1187 Change noise reduction. Argument is single float number.
1188 Syntax for the command is : "@var{noise_reduction}"
1190 @item noise_floor, nf
1191 Change noise floor. Argument is single float number.
1192 Syntax for the command is : "@var{noise_floor}"
1194 @item output_mode, om
1195 Change output mode operation.
1196 Syntax for the command is : "i", "o" or "n" string.
1200 Apply arbitrary expressions to samples in frequency domain.
1204 Set frequency domain real expression for each separate channel separated
1205 by '|'. Default is "re".
1206 If the number of input channels is greater than the number of
1207 expressions, the last specified expression is used for the remaining
1211 Set frequency domain imaginary expression for each separate channel
1212 separated by '|'. Default is "im".
1214 Each expression in @var{real} and @var{imag} can contain the following
1215 constants and functions:
1222 current frequency bin number
1225 number of available bins
1228 channel number of the current expression
1237 current real part of frequency bin of current channel
1240 current imaginary part of frequency bin of current channel
1243 Return the value of real part of frequency bin at location (@var{bin},@var{channel})
1246 Return the value of imaginary part of frequency bin at location (@var{bin},@var{channel})
1250 Set window size. Allowed range is from 16 to 131072.
1251 Default is @code{4096}
1254 Set window function. Default is @code{hann}.
1257 Set window overlap. If set to 1, the recommended overlap for selected
1258 window function will be picked. Default is @code{0.75}.
1261 @subsection Examples
1265 Leave almost only low frequencies in audio:
1267 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1271 Apply robotize effect:
1273 afftfilt="real='hypot(re,im)*sin(0)':imag='hypot(re,im)*cos(0)':win_size=512:overlap=0.75"
1277 Apply whisper effect:
1279 afftfilt="real='hypot(re,im)*cos((random(0)*2-1)*2*3.14)':imag='hypot(re,im)*sin((random(1)*2-1)*2*3.14)':win_size=128:overlap=0.8"
1286 Apply an arbitrary Finite Impulse Response filter.
1288 This filter is designed for applying long FIR filters,
1289 up to 60 seconds long.
1291 It can be used as component for digital crossover filters,
1292 room equalization, cross talk cancellation, wavefield synthesis,
1293 auralization, ambiophonics, ambisonics and spatialization.
1295 This filter uses the streams higher than first one as FIR coefficients.
1296 If the non-first stream holds a single channel, it will be used
1297 for all input channels in the first stream, otherwise
1298 the number of channels in the non-first stream must be same as
1299 the number of channels in the first stream.
1301 It accepts the following parameters:
1305 Set dry gain. This sets input gain.
1308 Set wet gain. This sets final output gain.
1311 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1314 Enable applying gain measured from power of IR.
1316 Set which approach to use for auto gain measurement.
1320 Do not apply any gain.
1323 select peak gain, very conservative approach. This is default value.
1326 select DC gain, limited application.
1329 select gain to noise approach, this is most popular one.
1333 Set gain to be applied to IR coefficients before filtering.
1334 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1337 Set format of IR stream. Can be @code{mono} or @code{input}.
1338 Default is @code{input}.
1341 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1342 Allowed range is 0.1 to 60 seconds.
1345 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1346 By default it is disabled.
1349 Set for which IR channel to display frequency response. By default is first channel
1350 displayed. This option is used only when @var{response} is enabled.
1353 Set video stream size. This option is used only when @var{response} is enabled.
1356 Set video stream frame rate. This option is used only when @var{response} is enabled.
1359 Set minimal partition size used for convolution. Default is @var{8192}.
1360 Allowed range is from @var{1} to @var{32768}.
1361 Lower values decreases latency at cost of higher CPU usage.
1364 Set maximal partition size used for convolution. Default is @var{8192}.
1365 Allowed range is from @var{8} to @var{32768}.
1366 Lower values may increase CPU usage.
1369 Set number of input impulse responses streams which will be switchable at runtime.
1370 Allowed range is from @var{1} to @var{32}. Default is @var{1}.
1373 Set IR stream which will be used for convolution, starting from @var{0}, should always be
1374 lower than supplied value by @code{nbirs} option. Default is @var{0}.
1375 This option can be changed at runtime via @ref{commands}.
1378 @subsection Examples
1382 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1384 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1391 Set output format constraints for the input audio. The framework will
1392 negotiate the most appropriate format to minimize conversions.
1394 It accepts the following parameters:
1397 @item sample_fmts, f
1398 A '|'-separated list of requested sample formats.
1400 @item sample_rates, r
1401 A '|'-separated list of requested sample rates.
1403 @item channel_layouts, cl
1404 A '|'-separated list of requested channel layouts.
1406 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1407 for the required syntax.
1410 If a parameter is omitted, all values are allowed.
1412 Force the output to either unsigned 8-bit or signed 16-bit stereo
1414 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1418 Apply frequency shift to input audio samples.
1420 The filter accepts the following options:
1424 Specify frequency shift. Allowed range is -INT_MAX to INT_MAX.
1425 Default value is 0.0.
1428 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
1429 Default value is 1.0.
1432 @subsection Commands
1434 This filter supports the all above options as @ref{commands}.
1438 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1439 processing reduces disturbing noise between useful signals.
1441 Gating is done by detecting the volume below a chosen level @var{threshold}
1442 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1443 floor is set via @var{range}. Because an exact manipulation of the signal
1444 would cause distortion of the waveform the reduction can be levelled over
1445 time. This is done by setting @var{attack} and @var{release}.
1447 @var{attack} determines how long the signal has to fall below the threshold
1448 before any reduction will occur and @var{release} sets the time the signal
1449 has to rise above the threshold to reduce the reduction again.
1450 Shorter signals than the chosen attack time will be left untouched.
1454 Set input level before filtering.
1455 Default is 1. Allowed range is from 0.015625 to 64.
1458 Set the mode of operation. Can be @code{upward} or @code{downward}.
1459 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
1460 will be amplified, expanding dynamic range in upward direction.
1461 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
1464 Set the level of gain reduction when the signal is below the threshold.
1465 Default is 0.06125. Allowed range is from 0 to 1.
1466 Setting this to 0 disables reduction and then filter behaves like expander.
1469 If a signal rises above this level the gain reduction is released.
1470 Default is 0.125. Allowed range is from 0 to 1.
1473 Set a ratio by which the signal is reduced.
1474 Default is 2. Allowed range is from 1 to 9000.
1477 Amount of milliseconds the signal has to rise above the threshold before gain
1479 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1482 Amount of milliseconds the signal has to fall below the threshold before the
1483 reduction is increased again. Default is 250 milliseconds.
1484 Allowed range is from 0.01 to 9000.
1487 Set amount of amplification of signal after processing.
1488 Default is 1. Allowed range is from 1 to 64.
1491 Curve the sharp knee around the threshold to enter gain reduction more softly.
1492 Default is 2.828427125. Allowed range is from 1 to 8.
1495 Choose if exact signal should be taken for detection or an RMS like one.
1496 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1499 Choose if the average level between all channels or the louder channel affects
1501 Default is @code{average}. Can be @code{average} or @code{maximum}.
1504 @subsection Commands
1506 This filter supports the all above options as @ref{commands}.
1510 Apply an arbitrary Infinite Impulse Response filter.
1512 It accepts the following parameters:
1516 Set B/numerator/zeros/reflection coefficients.
1519 Set A/denominator/poles/ladder coefficients.
1531 Set coefficients format.
1535 lattice-ladder function
1537 analog transfer function
1539 digital transfer function
1541 Z-plane zeros/poles, cartesian (default)
1543 Z-plane zeros/poles, polar radians
1545 Z-plane zeros/poles, polar degrees
1551 Set type of processing.
1563 Set filtering precision.
1567 double-precision floating-point (default)
1569 single-precision floating-point
1577 Normalize filter coefficients, by default is enabled.
1578 Enabling it will normalize magnitude response at DC to 0dB.
1581 How much to use filtered signal in output. Default is 1.
1582 Range is between 0 and 1.
1585 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1586 By default it is disabled.
1589 Set for which IR channel to display frequency response. By default is first channel
1590 displayed. This option is used only when @var{response} is enabled.
1593 Set video stream size. This option is used only when @var{response} is enabled.
1596 Coefficients in @code{tf} and @code{sf} format are separated by spaces and are in ascending
1599 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1600 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1603 Different coefficients and gains can be provided for every channel, in such case
1604 use '|' to separate coefficients or gains. Last provided coefficients will be
1605 used for all remaining channels.
1607 @subsection Examples
1611 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1613 aiir=k=1:z=7.957584807809675810E-1 -2.575128568908332300 3.674839853930788710 -2.57512875289799137 7.957586296317130880E-1:p=1 -2.86950072432325953 3.63022088054647218 -2.28075678147272232 6.361362326477423500E-1:f=tf:r=d
1617 Same as above but in @code{zp} format:
1619 aiir=k=0.79575848078096756:z=0.80918701+0.58773007i 0.80918701-0.58773007i 0.80884700+0.58784055i 0.80884700-0.58784055i:p=0.63892345+0.59951235i 0.63892345-0.59951235i 0.79582691+0.44198673i 0.79582691-0.44198673i:f=zp:r=s
1623 Apply 3-rd order analog normalized Butterworth low-pass filter, using analog transfer function format:
1625 aiir=z=1.3057 0 0 0:p=1.3057 2.3892 2.1860 1:f=sf:r=d
1631 The limiter prevents an input signal from rising over a desired threshold.
1632 This limiter uses lookahead technology to prevent your signal from distorting.
1633 It means that there is a small delay after the signal is processed. Keep in mind
1634 that the delay it produces is the attack time you set.
1636 The filter accepts the following options:
1640 Set input gain. Default is 1.
1643 Set output gain. Default is 1.
1646 Don't let signals above this level pass the limiter. Default is 1.
1649 The limiter will reach its attenuation level in this amount of time in
1650 milliseconds. Default is 5 milliseconds.
1653 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1654 Default is 50 milliseconds.
1657 When gain reduction is always needed ASC takes care of releasing to an
1658 average reduction level rather than reaching a reduction of 0 in the release
1662 Select how much the release time is affected by ASC, 0 means nearly no changes
1663 in release time while 1 produces higher release times.
1666 Auto level output signal. Default is enabled.
1667 This normalizes audio back to 0dB if enabled.
1670 Depending on picked setting it is recommended to upsample input 2x or 4x times
1671 with @ref{aresample} before applying this filter.
1675 Apply a two-pole all-pass filter with central frequency (in Hz)
1676 @var{frequency}, and filter-width @var{width}.
1677 An all-pass filter changes the audio's frequency to phase relationship
1678 without changing its frequency to amplitude relationship.
1680 The filter accepts the following options:
1684 Set frequency in Hz.
1687 Set method to specify band-width of filter.
1702 Specify the band-width of a filter in width_type units.
1705 How much to use filtered signal in output. Default is 1.
1706 Range is between 0 and 1.
1709 Specify which channels to filter, by default all available are filtered.
1712 Normalize biquad coefficients, by default is disabled.
1713 Enabling it will normalize magnitude response at DC to 0dB.
1716 Set the filter order, can be 1 or 2. Default is 2.
1719 Set transform type of IIR filter.
1728 Set precison of filtering.
1731 Pick automatic sample format depending on surround filters.
1733 Always use signed 16-bit.
1735 Always use signed 32-bit.
1737 Always use float 32-bit.
1739 Always use float 64-bit.
1743 @subsection Commands
1745 This filter supports the following commands:
1748 Change allpass frequency.
1749 Syntax for the command is : "@var{frequency}"
1752 Change allpass width_type.
1753 Syntax for the command is : "@var{width_type}"
1756 Change allpass width.
1757 Syntax for the command is : "@var{width}"
1761 Syntax for the command is : "@var{mix}"
1768 The filter accepts the following options:
1772 Set the number of loops. Setting this value to -1 will result in infinite loops.
1776 Set maximal number of samples. Default is 0.
1779 Set first sample of loop. Default is 0.
1785 Merge two or more audio streams into a single multi-channel stream.
1787 The filter accepts the following options:
1792 Set the number of inputs. Default is 2.
1796 If the channel layouts of the inputs are disjoint, and therefore compatible,
1797 the channel layout of the output will be set accordingly and the channels
1798 will be reordered as necessary. If the channel layouts of the inputs are not
1799 disjoint, the output will have all the channels of the first input then all
1800 the channels of the second input, in that order, and the channel layout of
1801 the output will be the default value corresponding to the total number of
1804 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1805 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1806 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1807 first input, b1 is the first channel of the second input).
1809 On the other hand, if both input are in stereo, the output channels will be
1810 in the default order: a1, a2, b1, b2, and the channel layout will be
1811 arbitrarily set to 4.0, which may or may not be the expected value.
1813 All inputs must have the same sample rate, and format.
1815 If inputs do not have the same duration, the output will stop with the
1818 @subsection Examples
1822 Merge two mono files into a stereo stream:
1824 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1828 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1830 ffmpeg -i input.mkv -filter_complex "[0:1][0:2][0:3][0:4][0:5][0:6] amerge=inputs=6" -c:a pcm_s16le output.mkv
1836 Mixes multiple audio inputs into a single output.
1838 Note that this filter only supports float samples (the @var{amerge}
1839 and @var{pan} audio filters support many formats). If the @var{amix}
1840 input has integer samples then @ref{aresample} will be automatically
1841 inserted to perform the conversion to float samples.
1845 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1847 will mix 3 input audio streams to a single output with the same duration as the
1848 first input and a dropout transition time of 3 seconds.
1850 It accepts the following parameters:
1854 The number of inputs. If unspecified, it defaults to 2.
1857 How to determine the end-of-stream.
1861 The duration of the longest input. (default)
1864 The duration of the shortest input.
1867 The duration of the first input.
1871 @item dropout_transition
1872 The transition time, in seconds, for volume renormalization when an input
1873 stream ends. The default value is 2 seconds.
1876 Specify weight of each input audio stream as sequence.
1877 Each weight is separated by space. By default all inputs have same weight.
1880 @subsection Commands
1882 This filter supports the following commands:
1885 Syntax is same as option with same name.
1890 Multiply first audio stream with second audio stream and store result
1891 in output audio stream. Multiplication is done by multiplying each
1892 sample from first stream with sample at same position from second stream.
1894 With this element-wise multiplication one can create amplitude fades and
1895 amplitude modulations.
1897 @section anequalizer
1899 High-order parametric multiband equalizer for each channel.
1901 It accepts the following parameters:
1905 This option string is in format:
1906 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1907 Each equalizer band is separated by '|'.
1911 Set channel number to which equalization will be applied.
1912 If input doesn't have that channel the entry is ignored.
1915 Set central frequency for band.
1916 If input doesn't have that frequency the entry is ignored.
1919 Set band width in Hertz.
1922 Set band gain in dB.
1925 Set filter type for band, optional, can be:
1929 Butterworth, this is default.
1940 With this option activated frequency response of anequalizer is displayed
1944 Set video stream size. Only useful if curves option is activated.
1947 Set max gain that will be displayed. Only useful if curves option is activated.
1948 Setting this to a reasonable value makes it possible to display gain which is derived from
1949 neighbour bands which are too close to each other and thus produce higher gain
1950 when both are activated.
1953 Set frequency scale used to draw frequency response in video output.
1954 Can be linear or logarithmic. Default is logarithmic.
1957 Set color for each channel curve which is going to be displayed in video stream.
1958 This is list of color names separated by space or by '|'.
1959 Unrecognised or missing colors will be replaced by white color.
1962 @subsection Examples
1966 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1967 for first 2 channels using Chebyshev type 1 filter:
1969 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1973 @subsection Commands
1975 This filter supports the following commands:
1978 Alter existing filter parameters.
1979 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1981 @var{fN} is existing filter number, starting from 0, if no such filter is available
1983 @var{freq} set new frequency parameter.
1984 @var{width} set new width parameter in Hertz.
1985 @var{gain} set new gain parameter in dB.
1987 Full filter invocation with asendcmd may look like this:
1988 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1993 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1995 Each sample is adjusted by looking for other samples with similar contexts. This
1996 context similarity is defined by comparing their surrounding patches of size
1997 @option{p}. Patches are searched in an area of @option{r} around the sample.
1999 The filter accepts the following options:
2003 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
2006 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
2007 Default value is 2 milliseconds.
2010 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
2011 Default value is 6 milliseconds.
2014 Set the output mode.
2016 It accepts the following values:
2019 Pass input unchanged.
2022 Pass noise filtered out.
2027 Default value is @var{o}.
2031 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
2034 @subsection Commands
2036 This filter supports the all above options as @ref{commands}.
2039 Apply Normalized Least-Mean-Squares algorithm to the first audio stream using the second audio stream.
2041 This adaptive filter is used to mimic a desired filter by finding the filter coefficients that
2042 relate to producing the least mean square of the error signal (difference between the desired,
2043 2nd input audio stream and the actual signal, the 1st input audio stream).
2045 A description of the accepted options follows.
2058 Set the filter leakage.
2061 It accepts the following values:
2070 Pass filtered samples.
2073 Pass difference between desired and filtered samples.
2075 Default value is @var{o}.
2079 @subsection Examples
2083 One of many usages of this filter is noise reduction, input audio is filtered
2084 with same samples that are delayed by fixed amount, one such example for stereo audio is:
2086 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
2090 @subsection Commands
2092 This filter supports the same commands as options, excluding option @code{order}.
2096 Pass the audio source unchanged to the output.
2100 Pad the end of an audio stream with silence.
2102 This can be used together with @command{ffmpeg} @option{-shortest} to
2103 extend audio streams to the same length as the video stream.
2105 A description of the accepted options follows.
2109 Set silence packet size. Default value is 4096.
2112 Set the number of samples of silence to add to the end. After the
2113 value is reached, the stream is terminated. This option is mutually
2114 exclusive with @option{whole_len}.
2117 Set the minimum total number of samples in the output audio stream. If
2118 the value is longer than the input audio length, silence is added to
2119 the end, until the value is reached. This option is mutually exclusive
2120 with @option{pad_len}.
2123 Specify the duration of samples of silence to add. See
2124 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2125 for the accepted syntax. Used only if set to non-zero value.
2128 Specify the minimum total duration in the output audio stream. See
2129 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2130 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
2131 the input audio length, silence is added to the end, until the value is reached.
2132 This option is mutually exclusive with @option{pad_dur}
2135 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
2136 nor @option{whole_dur} option is set, the filter will add silence to the end of
2137 the input stream indefinitely.
2139 @subsection Examples
2143 Add 1024 samples of silence to the end of the input:
2149 Make sure the audio output will contain at least 10000 samples, pad
2150 the input with silence if required:
2152 apad=whole_len=10000
2156 Use @command{ffmpeg} to pad the audio input with silence, so that the
2157 video stream will always result the shortest and will be converted
2158 until the end in the output file when using the @option{shortest}
2161 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
2166 Add a phasing effect to the input audio.
2168 A phaser filter creates series of peaks and troughs in the frequency spectrum.
2169 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
2171 A description of the accepted parameters follows.
2175 Set input gain. Default is 0.4.
2178 Set output gain. Default is 0.74
2181 Set delay in milliseconds. Default is 3.0.
2184 Set decay. Default is 0.4.
2187 Set modulation speed in Hz. Default is 0.5.
2190 Set modulation type. Default is triangular.
2192 It accepts the following values:
2199 @section aphaseshift
2200 Apply phase shift to input audio samples.
2202 The filter accepts the following options:
2206 Specify phase shift. Allowed range is from -1.0 to 1.0.
2207 Default value is 0.0.
2210 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
2211 Default value is 1.0.
2214 @subsection Commands
2216 This filter supports the all above options as @ref{commands}.
2220 Audio pulsator is something between an autopanner and a tremolo.
2221 But it can produce funny stereo effects as well. Pulsator changes the volume
2222 of the left and right channel based on a LFO (low frequency oscillator) with
2223 different waveforms and shifted phases.
2224 This filter have the ability to define an offset between left and right
2225 channel. An offset of 0 means that both LFO shapes match each other.
2226 The left and right channel are altered equally - a conventional tremolo.
2227 An offset of 50% means that the shape of the right channel is exactly shifted
2228 in phase (or moved backwards about half of the frequency) - pulsator acts as
2229 an autopanner. At 1 both curves match again. Every setting in between moves the
2230 phase shift gapless between all stages and produces some "bypassing" sounds with
2231 sine and triangle waveforms. The more you set the offset near 1 (starting from
2232 the 0.5) the faster the signal passes from the left to the right speaker.
2234 The filter accepts the following options:
2238 Set input gain. By default it is 1. Range is [0.015625 - 64].
2241 Set output gain. By default it is 1. Range is [0.015625 - 64].
2244 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
2245 sawup or sawdown. Default is sine.
2248 Set modulation. Define how much of original signal is affected by the LFO.
2251 Set left channel offset. Default is 0. Allowed range is [0 - 1].
2254 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
2257 Set pulse width. Default is 1. Allowed range is [0 - 2].
2260 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
2263 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
2267 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
2271 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
2272 if timing is set to hz.
2278 Resample the input audio to the specified parameters, using the
2279 libswresample library. If none are specified then the filter will
2280 automatically convert between its input and output.
2282 This filter is also able to stretch/squeeze the audio data to make it match
2283 the timestamps or to inject silence / cut out audio to make it match the
2284 timestamps, do a combination of both or do neither.
2286 The filter accepts the syntax
2287 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
2288 expresses a sample rate and @var{resampler_options} is a list of
2289 @var{key}=@var{value} pairs, separated by ":". See the
2290 @ref{Resampler Options,,"Resampler Options" section in the
2291 ffmpeg-resampler(1) manual,ffmpeg-resampler}
2292 for the complete list of supported options.
2294 @subsection Examples
2298 Resample the input audio to 44100Hz:
2304 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2305 samples per second compensation:
2307 aresample=async=1000
2313 Reverse an audio clip.
2315 Warning: This filter requires memory to buffer the entire clip, so trimming
2318 @subsection Examples
2322 Take the first 5 seconds of a clip, and reverse it.
2324 atrim=end=5,areverse
2330 Reduce noise from speech using Recurrent Neural Networks.
2332 This filter accepts the following options:
2336 Set train model file to load. This option is always required.
2339 Set how much to mix filtered samples into final output.
2340 Allowed range is from -1 to 1. Default value is 1.
2341 Negative values are special, they set how much to keep filtered noise
2342 in the final filter output. Set this option to -1 to hear actual
2343 noise removed from input signal.
2346 @section asetnsamples
2348 Set the number of samples per each output audio frame.
2350 The last output packet may contain a different number of samples, as
2351 the filter will flush all the remaining samples when the input audio
2354 The filter accepts the following options:
2358 @item nb_out_samples, n
2359 Set the number of frames per each output audio frame. The number is
2360 intended as the number of samples @emph{per each channel}.
2361 Default value is 1024.
2364 If set to 1, the filter will pad the last audio frame with zeroes, so
2365 that the last frame will contain the same number of samples as the
2366 previous ones. Default value is 1.
2369 For example, to set the number of per-frame samples to 1234 and
2370 disable padding for the last frame, use:
2372 asetnsamples=n=1234:p=0
2377 Set the sample rate without altering the PCM data.
2378 This will result in a change of speed and pitch.
2380 The filter accepts the following options:
2383 @item sample_rate, r
2384 Set the output sample rate. Default is 44100 Hz.
2389 Show a line containing various information for each input audio frame.
2390 The input audio is not modified.
2392 The shown line contains a sequence of key/value pairs of the form
2393 @var{key}:@var{value}.
2395 The following values are shown in the output:
2399 The (sequential) number of the input frame, starting from 0.
2402 The presentation timestamp of the input frame, in time base units; the time base
2403 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2406 The presentation timestamp of the input frame in seconds.
2409 position of the frame in the input stream, -1 if this information in
2410 unavailable and/or meaningless (for example in case of synthetic audio)
2419 The sample rate for the audio frame.
2422 The number of samples (per channel) in the frame.
2425 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2426 audio, the data is treated as if all the planes were concatenated.
2428 @item plane_checksums
2429 A list of Adler-32 checksums for each data plane.
2433 Apply audio soft clipping.
2435 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2436 along a smooth curve, rather than the abrupt shape of hard-clipping.
2438 This filter accepts the following options:
2442 Set type of soft-clipping.
2444 It accepts the following values:
2458 Set threshold from where to start clipping. Default value is 0dB or 1.
2461 Set gain applied to output. Default value is 0dB or 1.
2464 Set additional parameter which controls sigmoid function.
2467 Set oversampling factor.
2470 @subsection Commands
2472 This filter supports the all above options as @ref{commands}.
2475 Automatic Speech Recognition
2477 This filter uses PocketSphinx for speech recognition. To enable
2478 compilation of this filter, you need to configure FFmpeg with
2479 @code{--enable-pocketsphinx}.
2481 It accepts the following options:
2485 Set sampling rate of input audio. Defaults is @code{16000}.
2486 This need to match speech models, otherwise one will get poor results.
2489 Set dictionary containing acoustic model files.
2492 Set pronunciation dictionary.
2495 Set language model file.
2498 Set language model set.
2501 Set which language model to use.
2504 Set output for log messages.
2507 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2512 Display time domain statistical information about the audio channels.
2513 Statistics are calculated and displayed for each audio channel and,
2514 where applicable, an overall figure is also given.
2516 It accepts the following option:
2519 Short window length in seconds, used for peak and trough RMS measurement.
2520 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2524 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2525 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2528 Available keys for each channel are:
2574 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2575 this @code{lavfi.astats.Overall.Peak_count}.
2577 For description what each key means read below.
2580 Set number of frame after which stats are going to be recalculated.
2581 Default is disabled.
2583 @item measure_perchannel
2584 Select the entries which need to be measured per channel. The metadata keys can
2585 be used as flags, default is @option{all} which measures everything.
2586 @option{none} disables all per channel measurement.
2588 @item measure_overall
2589 Select the entries which need to be measured overall. The metadata keys can
2590 be used as flags, default is @option{all} which measures everything.
2591 @option{none} disables all overall measurement.
2595 A description of each shown parameter follows:
2599 Mean amplitude displacement from zero.
2602 Minimal sample level.
2605 Maximal sample level.
2607 @item Min difference
2608 Minimal difference between two consecutive samples.
2610 @item Max difference
2611 Maximal difference between two consecutive samples.
2613 @item Mean difference
2614 Mean difference between two consecutive samples.
2615 The average of each difference between two consecutive samples.
2617 @item RMS difference
2618 Root Mean Square difference between two consecutive samples.
2622 Standard peak and RMS level measured in dBFS.
2626 Peak and trough values for RMS level measured over a short window.
2629 Standard ratio of peak to RMS level (note: not in dB).
2632 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2633 (i.e. either @var{Min level} or @var{Max level}).
2636 Number of occasions (not the number of samples) that the signal attained either
2637 @var{Min level} or @var{Max level}.
2639 @item Noise floor dB
2640 Minimum local peak measured in dBFS over a short window.
2642 @item Noise floor count
2643 Number of occasions (not the number of samples) that the signal attained
2647 Overall bit depth of audio. Number of bits used for each sample.
2650 Measured dynamic range of audio in dB.
2652 @item Zero crossings
2653 Number of points where the waveform crosses the zero level axis.
2655 @item Zero crossings rate
2656 Rate of Zero crossings and number of audio samples.
2660 Boost subwoofer frequencies.
2662 The filter accepts the following options:
2666 Set dry gain, how much of original signal is kept. Allowed range is from 0 to 1.
2667 Default value is 0.7.
2670 Set wet gain, how much of filtered signal is kept. Allowed range is from 0 to 1.
2671 Default value is 0.7.
2674 Set delay line decay gain value. Allowed range is from 0 to 1.
2675 Default value is 0.7.
2678 Set delay line feedback gain value. Allowed range is from 0 to 1.
2679 Default value is 0.9.
2682 Set cutoff frequency in Hertz. Allowed range is 50 to 900.
2683 Default value is 100.
2686 Set slope amount for cutoff frequency. Allowed range is 0.0001 to 1.
2687 Default value is 0.5.
2690 Set delay. Allowed range is from 1 to 100.
2691 Default value is 20.
2694 @subsection Commands
2696 This filter supports the all above options as @ref{commands}.
2699 Cut subwoofer frequencies.
2701 This filter allows to set custom, steeper
2702 roll off than highpass filter, and thus is able to more attenuate
2703 frequency content in stop-band.
2705 The filter accepts the following options:
2709 Set cutoff frequency in Hertz. Allowed range is 2 to 200.
2710 Default value is 20.
2713 Set filter order. Available values are from 3 to 20.
2714 Default value is 10.
2717 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2720 @subsection Commands
2722 This filter supports the all above options as @ref{commands}.
2725 Cut super frequencies.
2727 The filter accepts the following options:
2731 Set cutoff frequency in Hertz. Allowed range is 20000 to 192000.
2732 Default value is 20000.
2735 Set filter order. Available values are from 3 to 20.
2736 Default value is 10.
2739 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2742 @subsection Commands
2744 This filter supports the all above options as @ref{commands}.
2747 Apply high order Butterworth band-pass filter.
2749 The filter accepts the following options:
2753 Set center frequency in Hertz. Allowed range is 2 to 999999.
2754 Default value is 1000.
2757 Set filter order. Available values are from 4 to 20.
2761 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2764 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2767 @subsection Commands
2769 This filter supports the all above options as @ref{commands}.
2772 Apply high order Butterworth band-stop filter.
2774 The filter accepts the following options:
2778 Set center frequency in Hertz. Allowed range is 2 to 999999.
2779 Default value is 1000.
2782 Set filter order. Available values are from 4 to 20.
2786 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2789 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2792 @subsection Commands
2794 This filter supports the all above options as @ref{commands}.
2800 The filter accepts exactly one parameter, the audio tempo. If not
2801 specified then the filter will assume nominal 1.0 tempo. Tempo must
2802 be in the [0.5, 100.0] range.
2804 Note that tempo greater than 2 will skip some samples rather than
2805 blend them in. If for any reason this is a concern it is always
2806 possible to daisy-chain several instances of atempo to achieve the
2807 desired product tempo.
2809 @subsection Examples
2813 Slow down audio to 80% tempo:
2819 To speed up audio to 300% tempo:
2825 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2827 atempo=sqrt(3),atempo=sqrt(3)
2831 @subsection Commands
2833 This filter supports the following commands:
2836 Change filter tempo scale factor.
2837 Syntax for the command is : "@var{tempo}"
2842 Trim the input so that the output contains one continuous subpart of the input.
2844 It accepts the following parameters:
2847 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2848 sample with the timestamp @var{start} will be the first sample in the output.
2851 Specify time of the first audio sample that will be dropped, i.e. the
2852 audio sample immediately preceding the one with the timestamp @var{end} will be
2853 the last sample in the output.
2856 Same as @var{start}, except this option sets the start timestamp in samples
2860 Same as @var{end}, except this option sets the end timestamp in samples instead
2864 The maximum duration of the output in seconds.
2867 The number of the first sample that should be output.
2870 The number of the first sample that should be dropped.
2873 @option{start}, @option{end}, and @option{duration} are expressed as time
2874 duration specifications; see
2875 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2877 Note that the first two sets of the start/end options and the @option{duration}
2878 option look at the frame timestamp, while the _sample options simply count the
2879 samples that pass through the filter. So start/end_pts and start/end_sample will
2880 give different results when the timestamps are wrong, inexact or do not start at
2881 zero. Also note that this filter does not modify the timestamps. If you wish
2882 to have the output timestamps start at zero, insert the asetpts filter after the
2885 If multiple start or end options are set, this filter tries to be greedy and
2886 keep all samples that match at least one of the specified constraints. To keep
2887 only the part that matches all the constraints at once, chain multiple atrim
2890 The defaults are such that all the input is kept. So it is possible to set e.g.
2891 just the end values to keep everything before the specified time.
2896 Drop everything except the second minute of input:
2898 ffmpeg -i INPUT -af atrim=60:120
2902 Keep only the first 1000 samples:
2904 ffmpeg -i INPUT -af atrim=end_sample=1000
2909 @section axcorrelate
2910 Calculate normalized cross-correlation between two input audio streams.
2912 Resulted samples are always between -1 and 1 inclusive.
2913 If result is 1 it means two input samples are highly correlated in that selected segment.
2914 Result 0 means they are not correlated at all.
2915 If result is -1 it means two input samples are out of phase, which means they cancel each
2918 The filter accepts the following options:
2922 Set size of segment over which cross-correlation is calculated.
2923 Default is 256. Allowed range is from 2 to 131072.
2926 Set algorithm for cross-correlation. Can be @code{slow} or @code{fast}.
2927 Default is @code{slow}. Fast algorithm assumes mean values over any given segment
2928 are always zero and thus need much less calculations to make.
2929 This is generally not true, but is valid for typical audio streams.
2932 @subsection Examples
2936 Calculate correlation between channels in stereo audio stream:
2938 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
2944 Apply a two-pole Butterworth band-pass filter with central
2945 frequency @var{frequency}, and (3dB-point) band-width width.
2946 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2947 instead of the default: constant 0dB peak gain.
2948 The filter roll off at 6dB per octave (20dB per decade).
2950 The filter accepts the following options:
2954 Set the filter's central frequency. Default is @code{3000}.
2957 Constant skirt gain if set to 1. Defaults to 0.
2960 Set method to specify band-width of filter.
2975 Specify the band-width of a filter in width_type units.
2978 How much to use filtered signal in output. Default is 1.
2979 Range is between 0 and 1.
2982 Specify which channels to filter, by default all available are filtered.
2985 Normalize biquad coefficients, by default is disabled.
2986 Enabling it will normalize magnitude response at DC to 0dB.
2989 Set transform type of IIR filter.
2998 Set precison of filtering.
3001 Pick automatic sample format depending on surround filters.
3003 Always use signed 16-bit.
3005 Always use signed 32-bit.
3007 Always use float 32-bit.
3009 Always use float 64-bit.
3013 @subsection Commands
3015 This filter supports the following commands:
3018 Change bandpass frequency.
3019 Syntax for the command is : "@var{frequency}"
3022 Change bandpass width_type.
3023 Syntax for the command is : "@var{width_type}"
3026 Change bandpass width.
3027 Syntax for the command is : "@var{width}"
3030 Change bandpass mix.
3031 Syntax for the command is : "@var{mix}"
3036 Apply a two-pole Butterworth band-reject filter with central
3037 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
3038 The filter roll off at 6dB per octave (20dB per decade).
3040 The filter accepts the following options:
3044 Set the filter's central frequency. Default is @code{3000}.
3047 Set method to specify band-width of filter.
3062 Specify the band-width of a filter in width_type units.
3065 How much to use filtered signal in output. Default is 1.
3066 Range is between 0 and 1.
3069 Specify which channels to filter, by default all available are filtered.
3072 Normalize biquad coefficients, by default is disabled.
3073 Enabling it will normalize magnitude response at DC to 0dB.
3076 Set transform type of IIR filter.
3085 Set precison of filtering.
3088 Pick automatic sample format depending on surround filters.
3090 Always use signed 16-bit.
3092 Always use signed 32-bit.
3094 Always use float 32-bit.
3096 Always use float 64-bit.
3100 @subsection Commands
3102 This filter supports the following commands:
3105 Change bandreject frequency.
3106 Syntax for the command is : "@var{frequency}"
3109 Change bandreject width_type.
3110 Syntax for the command is : "@var{width_type}"
3113 Change bandreject width.
3114 Syntax for the command is : "@var{width}"
3117 Change bandreject mix.
3118 Syntax for the command is : "@var{mix}"
3121 @section bass, lowshelf
3123 Boost or cut the bass (lower) frequencies of the audio using a two-pole
3124 shelving filter with a response similar to that of a standard
3125 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
3127 The filter accepts the following options:
3131 Give the gain at 0 Hz. Its useful range is about -20
3132 (for a large cut) to +20 (for a large boost).
3133 Beware of clipping when using a positive gain.
3136 Set the filter's central frequency and so can be used
3137 to extend or reduce the frequency range to be boosted or cut.
3138 The default value is @code{100} Hz.
3141 Set method to specify band-width of filter.
3156 Determine how steep is the filter's shelf transition.
3159 Set number of poles. Default is 2.
3162 How much to use filtered signal in output. Default is 1.
3163 Range is between 0 and 1.
3166 Specify which channels to filter, by default all available are filtered.
3169 Normalize biquad coefficients, by default is disabled.
3170 Enabling it will normalize magnitude response at DC to 0dB.
3173 Set transform type of IIR filter.
3182 Set precison of filtering.
3185 Pick automatic sample format depending on surround filters.
3187 Always use signed 16-bit.
3189 Always use signed 32-bit.
3191 Always use float 32-bit.
3193 Always use float 64-bit.
3197 @subsection Commands
3199 This filter supports the following commands:
3202 Change bass frequency.
3203 Syntax for the command is : "@var{frequency}"
3206 Change bass width_type.
3207 Syntax for the command is : "@var{width_type}"
3211 Syntax for the command is : "@var{width}"
3215 Syntax for the command is : "@var{gain}"
3219 Syntax for the command is : "@var{mix}"
3224 Apply a biquad IIR filter with the given coefficients.
3225 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
3226 are the numerator and denominator coefficients respectively.
3227 and @var{channels}, @var{c} specify which channels to filter, by default all
3228 available are filtered.
3230 @subsection Commands
3232 This filter supports the following commands:
3240 Change biquad parameter.
3241 Syntax for the command is : "@var{value}"
3244 How much to use filtered signal in output. Default is 1.
3245 Range is between 0 and 1.
3248 Specify which channels to filter, by default all available are filtered.
3251 Normalize biquad coefficients, by default is disabled.
3252 Enabling it will normalize magnitude response at DC to 0dB.
3255 Set transform type of IIR filter.
3264 Set precison of filtering.
3267 Pick automatic sample format depending on surround filters.
3269 Always use signed 16-bit.
3271 Always use signed 32-bit.
3273 Always use float 32-bit.
3275 Always use float 64-bit.
3280 Bauer stereo to binaural transformation, which improves headphone listening of
3281 stereo audio records.
3283 To enable compilation of this filter you need to configure FFmpeg with
3284 @code{--enable-libbs2b}.
3286 It accepts the following parameters:
3290 Pre-defined crossfeed level.
3294 Default level (fcut=700, feed=50).
3297 Chu Moy circuit (fcut=700, feed=60).
3300 Jan Meier circuit (fcut=650, feed=95).
3305 Cut frequency (in Hz).
3314 Remap input channels to new locations.
3316 It accepts the following parameters:
3319 Map channels from input to output. The argument is a '|'-separated list of
3320 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
3321 @var{in_channel} form. @var{in_channel} can be either the name of the input
3322 channel (e.g. FL for front left) or its index in the input channel layout.
3323 @var{out_channel} is the name of the output channel or its index in the output
3324 channel layout. If @var{out_channel} is not given then it is implicitly an
3325 index, starting with zero and increasing by one for each mapping.
3327 @item channel_layout
3328 The channel layout of the output stream.
3331 If no mapping is present, the filter will implicitly map input channels to
3332 output channels, preserving indices.
3334 @subsection Examples
3338 For example, assuming a 5.1+downmix input MOV file,
3340 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
3342 will create an output WAV file tagged as stereo from the downmix channels of
3346 To fix a 5.1 WAV improperly encoded in AAC's native channel order
3348 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
3352 @section channelsplit
3354 Split each channel from an input audio stream into a separate output stream.
3356 It accepts the following parameters:
3358 @item channel_layout
3359 The channel layout of the input stream. The default is "stereo".
3361 A channel layout describing the channels to be extracted as separate output streams
3362 or "all" to extract each input channel as a separate stream. The default is "all".
3364 Choosing channels not present in channel layout in the input will result in an error.
3367 @subsection Examples
3371 For example, assuming a stereo input MP3 file,
3373 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
3375 will create an output Matroska file with two audio streams, one containing only
3376 the left channel and the other the right channel.
3379 Split a 5.1 WAV file into per-channel files:
3381 ffmpeg -i in.wav -filter_complex
3382 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
3383 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
3384 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
3389 Extract only LFE from a 5.1 WAV file:
3391 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
3392 -map '[LFE]' lfe.wav
3397 Add a chorus effect to the audio.
3399 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
3401 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
3402 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
3403 The modulation depth defines the range the modulated delay is played before or after
3404 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
3405 sound tuned around the original one, like in a chorus where some vocals are slightly
3408 It accepts the following parameters:
3411 Set input gain. Default is 0.4.
3414 Set output gain. Default is 0.4.
3417 Set delays. A typical delay is around 40ms to 60ms.
3429 @subsection Examples
3435 chorus=0.7:0.9:55:0.4:0.25:2
3441 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
3445 Fuller sounding chorus with three delays:
3447 chorus=0.5:0.9:50|60|40:0.4|0.32|0.3:0.25|0.4|0.3:2|2.3|1.3
3452 Compress or expand the audio's dynamic range.
3454 It accepts the following parameters:
3460 A list of times in seconds for each channel over which the instantaneous level
3461 of the input signal is averaged to determine its volume. @var{attacks} refers to
3462 increase of volume and @var{decays} refers to decrease of volume. For most
3463 situations, the attack time (response to the audio getting louder) should be
3464 shorter than the decay time, because the human ear is more sensitive to sudden
3465 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
3466 a typical value for decay is 0.8 seconds.
3467 If specified number of attacks & decays is lower than number of channels, the last
3468 set attack/decay will be used for all remaining channels.
3471 A list of points for the transfer function, specified in dB relative to the
3472 maximum possible signal amplitude. Each key points list must be defined using
3473 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
3474 @code{x0/y0 x1/y1 x2/y2 ....}
3476 The input values must be in strictly increasing order but the transfer function
3477 does not have to be monotonically rising. The point @code{0/0} is assumed but
3478 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
3479 function are @code{-70/-70|-60/-20|1/0}.
3482 Set the curve radius in dB for all joints. It defaults to 0.01.
3485 Set the additional gain in dB to be applied at all points on the transfer
3486 function. This allows for easy adjustment of the overall gain.
3490 Set an initial volume, in dB, to be assumed for each channel when filtering
3491 starts. This permits the user to supply a nominal level initially, so that, for
3492 example, a very large gain is not applied to initial signal levels before the
3493 companding has begun to operate. A typical value for audio which is initially
3494 quiet is -90 dB. It defaults to 0.
3497 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
3498 delayed before being fed to the volume adjuster. Specifying a delay
3499 approximately equal to the attack/decay times allows the filter to effectively
3500 operate in predictive rather than reactive mode. It defaults to 0.
3504 @subsection Examples
3508 Make music with both quiet and loud passages suitable for listening to in a
3511 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
3514 Another example for audio with whisper and explosion parts:
3516 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
3520 A noise gate for when the noise is at a lower level than the signal:
3522 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
3526 Here is another noise gate, this time for when the noise is at a higher level
3527 than the signal (making it, in some ways, similar to squelch):
3529 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3533 2:1 compression starting at -6dB:
3535 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3539 2:1 compression starting at -9dB:
3541 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3545 2:1 compression starting at -12dB:
3547 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3551 2:1 compression starting at -18dB:
3553 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3557 3:1 compression starting at -15dB:
3559 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3565 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3571 compand=attacks=0:points=-80/-169|-54/-80|-49.5/-64.6|-41.1/-41.1|-25.8/-15|-10.8/-4.5|0/0|20/8.3
3575 Hard limiter at -6dB:
3577 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3581 Hard limiter at -12dB:
3583 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3587 Hard noise gate at -35 dB:
3589 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3595 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3599 @section compensationdelay
3601 Compensation Delay Line is a metric based delay to compensate differing
3602 positions of microphones or speakers.
3604 For example, you have recorded guitar with two microphones placed in
3605 different locations. Because the front of sound wave has fixed speed in
3606 normal conditions, the phasing of microphones can vary and depends on
3607 their location and interposition. The best sound mix can be achieved when
3608 these microphones are in phase (synchronized). Note that a distance of
3609 ~30 cm between microphones makes one microphone capture the signal in
3610 antiphase to the other microphone. That makes the final mix sound moody.
3611 This filter helps to solve phasing problems by adding different delays
3612 to each microphone track and make them synchronized.
3614 The best result can be reached when you take one track as base and
3615 synchronize other tracks one by one with it.
3616 Remember that synchronization/delay tolerance depends on sample rate, too.
3617 Higher sample rates will give more tolerance.
3619 The filter accepts the following parameters:
3623 Set millimeters distance. This is compensation distance for fine tuning.
3627 Set cm distance. This is compensation distance for tightening distance setup.
3631 Set meters distance. This is compensation distance for hard distance setup.
3635 Set dry amount. Amount of unprocessed (dry) signal.
3639 Set wet amount. Amount of processed (wet) signal.
3643 Set temperature in degrees Celsius. This is the temperature of the environment.
3648 Apply headphone crossfeed filter.
3650 Crossfeed is the process of blending the left and right channels of stereo
3652 It is mainly used to reduce extreme stereo separation of low frequencies.
3654 The intent is to produce more speaker like sound to the listener.
3656 The filter accepts the following options:
3660 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3661 This sets gain of low shelf filter for side part of stereo image.
3662 Default is -6dB. Max allowed is -30db when strength is set to 1.
3665 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3666 This sets cut off frequency of low shelf filter. Default is cut off near
3667 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3670 Set curve slope of low shelf filter. Default is 0.5.
3671 Allowed range is from 0.01 to 1.
3674 Set input gain. Default is 0.9.
3677 Set output gain. Default is 1.
3680 @subsection Commands
3682 This filter supports the all above options as @ref{commands}.
3684 @section crystalizer
3685 Simple algorithm for audio noise sharpening.
3687 This filter linearly increases differences betweeen each audio sample.
3689 The filter accepts the following options:
3693 Sets the intensity of effect (default: 2.0). Must be in range between -10.0 to 0
3694 (unchanged sound) to 10.0 (maximum effect).
3695 To inverse filtering use negative value.
3698 Enable clipping. By default is enabled.
3701 @subsection Commands
3703 This filter supports the all above options as @ref{commands}.
3706 Apply a DC shift to the audio.
3708 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3709 in the recording chain) from the audio. The effect of a DC offset is reduced
3710 headroom and hence volume. The @ref{astats} filter can be used to determine if
3711 a signal has a DC offset.
3715 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3719 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3720 used to prevent clipping.
3725 Apply de-essing to the audio samples.
3729 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3733 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3737 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3741 Set the output mode.
3743 It accepts the following values:
3746 Pass input unchanged.
3749 Pass ess filtered out.
3754 Default value is @var{o}.
3760 Measure audio dynamic range.
3762 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3763 is found in transition material. And anything less that 8 have very poor dynamics
3764 and is very compressed.
3766 The filter accepts the following options:
3770 Set window length in seconds used to split audio into segments of equal length.
3771 Default is 3 seconds.
3775 Dynamic Audio Normalizer.
3777 This filter applies a certain amount of gain to the input audio in order
3778 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3779 contrast to more "simple" normalization algorithms, the Dynamic Audio
3780 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3781 This allows for applying extra gain to the "quiet" sections of the audio
3782 while avoiding distortions or clipping the "loud" sections. In other words:
3783 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3784 sections, in the sense that the volume of each section is brought to the
3785 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3786 this goal *without* applying "dynamic range compressing". It will retain 100%
3787 of the dynamic range *within* each section of the audio file.
3791 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3792 Default is 500 milliseconds.
3793 The Dynamic Audio Normalizer processes the input audio in small chunks,
3794 referred to as frames. This is required, because a peak magnitude has no
3795 meaning for just a single sample value. Instead, we need to determine the
3796 peak magnitude for a contiguous sequence of sample values. While a "standard"
3797 normalizer would simply use the peak magnitude of the complete file, the
3798 Dynamic Audio Normalizer determines the peak magnitude individually for each
3799 frame. The length of a frame is specified in milliseconds. By default, the
3800 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3801 been found to give good results with most files.
3802 Note that the exact frame length, in number of samples, will be determined
3803 automatically, based on the sampling rate of the individual input audio file.
3806 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3807 number. Default is 31.
3808 Probably the most important parameter of the Dynamic Audio Normalizer is the
3809 @code{window size} of the Gaussian smoothing filter. The filter's window size
3810 is specified in frames, centered around the current frame. For the sake of
3811 simplicity, this must be an odd number. Consequently, the default value of 31
3812 takes into account the current frame, as well as the 15 preceding frames and
3813 the 15 subsequent frames. Using a larger window results in a stronger
3814 smoothing effect and thus in less gain variation, i.e. slower gain
3815 adaptation. Conversely, using a smaller window results in a weaker smoothing
3816 effect and thus in more gain variation, i.e. faster gain adaptation.
3817 In other words, the more you increase this value, the more the Dynamic Audio
3818 Normalizer will behave like a "traditional" normalization filter. On the
3819 contrary, the more you decrease this value, the more the Dynamic Audio
3820 Normalizer will behave like a dynamic range compressor.
3823 Set the target peak value. This specifies the highest permissible magnitude
3824 level for the normalized audio input. This filter will try to approach the
3825 target peak magnitude as closely as possible, but at the same time it also
3826 makes sure that the normalized signal will never exceed the peak magnitude.
3827 A frame's maximum local gain factor is imposed directly by the target peak
3828 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3829 It is not recommended to go above this value.
3832 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3833 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3834 factor for each input frame, i.e. the maximum gain factor that does not
3835 result in clipping or distortion. The maximum gain factor is determined by
3836 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3837 additionally bounds the frame's maximum gain factor by a predetermined
3838 (global) maximum gain factor. This is done in order to avoid excessive gain
3839 factors in "silent" or almost silent frames. By default, the maximum gain
3840 factor is 10.0, For most inputs the default value should be sufficient and
3841 it usually is not recommended to increase this value. Though, for input
3842 with an extremely low overall volume level, it may be necessary to allow even
3843 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3844 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3845 Instead, a "sigmoid" threshold function will be applied. This way, the
3846 gain factors will smoothly approach the threshold value, but never exceed that
3850 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3851 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3852 This means that the maximum local gain factor for each frame is defined
3853 (only) by the frame's highest magnitude sample. This way, the samples can
3854 be amplified as much as possible without exceeding the maximum signal
3855 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3856 Normalizer can also take into account the frame's root mean square,
3857 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3858 determine the power of a time-varying signal. It is therefore considered
3859 that the RMS is a better approximation of the "perceived loudness" than
3860 just looking at the signal's peak magnitude. Consequently, by adjusting all
3861 frames to a constant RMS value, a uniform "perceived loudness" can be
3862 established. If a target RMS value has been specified, a frame's local gain
3863 factor is defined as the factor that would result in exactly that RMS value.
3864 Note, however, that the maximum local gain factor is still restricted by the
3865 frame's highest magnitude sample, in order to prevent clipping.
3868 Enable channels coupling. By default is enabled.
3869 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3870 amount. This means the same gain factor will be applied to all channels, i.e.
3871 the maximum possible gain factor is determined by the "loudest" channel.
3872 However, in some recordings, it may happen that the volume of the different
3873 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3874 In this case, this option can be used to disable the channel coupling. This way,
3875 the gain factor will be determined independently for each channel, depending
3876 only on the individual channel's highest magnitude sample. This allows for
3877 harmonizing the volume of the different channels.
3880 Enable DC bias correction. By default is disabled.
3881 An audio signal (in the time domain) is a sequence of sample values.
3882 In the Dynamic Audio Normalizer these sample values are represented in the
3883 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3884 audio signal, or "waveform", should be centered around the zero point.
3885 That means if we calculate the mean value of all samples in a file, or in a
3886 single frame, then the result should be 0.0 or at least very close to that
3887 value. If, however, there is a significant deviation of the mean value from
3888 0.0, in either positive or negative direction, this is referred to as a
3889 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3890 Audio Normalizer provides optional DC bias correction.
3891 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3892 the mean value, or "DC correction" offset, of each input frame and subtract
3893 that value from all of the frame's sample values which ensures those samples
3894 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3895 boundaries, the DC correction offset values will be interpolated smoothly
3896 between neighbouring frames.
3898 @item altboundary, b
3899 Enable alternative boundary mode. By default is disabled.
3900 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3901 around each frame. This includes the preceding frames as well as the
3902 subsequent frames. However, for the "boundary" frames, located at the very
3903 beginning and at the very end of the audio file, not all neighbouring
3904 frames are available. In particular, for the first few frames in the audio
3905 file, the preceding frames are not known. And, similarly, for the last few
3906 frames in the audio file, the subsequent frames are not known. Thus, the
3907 question arises which gain factors should be assumed for the missing frames
3908 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3909 to deal with this situation. The default boundary mode assumes a gain factor
3910 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3911 "fade out" at the beginning and at the end of the input, respectively.
3914 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3915 By default, the Dynamic Audio Normalizer does not apply "traditional"
3916 compression. This means that signal peaks will not be pruned and thus the
3917 full dynamic range will be retained within each local neighbourhood. However,
3918 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3919 normalization algorithm with a more "traditional" compression.
3920 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3921 (thresholding) function. If (and only if) the compression feature is enabled,
3922 all input frames will be processed by a soft knee thresholding function prior
3923 to the actual normalization process. Put simply, the thresholding function is
3924 going to prune all samples whose magnitude exceeds a certain threshold value.
3925 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3926 value. Instead, the threshold value will be adjusted for each individual
3928 In general, smaller parameters result in stronger compression, and vice versa.
3929 Values below 3.0 are not recommended, because audible distortion may appear.
3932 Set the target threshold value. This specifies the lowest permissible
3933 magnitude level for the audio input which will be normalized.
3934 If input frame volume is above this value frame will be normalized.
3935 Otherwise frame may not be normalized at all. The default value is set
3936 to 0, which means all input frames will be normalized.
3937 This option is mostly useful if digital noise is not wanted to be amplified.
3940 @subsection Commands
3942 This filter supports the all above options as @ref{commands}.
3946 Make audio easier to listen to on headphones.
3948 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3949 so that when listened to on headphones the stereo image is moved from
3950 inside your head (standard for headphones) to outside and in front of
3951 the listener (standard for speakers).
3957 Apply a two-pole peaking equalisation (EQ) filter. With this
3958 filter, the signal-level at and around a selected frequency can
3959 be increased or decreased, whilst (unlike bandpass and bandreject
3960 filters) that at all other frequencies is unchanged.
3962 In order to produce complex equalisation curves, this filter can
3963 be given several times, each with a different central frequency.
3965 The filter accepts the following options:
3969 Set the filter's central frequency in Hz.
3972 Set method to specify band-width of filter.
3987 Specify the band-width of a filter in width_type units.
3990 Set the required gain or attenuation in dB.
3991 Beware of clipping when using a positive gain.
3994 How much to use filtered signal in output. Default is 1.
3995 Range is between 0 and 1.
3998 Specify which channels to filter, by default all available are filtered.
4001 Normalize biquad coefficients, by default is disabled.
4002 Enabling it will normalize magnitude response at DC to 0dB.
4005 Set transform type of IIR filter.
4014 Set precison of filtering.
4017 Pick automatic sample format depending on surround filters.
4019 Always use signed 16-bit.
4021 Always use signed 32-bit.
4023 Always use float 32-bit.
4025 Always use float 64-bit.
4029 @subsection Examples
4032 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
4034 equalizer=f=1000:t=h:width=200:g=-10
4038 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
4040 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
4044 @subsection Commands
4046 This filter supports the following commands:
4049 Change equalizer frequency.
4050 Syntax for the command is : "@var{frequency}"
4053 Change equalizer width_type.
4054 Syntax for the command is : "@var{width_type}"
4057 Change equalizer width.
4058 Syntax for the command is : "@var{width}"
4061 Change equalizer gain.
4062 Syntax for the command is : "@var{gain}"
4065 Change equalizer mix.
4066 Syntax for the command is : "@var{mix}"
4069 @section extrastereo
4071 Linearly increases the difference between left and right channels which
4072 adds some sort of "live" effect to playback.
4074 The filter accepts the following options:
4078 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
4079 (average of both channels), with 1.0 sound will be unchanged, with
4080 -1.0 left and right channels will be swapped.
4083 Enable clipping. By default is enabled.
4086 @subsection Commands
4088 This filter supports the all above options as @ref{commands}.
4090 @section firequalizer
4091 Apply FIR Equalization using arbitrary frequency response.
4093 The filter accepts the following option:
4097 Set gain curve equation (in dB). The expression can contain variables:
4100 the evaluated frequency
4104 channel number, set to 0 when multichannels evaluation is disabled
4106 channel id, see libavutil/channel_layout.h, set to the first channel id when
4107 multichannels evaluation is disabled
4111 channel_layout, see libavutil/channel_layout.h
4116 @item gain_interpolate(f)
4117 interpolate gain on frequency f based on gain_entry
4118 @item cubic_interpolate(f)
4119 same as gain_interpolate, but smoother
4121 This option is also available as command. Default is @code{gain_interpolate(f)}.
4124 Set gain entry for gain_interpolate function. The expression can
4128 store gain entry at frequency f with value g
4130 This option is also available as command.
4133 Set filter delay in seconds. Higher value means more accurate.
4134 Default is @code{0.01}.
4137 Set filter accuracy in Hz. Lower value means more accurate.
4138 Default is @code{5}.
4141 Set window function. Acceptable values are:
4144 rectangular window, useful when gain curve is already smooth
4146 hann window (default)
4152 3-terms continuous 1st derivative nuttall window
4154 minimum 3-terms discontinuous nuttall window
4156 4-terms continuous 1st derivative nuttall window
4158 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
4160 blackman-harris window
4166 If enabled, use fixed number of audio samples. This improves speed when
4167 filtering with large delay. Default is disabled.
4170 Enable multichannels evaluation on gain. Default is disabled.
4173 Enable zero phase mode by subtracting timestamp to compensate delay.
4174 Default is disabled.
4177 Set scale used by gain. Acceptable values are:
4180 linear frequency, linear gain
4182 linear frequency, logarithmic (in dB) gain (default)
4184 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
4186 logarithmic frequency, logarithmic gain
4190 Set file for dumping, suitable for gnuplot.
4193 Set scale for dumpfile. Acceptable values are same with scale option.
4197 Enable 2-channel convolution using complex FFT. This improves speed significantly.
4198 Default is disabled.
4201 Enable minimum phase impulse response. Default is disabled.
4204 @subsection Examples
4209 firequalizer=gain='if(lt(f,1000), 0, -INF)'
4212 lowpass at 1000 Hz with gain_entry:
4214 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
4217 custom equalization:
4219 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
4222 higher delay with zero phase to compensate delay:
4224 firequalizer=delay=0.1:fixed=on:zero_phase=on
4227 lowpass on left channel, highpass on right channel:
4229 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
4230 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
4235 Apply a flanging effect to the audio.
4237 The filter accepts the following options:
4241 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
4244 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
4247 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
4251 Set percentage of delayed signal mixed with original. Range from 0 to 100.
4252 Default value is 71.
4255 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
4258 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
4259 Default value is @var{sinusoidal}.
4262 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
4263 Default value is 25.
4266 Set delay-line interpolation, @var{linear} or @var{quadratic}.
4267 Default is @var{linear}.
4271 Apply Haas effect to audio.
4273 Note that this makes most sense to apply on mono signals.
4274 With this filter applied to mono signals it give some directionality and
4275 stretches its stereo image.
4277 The filter accepts the following options:
4281 Set input level. By default is @var{1}, or 0dB
4284 Set output level. By default is @var{1}, or 0dB.
4287 Set gain applied to side part of signal. By default is @var{1}.
4290 Set kind of middle source. Can be one of the following:
4300 Pick middle part signal of stereo image.
4303 Pick side part signal of stereo image.
4307 Change middle phase. By default is disabled.
4310 Set left channel delay. By default is @var{2.05} milliseconds.
4313 Set left channel balance. By default is @var{-1}.
4316 Set left channel gain. By default is @var{1}.
4319 Change left phase. By default is disabled.
4322 Set right channel delay. By defaults is @var{2.12} milliseconds.
4325 Set right channel balance. By default is @var{1}.
4328 Set right channel gain. By default is @var{1}.
4331 Change right phase. By default is enabled.
4336 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
4337 embedded HDCD codes is expanded into a 20-bit PCM stream.
4339 The filter supports the Peak Extend and Low-level Gain Adjustment features
4340 of HDCD, and detects the Transient Filter flag.
4343 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
4346 When using the filter with wav, note the default encoding for wav is 16-bit,
4347 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
4348 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
4350 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
4351 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
4354 The filter accepts the following options:
4357 @item disable_autoconvert
4358 Disable any automatic format conversion or resampling in the filter graph.
4360 @item process_stereo
4361 Process the stereo channels together. If target_gain does not match between
4362 channels, consider it invalid and use the last valid target_gain.
4365 Set the code detect timer period in ms.
4368 Always extend peaks above -3dBFS even if PE isn't signaled.
4371 Replace audio with a solid tone and adjust the amplitude to signal some
4372 specific aspect of the decoding process. The output file can be loaded in
4373 an audio editor alongside the original to aid analysis.
4375 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
4382 Gain adjustment level at each sample
4384 Samples where peak extend occurs
4386 Samples where the code detect timer is active
4388 Samples where the target gain does not match between channels
4394 Apply head-related transfer functions (HRTFs) to create virtual
4395 loudspeakers around the user for binaural listening via headphones.
4396 The HRIRs are provided via additional streams, for each channel
4397 one stereo input stream is needed.
4399 The filter accepts the following options:
4403 Set mapping of input streams for convolution.
4404 The argument is a '|'-separated list of channel names in order as they
4405 are given as additional stream inputs for filter.
4406 This also specify number of input streams. Number of input streams
4407 must be not less than number of channels in first stream plus one.
4410 Set gain applied to audio. Value is in dB. Default is 0.
4413 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4414 processing audio in time domain which is slow.
4415 @var{freq} is processing audio in frequency domain which is fast.
4416 Default is @var{freq}.
4419 Set custom gain for LFE channels. Value is in dB. Default is 0.
4422 Set size of frame in number of samples which will be processed at once.
4423 Default value is @var{1024}. Allowed range is from 1024 to 96000.
4426 Set format of hrir stream.
4427 Default value is @var{stereo}. Alternative value is @var{multich}.
4428 If value is set to @var{stereo}, number of additional streams should
4429 be greater or equal to number of input channels in first input stream.
4430 Also each additional stream should have stereo number of channels.
4431 If value is set to @var{multich}, number of additional streams should
4432 be exactly one. Also number of input channels of additional stream
4433 should be equal or greater than twice number of channels of first input
4437 @subsection Examples
4441 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4442 each amovie filter use stereo file with IR coefficients as input.
4443 The files give coefficients for each position of virtual loudspeaker:
4446 -filter_complex "amovie=azi_270_ele_0_DFC.wav[sr];amovie=azi_90_ele_0_DFC.wav[sl];amovie=azi_225_ele_0_DFC.wav[br];amovie=azi_135_ele_0_DFC.wav[bl];amovie=azi_0_ele_0_DFC.wav,asplit[fc][lfe];amovie=azi_35_ele_0_DFC.wav[fl];amovie=azi_325_ele_0_DFC.wav[fr];[0:a][fl][fr][fc][lfe][bl][br][sl][sr]headphone=FL|FR|FC|LFE|BL|BR|SL|SR"
4451 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4452 but now in @var{multich} @var{hrir} format.
4454 ffmpeg -i input.wav -filter_complex "amovie=minp.wav[hrirs];[0:a][hrirs]headphone=map=FL|FR|FC|LFE|BL|BR|SL|SR:hrir=multich"
4461 Apply a high-pass filter with 3dB point frequency.
4462 The filter can be either single-pole, or double-pole (the default).
4463 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4465 The filter accepts the following options:
4469 Set frequency in Hz. Default is 3000.
4472 Set number of poles. Default is 2.
4475 Set method to specify band-width of filter.
4490 Specify the band-width of a filter in width_type units.
4491 Applies only to double-pole filter.
4492 The default is 0.707q and gives a Butterworth response.
4495 How much to use filtered signal in output. Default is 1.
4496 Range is between 0 and 1.
4499 Specify which channels to filter, by default all available are filtered.
4502 Normalize biquad coefficients, by default is disabled.
4503 Enabling it will normalize magnitude response at DC to 0dB.
4506 Set transform type of IIR filter.
4515 Set precison of filtering.
4518 Pick automatic sample format depending on surround filters.
4520 Always use signed 16-bit.
4522 Always use signed 32-bit.
4524 Always use float 32-bit.
4526 Always use float 64-bit.
4530 @subsection Commands
4532 This filter supports the following commands:
4535 Change highpass frequency.
4536 Syntax for the command is : "@var{frequency}"
4539 Change highpass width_type.
4540 Syntax for the command is : "@var{width_type}"
4543 Change highpass width.
4544 Syntax for the command is : "@var{width}"
4547 Change highpass mix.
4548 Syntax for the command is : "@var{mix}"
4553 Join multiple input streams into one multi-channel stream.
4555 It accepts the following parameters:
4559 The number of input streams. It defaults to 2.
4561 @item channel_layout
4562 The desired output channel layout. It defaults to stereo.
4565 Map channels from inputs to output. The argument is a '|'-separated list of
4566 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4567 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4568 can be either the name of the input channel (e.g. FL for front left) or its
4569 index in the specified input stream. @var{out_channel} is the name of the output
4573 The filter will attempt to guess the mappings when they are not specified
4574 explicitly. It does so by first trying to find an unused matching input channel
4575 and if that fails it picks the first unused input channel.
4577 Join 3 inputs (with properly set channel layouts):
4579 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4582 Build a 5.1 output from 6 single-channel streams:
4584 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4585 'join=inputs=6:channel_layout=5.1:map=0.0-FL|1.0-FR|2.0-FC|3.0-SL|4.0-SR|5.0-LFE'
4591 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4593 To enable compilation of this filter you need to configure FFmpeg with
4594 @code{--enable-ladspa}.
4598 Specifies the name of LADSPA plugin library to load. If the environment
4599 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4600 each one of the directories specified by the colon separated list in
4601 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4602 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4603 @file{/usr/lib/ladspa/}.
4606 Specifies the plugin within the library. Some libraries contain only
4607 one plugin, but others contain many of them. If this is not set filter
4608 will list all available plugins within the specified library.
4611 Set the '|' separated list of controls which are zero or more floating point
4612 values that determine the behavior of the loaded plugin (for example delay,
4614 Controls need to be defined using the following syntax:
4615 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4616 @var{valuei} is the value set on the @var{i}-th control.
4617 Alternatively they can be also defined using the following syntax:
4618 @var{value0}|@var{value1}|@var{value2}|..., where
4619 @var{valuei} is the value set on the @var{i}-th control.
4620 If @option{controls} is set to @code{help}, all available controls and
4621 their valid ranges are printed.
4623 @item sample_rate, s
4624 Specify the sample rate, default to 44100. Only used if plugin have
4628 Set the number of samples per channel per each output frame, default
4629 is 1024. Only used if plugin have zero inputs.
4632 Set the minimum duration of the sourced audio. See
4633 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4634 for the accepted syntax.
4635 Note that the resulting duration may be greater than the specified duration,
4636 as the generated audio is always cut at the end of a complete frame.
4637 If not specified, or the expressed duration is negative, the audio is
4638 supposed to be generated forever.
4639 Only used if plugin have zero inputs.
4642 Enable latency compensation, by default is disabled.
4643 Only used if plugin have inputs.
4646 @subsection Examples
4650 List all available plugins within amp (LADSPA example plugin) library:
4656 List all available controls and their valid ranges for @code{vcf_notch}
4657 plugin from @code{VCF} library:
4659 ladspa=f=vcf:p=vcf_notch:c=help
4663 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4666 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4670 Add reverberation to the audio using TAP-plugins
4671 (Tom's Audio Processing plugins):
4673 ladspa=file=tap_reverb:tap_reverb
4677 Generate white noise, with 0.2 amplitude:
4679 ladspa=file=cmt:noise_source_white:c=c0=.2
4683 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4684 @code{C* Audio Plugin Suite} (CAPS) library:
4686 ladspa=file=caps:Click:c=c1=20'
4690 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4692 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4696 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4697 @code{SWH Plugins} collection:
4699 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4703 Attenuate low frequencies using Multiband EQ from Steve Harris
4704 @code{SWH Plugins} collection:
4706 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4710 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4713 ladspa=caps:Narrower
4717 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4719 ladspa=caps:White:.2
4723 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4725 ladspa=caps:Fractal:c=c1=1
4729 Dynamic volume normalization using @code{VLevel} plugin:
4731 ladspa=vlevel-ladspa:vlevel_mono
4735 @subsection Commands
4737 This filter supports the following commands:
4740 Modify the @var{N}-th control value.
4742 If the specified value is not valid, it is ignored and prior one is kept.
4747 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4748 Support for both single pass (livestreams, files) and double pass (files) modes.
4749 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4750 detect true peaks, the audio stream will be upsampled to 192 kHz.
4751 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4753 The filter accepts the following options:
4757 Set integrated loudness target.
4758 Range is -70.0 - -5.0. Default value is -24.0.
4761 Set loudness range target.
4762 Range is 1.0 - 20.0. Default value is 7.0.
4765 Set maximum true peak.
4766 Range is -9.0 - +0.0. Default value is -2.0.
4768 @item measured_I, measured_i
4769 Measured IL of input file.
4770 Range is -99.0 - +0.0.
4772 @item measured_LRA, measured_lra
4773 Measured LRA of input file.
4774 Range is 0.0 - 99.0.
4776 @item measured_TP, measured_tp
4777 Measured true peak of input file.
4778 Range is -99.0 - +99.0.
4780 @item measured_thresh
4781 Measured threshold of input file.
4782 Range is -99.0 - +0.0.
4785 Set offset gain. Gain is applied before the true-peak limiter.
4786 Range is -99.0 - +99.0. Default is +0.0.
4789 Normalize by linearly scaling the source audio.
4790 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4791 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4792 be lower than source LRA and the change in integrated loudness shouldn't
4793 result in a true peak which exceeds the target TP. If any of these
4794 conditions aren't met, normalization mode will revert to @var{dynamic}.
4795 Options are @code{true} or @code{false}. Default is @code{true}.
4798 Treat mono input files as "dual-mono". If a mono file is intended for playback
4799 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4800 If set to @code{true}, this option will compensate for this effect.
4801 Multi-channel input files are not affected by this option.
4802 Options are true or false. Default is false.
4805 Set print format for stats. Options are summary, json, or none.
4806 Default value is none.
4811 Apply a low-pass filter with 3dB point frequency.
4812 The filter can be either single-pole or double-pole (the default).
4813 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4815 The filter accepts the following options:
4819 Set frequency in Hz. Default is 500.
4822 Set number of poles. Default is 2.
4825 Set method to specify band-width of filter.
4840 Specify the band-width of a filter in width_type units.
4841 Applies only to double-pole filter.
4842 The default is 0.707q and gives a Butterworth response.
4845 How much to use filtered signal in output. Default is 1.
4846 Range is between 0 and 1.
4849 Specify which channels to filter, by default all available are filtered.
4852 Normalize biquad coefficients, by default is disabled.
4853 Enabling it will normalize magnitude response at DC to 0dB.
4856 Set transform type of IIR filter.
4865 Set precison of filtering.
4868 Pick automatic sample format depending on surround filters.
4870 Always use signed 16-bit.
4872 Always use signed 32-bit.
4874 Always use float 32-bit.
4876 Always use float 64-bit.
4880 @subsection Examples
4883 Lowpass only LFE channel, it LFE is not present it does nothing:
4889 @subsection Commands
4891 This filter supports the following commands:
4894 Change lowpass frequency.
4895 Syntax for the command is : "@var{frequency}"
4898 Change lowpass width_type.
4899 Syntax for the command is : "@var{width_type}"
4902 Change lowpass width.
4903 Syntax for the command is : "@var{width}"
4907 Syntax for the command is : "@var{mix}"
4912 Load a LV2 (LADSPA Version 2) plugin.
4914 To enable compilation of this filter you need to configure FFmpeg with
4915 @code{--enable-lv2}.
4919 Specifies the plugin URI. You may need to escape ':'.
4922 Set the '|' separated list of controls which are zero or more floating point
4923 values that determine the behavior of the loaded plugin (for example delay,
4925 If @option{controls} is set to @code{help}, all available controls and
4926 their valid ranges are printed.
4928 @item sample_rate, s
4929 Specify the sample rate, default to 44100. Only used if plugin have
4933 Set the number of samples per channel per each output frame, default
4934 is 1024. Only used if plugin have zero inputs.
4937 Set the minimum duration of the sourced audio. See
4938 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4939 for the accepted syntax.
4940 Note that the resulting duration may be greater than the specified duration,
4941 as the generated audio is always cut at the end of a complete frame.
4942 If not specified, or the expressed duration is negative, the audio is
4943 supposed to be generated forever.
4944 Only used if plugin have zero inputs.
4947 @subsection Examples
4951 Apply bass enhancer plugin from Calf:
4953 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4957 Apply vinyl plugin from Calf:
4959 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4963 Apply bit crusher plugin from ArtyFX:
4965 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4970 Multiband Compress or expand the audio's dynamic range.
4972 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4973 This is akin to the crossover of a loudspeaker, and results in flat frequency
4974 response when absent compander action.
4976 It accepts the following parameters:
4980 This option syntax is:
4981 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4982 For explanation of each item refer to compand filter documentation.
4988 Mix channels with specific gain levels. The filter accepts the output
4989 channel layout followed by a set of channels definitions.
4991 This filter is also designed to efficiently remap the channels of an audio
4994 The filter accepts parameters of the form:
4995 "@var{l}|@var{outdef}|@var{outdef}|..."
4999 output channel layout or number of channels
5002 output channel specification, of the form:
5003 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
5006 output channel to define, either a channel name (FL, FR, etc.) or a channel
5007 number (c0, c1, etc.)
5010 multiplicative coefficient for the channel, 1 leaving the volume unchanged
5013 input channel to use, see out_name for details; it is not possible to mix
5014 named and numbered input channels
5017 If the `=' in a channel specification is replaced by `<', then the gains for
5018 that specification will be renormalized so that the total is 1, thus
5019 avoiding clipping noise.
5021 @subsection Mixing examples
5023 For example, if you want to down-mix from stereo to mono, but with a bigger
5024 factor for the left channel:
5026 pan=1c|c0=0.9*c0+0.1*c1
5029 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
5030 7-channels surround:
5032 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
5035 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
5036 that should be preferred (see "-ac" option) unless you have very specific
5039 @subsection Remapping examples
5041 The channel remapping will be effective if, and only if:
5044 @item gain coefficients are zeroes or ones,
5045 @item only one input per channel output,
5048 If all these conditions are satisfied, the filter will notify the user ("Pure
5049 channel mapping detected"), and use an optimized and lossless method to do the
5052 For example, if you have a 5.1 source and want a stereo audio stream by
5053 dropping the extra channels:
5055 pan="stereo| c0=FL | c1=FR"
5058 Given the same source, you can also switch front left and front right channels
5059 and keep the input channel layout:
5061 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
5064 If the input is a stereo audio stream, you can mute the front left channel (and
5065 still keep the stereo channel layout) with:
5070 Still with a stereo audio stream input, you can copy the right channel in both
5071 front left and right:
5073 pan="stereo| c0=FR | c1=FR"
5078 ReplayGain scanner filter. This filter takes an audio stream as an input and
5079 outputs it unchanged.
5080 At end of filtering it displays @code{track_gain} and @code{track_peak}.
5084 Convert the audio sample format, sample rate and channel layout. It is
5085 not meant to be used directly.
5088 Apply time-stretching and pitch-shifting with librubberband.
5090 To enable compilation of this filter, you need to configure FFmpeg with
5091 @code{--enable-librubberband}.
5093 The filter accepts the following options:
5097 Set tempo scale factor.
5100 Set pitch scale factor.
5103 Set transients detector.
5104 Possible values are:
5113 Possible values are:
5122 Possible values are:
5129 Set processing window size.
5130 Possible values are:
5139 Possible values are:
5146 Enable formant preservation when shift pitching.
5147 Possible values are:
5155 Possible values are:
5164 Possible values are:
5171 @subsection Commands
5173 This filter supports the following commands:
5176 Change filter tempo scale factor.
5177 Syntax for the command is : "@var{tempo}"
5180 Change filter pitch scale factor.
5181 Syntax for the command is : "@var{pitch}"
5184 @section sidechaincompress
5186 This filter acts like normal compressor but has the ability to compress
5187 detected signal using second input signal.
5188 It needs two input streams and returns one output stream.
5189 First input stream will be processed depending on second stream signal.
5190 The filtered signal then can be filtered with other filters in later stages of
5191 processing. See @ref{pan} and @ref{amerge} filter.
5193 The filter accepts the following options:
5197 Set input gain. Default is 1. Range is between 0.015625 and 64.
5200 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
5201 Default is @code{downward}.
5204 If a signal of second stream raises above this level it will affect the gain
5205 reduction of first stream.
5206 By default is 0.125. Range is between 0.00097563 and 1.
5209 Set a ratio about which the signal is reduced. 1:2 means that if the level
5210 raised 4dB above the threshold, it will be only 2dB above after the reduction.
5211 Default is 2. Range is between 1 and 20.
5214 Amount of milliseconds the signal has to rise above the threshold before gain
5215 reduction starts. Default is 20. Range is between 0.01 and 2000.
5218 Amount of milliseconds the signal has to fall below the threshold before
5219 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
5222 Set the amount by how much signal will be amplified after processing.
5223 Default is 1. Range is from 1 to 64.
5226 Curve the sharp knee around the threshold to enter gain reduction more softly.
5227 Default is 2.82843. Range is between 1 and 8.
5230 Choose if the @code{average} level between all channels of side-chain stream
5231 or the louder(@code{maximum}) channel of side-chain stream affects the
5232 reduction. Default is @code{average}.
5235 Should the exact signal be taken in case of @code{peak} or an RMS one in case
5236 of @code{rms}. Default is @code{rms} which is mainly smoother.
5239 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
5242 How much to use compressed signal in output. Default is 1.
5243 Range is between 0 and 1.
5246 @subsection Commands
5248 This filter supports the all above options as @ref{commands}.
5250 @subsection Examples
5254 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
5255 depending on the signal of 2nd input and later compressed signal to be
5256 merged with 2nd input:
5258 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
5262 @section sidechaingate
5264 A sidechain gate acts like a normal (wideband) gate but has the ability to
5265 filter the detected signal before sending it to the gain reduction stage.
5266 Normally a gate uses the full range signal to detect a level above the
5268 For example: If you cut all lower frequencies from your sidechain signal
5269 the gate will decrease the volume of your track only if not enough highs
5270 appear. With this technique you are able to reduce the resonation of a
5271 natural drum or remove "rumbling" of muted strokes from a heavily distorted
5273 It needs two input streams and returns one output stream.
5274 First input stream will be processed depending on second stream signal.
5276 The filter accepts the following options:
5280 Set input level before filtering.
5281 Default is 1. Allowed range is from 0.015625 to 64.
5284 Set the mode of operation. Can be @code{upward} or @code{downward}.
5285 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
5286 will be amplified, expanding dynamic range in upward direction.
5287 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
5290 Set the level of gain reduction when the signal is below the threshold.
5291 Default is 0.06125. Allowed range is from 0 to 1.
5292 Setting this to 0 disables reduction and then filter behaves like expander.
5295 If a signal rises above this level the gain reduction is released.
5296 Default is 0.125. Allowed range is from 0 to 1.
5299 Set a ratio about which the signal is reduced.
5300 Default is 2. Allowed range is from 1 to 9000.
5303 Amount of milliseconds the signal has to rise above the threshold before gain
5305 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
5308 Amount of milliseconds the signal has to fall below the threshold before the
5309 reduction is increased again. Default is 250 milliseconds.
5310 Allowed range is from 0.01 to 9000.
5313 Set amount of amplification of signal after processing.
5314 Default is 1. Allowed range is from 1 to 64.
5317 Curve the sharp knee around the threshold to enter gain reduction more softly.
5318 Default is 2.828427125. Allowed range is from 1 to 8.
5321 Choose if exact signal should be taken for detection or an RMS like one.
5322 Default is rms. Can be peak or rms.
5325 Choose if the average level between all channels or the louder channel affects
5327 Default is average. Can be average or maximum.
5330 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
5333 @subsection Commands
5335 This filter supports the all above options as @ref{commands}.
5337 @section silencedetect
5339 Detect silence in an audio stream.
5341 This filter logs a message when it detects that the input audio volume is less
5342 or equal to a noise tolerance value for a duration greater or equal to the
5343 minimum detected noise duration.
5345 The printed times and duration are expressed in seconds. The
5346 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
5347 is set on the first frame whose timestamp equals or exceeds the detection
5348 duration and it contains the timestamp of the first frame of the silence.
5350 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
5351 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
5352 keys are set on the first frame after the silence. If @option{mono} is
5353 enabled, and each channel is evaluated separately, the @code{.X}
5354 suffixed keys are used, and @code{X} corresponds to the channel number.
5356 The filter accepts the following options:
5360 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
5361 specified value) or amplitude ratio. Default is -60dB, or 0.001.
5364 Set silence duration until notification (default is 2 seconds). See
5365 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5366 for the accepted syntax.
5369 Process each channel separately, instead of combined. By default is disabled.
5372 @subsection Examples
5376 Detect 5 seconds of silence with -50dB noise tolerance:
5378 silencedetect=n=-50dB:d=5
5382 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
5383 tolerance in @file{silence.mp3}:
5385 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
5389 @section silenceremove
5391 Remove silence from the beginning, middle or end of the audio.
5393 The filter accepts the following options:
5397 This value is used to indicate if audio should be trimmed at beginning of
5398 the audio. A value of zero indicates no silence should be trimmed from the
5399 beginning. When specifying a non-zero value, it trims audio up until it
5400 finds non-silence. Normally, when trimming silence from beginning of audio
5401 the @var{start_periods} will be @code{1} but it can be increased to higher
5402 values to trim all audio up to specific count of non-silence periods.
5403 Default value is @code{0}.
5405 @item start_duration
5406 Specify the amount of time that non-silence must be detected before it stops
5407 trimming audio. By increasing the duration, bursts of noises can be treated
5408 as silence and trimmed off. Default value is @code{0}.
5410 @item start_threshold
5411 This indicates what sample value should be treated as silence. For digital
5412 audio, a value of @code{0} may be fine but for audio recorded from analog,
5413 you may wish to increase the value to account for background noise.
5414 Can be specified in dB (in case "dB" is appended to the specified value)
5415 or amplitude ratio. Default value is @code{0}.
5418 Specify max duration of silence at beginning that will be kept after
5419 trimming. Default is 0, which is equal to trimming all samples detected
5423 Specify mode of detection of silence end in start of multi-channel audio.
5424 Can be @var{any} or @var{all}. Default is @var{any}.
5425 With @var{any}, any sample that is detected as non-silence will cause
5426 stopped trimming of silence.
5427 With @var{all}, only if all channels are detected as non-silence will cause
5428 stopped trimming of silence.
5431 Set the count for trimming silence from the end of audio.
5432 To remove silence from the middle of a file, specify a @var{stop_periods}
5433 that is negative. This value is then treated as a positive value and is
5434 used to indicate the effect should restart processing as specified by
5435 @var{start_periods}, making it suitable for removing periods of silence
5436 in the middle of the audio.
5437 Default value is @code{0}.
5440 Specify a duration of silence that must exist before audio is not copied any
5441 more. By specifying a higher duration, silence that is wanted can be left in
5443 Default value is @code{0}.
5445 @item stop_threshold
5446 This is the same as @option{start_threshold} but for trimming silence from
5448 Can be specified in dB (in case "dB" is appended to the specified value)
5449 or amplitude ratio. Default value is @code{0}.
5452 Specify max duration of silence at end that will be kept after
5453 trimming. Default is 0, which is equal to trimming all samples detected
5457 Specify mode of detection of silence start in end of multi-channel audio.
5458 Can be @var{any} or @var{all}. Default is @var{any}.
5459 With @var{any}, any sample that is detected as non-silence will cause
5460 stopped trimming of silence.
5461 With @var{all}, only if all channels are detected as non-silence will cause
5462 stopped trimming of silence.
5465 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
5466 and works better with digital silence which is exactly 0.
5467 Default value is @code{rms}.
5470 Set duration in number of seconds used to calculate size of window in number
5471 of samples for detecting silence.
5472 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
5475 @subsection Examples
5479 The following example shows how this filter can be used to start a recording
5480 that does not contain the delay at the start which usually occurs between
5481 pressing the record button and the start of the performance:
5483 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
5487 Trim all silence encountered from beginning to end where there is more than 1
5488 second of silence in audio:
5490 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
5494 Trim all digital silence samples, using peak detection, from beginning to end
5495 where there is more than 0 samples of digital silence in audio and digital
5496 silence is detected in all channels at same positions in stream:
5498 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
5504 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
5505 loudspeakers around the user for binaural listening via headphones (audio
5506 formats up to 9 channels supported).
5507 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
5508 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
5509 Austrian Academy of Sciences.
5511 To enable compilation of this filter you need to configure FFmpeg with
5512 @code{--enable-libmysofa}.
5514 The filter accepts the following options:
5518 Set the SOFA file used for rendering.
5521 Set gain applied to audio. Value is in dB. Default is 0.
5524 Set rotation of virtual loudspeakers in deg. Default is 0.
5527 Set elevation of virtual speakers in deg. Default is 0.
5530 Set distance in meters between loudspeakers and the listener with near-field
5531 HRTFs. Default is 1.
5534 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
5535 processing audio in time domain which is slow.
5536 @var{freq} is processing audio in frequency domain which is fast.
5537 Default is @var{freq}.
5540 Set custom positions of virtual loudspeakers. Syntax for this option is:
5541 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
5542 Each virtual loudspeaker is described with short channel name following with
5543 azimuth and elevation in degrees.
5544 Each virtual loudspeaker description is separated by '|'.
5545 For example to override front left and front right channel positions use:
5546 'speakers=FL 45 15|FR 345 15'.
5547 Descriptions with unrecognised channel names are ignored.
5550 Set custom gain for LFE channels. Value is in dB. Default is 0.
5553 Set custom frame size in number of samples. Default is 1024.
5554 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5555 is set to @var{freq}.
5558 Should all IRs be normalized upon importing SOFA file.
5559 By default is enabled.
5562 Should nearest IRs be interpolated with neighbor IRs if exact position
5563 does not match. By default is disabled.
5566 Minphase all IRs upon loading of SOFA file. By default is disabled.
5569 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5572 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5575 @subsection Examples
5579 Using ClubFritz6 sofa file:
5581 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5585 Using ClubFritz12 sofa file and bigger radius with small rotation:
5587 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5591 Similar as above but with custom speaker positions for front left, front right, back left and back right
5592 and also with custom gain:
5594 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5601 This filter expands or compresses each half-cycle of audio samples
5602 (local set of samples all above or all below zero and between two nearest zero crossings) depending
5603 on threshold value, so audio reaches target peak value under conditions controlled by below options.
5605 The filter accepts the following options:
5609 Set the expansion target peak value. This specifies the highest allowed absolute amplitude
5610 level for the normalized audio input. Default value is 0.95. Allowed range is from 0.0 to 1.0.
5613 Set the maximum expansion factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5614 This option controls maximum local half-cycle of samples expansion. The maximum expansion
5615 would be such that local peak value reaches target peak value but never to surpass it and that
5616 ratio between new and previous peak value does not surpass this option value.
5618 @item compression, c
5619 Set the maximum compression factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5620 This option controls maximum local half-cycle of samples compression. This option is used
5621 only if @option{threshold} option is set to value greater than 0.0, then in such cases
5622 when local peak is lower or same as value set by @option{threshold} all samples belonging to
5623 that peak's half-cycle will be compressed by current compression factor.
5626 Set the threshold value. Default value is 0.0. Allowed range is from 0.0 to 1.0.
5627 This option specifies which half-cycles of samples will be compressed and which will be expanded.
5628 Any half-cycle samples with their local peak value below or same as this option value will be
5629 compressed by current compression factor, otherwise, if greater than threshold value they will be
5630 expanded with expansion factor so that it could reach peak target value but never surpass it.
5633 Set the expansion raising amount per each half-cycle of samples. Default value is 0.001.
5634 Allowed range is from 0.0 to 1.0. This controls how fast expansion factor is raised per
5635 each new half-cycle until it reaches @option{expansion} value.
5636 Setting this options too high may lead to distortions.
5639 Set the compression raising amount per each half-cycle of samples. Default value is 0.001.
5640 Allowed range is from 0.0 to 1.0. This controls how fast compression factor is raised per
5641 each new half-cycle until it reaches @option{compression} value.
5644 Specify which channels to filter, by default all available channels are filtered.
5647 Enable inverted filtering, by default is disabled. This inverts interpretation of @option{threshold}
5648 option. When enabled any half-cycle of samples with their local peak value below or same as
5649 @option{threshold} option will be expanded otherwise it will be compressed.
5652 Link channels when calculating gain applied to each filtered channel sample, by default is disabled.
5653 When disabled each filtered channel gain calculation is independent, otherwise when this option
5654 is enabled the minimum of all possible gains for each filtered channel is used.
5657 @subsection Commands
5659 This filter supports the all above options as @ref{commands}.
5661 @section stereotools
5663 This filter has some handy utilities to manage stereo signals, for converting
5664 M/S stereo recordings to L/R signal while having control over the parameters
5665 or spreading the stereo image of master track.
5667 The filter accepts the following options:
5671 Set input level before filtering for both channels. Defaults is 1.
5672 Allowed range is from 0.015625 to 64.
5675 Set output level after filtering for both channels. Defaults is 1.
5676 Allowed range is from 0.015625 to 64.
5679 Set input balance between both channels. Default is 0.
5680 Allowed range is from -1 to 1.
5683 Set output balance between both channels. Default is 0.
5684 Allowed range is from -1 to 1.
5687 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5688 clipping. Disabled by default.
5691 Mute the left channel. Disabled by default.
5694 Mute the right channel. Disabled by default.
5697 Change the phase of the left channel. Disabled by default.
5700 Change the phase of the right channel. Disabled by default.
5703 Set stereo mode. Available values are:
5707 Left/Right to Left/Right, this is default.
5710 Left/Right to Mid/Side.
5713 Mid/Side to Left/Right.
5716 Left/Right to Left/Left.
5719 Left/Right to Right/Right.
5722 Left/Right to Left + Right.
5725 Left/Right to Right/Left.
5728 Mid/Side to Left/Left.
5731 Mid/Side to Right/Right.
5734 Mid/Side to Right/Left.
5737 Left/Right to Left - Right.
5741 Set level of side signal. Default is 1.
5742 Allowed range is from 0.015625 to 64.
5745 Set balance of side signal. Default is 0.
5746 Allowed range is from -1 to 1.
5749 Set level of the middle signal. Default is 1.
5750 Allowed range is from 0.015625 to 64.
5753 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5756 Set stereo base between mono and inversed channels. Default is 0.
5757 Allowed range is from -1 to 1.
5760 Set delay in milliseconds how much to delay left from right channel and
5761 vice versa. Default is 0. Allowed range is from -20 to 20.
5764 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5767 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5769 @item bmode_in, bmode_out
5770 Set balance mode for balance_in/balance_out option.
5772 Can be one of the following:
5776 Classic balance mode. Attenuate one channel at time.
5777 Gain is raised up to 1.
5780 Similar as classic mode above but gain is raised up to 2.
5783 Equal power distribution, from -6dB to +6dB range.
5787 @subsection Commands
5789 This filter supports the all above options as @ref{commands}.
5791 @subsection Examples
5795 Apply karaoke like effect:
5797 stereotools=mlev=0.015625
5801 Convert M/S signal to L/R:
5803 "stereotools=mode=ms>lr"
5807 @section stereowiden
5809 This filter enhance the stereo effect by suppressing signal common to both
5810 channels and by delaying the signal of left into right and vice versa,
5811 thereby widening the stereo effect.
5813 The filter accepts the following options:
5817 Time in milliseconds of the delay of left signal into right and vice versa.
5818 Default is 20 milliseconds.
5821 Amount of gain in delayed signal into right and vice versa. Gives a delay
5822 effect of left signal in right output and vice versa which gives widening
5823 effect. Default is 0.3.
5826 Cross feed of left into right with inverted phase. This helps in suppressing
5827 the mono. If the value is 1 it will cancel all the signal common to both
5828 channels. Default is 0.3.
5831 Set level of input signal of original channel. Default is 0.8.
5834 @subsection Commands
5836 This filter supports the all above options except @code{delay} as @ref{commands}.
5838 @section superequalizer
5839 Apply 18 band equalizer.
5841 The filter accepts the following options:
5848 Set 131Hz band gain.
5850 Set 185Hz band gain.
5852 Set 262Hz band gain.
5854 Set 370Hz band gain.
5856 Set 523Hz band gain.
5858 Set 740Hz band gain.
5860 Set 1047Hz band gain.
5862 Set 1480Hz band gain.
5864 Set 2093Hz band gain.
5866 Set 2960Hz band gain.
5868 Set 4186Hz band gain.
5870 Set 5920Hz band gain.
5872 Set 8372Hz band gain.
5874 Set 11840Hz band gain.
5876 Set 16744Hz band gain.
5878 Set 20000Hz band gain.
5882 Apply audio surround upmix filter.
5884 This filter allows to produce multichannel output from audio stream.
5886 The filter accepts the following options:
5890 Set output channel layout. By default, this is @var{5.1}.
5892 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5893 for the required syntax.
5896 Set input channel layout. By default, this is @var{stereo}.
5898 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5899 for the required syntax.
5902 Set input volume level. By default, this is @var{1}.
5905 Set output volume level. By default, this is @var{1}.
5908 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5911 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5914 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5917 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5918 In @var{add} mode, LFE channel is created from input audio and added to output.
5919 In @var{sub} mode, LFE channel is created from input audio and added to output but
5920 also all non-LFE output channels are subtracted with output LFE channel.
5923 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5924 Default is @var{90}.
5927 Set front center input volume. By default, this is @var{1}.
5930 Set front center output volume. By default, this is @var{1}.
5933 Set front left input volume. By default, this is @var{1}.
5936 Set front left output volume. By default, this is @var{1}.
5939 Set front right input volume. By default, this is @var{1}.
5942 Set front right output volume. By default, this is @var{1}.
5945 Set side left input volume. By default, this is @var{1}.
5948 Set side left output volume. By default, this is @var{1}.
5951 Set side right input volume. By default, this is @var{1}.
5954 Set side right output volume. By default, this is @var{1}.
5957 Set back left input volume. By default, this is @var{1}.
5960 Set back left output volume. By default, this is @var{1}.
5963 Set back right input volume. By default, this is @var{1}.
5966 Set back right output volume. By default, this is @var{1}.
5969 Set back center input volume. By default, this is @var{1}.
5972 Set back center output volume. By default, this is @var{1}.
5975 Set LFE input volume. By default, this is @var{1}.
5978 Set LFE output volume. By default, this is @var{1}.
5981 Set spread usage of stereo image across X axis for all channels.
5984 Set spread usage of stereo image across Y axis for all channels.
5986 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5987 Set spread usage of stereo image across X axis for each channel.
5989 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5990 Set spread usage of stereo image across Y axis for each channel.
5993 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5996 Set window function.
5998 It accepts the following values:
6021 Default is @code{hann}.
6024 Set window overlap. If set to 1, the recommended overlap for selected
6025 window function will be picked. Default is @code{0.5}.
6028 @section treble, highshelf
6030 Boost or cut treble (upper) frequencies of the audio using a two-pole
6031 shelving filter with a response similar to that of a standard
6032 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
6034 The filter accepts the following options:
6038 Give the gain at whichever is the lower of ~22 kHz and the
6039 Nyquist frequency. Its useful range is about -20 (for a large cut)
6040 to +20 (for a large boost). Beware of clipping when using a positive gain.
6043 Set the filter's central frequency and so can be used
6044 to extend or reduce the frequency range to be boosted or cut.
6045 The default value is @code{3000} Hz.
6048 Set method to specify band-width of filter.
6063 Determine how steep is the filter's shelf transition.
6066 Set number of poles. Default is 2.
6069 How much to use filtered signal in output. Default is 1.
6070 Range is between 0 and 1.
6073 Specify which channels to filter, by default all available are filtered.
6076 Normalize biquad coefficients, by default is disabled.
6077 Enabling it will normalize magnitude response at DC to 0dB.
6080 Set transform type of IIR filter.
6089 Set precison of filtering.
6092 Pick automatic sample format depending on surround filters.
6094 Always use signed 16-bit.
6096 Always use signed 32-bit.
6098 Always use float 32-bit.
6100 Always use float 64-bit.
6104 @subsection Commands
6106 This filter supports the following commands:
6109 Change treble frequency.
6110 Syntax for the command is : "@var{frequency}"
6113 Change treble width_type.
6114 Syntax for the command is : "@var{width_type}"
6117 Change treble width.
6118 Syntax for the command is : "@var{width}"
6122 Syntax for the command is : "@var{gain}"
6126 Syntax for the command is : "@var{mix}"
6131 Sinusoidal amplitude modulation.
6133 The filter accepts the following options:
6137 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
6138 (20 Hz or lower) will result in a tremolo effect.
6139 This filter may also be used as a ring modulator by specifying
6140 a modulation frequency higher than 20 Hz.
6141 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6144 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6145 Default value is 0.5.
6150 Sinusoidal phase modulation.
6152 The filter accepts the following options:
6156 Modulation frequency in Hertz.
6157 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6160 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6161 Default value is 0.5.
6166 Adjust the input audio volume.
6168 It accepts the following parameters:
6172 Set audio volume expression.
6174 Output values are clipped to the maximum value.
6176 The output audio volume is given by the relation:
6178 @var{output_volume} = @var{volume} * @var{input_volume}
6181 The default value for @var{volume} is "1.0".
6184 This parameter represents the mathematical precision.
6186 It determines which input sample formats will be allowed, which affects the
6187 precision of the volume scaling.
6191 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
6193 32-bit floating-point; this limits input sample format to FLT. (default)
6195 64-bit floating-point; this limits input sample format to DBL.
6199 Choose the behaviour on encountering ReplayGain side data in input frames.
6203 Remove ReplayGain side data, ignoring its contents (the default).
6206 Ignore ReplayGain side data, but leave it in the frame.
6209 Prefer the track gain, if present.
6212 Prefer the album gain, if present.
6215 @item replaygain_preamp
6216 Pre-amplification gain in dB to apply to the selected replaygain gain.
6218 Default value for @var{replaygain_preamp} is 0.0.
6220 @item replaygain_noclip
6221 Prevent clipping by limiting the gain applied.
6223 Default value for @var{replaygain_noclip} is 1.
6226 Set when the volume expression is evaluated.
6228 It accepts the following values:
6231 only evaluate expression once during the filter initialization, or
6232 when the @samp{volume} command is sent
6235 evaluate expression for each incoming frame
6238 Default value is @samp{once}.
6241 The volume expression can contain the following parameters.
6245 frame number (starting at zero)
6248 @item nb_consumed_samples
6249 number of samples consumed by the filter
6251 number of samples in the current frame
6253 original frame position in the file
6259 PTS at start of stream
6261 time at start of stream
6267 last set volume value
6270 Note that when @option{eval} is set to @samp{once} only the
6271 @var{sample_rate} and @var{tb} variables are available, all other
6272 variables will evaluate to NAN.
6274 @subsection Commands
6276 This filter supports the following commands:
6279 Modify the volume expression.
6280 The command accepts the same syntax of the corresponding option.
6282 If the specified expression is not valid, it is kept at its current
6286 @subsection Examples
6290 Halve the input audio volume:
6294 volume=volume=-6.0206dB
6297 In all the above example the named key for @option{volume} can be
6298 omitted, for example like in:
6304 Increase input audio power by 6 decibels using fixed-point precision:
6306 volume=volume=6dB:precision=fixed
6310 Fade volume after time 10 with an annihilation period of 5 seconds:
6312 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
6316 @section volumedetect
6318 Detect the volume of the input video.
6320 The filter has no parameters. The input is not modified. Statistics about
6321 the volume will be printed in the log when the input stream end is reached.
6323 In particular it will show the mean volume (root mean square), maximum
6324 volume (on a per-sample basis), and the beginning of a histogram of the
6325 registered volume values (from the maximum value to a cumulated 1/1000 of
6328 All volumes are in decibels relative to the maximum PCM value.
6330 @subsection Examples
6332 Here is an excerpt of the output:
6334 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
6335 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
6336 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
6337 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
6338 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
6339 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
6340 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
6341 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
6342 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
6348 The mean square energy is approximately -27 dB, or 10^-2.7.
6350 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
6352 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
6355 In other words, raising the volume by +4 dB does not cause any clipping,
6356 raising it by +5 dB causes clipping for 6 samples, etc.
6358 @c man end AUDIO FILTERS
6360 @chapter Audio Sources
6361 @c man begin AUDIO SOURCES
6363 Below is a description of the currently available audio sources.
6367 Buffer audio frames, and make them available to the filter chain.
6369 This source is mainly intended for a programmatic use, in particular
6370 through the interface defined in @file{libavfilter/buffersrc.h}.
6372 It accepts the following parameters:
6376 The timebase which will be used for timestamps of submitted frames. It must be
6377 either a floating-point number or in @var{numerator}/@var{denominator} form.
6380 The sample rate of the incoming audio buffers.
6383 The sample format of the incoming audio buffers.
6384 Either a sample format name or its corresponding integer representation from
6385 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
6387 @item channel_layout
6388 The channel layout of the incoming audio buffers.
6389 Either a channel layout name from channel_layout_map in
6390 @file{libavutil/channel_layout.c} or its corresponding integer representation
6391 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
6394 The number of channels of the incoming audio buffers.
6395 If both @var{channels} and @var{channel_layout} are specified, then they
6400 @subsection Examples
6403 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
6406 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
6407 Since the sample format with name "s16p" corresponds to the number
6408 6 and the "stereo" channel layout corresponds to the value 0x3, this is
6411 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
6416 Generate an audio signal specified by an expression.
6418 This source accepts in input one or more expressions (one for each
6419 channel), which are evaluated and used to generate a corresponding
6422 This source accepts the following options:
6426 Set the '|'-separated expressions list for each separate channel. In case the
6427 @option{channel_layout} option is not specified, the selected channel layout
6428 depends on the number of provided expressions. Otherwise the last
6429 specified expression is applied to the remaining output channels.
6431 @item channel_layout, c
6432 Set the channel layout. The number of channels in the specified layout
6433 must be equal to the number of specified expressions.
6436 Set the minimum duration of the sourced audio. See
6437 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6438 for the accepted syntax.
6439 Note that the resulting duration may be greater than the specified
6440 duration, as the generated audio is always cut at the end of a
6443 If not specified, or the expressed duration is negative, the audio is
6444 supposed to be generated forever.
6447 Set the number of samples per channel per each output frame,
6450 @item sample_rate, s
6451 Specify the sample rate, default to 44100.
6454 Each expression in @var{exprs} can contain the following constants:
6458 number of the evaluated sample, starting from 0
6461 time of the evaluated sample expressed in seconds, starting from 0
6468 @subsection Examples
6478 Generate a sin signal with frequency of 440 Hz, set sample rate to
6481 aevalsrc="sin(440*2*PI*t):s=8000"
6485 Generate a two channels signal, specify the channel layout (Front
6486 Center + Back Center) explicitly:
6488 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
6492 Generate white noise:
6494 aevalsrc="-2+random(0)"
6498 Generate an amplitude modulated signal:
6500 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
6504 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
6506 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
6513 Generate a FIR coefficients using frequency sampling method.
6515 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6517 The filter accepts the following options:
6521 Set number of filter coefficents in output audio stream.
6522 Default value is 1025.
6525 Set frequency points from where magnitude and phase are set.
6526 This must be in non decreasing order, and first element must be 0, while last element
6527 must be 1. Elements are separated by white spaces.
6530 Set magnitude value for every frequency point set by @option{frequency}.
6531 Number of values must be same as number of frequency points.
6532 Values are separated by white spaces.
6535 Set phase value for every frequency point set by @option{frequency}.
6536 Number of values must be same as number of frequency points.
6537 Values are separated by white spaces.
6539 @item sample_rate, r
6540 Set sample rate, default is 44100.
6543 Set number of samples per each frame. Default is 1024.
6546 Set window function. Default is blackman.
6551 The null audio source, return unprocessed audio frames. It is mainly useful
6552 as a template and to be employed in analysis / debugging tools, or as
6553 the source for filters which ignore the input data (for example the sox
6556 This source accepts the following options:
6560 @item channel_layout, cl
6562 Specifies the channel layout, and can be either an integer or a string
6563 representing a channel layout. The default value of @var{channel_layout}
6566 Check the channel_layout_map definition in
6567 @file{libavutil/channel_layout.c} for the mapping between strings and
6568 channel layout values.
6570 @item sample_rate, r
6571 Specifies the sample rate, and defaults to 44100.
6574 Set the number of samples per requested frames.
6577 Set the duration of the sourced audio. See
6578 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6579 for the accepted syntax.
6581 If not specified, or the expressed duration is negative, the audio is
6582 supposed to be generated forever.
6585 @subsection Examples
6589 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
6591 anullsrc=r=48000:cl=4
6595 Do the same operation with a more obvious syntax:
6597 anullsrc=r=48000:cl=mono
6601 All the parameters need to be explicitly defined.
6605 Synthesize a voice utterance using the libflite library.
6607 To enable compilation of this filter you need to configure FFmpeg with
6608 @code{--enable-libflite}.
6610 Note that versions of the flite library prior to 2.0 are not thread-safe.
6612 The filter accepts the following options:
6617 If set to 1, list the names of the available voices and exit
6618 immediately. Default value is 0.
6621 Set the maximum number of samples per frame. Default value is 512.
6624 Set the filename containing the text to speak.
6627 Set the text to speak.
6630 Set the voice to use for the speech synthesis. Default value is
6631 @code{kal}. See also the @var{list_voices} option.
6634 @subsection Examples
6638 Read from file @file{speech.txt}, and synthesize the text using the
6639 standard flite voice:
6641 flite=textfile=speech.txt
6645 Read the specified text selecting the @code{slt} voice:
6647 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6651 Input text to ffmpeg:
6653 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6657 Make @file{ffplay} speak the specified text, using @code{flite} and
6658 the @code{lavfi} device:
6660 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
6664 For more information about libflite, check:
6665 @url{http://www.festvox.org/flite/}
6669 Generate a noise audio signal.
6671 The filter accepts the following options:
6674 @item sample_rate, r
6675 Specify the sample rate. Default value is 48000 Hz.
6678 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
6682 Specify the duration of the generated audio stream. Not specifying this option
6683 results in noise with an infinite length.
6685 @item color, colour, c
6686 Specify the color of noise. Available noise colors are white, pink, brown,
6687 blue, violet and velvet. Default color is white.
6690 Specify a value used to seed the PRNG.
6693 Set the number of samples per each output frame, default is 1024.
6696 @subsection Examples
6701 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6703 anoisesrc=d=60:c=pink:r=44100:a=0.5
6709 Generate odd-tap Hilbert transform FIR coefficients.
6711 The resulting stream can be used with @ref{afir} filter for phase-shifting
6712 the signal by 90 degrees.
6714 This is used in many matrix coding schemes and for analytic signal generation.
6715 The process is often written as a multiplication by i (or j), the imaginary unit.
6717 The filter accepts the following options:
6721 @item sample_rate, s
6722 Set sample rate, default is 44100.
6725 Set length of FIR filter, default is 22051.
6728 Set number of samples per each frame.
6731 Set window function to be used when generating FIR coefficients.
6736 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6738 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6740 The filter accepts the following options:
6743 @item sample_rate, r
6744 Set sample rate, default is 44100.
6747 Set number of samples per each frame. Default is 1024.
6750 Set high-pass frequency. Default is 0.
6753 Set low-pass frequency. Default is 0.
6754 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6755 is higher than 0 then filter will create band-pass filter coefficients,
6756 otherwise band-reject filter coefficients.
6759 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6762 Set Kaiser window beta.
6765 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6768 Enable rounding, by default is disabled.
6771 Set number of taps for high-pass filter.
6774 Set number of taps for low-pass filter.
6779 Generate an audio signal made of a sine wave with amplitude 1/8.
6781 The audio signal is bit-exact.
6783 The filter accepts the following options:
6788 Set the carrier frequency. Default is 440 Hz.
6790 @item beep_factor, b
6791 Enable a periodic beep every second with frequency @var{beep_factor} times
6792 the carrier frequency. Default is 0, meaning the beep is disabled.
6794 @item sample_rate, r
6795 Specify the sample rate, default is 44100.
6798 Specify the duration of the generated audio stream.
6800 @item samples_per_frame
6801 Set the number of samples per output frame.
6803 The expression can contain the following constants:
6807 The (sequential) number of the output audio frame, starting from 0.
6810 The PTS (Presentation TimeStamp) of the output audio frame,
6811 expressed in @var{TB} units.
6814 The PTS of the output audio frame, expressed in seconds.
6817 The timebase of the output audio frames.
6820 Default is @code{1024}.
6823 @subsection Examples
6828 Generate a simple 440 Hz sine wave:
6834 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6838 sine=frequency=220:beep_factor=4:duration=5
6842 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6845 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6849 @c man end AUDIO SOURCES
6851 @chapter Audio Sinks
6852 @c man begin AUDIO SINKS
6854 Below is a description of the currently available audio sinks.
6856 @section abuffersink
6858 Buffer audio frames, and make them available to the end of filter chain.
6860 This sink is mainly intended for programmatic use, in particular
6861 through the interface defined in @file{libavfilter/buffersink.h}
6862 or the options system.
6864 It accepts a pointer to an AVABufferSinkContext structure, which
6865 defines the incoming buffers' formats, to be passed as the opaque
6866 parameter to @code{avfilter_init_filter} for initialization.
6869 Null audio sink; do absolutely nothing with the input audio. It is
6870 mainly useful as a template and for use in analysis / debugging
6873 @c man end AUDIO SINKS
6875 @chapter Video Filters
6876 @c man begin VIDEO FILTERS
6878 When you configure your FFmpeg build, you can disable any of the
6879 existing filters using @code{--disable-filters}.
6880 The configure output will show the video filters included in your
6883 Below is a description of the currently available video filters.
6887 Mark a region of interest in a video frame.
6889 The frame data is passed through unchanged, but metadata is attached
6890 to the frame indicating regions of interest which can affect the
6891 behaviour of later encoding. Multiple regions can be marked by
6892 applying the filter multiple times.
6896 Region distance in pixels from the left edge of the frame.
6898 Region distance in pixels from the top edge of the frame.
6900 Region width in pixels.
6902 Region height in pixels.
6904 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6905 and may contain the following variables:
6908 Width of the input frame.
6910 Height of the input frame.
6914 Quantisation offset to apply within the region.
6916 This must be a real value in the range -1 to +1. A value of zero
6917 indicates no quality change. A negative value asks for better quality
6918 (less quantisation), while a positive value asks for worse quality
6919 (greater quantisation).
6921 The range is calibrated so that the extreme values indicate the
6922 largest possible offset - if the rest of the frame is encoded with the
6923 worst possible quality, an offset of -1 indicates that this region
6924 should be encoded with the best possible quality anyway. Intermediate
6925 values are then interpolated in some codec-dependent way.
6927 For example, in 10-bit H.264 the quantisation parameter varies between
6928 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6929 this region should be encoded with a QP around one-tenth of the full
6930 range better than the rest of the frame. So, if most of the frame
6931 were to be encoded with a QP of around 30, this region would get a QP
6932 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6933 An extreme value of -1 would indicate that this region should be
6934 encoded with the best possible quality regardless of the treatment of
6935 the rest of the frame - that is, should be encoded at a QP of -12.
6937 If set to true, remove any existing regions of interest marked on the
6938 frame before adding the new one.
6941 @subsection Examples
6945 Mark the centre quarter of the frame as interesting.
6947 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6950 Mark the 100-pixel-wide region on the left edge of the frame as very
6951 uninteresting (to be encoded at much lower quality than the rest of
6954 addroi=0:0:100:ih:+1/5
6958 @section alphaextract
6960 Extract the alpha component from the input as a grayscale video. This
6961 is especially useful with the @var{alphamerge} filter.
6965 Add or replace the alpha component of the primary input with the
6966 grayscale value of a second input. This is intended for use with
6967 @var{alphaextract} to allow the transmission or storage of frame
6968 sequences that have alpha in a format that doesn't support an alpha
6971 For example, to reconstruct full frames from a normal YUV-encoded video
6972 and a separate video created with @var{alphaextract}, you might use:
6974 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6979 Amplify differences between current pixel and pixels of adjacent frames in
6980 same pixel location.
6982 This filter accepts the following options:
6986 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6987 For example radius of 3 will instruct filter to calculate average of 7 frames.
6990 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6993 Set threshold for difference amplification. Any difference greater or equal to
6994 this value will not alter source pixel. Default is 10.
6995 Allowed range is from 0 to 65535.
6998 Set tolerance for difference amplification. Any difference lower to
6999 this value will not alter source pixel. Default is 0.
7000 Allowed range is from 0 to 65535.
7003 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
7004 This option controls maximum possible value that will decrease source pixel value.
7007 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
7008 This option controls maximum possible value that will increase source pixel value.
7011 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
7014 @subsection Commands
7016 This filter supports the following @ref{commands} that corresponds to option of same name:
7028 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
7029 and libavformat to work. On the other hand, it is limited to ASS (Advanced
7030 Substation Alpha) subtitles files.
7032 This filter accepts the following option in addition to the common options from
7033 the @ref{subtitles} filter:
7037 Set the shaping engine
7039 Available values are:
7042 The default libass shaping engine, which is the best available.
7044 Fast, font-agnostic shaper that can do only substitutions
7046 Slower shaper using OpenType for substitutions and positioning
7049 The default is @code{auto}.
7053 Apply an Adaptive Temporal Averaging Denoiser to the video input.
7055 The filter accepts the following options:
7059 Set threshold A for 1st plane. Default is 0.02.
7060 Valid range is 0 to 0.3.
7063 Set threshold B for 1st plane. Default is 0.04.
7064 Valid range is 0 to 5.
7067 Set threshold A for 2nd plane. Default is 0.02.
7068 Valid range is 0 to 0.3.
7071 Set threshold B for 2nd plane. Default is 0.04.
7072 Valid range is 0 to 5.
7075 Set threshold A for 3rd plane. Default is 0.02.
7076 Valid range is 0 to 0.3.
7079 Set threshold B for 3rd plane. Default is 0.04.
7080 Valid range is 0 to 5.
7082 Threshold A is designed to react on abrupt changes in the input signal and
7083 threshold B is designed to react on continuous changes in the input signal.
7086 Set number of frames filter will use for averaging. Default is 9. Must be odd
7087 number in range [5, 129].
7090 Set what planes of frame filter will use for averaging. Default is all.
7093 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
7094 Alternatively can be set to @code{s} serial.
7096 Parallel can be faster then serial, while other way around is never true.
7097 Parallel will abort early on first change being greater then thresholds, while serial
7098 will continue processing other side of frames if they are equal or below thresholds.
7101 @subsection Commands
7102 This filter supports same @ref{commands} as options except option @code{s}.
7103 The command accepts the same syntax of the corresponding option.
7107 Apply average blur filter.
7109 The filter accepts the following options:
7113 Set horizontal radius size.
7116 Set which planes to filter. By default all planes are filtered.
7119 Set vertical radius size, if zero it will be same as @code{sizeX}.
7120 Default is @code{0}.
7123 @subsection Commands
7124 This filter supports same commands as options.
7125 The command accepts the same syntax of the corresponding option.
7127 If the specified expression is not valid, it is kept at its current
7132 Compute the bounding box for the non-black pixels in the input frame
7135 This filter computes the bounding box containing all the pixels with a
7136 luminance value greater than the minimum allowed value.
7137 The parameters describing the bounding box are printed on the filter
7140 The filter accepts the following option:
7144 Set the minimal luminance value. Default is @code{16}.
7148 Apply bilateral filter, spatial smoothing while preserving edges.
7150 The filter accepts the following options:
7153 Set sigma of gaussian function to calculate spatial weight.
7154 Allowed range is 0 to 512. Default is 0.1.
7157 Set sigma of gaussian function to calculate range weight.
7158 Allowed range is 0 to 1. Default is 0.1.
7161 Set planes to filter. Default is first only.
7164 @subsection Commands
7166 This filter supports the all above options as @ref{commands}.
7168 @section bitplanenoise
7170 Show and measure bit plane noise.
7172 The filter accepts the following options:
7176 Set which plane to analyze. Default is @code{1}.
7179 Filter out noisy pixels from @code{bitplane} set above.
7180 Default is disabled.
7183 @section blackdetect
7185 Detect video intervals that are (almost) completely black. Can be
7186 useful to detect chapter transitions, commercials, or invalid
7189 The filter outputs its detection analysis to both the log as well as
7190 frame metadata. If a black segment of at least the specified minimum
7191 duration is found, a line with the start and end timestamps as well
7192 as duration is printed to the log with level @code{info}. In addition,
7193 a log line with level @code{debug} is printed per frame showing the
7194 black amount detected for that frame.
7196 The filter also attaches metadata to the first frame of a black
7197 segment with key @code{lavfi.black_start} and to the first frame
7198 after the black segment ends with key @code{lavfi.black_end}. The
7199 value is the frame's timestamp. This metadata is added regardless
7200 of the minimum duration specified.
7202 The filter accepts the following options:
7205 @item black_min_duration, d
7206 Set the minimum detected black duration expressed in seconds. It must
7207 be a non-negative floating point number.
7209 Default value is 2.0.
7211 @item picture_black_ratio_th, pic_th
7212 Set the threshold for considering a picture "black".
7213 Express the minimum value for the ratio:
7215 @var{nb_black_pixels} / @var{nb_pixels}
7218 for which a picture is considered black.
7219 Default value is 0.98.
7221 @item pixel_black_th, pix_th
7222 Set the threshold for considering a pixel "black".
7224 The threshold expresses the maximum pixel luminance value for which a
7225 pixel is considered "black". The provided value is scaled according to
7226 the following equation:
7228 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
7231 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
7232 the input video format, the range is [0-255] for YUV full-range
7233 formats and [16-235] for YUV non full-range formats.
7235 Default value is 0.10.
7238 The following example sets the maximum pixel threshold to the minimum
7239 value, and detects only black intervals of 2 or more seconds:
7241 blackdetect=d=2:pix_th=0.00
7246 Detect frames that are (almost) completely black. Can be useful to
7247 detect chapter transitions or commercials. Output lines consist of
7248 the frame number of the detected frame, the percentage of blackness,
7249 the position in the file if known or -1 and the timestamp in seconds.
7251 In order to display the output lines, you need to set the loglevel at
7252 least to the AV_LOG_INFO value.
7254 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
7255 The value represents the percentage of pixels in the picture that
7256 are below the threshold value.
7258 It accepts the following parameters:
7263 The percentage of the pixels that have to be below the threshold; it defaults to
7266 @item threshold, thresh
7267 The threshold below which a pixel value is considered black; it defaults to
7275 Blend two video frames into each other.
7277 The @code{blend} filter takes two input streams and outputs one
7278 stream, the first input is the "top" layer and second input is
7279 "bottom" layer. By default, the output terminates when the longest input terminates.
7281 The @code{tblend} (time blend) filter takes two consecutive frames
7282 from one single stream, and outputs the result obtained by blending
7283 the new frame on top of the old frame.
7285 A description of the accepted options follows.
7293 Set blend mode for specific pixel component or all pixel components in case
7294 of @var{all_mode}. Default value is @code{normal}.
7296 Available values for component modes are:
7338 Set blend opacity for specific pixel component or all pixel components in case
7339 of @var{all_opacity}. Only used in combination with pixel component blend modes.
7346 Set blend expression for specific pixel component or all pixel components in case
7347 of @var{all_expr}. Note that related mode options will be ignored if those are set.
7349 The expressions can use the following variables:
7353 The sequential number of the filtered frame, starting from @code{0}.
7357 the coordinates of the current sample
7361 the width and height of currently filtered plane
7365 Width and height scale for the plane being filtered. It is the
7366 ratio between the dimensions of the current plane to the luma plane,
7367 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
7368 the luma plane and @code{0.5,0.5} for the chroma planes.
7371 Time of the current frame, expressed in seconds.
7374 Value of pixel component at current location for first video frame (top layer).
7377 Value of pixel component at current location for second video frame (bottom layer).
7381 The @code{blend} filter also supports the @ref{framesync} options.
7383 @subsection Examples
7387 Apply transition from bottom layer to top layer in first 10 seconds:
7389 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
7393 Apply linear horizontal transition from top layer to bottom layer:
7395 blend=all_expr='A*(X/W)+B*(1-X/W)'
7399 Apply 1x1 checkerboard effect:
7401 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
7405 Apply uncover left effect:
7407 blend=all_expr='if(gte(N*SW+X,W),A,B)'
7411 Apply uncover down effect:
7413 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
7417 Apply uncover up-left effect:
7419 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
7423 Split diagonally video and shows top and bottom layer on each side:
7425 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
7429 Display differences between the current and the previous frame:
7431 tblend=all_mode=grainextract
7437 Denoise frames using Block-Matching 3D algorithm.
7439 The filter accepts the following options.
7443 Set denoising strength. Default value is 1.
7444 Allowed range is from 0 to 999.9.
7445 The denoising algorithm is very sensitive to sigma, so adjust it
7446 according to the source.
7449 Set local patch size. This sets dimensions in 2D.
7452 Set sliding step for processing blocks. Default value is 4.
7453 Allowed range is from 1 to 64.
7454 Smaller values allows processing more reference blocks and is slower.
7457 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
7458 When set to 1, no block matching is done. Larger values allows more blocks
7460 Allowed range is from 1 to 256.
7463 Set radius for search block matching. Default is 9.
7464 Allowed range is from 1 to INT32_MAX.
7467 Set step between two search locations for block matching. Default is 1.
7468 Allowed range is from 1 to 64. Smaller is slower.
7471 Set threshold of mean square error for block matching. Valid range is 0 to
7475 Set thresholding parameter for hard thresholding in 3D transformed domain.
7476 Larger values results in stronger hard-thresholding filtering in frequency
7480 Set filtering estimation mode. Can be @code{basic} or @code{final}.
7481 Default is @code{basic}.
7484 If enabled, filter will use 2nd stream for block matching.
7485 Default is disabled for @code{basic} value of @var{estim} option,
7486 and always enabled if value of @var{estim} is @code{final}.
7489 Set planes to filter. Default is all available except alpha.
7492 @subsection Examples
7496 Basic filtering with bm3d:
7498 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
7502 Same as above, but filtering only luma:
7504 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
7508 Same as above, but with both estimation modes:
7510 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
7514 Same as above, but prefilter with @ref{nlmeans} filter instead:
7516 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
7522 Apply a boxblur algorithm to the input video.
7524 It accepts the following parameters:
7528 @item luma_radius, lr
7529 @item luma_power, lp
7530 @item chroma_radius, cr
7531 @item chroma_power, cp
7532 @item alpha_radius, ar
7533 @item alpha_power, ap
7537 A description of the accepted options follows.
7540 @item luma_radius, lr
7541 @item chroma_radius, cr
7542 @item alpha_radius, ar
7543 Set an expression for the box radius in pixels used for blurring the
7544 corresponding input plane.
7546 The radius value must be a non-negative number, and must not be
7547 greater than the value of the expression @code{min(w,h)/2} for the
7548 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
7551 Default value for @option{luma_radius} is "2". If not specified,
7552 @option{chroma_radius} and @option{alpha_radius} default to the
7553 corresponding value set for @option{luma_radius}.
7555 The expressions can contain the following constants:
7559 The input width and height in pixels.
7563 The input chroma image width and height in pixels.
7567 The horizontal and vertical chroma subsample values. For example, for the
7568 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
7571 @item luma_power, lp
7572 @item chroma_power, cp
7573 @item alpha_power, ap
7574 Specify how many times the boxblur filter is applied to the
7575 corresponding plane.
7577 Default value for @option{luma_power} is 2. If not specified,
7578 @option{chroma_power} and @option{alpha_power} default to the
7579 corresponding value set for @option{luma_power}.
7581 A value of 0 will disable the effect.
7584 @subsection Examples
7588 Apply a boxblur filter with the luma, chroma, and alpha radii
7591 boxblur=luma_radius=2:luma_power=1
7596 Set the luma radius to 2, and alpha and chroma radius to 0:
7598 boxblur=2:1:cr=0:ar=0
7602 Set the luma and chroma radii to a fraction of the video dimension:
7604 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
7610 Deinterlace the input video ("bwdif" stands for "Bob Weaver
7611 Deinterlacing Filter").
7613 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
7614 interpolation algorithms.
7615 It accepts the following parameters:
7619 The interlacing mode to adopt. It accepts one of the following values:
7623 Output one frame for each frame.
7625 Output one frame for each field.
7628 The default value is @code{send_field}.
7631 The picture field parity assumed for the input interlaced video. It accepts one
7632 of the following values:
7636 Assume the top field is first.
7638 Assume the bottom field is first.
7640 Enable automatic detection of field parity.
7643 The default value is @code{auto}.
7644 If the interlacing is unknown or the decoder does not export this information,
7645 top field first will be assumed.
7648 Specify which frames to deinterlace. Accepts one of the following
7653 Deinterlace all frames.
7655 Only deinterlace frames marked as interlaced.
7658 The default value is @code{all}.
7663 Apply Contrast Adaptive Sharpen filter to video stream.
7665 The filter accepts the following options:
7669 Set the sharpening strength. Default value is 0.
7672 Set planes to filter. Default value is to filter all
7673 planes except alpha plane.
7676 @subsection Commands
7677 This filter supports same @ref{commands} as options.
7680 Remove all color information for all colors except for certain one.
7682 The filter accepts the following options:
7686 The color which will not be replaced with neutral chroma.
7689 Similarity percentage with the above color.
7690 0.01 matches only the exact key color, while 1.0 matches everything.
7694 0.0 makes pixels either fully gray, or not gray at all.
7695 Higher values result in more preserved color.
7698 Signals that the color passed is already in YUV instead of RGB.
7700 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7701 This can be used to pass exact YUV values as hexadecimal numbers.
7704 @subsection Commands
7705 This filter supports same @ref{commands} as options.
7706 The command accepts the same syntax of the corresponding option.
7708 If the specified expression is not valid, it is kept at its current
7712 YUV colorspace color/chroma keying.
7714 The filter accepts the following options:
7718 The color which will be replaced with transparency.
7721 Similarity percentage with the key color.
7723 0.01 matches only the exact key color, while 1.0 matches everything.
7728 0.0 makes pixels either fully transparent, or not transparent at all.
7730 Higher values result in semi-transparent pixels, with a higher transparency
7731 the more similar the pixels color is to the key color.
7734 Signals that the color passed is already in YUV instead of RGB.
7736 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7737 This can be used to pass exact YUV values as hexadecimal numbers.
7740 @subsection Commands
7741 This filter supports same @ref{commands} as options.
7742 The command accepts the same syntax of the corresponding option.
7744 If the specified expression is not valid, it is kept at its current
7747 @subsection Examples
7751 Make every green pixel in the input image transparent:
7753 ffmpeg -i input.png -vf chromakey=green out.png
7757 Overlay a greenscreen-video on top of a static black background.
7759 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
7764 Reduce chrominance noise.
7766 The filter accepts the following options:
7770 Set threshold for averaging chrominance values.
7771 Sum of absolute difference of Y, U and V pixel components of current
7772 pixel and neighbour pixels lower than this threshold will be used in
7773 averaging. Luma component is left unchanged and is copied to output.
7774 Default value is 30. Allowed range is from 1 to 200.
7777 Set horizontal radius of rectangle used for averaging.
7778 Allowed range is from 1 to 100. Default value is 5.
7781 Set vertical radius of rectangle used for averaging.
7782 Allowed range is from 1 to 100. Default value is 5.
7785 Set horizontal step when averaging. Default value is 1.
7786 Allowed range is from 1 to 50.
7787 Mostly useful to speed-up filtering.
7790 Set vertical step when averaging. Default value is 1.
7791 Allowed range is from 1 to 50.
7792 Mostly useful to speed-up filtering.
7795 Set Y threshold for averaging chrominance values.
7796 Set finer control for max allowed difference between Y components
7797 of current pixel and neigbour pixels.
7798 Default value is 200. Allowed range is from 1 to 200.
7801 Set U threshold for averaging chrominance values.
7802 Set finer control for max allowed difference between U components
7803 of current pixel and neigbour pixels.
7804 Default value is 200. Allowed range is from 1 to 200.
7807 Set V threshold for averaging chrominance values.
7808 Set finer control for max allowed difference between V components
7809 of current pixel and neigbour pixels.
7810 Default value is 200. Allowed range is from 1 to 200.
7813 @subsection Commands
7814 This filter supports same @ref{commands} as options.
7815 The command accepts the same syntax of the corresponding option.
7817 @section chromashift
7818 Shift chroma pixels horizontally and/or vertically.
7820 The filter accepts the following options:
7823 Set amount to shift chroma-blue horizontally.
7825 Set amount to shift chroma-blue vertically.
7827 Set amount to shift chroma-red horizontally.
7829 Set amount to shift chroma-red vertically.
7831 Set edge mode, can be @var{smear}, default, or @var{warp}.
7834 @subsection Commands
7836 This filter supports the all above options as @ref{commands}.
7840 Display CIE color diagram with pixels overlaid onto it.
7842 The filter accepts the following options:
7857 @item uhdtv, rec2020
7871 Set what gamuts to draw.
7873 See @code{system} option for available values.
7876 Set ciescope size, by default set to 512.
7879 Set intensity used to map input pixel values to CIE diagram.
7882 Set contrast used to draw tongue colors that are out of active color system gamut.
7885 Correct gamma displayed on scope, by default enabled.
7888 Show white point on CIE diagram, by default disabled.
7891 Set input gamma. Used only with XYZ input color space.
7896 Visualize information exported by some codecs.
7898 Some codecs can export information through frames using side-data or other
7899 means. For example, some MPEG based codecs export motion vectors through the
7900 @var{export_mvs} flag in the codec @option{flags2} option.
7902 The filter accepts the following option:
7906 Set motion vectors to visualize.
7908 Available flags for @var{mv} are:
7912 forward predicted MVs of P-frames
7914 forward predicted MVs of B-frames
7916 backward predicted MVs of B-frames
7920 Display quantization parameters using the chroma planes.
7923 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7925 Available flags for @var{mv_type} are:
7929 forward predicted MVs
7931 backward predicted MVs
7934 @item frame_type, ft
7935 Set frame type to visualize motion vectors of.
7937 Available flags for @var{frame_type} are:
7941 intra-coded frames (I-frames)
7943 predicted frames (P-frames)
7945 bi-directionally predicted frames (B-frames)
7949 @subsection Examples
7953 Visualize forward predicted MVs of all frames using @command{ffplay}:
7955 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7959 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7961 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7965 @section colorbalance
7966 Modify intensity of primary colors (red, green and blue) of input frames.
7968 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7969 regions for the red-cyan, green-magenta or blue-yellow balance.
7971 A positive adjustment value shifts the balance towards the primary color, a negative
7972 value towards the complementary color.
7974 The filter accepts the following options:
7980 Adjust red, green and blue shadows (darkest pixels).
7985 Adjust red, green and blue midtones (medium pixels).
7990 Adjust red, green and blue highlights (brightest pixels).
7992 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7995 Preserve lightness when changing color balance. Default is disabled.
7998 @subsection Examples
8002 Add red color cast to shadows:
8008 @subsection Commands
8010 This filter supports the all above options as @ref{commands}.
8012 @section colorchannelmixer
8014 Adjust video input frames by re-mixing color channels.
8016 This filter modifies a color channel by adding the values associated to
8017 the other channels of the same pixels. For example if the value to
8018 modify is red, the output value will be:
8020 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
8023 The filter accepts the following options:
8030 Adjust contribution of input red, green, blue and alpha channels for output red channel.
8031 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
8037 Adjust contribution of input red, green, blue and alpha channels for output green channel.
8038 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
8044 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
8045 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
8051 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
8052 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
8054 Allowed ranges for options are @code{[-2.0, 2.0]}.
8057 @subsection Examples
8061 Convert source to grayscale:
8063 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
8066 Simulate sepia tones:
8068 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
8072 @subsection Commands
8074 This filter supports the all above options as @ref{commands}.
8077 RGB colorspace color keying.
8079 The filter accepts the following options:
8083 The color which will be replaced with transparency.
8086 Similarity percentage with the key color.
8088 0.01 matches only the exact key color, while 1.0 matches everything.
8093 0.0 makes pixels either fully transparent, or not transparent at all.
8095 Higher values result in semi-transparent pixels, with a higher transparency
8096 the more similar the pixels color is to the key color.
8099 @subsection Examples
8103 Make every green pixel in the input image transparent:
8105 ffmpeg -i input.png -vf colorkey=green out.png
8109 Overlay a greenscreen-video on top of a static background image.
8111 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
8115 @subsection Commands
8116 This filter supports same @ref{commands} as options.
8117 The command accepts the same syntax of the corresponding option.
8119 If the specified expression is not valid, it is kept at its current
8123 Remove all color information for all RGB colors except for certain one.
8125 The filter accepts the following options:
8129 The color which will not be replaced with neutral gray.
8132 Similarity percentage with the above color.
8133 0.01 matches only the exact key color, while 1.0 matches everything.
8136 Blend percentage. 0.0 makes pixels fully gray.
8137 Higher values result in more preserved color.
8140 @subsection Commands
8141 This filter supports same @ref{commands} as options.
8142 The command accepts the same syntax of the corresponding option.
8144 If the specified expression is not valid, it is kept at its current
8147 @section colorlevels
8149 Adjust video input frames using levels.
8151 The filter accepts the following options:
8158 Adjust red, green, blue and alpha input black point.
8159 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
8165 Adjust red, green, blue and alpha input white point.
8166 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
8168 Input levels are used to lighten highlights (bright tones), darken shadows
8169 (dark tones), change the balance of bright and dark tones.
8175 Adjust red, green, blue and alpha output black point.
8176 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
8182 Adjust red, green, blue and alpha output white point.
8183 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
8185 Output levels allows manual selection of a constrained output level range.
8188 @subsection Examples
8192 Make video output darker:
8194 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
8200 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
8204 Make video output lighter:
8206 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
8210 Increase brightness:
8212 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
8216 @subsection Commands
8218 This filter supports the all above options as @ref{commands}.
8220 @section colormatrix
8222 Convert color matrix.
8224 The filter accepts the following options:
8229 Specify the source and destination color matrix. Both values must be
8232 The accepted values are:
8260 For example to convert from BT.601 to SMPTE-240M, use the command:
8262 colormatrix=bt601:smpte240m
8267 Convert colorspace, transfer characteristics or color primaries.
8268 Input video needs to have an even size.
8270 The filter accepts the following options:
8275 Specify all color properties at once.
8277 The accepted values are:
8307 Specify output colorspace.
8309 The accepted values are:
8318 BT.470BG or BT.601-6 625
8321 SMPTE-170M or BT.601-6 525
8330 BT.2020 with non-constant luminance
8336 Specify output transfer characteristics.
8338 The accepted values are:
8350 Constant gamma of 2.2
8353 Constant gamma of 2.8
8356 SMPTE-170M, BT.601-6 625 or BT.601-6 525
8374 BT.2020 for 10-bits content
8377 BT.2020 for 12-bits content
8383 Specify output color primaries.
8385 The accepted values are:
8394 BT.470BG or BT.601-6 625
8397 SMPTE-170M or BT.601-6 525
8421 Specify output color range.
8423 The accepted values are:
8426 TV (restricted) range
8429 MPEG (restricted) range
8440 Specify output color format.
8442 The accepted values are:
8445 YUV 4:2:0 planar 8-bits
8448 YUV 4:2:0 planar 10-bits
8451 YUV 4:2:0 planar 12-bits
8454 YUV 4:2:2 planar 8-bits
8457 YUV 4:2:2 planar 10-bits
8460 YUV 4:2:2 planar 12-bits
8463 YUV 4:4:4 planar 8-bits
8466 YUV 4:4:4 planar 10-bits
8469 YUV 4:4:4 planar 12-bits
8474 Do a fast conversion, which skips gamma/primary correction. This will take
8475 significantly less CPU, but will be mathematically incorrect. To get output
8476 compatible with that produced by the colormatrix filter, use fast=1.
8479 Specify dithering mode.
8481 The accepted values are:
8487 Floyd-Steinberg dithering
8491 Whitepoint adaptation mode.
8493 The accepted values are:
8496 Bradford whitepoint adaptation
8499 von Kries whitepoint adaptation
8502 identity whitepoint adaptation (i.e. no whitepoint adaptation)
8506 Override all input properties at once. Same accepted values as @ref{all}.
8509 Override input colorspace. Same accepted values as @ref{space}.
8512 Override input color primaries. Same accepted values as @ref{primaries}.
8515 Override input transfer characteristics. Same accepted values as @ref{trc}.
8518 Override input color range. Same accepted values as @ref{range}.
8522 The filter converts the transfer characteristics, color space and color
8523 primaries to the specified user values. The output value, if not specified,
8524 is set to a default value based on the "all" property. If that property is
8525 also not specified, the filter will log an error. The output color range and
8526 format default to the same value as the input color range and format. The
8527 input transfer characteristics, color space, color primaries and color range
8528 should be set on the input data. If any of these are missing, the filter will
8529 log an error and no conversion will take place.
8531 For example to convert the input to SMPTE-240M, use the command:
8533 colorspace=smpte240m
8536 @section convolution
8538 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
8540 The filter accepts the following options:
8547 Set matrix for each plane.
8548 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
8549 and from 1 to 49 odd number of signed integers in @var{row} mode.
8555 Set multiplier for calculated value for each plane.
8556 If unset or 0, it will be sum of all matrix elements.
8562 Set bias for each plane. This value is added to the result of the multiplication.
8563 Useful for making the overall image brighter or darker. Default is 0.0.
8569 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
8570 Default is @var{square}.
8573 @subsection Commands
8575 This filter supports the all above options as @ref{commands}.
8577 @subsection Examples
8583 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"
8589 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"
8595 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"
8601 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"
8605 Apply laplacian edge detector which includes diagonals:
8607 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"
8613 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"
8619 Apply 2D convolution of video stream in frequency domain using second stream
8622 The filter accepts the following options:
8626 Set which planes to process.
8629 Set which impulse video frames will be processed, can be @var{first}
8630 or @var{all}. Default is @var{all}.
8633 The @code{convolve} filter also supports the @ref{framesync} options.
8637 Copy the input video source unchanged to the output. This is mainly useful for
8642 Video filtering on GPU using Apple's CoreImage API on OSX.
8644 Hardware acceleration is based on an OpenGL context. Usually, this means it is
8645 processed by video hardware. However, software-based OpenGL implementations
8646 exist which means there is no guarantee for hardware processing. It depends on
8649 There are many filters and image generators provided by Apple that come with a
8650 large variety of options. The filter has to be referenced by its name along
8653 The coreimage filter accepts the following options:
8656 List all available filters and generators along with all their respective
8657 options as well as possible minimum and maximum values along with the default
8664 Specify all filters by their respective name and options.
8665 Use @var{list_filters} to determine all valid filter names and options.
8666 Numerical options are specified by a float value and are automatically clamped
8667 to their respective value range. Vector and color options have to be specified
8668 by a list of space separated float values. Character escaping has to be done.
8669 A special option name @code{default} is available to use default options for a
8672 It is required to specify either @code{default} or at least one of the filter options.
8673 All omitted options are used with their default values.
8674 The syntax of the filter string is as follows:
8676 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
8680 Specify a rectangle where the output of the filter chain is copied into the
8681 input image. It is given by a list of space separated float values:
8683 output_rect=x\ y\ width\ height
8685 If not given, the output rectangle equals the dimensions of the input image.
8686 The output rectangle is automatically cropped at the borders of the input
8687 image. Negative values are valid for each component.
8689 output_rect=25\ 25\ 100\ 100
8693 Several filters can be chained for successive processing without GPU-HOST
8694 transfers allowing for fast processing of complex filter chains.
8695 Currently, only filters with zero (generators) or exactly one (filters) input
8696 image and one output image are supported. Also, transition filters are not yet
8699 Some filters generate output images with additional padding depending on the
8700 respective filter kernel. The padding is automatically removed to ensure the
8701 filter output has the same size as the input image.
8703 For image generators, the size of the output image is determined by the
8704 previous output image of the filter chain or the input image of the whole
8705 filterchain, respectively. The generators do not use the pixel information of
8706 this image to generate their output. However, the generated output is
8707 blended onto this image, resulting in partial or complete coverage of the
8710 The @ref{coreimagesrc} video source can be used for generating input images
8711 which are directly fed into the filter chain. By using it, providing input
8712 images by another video source or an input video is not required.
8714 @subsection Examples
8719 List all filters available:
8721 coreimage=list_filters=true
8725 Use the CIBoxBlur filter with default options to blur an image:
8727 coreimage=filter=CIBoxBlur@@default
8731 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8732 its center at 100x100 and a radius of 50 pixels:
8734 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8738 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8739 given as complete and escaped command-line for Apple's standard bash shell:
8741 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8747 Cover a rectangular object
8749 It accepts the following options:
8753 Filepath of the optional cover image, needs to be in yuv420.
8758 It accepts the following values:
8761 cover it by the supplied image
8763 cover it by interpolating the surrounding pixels
8766 Default value is @var{blur}.
8769 @subsection Examples
8773 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8775 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8781 Crop the input video to given dimensions.
8783 It accepts the following parameters:
8787 The width of the output video. It defaults to @code{iw}.
8788 This expression is evaluated only once during the filter
8789 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8792 The height of the output video. It defaults to @code{ih}.
8793 This expression is evaluated only once during the filter
8794 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8797 The horizontal position, in the input video, of the left edge of the output
8798 video. It defaults to @code{(in_w-out_w)/2}.
8799 This expression is evaluated per-frame.
8802 The vertical position, in the input video, of the top edge of the output video.
8803 It defaults to @code{(in_h-out_h)/2}.
8804 This expression is evaluated per-frame.
8807 If set to 1 will force the output display aspect ratio
8808 to be the same of the input, by changing the output sample aspect
8809 ratio. It defaults to 0.
8812 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8813 width/height/x/y as specified and will not be rounded to nearest smaller value.
8817 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8818 expressions containing the following constants:
8823 The computed values for @var{x} and @var{y}. They are evaluated for
8828 The input width and height.
8832 These are the same as @var{in_w} and @var{in_h}.
8836 The output (cropped) width and height.
8840 These are the same as @var{out_w} and @var{out_h}.
8843 same as @var{iw} / @var{ih}
8846 input sample aspect ratio
8849 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8853 horizontal and vertical chroma subsample values. For example for the
8854 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8857 The number of the input frame, starting from 0.
8860 the position in the file of the input frame, NAN if unknown
8863 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8867 The expression for @var{out_w} may depend on the value of @var{out_h},
8868 and the expression for @var{out_h} may depend on @var{out_w}, but they
8869 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8870 evaluated after @var{out_w} and @var{out_h}.
8872 The @var{x} and @var{y} parameters specify the expressions for the
8873 position of the top-left corner of the output (non-cropped) area. They
8874 are evaluated for each frame. If the evaluated value is not valid, it
8875 is approximated to the nearest valid value.
8877 The expression for @var{x} may depend on @var{y}, and the expression
8878 for @var{y} may depend on @var{x}.
8880 @subsection Examples
8884 Crop area with size 100x100 at position (12,34).
8889 Using named options, the example above becomes:
8891 crop=w=100:h=100:x=12:y=34
8895 Crop the central input area with size 100x100:
8901 Crop the central input area with size 2/3 of the input video:
8903 crop=2/3*in_w:2/3*in_h
8907 Crop the input video central square:
8914 Delimit the rectangle with the top-left corner placed at position
8915 100:100 and the right-bottom corner corresponding to the right-bottom
8916 corner of the input image.
8918 crop=in_w-100:in_h-100:100:100
8922 Crop 10 pixels from the left and right borders, and 20 pixels from
8923 the top and bottom borders
8925 crop=in_w-2*10:in_h-2*20
8929 Keep only the bottom right quarter of the input image:
8931 crop=in_w/2:in_h/2:in_w/2:in_h/2
8935 Crop height for getting Greek harmony:
8937 crop=in_w:1/PHI*in_w
8941 Apply trembling effect:
8943 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)
8947 Apply erratic camera effect depending on timestamp:
8949 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)"
8953 Set x depending on the value of y:
8955 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8959 @subsection Commands
8961 This filter supports the following commands:
8967 Set width/height of the output video and the horizontal/vertical position
8969 The command accepts the same syntax of the corresponding option.
8971 If the specified expression is not valid, it is kept at its current
8977 Auto-detect the crop size.
8979 It calculates the necessary cropping parameters and prints the
8980 recommended parameters via the logging system. The detected dimensions
8981 correspond to the non-black area of the input video.
8983 It accepts the following parameters:
8988 Set higher black value threshold, which can be optionally specified
8989 from nothing (0) to everything (255 for 8-bit based formats). An intensity
8990 value greater to the set value is considered non-black. It defaults to 24.
8991 You can also specify a value between 0.0 and 1.0 which will be scaled depending
8992 on the bitdepth of the pixel format.
8995 The value which the width/height should be divisible by. It defaults to
8996 16. The offset is automatically adjusted to center the video. Use 2 to
8997 get only even dimensions (needed for 4:2:2 video). 16 is best when
8998 encoding to most video codecs.
9001 Set the number of initial frames for which evaluation is skipped.
9002 Default is 2. Range is 0 to INT_MAX.
9004 @item reset_count, reset
9005 Set the counter that determines after how many frames cropdetect will
9006 reset the previously detected largest video area and start over to
9007 detect the current optimal crop area. Default value is 0.
9009 This can be useful when channel logos distort the video area. 0
9010 indicates 'never reset', and returns the largest area encountered during
9017 Delay video filtering until a given wallclock timestamp. The filter first
9018 passes on @option{preroll} amount of frames, then it buffers at most
9019 @option{buffer} amount of frames and waits for the cue. After reaching the cue
9020 it forwards the buffered frames and also any subsequent frames coming in its
9023 The filter can be used synchronize the output of multiple ffmpeg processes for
9024 realtime output devices like decklink. By putting the delay in the filtering
9025 chain and pre-buffering frames the process can pass on data to output almost
9026 immediately after the target wallclock timestamp is reached.
9028 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
9034 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
9037 The duration of content to pass on as preroll expressed in seconds. Default is 0.
9040 The maximum duration of content to buffer before waiting for the cue expressed
9041 in seconds. Default is 0.
9048 Apply color adjustments using curves.
9050 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
9051 component (red, green and blue) has its values defined by @var{N} key points
9052 tied from each other using a smooth curve. The x-axis represents the pixel
9053 values from the input frame, and the y-axis the new pixel values to be set for
9056 By default, a component curve is defined by the two points @var{(0;0)} and
9057 @var{(1;1)}. This creates a straight line where each original pixel value is
9058 "adjusted" to its own value, which means no change to the image.
9060 The filter allows you to redefine these two points and add some more. A new
9061 curve (using a natural cubic spline interpolation) will be define to pass
9062 smoothly through all these new coordinates. The new defined points needs to be
9063 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
9064 be in the @var{[0;1]} interval. If the computed curves happened to go outside
9065 the vector spaces, the values will be clipped accordingly.
9067 The filter accepts the following options:
9071 Select one of the available color presets. This option can be used in addition
9072 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
9073 options takes priority on the preset values.
9074 Available presets are:
9077 @item color_negative
9080 @item increase_contrast
9082 @item linear_contrast
9083 @item medium_contrast
9085 @item strong_contrast
9088 Default is @code{none}.
9090 Set the master key points. These points will define a second pass mapping. It
9091 is sometimes called a "luminance" or "value" mapping. It can be used with
9092 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
9093 post-processing LUT.
9095 Set the key points for the red component.
9097 Set the key points for the green component.
9099 Set the key points for the blue component.
9101 Set the key points for all components (not including master).
9102 Can be used in addition to the other key points component
9103 options. In this case, the unset component(s) will fallback on this
9104 @option{all} setting.
9106 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
9108 Save Gnuplot script of the curves in specified file.
9111 To avoid some filtergraph syntax conflicts, each key points list need to be
9112 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
9114 @subsection Examples
9118 Increase slightly the middle level of blue:
9120 curves=blue='0/0 0.5/0.58 1/1'
9126 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'
9128 Here we obtain the following coordinates for each components:
9131 @code{(0;0.11) (0.42;0.51) (1;0.95)}
9133 @code{(0;0) (0.50;0.48) (1;1)}
9135 @code{(0;0.22) (0.49;0.44) (1;0.80)}
9139 The previous example can also be achieved with the associated built-in preset:
9141 curves=preset=vintage
9151 Use a Photoshop preset and redefine the points of the green component:
9153 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
9157 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
9158 and @command{gnuplot}:
9160 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
9161 gnuplot -p /tmp/curves.plt
9167 Video data analysis filter.
9169 This filter shows hexadecimal pixel values of part of video.
9171 The filter accepts the following options:
9175 Set output video size.
9178 Set x offset from where to pick pixels.
9181 Set y offset from where to pick pixels.
9184 Set scope mode, can be one of the following:
9187 Draw hexadecimal pixel values with white color on black background.
9190 Draw hexadecimal pixel values with input video pixel color on black
9194 Draw hexadecimal pixel values on color background picked from input video,
9195 the text color is picked in such way so its always visible.
9199 Draw rows and columns numbers on left and top of video.
9202 Set background opacity.
9205 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
9209 Apply Directional blur filter.
9211 The filter accepts the following options:
9215 Set angle of directional blur. Default is @code{45}.
9218 Set radius of directional blur. Default is @code{5}.
9221 Set which planes to filter. By default all planes are filtered.
9224 @subsection Commands
9225 This filter supports same @ref{commands} as options.
9226 The command accepts the same syntax of the corresponding option.
9228 If the specified expression is not valid, it is kept at its current
9233 Denoise frames using 2D DCT (frequency domain filtering).
9235 This filter is not designed for real time.
9237 The filter accepts the following options:
9241 Set the noise sigma constant.
9243 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
9244 coefficient (absolute value) below this threshold with be dropped.
9246 If you need a more advanced filtering, see @option{expr}.
9248 Default is @code{0}.
9251 Set number overlapping pixels for each block. Since the filter can be slow, you
9252 may want to reduce this value, at the cost of a less effective filter and the
9253 risk of various artefacts.
9255 If the overlapping value doesn't permit processing the whole input width or
9256 height, a warning will be displayed and according borders won't be denoised.
9258 Default value is @var{blocksize}-1, which is the best possible setting.
9261 Set the coefficient factor expression.
9263 For each coefficient of a DCT block, this expression will be evaluated as a
9264 multiplier value for the coefficient.
9266 If this is option is set, the @option{sigma} option will be ignored.
9268 The absolute value of the coefficient can be accessed through the @var{c}
9272 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
9273 @var{blocksize}, which is the width and height of the processed blocks.
9275 The default value is @var{3} (8x8) and can be raised to @var{4} for a
9276 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
9277 on the speed processing. Also, a larger block size does not necessarily means a
9281 @subsection Examples
9283 Apply a denoise with a @option{sigma} of @code{4.5}:
9288 The same operation can be achieved using the expression system:
9290 dctdnoiz=e='gte(c, 4.5*3)'
9293 Violent denoise using a block size of @code{16x16}:
9300 Remove banding artifacts from input video.
9301 It works by replacing banded pixels with average value of referenced pixels.
9303 The filter accepts the following options:
9310 Set banding detection threshold for each plane. Default is 0.02.
9311 Valid range is 0.00003 to 0.5.
9312 If difference between current pixel and reference pixel is less than threshold,
9313 it will be considered as banded.
9316 Banding detection range in pixels. Default is 16. If positive, random number
9317 in range 0 to set value will be used. If negative, exact absolute value
9319 The range defines square of four pixels around current pixel.
9322 Set direction in radians from which four pixel will be compared. If positive,
9323 random direction from 0 to set direction will be picked. If negative, exact of
9324 absolute value will be picked. For example direction 0, -PI or -2*PI radians
9325 will pick only pixels on same row and -PI/2 will pick only pixels on same
9329 If enabled, current pixel is compared with average value of all four
9330 surrounding pixels. The default is enabled. If disabled current pixel is
9331 compared with all four surrounding pixels. The pixel is considered banded
9332 if only all four differences with surrounding pixels are less than threshold.
9335 If enabled, current pixel is changed if and only if all pixel components are banded,
9336 e.g. banding detection threshold is triggered for all color components.
9337 The default is disabled.
9342 Remove blocking artifacts from input video.
9344 The filter accepts the following options:
9348 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
9349 This controls what kind of deblocking is applied.
9352 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
9358 Set blocking detection thresholds. Allowed range is 0 to 1.
9359 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
9360 Using higher threshold gives more deblocking strength.
9361 Setting @var{alpha} controls threshold detection at exact edge of block.
9362 Remaining options controls threshold detection near the edge. Each one for
9363 below/above or left/right. Setting any of those to @var{0} disables
9367 Set planes to filter. Default is to filter all available planes.
9370 @subsection Examples
9374 Deblock using weak filter and block size of 4 pixels.
9376 deblock=filter=weak:block=4
9380 Deblock using strong filter, block size of 4 pixels and custom thresholds for
9381 deblocking more edges.
9383 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
9387 Similar as above, but filter only first plane.
9389 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
9393 Similar as above, but filter only second and third plane.
9395 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
9402 Drop duplicated frames at regular intervals.
9404 The filter accepts the following options:
9408 Set the number of frames from which one will be dropped. Setting this to
9409 @var{N} means one frame in every batch of @var{N} frames will be dropped.
9410 Default is @code{5}.
9413 Set the threshold for duplicate detection. If the difference metric for a frame
9414 is less than or equal to this value, then it is declared as duplicate. Default
9418 Set scene change threshold. Default is @code{15}.
9422 Set the size of the x and y-axis blocks used during metric calculations.
9423 Larger blocks give better noise suppression, but also give worse detection of
9424 small movements. Must be a power of two. Default is @code{32}.
9427 Mark main input as a pre-processed input and activate clean source input
9428 stream. This allows the input to be pre-processed with various filters to help
9429 the metrics calculation while keeping the frame selection lossless. When set to
9430 @code{1}, the first stream is for the pre-processed input, and the second
9431 stream is the clean source from where the kept frames are chosen. Default is
9435 Set whether or not chroma is considered in the metric calculations. Default is
9441 Apply 2D deconvolution of video stream in frequency domain using second stream
9444 The filter accepts the following options:
9448 Set which planes to process.
9451 Set which impulse video frames will be processed, can be @var{first}
9452 or @var{all}. Default is @var{all}.
9455 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
9456 and height are not same and not power of 2 or if stream prior to convolving
9460 The @code{deconvolve} filter also supports the @ref{framesync} options.
9464 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
9466 It accepts the following options:
9470 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
9471 @var{rainbows} for cross-color reduction.
9474 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
9477 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
9480 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
9483 Set temporal chroma threshold. Lower values increases reduction of cross-color.
9488 Apply deflate effect to the video.
9490 This filter replaces the pixel by the local(3x3) average by taking into account
9491 only values lower than the pixel.
9493 It accepts the following options:
9500 Limit the maximum change for each plane, default is 65535.
9501 If 0, plane will remain unchanged.
9504 @subsection Commands
9506 This filter supports the all above options as @ref{commands}.
9510 Remove temporal frame luminance variations.
9512 It accepts the following options:
9516 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
9519 Set averaging mode to smooth temporal luminance variations.
9521 Available values are:
9546 Do not actually modify frame. Useful when one only wants metadata.
9551 Remove judder produced by partially interlaced telecined content.
9553 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
9554 source was partially telecined content then the output of @code{pullup,dejudder}
9555 will have a variable frame rate. May change the recorded frame rate of the
9556 container. Aside from that change, this filter will not affect constant frame
9559 The option available in this filter is:
9563 Specify the length of the window over which the judder repeats.
9565 Accepts any integer greater than 1. Useful values are:
9569 If the original was telecined from 24 to 30 fps (Film to NTSC).
9572 If the original was telecined from 25 to 30 fps (PAL to NTSC).
9575 If a mixture of the two.
9578 The default is @samp{4}.
9583 Suppress a TV station logo by a simple interpolation of the surrounding
9584 pixels. Just set a rectangle covering the logo and watch it disappear
9585 (and sometimes something even uglier appear - your mileage may vary).
9587 It accepts the following parameters:
9592 Specify the top left corner coordinates of the logo. They must be
9597 Specify the width and height of the logo to clear. They must be
9601 Specify the thickness of the fuzzy edge of the rectangle (added to
9602 @var{w} and @var{h}). The default value is 1. This option is
9603 deprecated, setting higher values should no longer be necessary and
9607 When set to 1, a green rectangle is drawn on the screen to simplify
9608 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
9609 The default value is 0.
9611 The rectangle is drawn on the outermost pixels which will be (partly)
9612 replaced with interpolated values. The values of the next pixels
9613 immediately outside this rectangle in each direction will be used to
9614 compute the interpolated pixel values inside the rectangle.
9618 @subsection Examples
9622 Set a rectangle covering the area with top left corner coordinates 0,0
9623 and size 100x77, and a band of size 10:
9625 delogo=x=0:y=0:w=100:h=77:band=10
9633 Remove the rain in the input image/video by applying the derain methods based on
9634 convolutional neural networks. Supported models:
9638 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
9639 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
9642 Training as well as model generation scripts are provided in
9643 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
9645 Native model files (.model) can be generated from TensorFlow model
9646 files (.pb) by using tools/python/convert.py
9648 The filter accepts the following options:
9652 Specify which filter to use. This option accepts the following values:
9656 Derain filter. To conduct derain filter, you need to use a derain model.
9659 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
9661 Default value is @samp{derain}.
9664 Specify which DNN backend to use for model loading and execution. This option accepts
9665 the following values:
9669 Native implementation of DNN loading and execution.
9672 TensorFlow backend. To enable this backend you
9673 need to install the TensorFlow for C library (see
9674 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9675 @code{--enable-libtensorflow}
9677 Default value is @samp{native}.
9680 Set path to model file specifying network architecture and its parameters.
9681 Note that different backends use different file formats. TensorFlow and native
9682 backend can load files for only its format.
9685 It can also be finished with @ref{dnn_processing} filter.
9689 Attempt to fix small changes in horizontal and/or vertical shift. This
9690 filter helps remove camera shake from hand-holding a camera, bumping a
9691 tripod, moving on a vehicle, etc.
9693 The filter accepts the following options:
9701 Specify a rectangular area where to limit the search for motion
9703 If desired the search for motion vectors can be limited to a
9704 rectangular area of the frame defined by its top left corner, width
9705 and height. These parameters have the same meaning as the drawbox
9706 filter which can be used to visualise the position of the bounding
9709 This is useful when simultaneous movement of subjects within the frame
9710 might be confused for camera motion by the motion vector search.
9712 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
9713 then the full frame is used. This allows later options to be set
9714 without specifying the bounding box for the motion vector search.
9716 Default - search the whole frame.
9720 Specify the maximum extent of movement in x and y directions in the
9721 range 0-64 pixels. Default 16.
9724 Specify how to generate pixels to fill blanks at the edge of the
9725 frame. Available values are:
9728 Fill zeroes at blank locations
9730 Original image at blank locations
9732 Extruded edge value at blank locations
9734 Mirrored edge at blank locations
9736 Default value is @samp{mirror}.
9739 Specify the blocksize to use for motion search. Range 4-128 pixels,
9743 Specify the contrast threshold for blocks. Only blocks with more than
9744 the specified contrast (difference between darkest and lightest
9745 pixels) will be considered. Range 1-255, default 125.
9748 Specify the search strategy. Available values are:
9751 Set exhaustive search
9753 Set less exhaustive search.
9755 Default value is @samp{exhaustive}.
9758 If set then a detailed log of the motion search is written to the
9765 Remove unwanted contamination of foreground colors, caused by reflected color of
9766 greenscreen or bluescreen.
9768 This filter accepts the following options:
9772 Set what type of despill to use.
9775 Set how spillmap will be generated.
9778 Set how much to get rid of still remaining spill.
9781 Controls amount of red in spill area.
9784 Controls amount of green in spill area.
9785 Should be -1 for greenscreen.
9788 Controls amount of blue in spill area.
9789 Should be -1 for bluescreen.
9792 Controls brightness of spill area, preserving colors.
9795 Modify alpha from generated spillmap.
9798 @subsection Commands
9800 This filter supports the all above options as @ref{commands}.
9804 Apply an exact inverse of the telecine operation. It requires a predefined
9805 pattern specified using the pattern option which must be the same as that passed
9806 to the telecine filter.
9808 This filter accepts the following options:
9817 The default value is @code{top}.
9821 A string of numbers representing the pulldown pattern you wish to apply.
9822 The default value is @code{23}.
9825 A number representing position of the first frame with respect to the telecine
9826 pattern. This is to be used if the stream is cut. The default value is @code{0}.
9831 Apply dilation effect to the video.
9833 This filter replaces the pixel by the local(3x3) maximum.
9835 It accepts the following options:
9842 Limit the maximum change for each plane, default is 65535.
9843 If 0, plane will remain unchanged.
9846 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9849 Flags to local 3x3 coordinates maps like this:
9856 @subsection Commands
9858 This filter supports the all above options as @ref{commands}.
9862 Displace pixels as indicated by second and third input stream.
9864 It takes three input streams and outputs one stream, the first input is the
9865 source, and second and third input are displacement maps.
9867 The second input specifies how much to displace pixels along the
9868 x-axis, while the third input specifies how much to displace pixels
9870 If one of displacement map streams terminates, last frame from that
9871 displacement map will be used.
9873 Note that once generated, displacements maps can be reused over and over again.
9875 A description of the accepted options follows.
9879 Set displace behavior for pixels that are out of range.
9881 Available values are:
9884 Missing pixels are replaced by black pixels.
9887 Adjacent pixels will spread out to replace missing pixels.
9890 Out of range pixels are wrapped so they point to pixels of other side.
9893 Out of range pixels will be replaced with mirrored pixels.
9895 Default is @samp{smear}.
9899 @subsection Examples
9903 Add ripple effect to rgb input of video size hd720:
9905 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
9909 Add wave effect to rgb input of video size hd720:
9911 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
9915 @anchor{dnn_processing}
9916 @section dnn_processing
9918 Do image processing with deep neural networks. It works together with another filter
9919 which converts the pixel format of the Frame to what the dnn network requires.
9921 The filter accepts the following options:
9925 Specify which DNN backend to use for model loading and execution. This option accepts
9926 the following values:
9930 Native implementation of DNN loading and execution.
9933 TensorFlow backend. To enable this backend you
9934 need to install the TensorFlow for C library (see
9935 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9936 @code{--enable-libtensorflow}
9939 OpenVINO backend. To enable this backend you
9940 need to build and install the OpenVINO for C library (see
9941 @url{https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md}) and configure FFmpeg with
9942 @code{--enable-libopenvino} (--extra-cflags=-I... --extra-ldflags=-L... might
9943 be needed if the header files and libraries are not installed into system path)
9947 Default value is @samp{native}.
9950 Set path to model file specifying network architecture and its parameters.
9951 Note that different backends use different file formats. TensorFlow, OpenVINO and native
9952 backend can load files for only its format.
9954 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9957 Set the input name of the dnn network.
9960 Set the output name of the dnn network.
9963 use DNN async execution if set (default: set),
9964 roll back to sync execution if the backend does not support async.
9968 @subsection Examples
9972 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
9974 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
9978 Halve the pixel value of the frame with format gray32f:
9980 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
9984 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
9986 ./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
9990 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
9992 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
9999 Draw a colored box on the input image.
10001 It accepts the following parameters:
10006 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
10010 The expressions which specify the width and height of the box; if 0 they are interpreted as
10011 the input width and height. It defaults to 0.
10014 Specify the color of the box to write. For the general syntax of this option,
10015 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10016 value @code{invert} is used, the box edge color is the same as the
10017 video with inverted luma.
10020 The expression which sets the thickness of the box edge.
10021 A value of @code{fill} will create a filled box. Default value is @code{3}.
10023 See below for the list of accepted constants.
10026 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
10027 will overwrite the video's color and alpha pixels.
10028 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
10031 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10032 following constants:
10036 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10040 horizontal and vertical chroma subsample values. For example for the
10041 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10045 The input width and height.
10048 The input sample aspect ratio.
10052 The x and y offset coordinates where the box is drawn.
10056 The width and height of the drawn box.
10059 The thickness of the drawn box.
10061 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10062 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10066 @subsection Examples
10070 Draw a black box around the edge of the input image:
10076 Draw a box with color red and an opacity of 50%:
10078 drawbox=10:20:200:60:red@@0.5
10081 The previous example can be specified as:
10083 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
10087 Fill the box with pink color:
10089 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
10093 Draw a 2-pixel red 2.40:1 mask:
10095 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
10099 @subsection Commands
10100 This filter supports same commands as options.
10101 The command accepts the same syntax of the corresponding option.
10103 If the specified expression is not valid, it is kept at its current
10108 Draw a graph using input video metadata.
10110 It accepts the following parameters:
10114 Set 1st frame metadata key from which metadata values will be used to draw a graph.
10117 Set 1st foreground color expression.
10120 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
10123 Set 2nd foreground color expression.
10126 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
10129 Set 3rd foreground color expression.
10132 Set 4th frame metadata key from which metadata values will be used to draw a graph.
10135 Set 4th foreground color expression.
10138 Set minimal value of metadata value.
10141 Set maximal value of metadata value.
10144 Set graph background color. Default is white.
10149 Available values for mode is:
10156 Default is @code{line}.
10161 Available values for slide is:
10164 Draw new frame when right border is reached.
10167 Replace old columns with new ones.
10170 Scroll from right to left.
10173 Scroll from left to right.
10176 Draw single picture.
10179 Default is @code{frame}.
10182 Set size of graph video. For the syntax of this option, check the
10183 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
10184 The default value is @code{900x256}.
10187 Set the output frame rate. Default value is @code{25}.
10189 The foreground color expressions can use the following variables:
10192 Minimal value of metadata value.
10195 Maximal value of metadata value.
10198 Current metadata key value.
10201 The color is defined as 0xAABBGGRR.
10204 Example using metadata from @ref{signalstats} filter:
10206 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
10209 Example using metadata from @ref{ebur128} filter:
10211 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
10216 Draw a grid on the input image.
10218 It accepts the following parameters:
10223 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
10227 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
10228 input width and height, respectively, minus @code{thickness}, so image gets
10229 framed. Default to 0.
10232 Specify the color of the grid. For the general syntax of this option,
10233 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10234 value @code{invert} is used, the grid color is the same as the
10235 video with inverted luma.
10238 The expression which sets the thickness of the grid line. Default value is @code{1}.
10240 See below for the list of accepted constants.
10243 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
10244 will overwrite the video's color and alpha pixels.
10245 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
10248 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10249 following constants:
10253 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10257 horizontal and vertical chroma subsample values. For example for the
10258 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10262 The input grid cell width and height.
10265 The input sample aspect ratio.
10269 The x and y coordinates of some point of grid intersection (meant to configure offset).
10273 The width and height of the drawn cell.
10276 The thickness of the drawn cell.
10278 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10279 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10283 @subsection Examples
10287 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
10289 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
10293 Draw a white 3x3 grid with an opacity of 50%:
10295 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
10299 @subsection Commands
10300 This filter supports same commands as options.
10301 The command accepts the same syntax of the corresponding option.
10303 If the specified expression is not valid, it is kept at its current
10309 Draw a text string or text from a specified file on top of a video, using the
10310 libfreetype library.
10312 To enable compilation of this filter, you need to configure FFmpeg with
10313 @code{--enable-libfreetype}.
10314 To enable default font fallback and the @var{font} option you need to
10315 configure FFmpeg with @code{--enable-libfontconfig}.
10316 To enable the @var{text_shaping} option, you need to configure FFmpeg with
10317 @code{--enable-libfribidi}.
10321 It accepts the following parameters:
10326 Used to draw a box around text using the background color.
10327 The value must be either 1 (enable) or 0 (disable).
10328 The default value of @var{box} is 0.
10331 Set the width of the border to be drawn around the box using @var{boxcolor}.
10332 The default value of @var{boxborderw} is 0.
10335 The color to be used for drawing box around text. For the syntax of this
10336 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10338 The default value of @var{boxcolor} is "white".
10341 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
10342 The default value of @var{line_spacing} is 0.
10345 Set the width of the border to be drawn around the text using @var{bordercolor}.
10346 The default value of @var{borderw} is 0.
10349 Set the color to be used for drawing border around text. For the syntax of this
10350 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10352 The default value of @var{bordercolor} is "black".
10355 Select how the @var{text} is expanded. Can be either @code{none},
10356 @code{strftime} (deprecated) or
10357 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
10361 Set a start time for the count. Value is in microseconds. Only applied
10362 in the deprecated strftime expansion mode. To emulate in normal expansion
10363 mode use the @code{pts} function, supplying the start time (in seconds)
10364 as the second argument.
10367 If true, check and fix text coords to avoid clipping.
10370 The color to be used for drawing fonts. For the syntax of this option, check
10371 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10373 The default value of @var{fontcolor} is "black".
10375 @item fontcolor_expr
10376 String which is expanded the same way as @var{text} to obtain dynamic
10377 @var{fontcolor} value. By default this option has empty value and is not
10378 processed. When this option is set, it overrides @var{fontcolor} option.
10381 The font family to be used for drawing text. By default Sans.
10384 The font file to be used for drawing text. The path must be included.
10385 This parameter is mandatory if the fontconfig support is disabled.
10388 Draw the text applying alpha blending. The value can
10389 be a number between 0.0 and 1.0.
10390 The expression accepts the same variables @var{x, y} as well.
10391 The default value is 1.
10392 Please see @var{fontcolor_expr}.
10395 The font size to be used for drawing text.
10396 The default value of @var{fontsize} is 16.
10399 If set to 1, attempt to shape the text (for example, reverse the order of
10400 right-to-left text and join Arabic characters) before drawing it.
10401 Otherwise, just draw the text exactly as given.
10402 By default 1 (if supported).
10404 @item ft_load_flags
10405 The flags to be used for loading the fonts.
10407 The flags map the corresponding flags supported by libfreetype, and are
10408 a combination of the following values:
10415 @item vertical_layout
10416 @item force_autohint
10419 @item ignore_global_advance_width
10421 @item ignore_transform
10423 @item linear_design
10427 Default value is "default".
10429 For more information consult the documentation for the FT_LOAD_*
10433 The color to be used for drawing a shadow behind the drawn text. For the
10434 syntax of this option, check the @ref{color syntax,,"Color" section in the
10435 ffmpeg-utils manual,ffmpeg-utils}.
10437 The default value of @var{shadowcolor} is "black".
10441 The x and y offsets for the text shadow position with respect to the
10442 position of the text. They can be either positive or negative
10443 values. The default value for both is "0".
10446 The starting frame number for the n/frame_num variable. The default value
10450 The size in number of spaces to use for rendering the tab.
10451 Default value is 4.
10454 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
10455 format. It can be used with or without text parameter. @var{timecode_rate}
10456 option must be specified.
10458 @item timecode_rate, rate, r
10459 Set the timecode frame rate (timecode only). Value will be rounded to nearest
10460 integer. Minimum value is "1".
10461 Drop-frame timecode is supported for frame rates 30 & 60.
10464 If set to 1, the output of the timecode option will wrap around at 24 hours.
10465 Default is 0 (disabled).
10468 The text string to be drawn. The text must be a sequence of UTF-8
10469 encoded characters.
10470 This parameter is mandatory if no file is specified with the parameter
10474 A text file containing text to be drawn. The text must be a sequence
10475 of UTF-8 encoded characters.
10477 This parameter is mandatory if no text string is specified with the
10478 parameter @var{text}.
10480 If both @var{text} and @var{textfile} are specified, an error is thrown.
10483 If set to 1, the @var{textfile} will be reloaded before each frame.
10484 Be sure to update it atomically, or it may be read partially, or even fail.
10488 The expressions which specify the offsets where text will be drawn
10489 within the video frame. They are relative to the top/left border of the
10492 The default value of @var{x} and @var{y} is "0".
10494 See below for the list of accepted constants and functions.
10497 The parameters for @var{x} and @var{y} are expressions containing the
10498 following constants and functions:
10502 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
10506 horizontal and vertical chroma subsample values. For example for the
10507 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10510 the height of each text line
10518 @item max_glyph_a, ascent
10519 the maximum distance from the baseline to the highest/upper grid
10520 coordinate used to place a glyph outline point, for all the rendered
10522 It is a positive value, due to the grid's orientation with the Y axis
10525 @item max_glyph_d, descent
10526 the maximum distance from the baseline to the lowest grid coordinate
10527 used to place a glyph outline point, for all the rendered glyphs.
10528 This is a negative value, due to the grid's orientation, with the Y axis
10532 maximum glyph height, that is the maximum height for all the glyphs
10533 contained in the rendered text, it is equivalent to @var{ascent} -
10537 maximum glyph width, that is the maximum width for all the glyphs
10538 contained in the rendered text
10541 the number of input frame, starting from 0
10543 @item rand(min, max)
10544 return a random number included between @var{min} and @var{max}
10547 The input sample aspect ratio.
10550 timestamp expressed in seconds, NAN if the input timestamp is unknown
10553 the height of the rendered text
10556 the width of the rendered text
10560 the x and y offset coordinates where the text is drawn.
10562 These parameters allow the @var{x} and @var{y} expressions to refer
10563 to each other, so you can for example specify @code{y=x/dar}.
10566 A one character description of the current frame's picture type.
10569 The current packet's position in the input file or stream
10570 (in bytes, from the start of the input). A value of -1 indicates
10571 this info is not available.
10574 The current packet's duration, in seconds.
10577 The current packet's size (in bytes).
10580 @anchor{drawtext_expansion}
10581 @subsection Text expansion
10583 If @option{expansion} is set to @code{strftime},
10584 the filter recognizes strftime() sequences in the provided text and
10585 expands them accordingly. Check the documentation of strftime(). This
10586 feature is deprecated.
10588 If @option{expansion} is set to @code{none}, the text is printed verbatim.
10590 If @option{expansion} is set to @code{normal} (which is the default),
10591 the following expansion mechanism is used.
10593 The backslash character @samp{\}, followed by any character, always expands to
10594 the second character.
10596 Sequences of the form @code{%@{...@}} are expanded. The text between the
10597 braces is a function name, possibly followed by arguments separated by ':'.
10598 If the arguments contain special characters or delimiters (':' or '@}'),
10599 they should be escaped.
10601 Note that they probably must also be escaped as the value for the
10602 @option{text} option in the filter argument string and as the filter
10603 argument in the filtergraph description, and possibly also for the shell,
10604 that makes up to four levels of escaping; using a text file avoids these
10607 The following functions are available:
10612 The expression evaluation result.
10614 It must take one argument specifying the expression to be evaluated,
10615 which accepts the same constants and functions as the @var{x} and
10616 @var{y} values. Note that not all constants should be used, for
10617 example the text size is not known when evaluating the expression, so
10618 the constants @var{text_w} and @var{text_h} will have an undefined
10621 @item expr_int_format, eif
10622 Evaluate the expression's value and output as formatted integer.
10624 The first argument is the expression to be evaluated, just as for the @var{expr} function.
10625 The second argument specifies the output format. Allowed values are @samp{x},
10626 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
10627 @code{printf} function.
10628 The third parameter is optional and sets the number of positions taken by the output.
10629 It can be used to add padding with zeros from the left.
10632 The time at which the filter is running, expressed in UTC.
10633 It can accept an argument: a strftime() format string.
10636 The time at which the filter is running, expressed in the local time zone.
10637 It can accept an argument: a strftime() format string.
10640 Frame metadata. Takes one or two arguments.
10642 The first argument is mandatory and specifies the metadata key.
10644 The second argument is optional and specifies a default value, used when the
10645 metadata key is not found or empty.
10647 Available metadata can be identified by inspecting entries
10648 starting with TAG included within each frame section
10649 printed by running @code{ffprobe -show_frames}.
10651 String metadata generated in filters leading to
10652 the drawtext filter are also available.
10655 The frame number, starting from 0.
10658 A one character description of the current picture type.
10661 The timestamp of the current frame.
10662 It can take up to three arguments.
10664 The first argument is the format of the timestamp; it defaults to @code{flt}
10665 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
10666 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
10667 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
10668 @code{localtime} stands for the timestamp of the frame formatted as
10669 local time zone time.
10671 The second argument is an offset added to the timestamp.
10673 If the format is set to @code{hms}, a third argument @code{24HH} may be
10674 supplied to present the hour part of the formatted timestamp in 24h format
10677 If the format is set to @code{localtime} or @code{gmtime},
10678 a third argument may be supplied: a strftime() format string.
10679 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
10682 @subsection Commands
10684 This filter supports altering parameters via commands:
10687 Alter existing filter parameters.
10689 Syntax for the argument is the same as for filter invocation, e.g.
10692 fontsize=56:fontcolor=green:text='Hello World'
10695 Full filter invocation with sendcmd would look like this:
10698 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
10702 If the entire argument can't be parsed or applied as valid values then the filter will
10703 continue with its existing parameters.
10705 @subsection Examples
10709 Draw "Test Text" with font FreeSerif, using the default values for the
10710 optional parameters.
10713 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
10717 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
10718 and y=50 (counting from the top-left corner of the screen), text is
10719 yellow with a red box around it. Both the text and the box have an
10723 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
10724 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
10727 Note that the double quotes are not necessary if spaces are not used
10728 within the parameter list.
10731 Show the text at the center of the video frame:
10733 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
10737 Show the text at a random position, switching to a new position every 30 seconds:
10739 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)"
10743 Show a text line sliding from right to left in the last row of the video
10744 frame. The file @file{LONG_LINE} is assumed to contain a single line
10747 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
10751 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
10753 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
10757 Draw a single green letter "g", at the center of the input video.
10758 The glyph baseline is placed at half screen height.
10760 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
10764 Show text for 1 second every 3 seconds:
10766 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
10770 Use fontconfig to set the font. Note that the colons need to be escaped.
10772 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
10776 Draw "Test Text" with font size dependent on height of the video.
10778 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
10782 Print the date of a real-time encoding (see strftime(3)):
10784 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
10788 Show text fading in and out (appearing/disappearing):
10791 DS=1.0 # display start
10792 DE=10.0 # display end
10793 FID=1.5 # fade in duration
10794 FOD=5 # fade out duration
10795 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 @}"
10799 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
10800 and the @option{fontsize} value are included in the @option{y} offset.
10802 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
10803 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
10807 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
10808 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
10809 must have option @option{-export_path_metadata 1} for the special metadata fields
10810 to be available for filters.
10812 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
10817 For more information about libfreetype, check:
10818 @url{http://www.freetype.org/}.
10820 For more information about fontconfig, check:
10821 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
10823 For more information about libfribidi, check:
10824 @url{http://fribidi.org/}.
10826 @section edgedetect
10828 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
10830 The filter accepts the following options:
10835 Set low and high threshold values used by the Canny thresholding
10838 The high threshold selects the "strong" edge pixels, which are then
10839 connected through 8-connectivity with the "weak" edge pixels selected
10840 by the low threshold.
10842 @var{low} and @var{high} threshold values must be chosen in the range
10843 [0,1], and @var{low} should be lesser or equal to @var{high}.
10845 Default value for @var{low} is @code{20/255}, and default value for @var{high}
10849 Define the drawing mode.
10853 Draw white/gray wires on black background.
10856 Mix the colors to create a paint/cartoon effect.
10859 Apply Canny edge detector on all selected planes.
10861 Default value is @var{wires}.
10864 Select planes for filtering. By default all available planes are filtered.
10867 @subsection Examples
10871 Standard edge detection with custom values for the hysteresis thresholding:
10873 edgedetect=low=0.1:high=0.4
10877 Painting effect without thresholding:
10879 edgedetect=mode=colormix:high=0
10885 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10887 For each input image, the filter will compute the optimal mapping from
10888 the input to the output given the codebook length, that is the number
10889 of distinct output colors.
10891 This filter accepts the following options.
10894 @item codebook_length, l
10895 Set codebook length. The value must be a positive integer, and
10896 represents the number of distinct output colors. Default value is 256.
10899 Set the maximum number of iterations to apply for computing the optimal
10900 mapping. The higher the value the better the result and the higher the
10901 computation time. Default value is 1.
10904 Set a random seed, must be an integer included between 0 and
10905 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10906 will try to use a good random seed on a best effort basis.
10909 Set pal8 output pixel format. This option does not work with codebook
10910 length greater than 256.
10915 Measure graylevel entropy in histogram of color channels of video frames.
10917 It accepts the following parameters:
10921 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10923 @var{diff} mode measures entropy of histogram delta values, absolute differences
10924 between neighbour histogram values.
10928 Set brightness, contrast, saturation and approximate gamma adjustment.
10930 The filter accepts the following options:
10934 Set the contrast expression. The value must be a float value in range
10935 @code{-1000.0} to @code{1000.0}. The default value is "1".
10938 Set the brightness expression. The value must be a float value in
10939 range @code{-1.0} to @code{1.0}. The default value is "0".
10942 Set the saturation expression. The value must be a float in
10943 range @code{0.0} to @code{3.0}. The default value is "1".
10946 Set the gamma expression. The value must be a float in range
10947 @code{0.1} to @code{10.0}. The default value is "1".
10950 Set the gamma expression for red. The value must be a float in
10951 range @code{0.1} to @code{10.0}. The default value is "1".
10954 Set the gamma expression for green. The value must be a float in range
10955 @code{0.1} to @code{10.0}. The default value is "1".
10958 Set the gamma expression for blue. The value must be a float in range
10959 @code{0.1} to @code{10.0}. The default value is "1".
10962 Set the gamma weight expression. It can be used to reduce the effect
10963 of a high gamma value on bright image areas, e.g. keep them from
10964 getting overamplified and just plain white. The value must be a float
10965 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
10966 gamma correction all the way down while @code{1.0} leaves it at its
10967 full strength. Default is "1".
10970 Set when the expressions for brightness, contrast, saturation and
10971 gamma expressions are evaluated.
10973 It accepts the following values:
10976 only evaluate expressions once during the filter initialization or
10977 when a command is processed
10980 evaluate expressions for each incoming frame
10983 Default value is @samp{init}.
10986 The expressions accept the following parameters:
10989 frame count of the input frame starting from 0
10992 byte position of the corresponding packet in the input file, NAN if
10996 frame rate of the input video, NAN if the input frame rate is unknown
10999 timestamp expressed in seconds, NAN if the input timestamp is unknown
11002 @subsection Commands
11003 The filter supports the following commands:
11007 Set the contrast expression.
11010 Set the brightness expression.
11013 Set the saturation expression.
11016 Set the gamma expression.
11019 Set the gamma_r expression.
11022 Set gamma_g expression.
11025 Set gamma_b expression.
11028 Set gamma_weight expression.
11030 The command accepts the same syntax of the corresponding option.
11032 If the specified expression is not valid, it is kept at its current
11039 Apply erosion effect to the video.
11041 This filter replaces the pixel by the local(3x3) minimum.
11043 It accepts the following options:
11050 Limit the maximum change for each plane, default is 65535.
11051 If 0, plane will remain unchanged.
11054 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
11057 Flags to local 3x3 coordinates maps like this:
11064 @subsection Commands
11066 This filter supports the all above options as @ref{commands}.
11068 @section extractplanes
11070 Extract color channel components from input video stream into
11071 separate grayscale video streams.
11073 The filter accepts the following option:
11077 Set plane(s) to extract.
11079 Available values for planes are:
11090 Choosing planes not available in the input will result in an error.
11091 That means you cannot select @code{r}, @code{g}, @code{b} planes
11092 with @code{y}, @code{u}, @code{v} planes at same time.
11095 @subsection Examples
11099 Extract luma, u and v color channel component from input video frame
11100 into 3 grayscale outputs:
11102 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
11108 Apply a fade-in/out effect to the input video.
11110 It accepts the following parameters:
11114 The effect type can be either "in" for a fade-in, or "out" for a fade-out
11116 Default is @code{in}.
11118 @item start_frame, s
11119 Specify the number of the frame to start applying the fade
11120 effect at. Default is 0.
11123 The number of frames that the fade effect lasts. At the end of the
11124 fade-in effect, the output video will have the same intensity as the input video.
11125 At the end of the fade-out transition, the output video will be filled with the
11126 selected @option{color}.
11130 If set to 1, fade only alpha channel, if one exists on the input.
11131 Default value is 0.
11133 @item start_time, st
11134 Specify the timestamp (in seconds) of the frame to start to apply the fade
11135 effect. If both start_frame and start_time are specified, the fade will start at
11136 whichever comes last. Default is 0.
11139 The number of seconds for which the fade effect has to last. At the end of the
11140 fade-in effect the output video will have the same intensity as the input video,
11141 at the end of the fade-out transition the output video will be filled with the
11142 selected @option{color}.
11143 If both duration and nb_frames are specified, duration is used. Default is 0
11144 (nb_frames is used by default).
11147 Specify the color of the fade. Default is "black".
11150 @subsection Examples
11154 Fade in the first 30 frames of video:
11159 The command above is equivalent to:
11165 Fade out the last 45 frames of a 200-frame video:
11168 fade=type=out:start_frame=155:nb_frames=45
11172 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
11174 fade=in:0:25, fade=out:975:25
11178 Make the first 5 frames yellow, then fade in from frame 5-24:
11180 fade=in:5:20:color=yellow
11184 Fade in alpha over first 25 frames of video:
11186 fade=in:0:25:alpha=1
11190 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
11192 fade=t=in:st=5.5:d=0.5
11198 Denoise frames using 3D FFT (frequency domain filtering).
11200 The filter accepts the following options:
11204 Set the noise sigma constant. This sets denoising strength.
11205 Default value is 1. Allowed range is from 0 to 30.
11206 Using very high sigma with low overlap may give blocking artifacts.
11209 Set amount of denoising. By default all detected noise is reduced.
11210 Default value is 1. Allowed range is from 0 to 1.
11213 Set size of block, Default is 4, can be 3, 4, 5 or 6.
11214 Actual size of block in pixels is 2 to power of @var{block}, so by default
11215 block size in pixels is 2^4 which is 16.
11218 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
11221 Set number of previous frames to use for denoising. By default is set to 0.
11224 Set number of next frames to to use for denoising. By default is set to 0.
11227 Set planes which will be filtered, by default are all available filtered
11232 Apply arbitrary expressions to samples in frequency domain
11236 Adjust the dc value (gain) of the luma plane of the image. The filter
11237 accepts an integer value in range @code{0} to @code{1000}. The default
11238 value is set to @code{0}.
11241 Adjust the dc value (gain) of the 1st chroma plane of the image. The
11242 filter accepts an integer value in range @code{0} to @code{1000}. The
11243 default value is set to @code{0}.
11246 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
11247 filter accepts an integer value in range @code{0} to @code{1000}. The
11248 default value is set to @code{0}.
11251 Set the frequency domain weight expression for the luma plane.
11254 Set the frequency domain weight expression for the 1st chroma plane.
11257 Set the frequency domain weight expression for the 2nd chroma plane.
11260 Set when the expressions are evaluated.
11262 It accepts the following values:
11265 Only evaluate expressions once during the filter initialization.
11268 Evaluate expressions for each incoming frame.
11271 Default value is @samp{init}.
11273 The filter accepts the following variables:
11276 The coordinates of the current sample.
11280 The width and height of the image.
11283 The number of input frame, starting from 0.
11286 @subsection Examples
11292 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
11298 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
11304 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
11310 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
11317 Extract a single field from an interlaced image using stride
11318 arithmetic to avoid wasting CPU time. The output frames are marked as
11321 The filter accepts the following options:
11325 Specify whether to extract the top (if the value is @code{0} or
11326 @code{top}) or the bottom field (if the value is @code{1} or
11332 Create new frames by copying the top and bottom fields from surrounding frames
11333 supplied as numbers by the hint file.
11337 Set file containing hints: absolute/relative frame numbers.
11339 There must be one line for each frame in a clip. Each line must contain two
11340 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
11341 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
11342 is current frame number for @code{absolute} mode or out of [-1, 1] range
11343 for @code{relative} mode. First number tells from which frame to pick up top
11344 field and second number tells from which frame to pick up bottom field.
11346 If optionally followed by @code{+} output frame will be marked as interlaced,
11347 else if followed by @code{-} output frame will be marked as progressive, else
11348 it will be marked same as input frame.
11349 If optionally followed by @code{t} output frame will use only top field, or in
11350 case of @code{b} it will use only bottom field.
11351 If line starts with @code{#} or @code{;} that line is skipped.
11354 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
11357 Example of first several lines of @code{hint} file for @code{relative} mode:
11359 0,0 - # first frame
11360 1,0 - # second frame, use third's frame top field and second's frame bottom field
11361 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
11376 @section fieldmatch
11378 Field matching filter for inverse telecine. It is meant to reconstruct the
11379 progressive frames from a telecined stream. The filter does not drop duplicated
11380 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
11381 followed by a decimation filter such as @ref{decimate} in the filtergraph.
11383 The separation of the field matching and the decimation is notably motivated by
11384 the possibility of inserting a de-interlacing filter fallback between the two.
11385 If the source has mixed telecined and real interlaced content,
11386 @code{fieldmatch} will not be able to match fields for the interlaced parts.
11387 But these remaining combed frames will be marked as interlaced, and thus can be
11388 de-interlaced by a later filter such as @ref{yadif} before decimation.
11390 In addition to the various configuration options, @code{fieldmatch} can take an
11391 optional second stream, activated through the @option{ppsrc} option. If
11392 enabled, the frames reconstruction will be based on the fields and frames from
11393 this second stream. This allows the first input to be pre-processed in order to
11394 help the various algorithms of the filter, while keeping the output lossless
11395 (assuming the fields are matched properly). Typically, a field-aware denoiser,
11396 or brightness/contrast adjustments can help.
11398 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
11399 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
11400 which @code{fieldmatch} is based on. While the semantic and usage are very
11401 close, some behaviour and options names can differ.
11403 The @ref{decimate} filter currently only works for constant frame rate input.
11404 If your input has mixed telecined (30fps) and progressive content with a lower
11405 framerate like 24fps use the following filterchain to produce the necessary cfr
11406 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
11408 The filter accepts the following options:
11412 Specify the assumed field order of the input stream. Available values are:
11416 Auto detect parity (use FFmpeg's internal parity value).
11418 Assume bottom field first.
11420 Assume top field first.
11423 Note that it is sometimes recommended not to trust the parity announced by the
11426 Default value is @var{auto}.
11429 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
11430 sense that it won't risk creating jerkiness due to duplicate frames when
11431 possible, but if there are bad edits or blended fields it will end up
11432 outputting combed frames when a good match might actually exist. On the other
11433 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
11434 but will almost always find a good frame if there is one. The other values are
11435 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
11436 jerkiness and creating duplicate frames versus finding good matches in sections
11437 with bad edits, orphaned fields, blended fields, etc.
11439 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
11441 Available values are:
11445 2-way matching (p/c)
11447 2-way matching, and trying 3rd match if still combed (p/c + n)
11449 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
11451 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
11452 still combed (p/c + n + u/b)
11454 3-way matching (p/c/n)
11456 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
11457 detected as combed (p/c/n + u/b)
11460 The parenthesis at the end indicate the matches that would be used for that
11461 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
11464 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
11467 Default value is @var{pc_n}.
11470 Mark the main input stream as a pre-processed input, and enable the secondary
11471 input stream as the clean source to pick the fields from. See the filter
11472 introduction for more details. It is similar to the @option{clip2} feature from
11475 Default value is @code{0} (disabled).
11478 Set the field to match from. It is recommended to set this to the same value as
11479 @option{order} unless you experience matching failures with that setting. In
11480 certain circumstances changing the field that is used to match from can have a
11481 large impact on matching performance. Available values are:
11485 Automatic (same value as @option{order}).
11487 Match from the bottom field.
11489 Match from the top field.
11492 Default value is @var{auto}.
11495 Set whether or not chroma is included during the match comparisons. In most
11496 cases it is recommended to leave this enabled. You should set this to @code{0}
11497 only if your clip has bad chroma problems such as heavy rainbowing or other
11498 artifacts. Setting this to @code{0} could also be used to speed things up at
11499 the cost of some accuracy.
11501 Default value is @code{1}.
11505 These define an exclusion band which excludes the lines between @option{y0} and
11506 @option{y1} from being included in the field matching decision. An exclusion
11507 band can be used to ignore subtitles, a logo, or other things that may
11508 interfere with the matching. @option{y0} sets the starting scan line and
11509 @option{y1} sets the ending line; all lines in between @option{y0} and
11510 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
11511 @option{y0} and @option{y1} to the same value will disable the feature.
11512 @option{y0} and @option{y1} defaults to @code{0}.
11515 Set the scene change detection threshold as a percentage of maximum change on
11516 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
11517 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
11518 @option{scthresh} is @code{[0.0, 100.0]}.
11520 Default value is @code{12.0}.
11523 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
11524 account the combed scores of matches when deciding what match to use as the
11525 final match. Available values are:
11529 No final matching based on combed scores.
11531 Combed scores are only used when a scene change is detected.
11533 Use combed scores all the time.
11536 Default is @var{sc}.
11539 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
11540 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
11541 Available values are:
11545 No forced calculation.
11547 Force p/c/n calculations.
11549 Force p/c/n/u/b calculations.
11552 Default value is @var{none}.
11555 This is the area combing threshold used for combed frame detection. This
11556 essentially controls how "strong" or "visible" combing must be to be detected.
11557 Larger values mean combing must be more visible and smaller values mean combing
11558 can be less visible or strong and still be detected. Valid settings are from
11559 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
11560 be detected as combed). This is basically a pixel difference value. A good
11561 range is @code{[8, 12]}.
11563 Default value is @code{9}.
11566 Sets whether or not chroma is considered in the combed frame decision. Only
11567 disable this if your source has chroma problems (rainbowing, etc.) that are
11568 causing problems for the combed frame detection with chroma enabled. Actually,
11569 using @option{chroma}=@var{0} is usually more reliable, except for the case
11570 where there is chroma only combing in the source.
11572 Default value is @code{0}.
11576 Respectively set the x-axis and y-axis size of the window used during combed
11577 frame detection. This has to do with the size of the area in which
11578 @option{combpel} pixels are required to be detected as combed for a frame to be
11579 declared combed. See the @option{combpel} parameter description for more info.
11580 Possible values are any number that is a power of 2 starting at 4 and going up
11583 Default value is @code{16}.
11586 The number of combed pixels inside any of the @option{blocky} by
11587 @option{blockx} size blocks on the frame for the frame to be detected as
11588 combed. While @option{cthresh} controls how "visible" the combing must be, this
11589 setting controls "how much" combing there must be in any localized area (a
11590 window defined by the @option{blockx} and @option{blocky} settings) on the
11591 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
11592 which point no frames will ever be detected as combed). This setting is known
11593 as @option{MI} in TFM/VFM vocabulary.
11595 Default value is @code{80}.
11598 @anchor{p/c/n/u/b meaning}
11599 @subsection p/c/n/u/b meaning
11601 @subsubsection p/c/n
11603 We assume the following telecined stream:
11606 Top fields: 1 2 2 3 4
11607 Bottom fields: 1 2 3 4 4
11610 The numbers correspond to the progressive frame the fields relate to. Here, the
11611 first two frames are progressive, the 3rd and 4th are combed, and so on.
11613 When @code{fieldmatch} is configured to run a matching from bottom
11614 (@option{field}=@var{bottom}) this is how this input stream get transformed:
11619 B 1 2 3 4 4 <-- matching reference
11628 As a result of the field matching, we can see that some frames get duplicated.
11629 To perform a complete inverse telecine, you need to rely on a decimation filter
11630 after this operation. See for instance the @ref{decimate} filter.
11632 The same operation now matching from top fields (@option{field}=@var{top})
11637 T 1 2 2 3 4 <-- matching reference
11647 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
11648 basically, they refer to the frame and field of the opposite parity:
11651 @item @var{p} matches the field of the opposite parity in the previous frame
11652 @item @var{c} matches the field of the opposite parity in the current frame
11653 @item @var{n} matches the field of the opposite parity in the next frame
11658 The @var{u} and @var{b} matching are a bit special in the sense that they match
11659 from the opposite parity flag. In the following examples, we assume that we are
11660 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
11661 'x' is placed above and below each matched fields.
11663 With bottom matching (@option{field}=@var{bottom}):
11668 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11669 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11677 With top matching (@option{field}=@var{top}):
11682 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11683 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11691 @subsection Examples
11693 Simple IVTC of a top field first telecined stream:
11695 fieldmatch=order=tff:combmatch=none, decimate
11698 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
11700 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
11703 @section fieldorder
11705 Transform the field order of the input video.
11707 It accepts the following parameters:
11712 The output field order. Valid values are @var{tff} for top field first or @var{bff}
11713 for bottom field first.
11716 The default value is @samp{tff}.
11718 The transformation is done by shifting the picture content up or down
11719 by one line, and filling the remaining line with appropriate picture content.
11720 This method is consistent with most broadcast field order converters.
11722 If the input video is not flagged as being interlaced, or it is already
11723 flagged as being of the required output field order, then this filter does
11724 not alter the incoming video.
11726 It is very useful when converting to or from PAL DV material,
11727 which is bottom field first.
11731 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
11734 @section fifo, afifo
11736 Buffer input images and send them when they are requested.
11738 It is mainly useful when auto-inserted by the libavfilter
11741 It does not take parameters.
11743 @section fillborders
11745 Fill borders of the input video, without changing video stream dimensions.
11746 Sometimes video can have garbage at the four edges and you may not want to
11747 crop video input to keep size multiple of some number.
11749 This filter accepts the following options:
11753 Number of pixels to fill from left border.
11756 Number of pixels to fill from right border.
11759 Number of pixels to fill from top border.
11762 Number of pixels to fill from bottom border.
11767 It accepts the following values:
11770 fill pixels using outermost pixels
11773 fill pixels using mirroring (half sample symmetric)
11776 fill pixels with constant value
11779 fill pixels using reflecting (whole sample symmetric)
11782 fill pixels using wrapping
11785 fade pixels to constant value
11788 Default is @var{smear}.
11791 Set color for pixels in fixed or fade mode. Default is @var{black}.
11794 @subsection Commands
11795 This filter supports same @ref{commands} as options.
11796 The command accepts the same syntax of the corresponding option.
11798 If the specified expression is not valid, it is kept at its current
11803 Find a rectangular object
11805 It accepts the following options:
11809 Filepath of the object image, needs to be in gray8.
11812 Detection threshold, default is 0.5.
11815 Number of mipmaps, default is 3.
11817 @item xmin, ymin, xmax, ymax
11818 Specifies the rectangle in which to search.
11821 @subsection Examples
11825 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
11827 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
11833 Flood area with values of same pixel components with another values.
11835 It accepts the following options:
11838 Set pixel x coordinate.
11841 Set pixel y coordinate.
11844 Set source #0 component value.
11847 Set source #1 component value.
11850 Set source #2 component value.
11853 Set source #3 component value.
11856 Set destination #0 component value.
11859 Set destination #1 component value.
11862 Set destination #2 component value.
11865 Set destination #3 component value.
11871 Convert the input video to one of the specified pixel formats.
11872 Libavfilter will try to pick one that is suitable as input to
11875 It accepts the following parameters:
11879 A '|'-separated list of pixel format names, such as
11880 "pix_fmts=yuv420p|monow|rgb24".
11884 @subsection Examples
11888 Convert the input video to the @var{yuv420p} format
11890 format=pix_fmts=yuv420p
11893 Convert the input video to any of the formats in the list
11895 format=pix_fmts=yuv420p|yuv444p|yuv410p
11902 Convert the video to specified constant frame rate by duplicating or dropping
11903 frames as necessary.
11905 It accepts the following parameters:
11909 The desired output frame rate. The default is @code{25}.
11912 Assume the first PTS should be the given value, in seconds. This allows for
11913 padding/trimming at the start of stream. By default, no assumption is made
11914 about the first frame's expected PTS, so no padding or trimming is done.
11915 For example, this could be set to 0 to pad the beginning with duplicates of
11916 the first frame if a video stream starts after the audio stream or to trim any
11917 frames with a negative PTS.
11920 Timestamp (PTS) rounding method.
11922 Possible values are:
11929 round towards -infinity
11931 round towards +infinity
11935 The default is @code{near}.
11938 Action performed when reading the last frame.
11940 Possible values are:
11943 Use same timestamp rounding method as used for other frames.
11945 Pass through last frame if input duration has not been reached yet.
11947 The default is @code{round}.
11951 Alternatively, the options can be specified as a flat string:
11952 @var{fps}[:@var{start_time}[:@var{round}]].
11954 See also the @ref{setpts} filter.
11956 @subsection Examples
11960 A typical usage in order to set the fps to 25:
11966 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
11968 fps=fps=film:round=near
11974 Pack two different video streams into a stereoscopic video, setting proper
11975 metadata on supported codecs. The two views should have the same size and
11976 framerate and processing will stop when the shorter video ends. Please note
11977 that you may conveniently adjust view properties with the @ref{scale} and
11980 It accepts the following parameters:
11984 The desired packing format. Supported values are:
11989 The views are next to each other (default).
11992 The views are on top of each other.
11995 The views are packed by line.
11998 The views are packed by column.
12001 The views are temporally interleaved.
12010 # Convert left and right views into a frame-sequential video
12011 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
12013 # Convert views into a side-by-side video with the same output resolution as the input
12014 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
12019 Change the frame rate by interpolating new video output frames from the source
12022 This filter is not designed to function correctly with interlaced media. If
12023 you wish to change the frame rate of interlaced media then you are required
12024 to deinterlace before this filter and re-interlace after this filter.
12026 A description of the accepted options follows.
12030 Specify the output frames per second. This option can also be specified
12031 as a value alone. The default is @code{50}.
12034 Specify the start of a range where the output frame will be created as a
12035 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12036 the default is @code{15}.
12039 Specify the end of a range where the output frame will be created as a
12040 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12041 the default is @code{240}.
12044 Specify the level at which a scene change is detected as a value between
12045 0 and 100 to indicate a new scene; a low value reflects a low
12046 probability for the current frame to introduce a new scene, while a higher
12047 value means the current frame is more likely to be one.
12048 The default is @code{8.2}.
12051 Specify flags influencing the filter process.
12053 Available value for @var{flags} is:
12056 @item scene_change_detect, scd
12057 Enable scene change detection using the value of the option @var{scene}.
12058 This flag is enabled by default.
12064 Select one frame every N-th frame.
12066 This filter accepts the following option:
12069 Select frame after every @code{step} frames.
12070 Allowed values are positive integers higher than 0. Default value is @code{1}.
12073 @section freezedetect
12075 Detect frozen video.
12077 This filter logs a message and sets frame metadata when it detects that the
12078 input video has no significant change in content during a specified duration.
12079 Video freeze detection calculates the mean average absolute difference of all
12080 the components of video frames and compares it to a noise floor.
12082 The printed times and duration are expressed in seconds. The
12083 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
12084 whose timestamp equals or exceeds the detection duration and it contains the
12085 timestamp of the first frame of the freeze. The
12086 @code{lavfi.freezedetect.freeze_duration} and
12087 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
12090 The filter accepts the following options:
12094 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
12095 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
12099 Set freeze duration until notification (default is 2 seconds).
12102 @section freezeframes
12104 Freeze video frames.
12106 This filter freezes video frames using frame from 2nd input.
12108 The filter accepts the following options:
12112 Set number of first frame from which to start freeze.
12115 Set number of last frame from which to end freeze.
12118 Set number of frame from 2nd input which will be used instead of replaced frames.
12124 Apply a frei0r effect to the input video.
12126 To enable the compilation of this filter, you need to install the frei0r
12127 header and configure FFmpeg with @code{--enable-frei0r}.
12129 It accepts the following parameters:
12134 The name of the frei0r effect to load. If the environment variable
12135 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
12136 directories specified by the colon-separated list in @env{FREI0R_PATH}.
12137 Otherwise, the standard frei0r paths are searched, in this order:
12138 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
12139 @file{/usr/lib/frei0r-1/}.
12141 @item filter_params
12142 A '|'-separated list of parameters to pass to the frei0r effect.
12146 A frei0r effect parameter can be a boolean (its value is either
12147 "y" or "n"), a double, a color (specified as
12148 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
12149 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
12150 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
12151 a position (specified as @var{X}/@var{Y}, where
12152 @var{X} and @var{Y} are floating point numbers) and/or a string.
12154 The number and types of parameters depend on the loaded effect. If an
12155 effect parameter is not specified, the default value is set.
12157 @subsection Examples
12161 Apply the distort0r effect, setting the first two double parameters:
12163 frei0r=filter_name=distort0r:filter_params=0.5|0.01
12167 Apply the colordistance effect, taking a color as the first parameter:
12169 frei0r=colordistance:0.2/0.3/0.4
12170 frei0r=colordistance:violet
12171 frei0r=colordistance:0x112233
12175 Apply the perspective effect, specifying the top left and top right image
12178 frei0r=perspective:0.2/0.2|0.8/0.2
12182 For more information, see
12183 @url{http://frei0r.dyne.org}
12185 @subsection Commands
12187 This filter supports the @option{filter_params} option as @ref{commands}.
12191 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
12193 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
12194 processing filter, one of them is performed once per block, not per pixel.
12195 This allows for much higher speed.
12197 The filter accepts the following options:
12201 Set quality. This option defines the number of levels for averaging. It accepts
12202 an integer in the range 4-5. Default value is @code{4}.
12205 Force a constant quantization parameter. It accepts an integer in range 0-63.
12206 If not set, the filter will use the QP from the video stream (if available).
12209 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
12210 more details but also more artifacts, while higher values make the image smoother
12211 but also blurrier. Default value is @code{0} − PSNR optimal.
12213 @item use_bframe_qp
12214 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
12215 option may cause flicker since the B-Frames have often larger QP. Default is
12216 @code{0} (not enabled).
12222 Apply Gaussian blur filter.
12224 The filter accepts the following options:
12228 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
12231 Set number of steps for Gaussian approximation. Default is @code{1}.
12234 Set which planes to filter. By default all planes are filtered.
12237 Set vertical sigma, if negative it will be same as @code{sigma}.
12238 Default is @code{-1}.
12241 @subsection Commands
12242 This filter supports same commands as options.
12243 The command accepts the same syntax of the corresponding option.
12245 If the specified expression is not valid, it is kept at its current
12250 Apply generic equation to each pixel.
12252 The filter accepts the following options:
12255 @item lum_expr, lum
12256 Set the luminance expression.
12258 Set the chrominance blue expression.
12260 Set the chrominance red expression.
12261 @item alpha_expr, a
12262 Set the alpha expression.
12264 Set the red expression.
12265 @item green_expr, g
12266 Set the green expression.
12268 Set the blue expression.
12271 The colorspace is selected according to the specified options. If one
12272 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
12273 options is specified, the filter will automatically select a YCbCr
12274 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
12275 @option{blue_expr} options is specified, it will select an RGB
12278 If one of the chrominance expression is not defined, it falls back on the other
12279 one. If no alpha expression is specified it will evaluate to opaque value.
12280 If none of chrominance expressions are specified, they will evaluate
12281 to the luminance expression.
12283 The expressions can use the following variables and functions:
12287 The sequential number of the filtered frame, starting from @code{0}.
12291 The coordinates of the current sample.
12295 The width and height of the image.
12299 Width and height scale depending on the currently filtered plane. It is the
12300 ratio between the corresponding luma plane number of pixels and the current
12301 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
12302 @code{0.5,0.5} for chroma planes.
12305 Time of the current frame, expressed in seconds.
12308 Return the value of the pixel at location (@var{x},@var{y}) of the current
12312 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
12316 Return the value of the pixel at location (@var{x},@var{y}) of the
12317 blue-difference chroma plane. Return 0 if there is no such plane.
12320 Return the value of the pixel at location (@var{x},@var{y}) of the
12321 red-difference chroma plane. Return 0 if there is no such plane.
12326 Return the value of the pixel at location (@var{x},@var{y}) of the
12327 red/green/blue component. Return 0 if there is no such component.
12330 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
12331 plane. Return 0 if there is no such plane.
12333 @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)
12334 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
12335 sums of samples within a rectangle. See the functions without the sum postfix.
12337 @item interpolation
12338 Set one of interpolation methods:
12343 Default is bilinear.
12346 For functions, if @var{x} and @var{y} are outside the area, the value will be
12347 automatically clipped to the closer edge.
12349 Please note that this filter can use multiple threads in which case each slice
12350 will have its own expression state. If you want to use only a single expression
12351 state because your expressions depend on previous state then you should limit
12352 the number of filter threads to 1.
12354 @subsection Examples
12358 Flip the image horizontally:
12364 Generate a bidimensional sine wave, with angle @code{PI/3} and a
12365 wavelength of 100 pixels:
12367 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
12371 Generate a fancy enigmatic moving light:
12373 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
12377 Generate a quick emboss effect:
12379 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
12383 Modify RGB components depending on pixel position:
12385 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
12389 Create a radial gradient that is the same size as the input (also see
12390 the @ref{vignette} filter):
12392 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
12398 Fix the banding artifacts that are sometimes introduced into nearly flat
12399 regions by truncation to 8-bit color depth.
12400 Interpolate the gradients that should go where the bands are, and
12403 It is designed for playback only. Do not use it prior to
12404 lossy compression, because compression tends to lose the dither and
12405 bring back the bands.
12407 It accepts the following parameters:
12412 The maximum amount by which the filter will change any one pixel. This is also
12413 the threshold for detecting nearly flat regions. Acceptable values range from
12414 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
12418 The neighborhood to fit the gradient to. A larger radius makes for smoother
12419 gradients, but also prevents the filter from modifying the pixels near detailed
12420 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
12421 values will be clipped to the valid range.
12425 Alternatively, the options can be specified as a flat string:
12426 @var{strength}[:@var{radius}]
12428 @subsection Examples
12432 Apply the filter with a @code{3.5} strength and radius of @code{8}:
12438 Specify radius, omitting the strength (which will fall-back to the default
12446 @anchor{graphmonitor}
12447 @section graphmonitor
12448 Show various filtergraph stats.
12450 With this filter one can debug complete filtergraph.
12451 Especially issues with links filling with queued frames.
12453 The filter accepts the following options:
12457 Set video output size. Default is @var{hd720}.
12460 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
12463 Set output mode, can be @var{fulll} or @var{compact}.
12464 In @var{compact} mode only filters with some queued frames have displayed stats.
12467 Set flags which enable which stats are shown in video.
12469 Available values for flags are:
12472 Display number of queued frames in each link.
12474 @item frame_count_in
12475 Display number of frames taken from filter.
12477 @item frame_count_out
12478 Display number of frames given out from filter.
12481 Display current filtered frame pts.
12484 Display current filtered frame time.
12487 Display time base for filter link.
12490 Display used format for filter link.
12493 Display video size or number of audio channels in case of audio used by filter link.
12496 Display video frame rate or sample rate in case of audio used by filter link.
12499 Display link output status.
12503 Set upper limit for video rate of output stream, Default value is @var{25}.
12504 This guarantee that output video frame rate will not be higher than this value.
12508 A color constancy variation filter which estimates scene illumination via grey edge algorithm
12509 and corrects the scene colors accordingly.
12511 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
12513 The filter accepts the following options:
12517 The order of differentiation to be applied on the scene. Must be chosen in the range
12518 [0,2] and default value is 1.
12521 The Minkowski parameter to be used for calculating the Minkowski distance. Must
12522 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
12523 max value instead of calculating Minkowski distance.
12526 The standard deviation of Gaussian blur to be applied on the scene. Must be
12527 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
12528 can't be equal to 0 if @var{difford} is greater than 0.
12531 @subsection Examples
12537 greyedge=difford=1:minknorm=5:sigma=2
12543 greyedge=difford=1:minknorm=0:sigma=2
12551 Apply a Hald CLUT to a video stream.
12553 First input is the video stream to process, and second one is the Hald CLUT.
12554 The Hald CLUT input can be a simple picture or a complete video stream.
12556 The filter accepts the following options:
12560 Force termination when the shortest input terminates. Default is @code{0}.
12562 Continue applying the last CLUT after the end of the stream. A value of
12563 @code{0} disable the filter after the last frame of the CLUT is reached.
12564 Default is @code{1}.
12567 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
12568 filters share the same internals).
12570 This filter also supports the @ref{framesync} options.
12572 More information about the Hald CLUT can be found on Eskil Steenberg's website
12573 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
12575 @subsection Workflow examples
12577 @subsubsection Hald CLUT video stream
12579 Generate an identity Hald CLUT stream altered with various effects:
12581 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
12584 Note: make sure you use a lossless codec.
12586 Then use it with @code{haldclut} to apply it on some random stream:
12588 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
12591 The Hald CLUT will be applied to the 10 first seconds (duration of
12592 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
12593 to the remaining frames of the @code{mandelbrot} stream.
12595 @subsubsection Hald CLUT with preview
12597 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
12598 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
12599 biggest possible square starting at the top left of the picture. The remaining
12600 padding pixels (bottom or right) will be ignored. This area can be used to add
12601 a preview of the Hald CLUT.
12603 Typically, the following generated Hald CLUT will be supported by the
12604 @code{haldclut} filter:
12607 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
12608 pad=iw+320 [padded_clut];
12609 smptebars=s=320x256, split [a][b];
12610 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
12611 [main][b] overlay=W-320" -frames:v 1 clut.png
12614 It contains the original and a preview of the effect of the CLUT: SMPTE color
12615 bars are displayed on the right-top, and below the same color bars processed by
12618 Then, the effect of this Hald CLUT can be visualized with:
12620 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
12625 Flip the input video horizontally.
12627 For example, to horizontally flip the input video with @command{ffmpeg}:
12629 ffmpeg -i in.avi -vf "hflip" out.avi
12633 This filter applies a global color histogram equalization on a
12636 It can be used to correct video that has a compressed range of pixel
12637 intensities. The filter redistributes the pixel intensities to
12638 equalize their distribution across the intensity range. It may be
12639 viewed as an "automatically adjusting contrast filter". This filter is
12640 useful only for correcting degraded or poorly captured source
12643 The filter accepts the following options:
12647 Determine the amount of equalization to be applied. As the strength
12648 is reduced, the distribution of pixel intensities more-and-more
12649 approaches that of the input frame. The value must be a float number
12650 in the range [0,1] and defaults to 0.200.
12653 Set the maximum intensity that can generated and scale the output
12654 values appropriately. The strength should be set as desired and then
12655 the intensity can be limited if needed to avoid washing-out. The value
12656 must be a float number in the range [0,1] and defaults to 0.210.
12659 Set the antibanding level. If enabled the filter will randomly vary
12660 the luminance of output pixels by a small amount to avoid banding of
12661 the histogram. Possible values are @code{none}, @code{weak} or
12662 @code{strong}. It defaults to @code{none}.
12668 Compute and draw a color distribution histogram for the input video.
12670 The computed histogram is a representation of the color component
12671 distribution in an image.
12673 Standard histogram displays the color components distribution in an image.
12674 Displays color graph for each color component. Shows distribution of
12675 the Y, U, V, A or R, G, B components, depending on input format, in the
12676 current frame. Below each graph a color component scale meter is shown.
12678 The filter accepts the following options:
12682 Set height of level. Default value is @code{200}.
12683 Allowed range is [50, 2048].
12686 Set height of color scale. Default value is @code{12}.
12687 Allowed range is [0, 40].
12691 It accepts the following values:
12694 Per color component graphs are placed below each other.
12697 Per color component graphs are placed side by side.
12700 Presents information identical to that in the @code{parade}, except
12701 that the graphs representing color components are superimposed directly
12704 Default is @code{stack}.
12707 Set mode. Can be either @code{linear}, or @code{logarithmic}.
12708 Default is @code{linear}.
12711 Set what color components to display.
12712 Default is @code{7}.
12715 Set foreground opacity. Default is @code{0.7}.
12718 Set background opacity. Default is @code{0.5}.
12721 @subsection Examples
12726 Calculate and draw histogram:
12728 ffplay -i input -vf histogram
12736 This is a high precision/quality 3d denoise filter. It aims to reduce
12737 image noise, producing smooth images and making still images really
12738 still. It should enhance compressibility.
12740 It accepts the following optional parameters:
12744 A non-negative floating point number which specifies spatial luma strength.
12745 It defaults to 4.0.
12747 @item chroma_spatial
12748 A non-negative floating point number which specifies spatial chroma strength.
12749 It defaults to 3.0*@var{luma_spatial}/4.0.
12752 A floating point number which specifies luma temporal strength. It defaults to
12753 6.0*@var{luma_spatial}/4.0.
12756 A floating point number which specifies chroma temporal strength. It defaults to
12757 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
12760 @subsection Commands
12761 This filter supports same @ref{commands} as options.
12762 The command accepts the same syntax of the corresponding option.
12764 If the specified expression is not valid, it is kept at its current
12767 @anchor{hwdownload}
12768 @section hwdownload
12770 Download hardware frames to system memory.
12772 The input must be in hardware frames, and the output a non-hardware format.
12773 Not all formats will be supported on the output - it may be necessary to insert
12774 an additional @option{format} filter immediately following in the graph to get
12775 the output in a supported format.
12779 Map hardware frames to system memory or to another device.
12781 This filter has several different modes of operation; which one is used depends
12782 on the input and output formats:
12785 Hardware frame input, normal frame output
12787 Map the input frames to system memory and pass them to the output. If the
12788 original hardware frame is later required (for example, after overlaying
12789 something else on part of it), the @option{hwmap} filter can be used again
12790 in the next mode to retrieve it.
12792 Normal frame input, hardware frame output
12794 If the input is actually a software-mapped hardware frame, then unmap it -
12795 that is, return the original hardware frame.
12797 Otherwise, a device must be provided. Create new hardware surfaces on that
12798 device for the output, then map them back to the software format at the input
12799 and give those frames to the preceding filter. This will then act like the
12800 @option{hwupload} filter, but may be able to avoid an additional copy when
12801 the input is already in a compatible format.
12803 Hardware frame input and output
12805 A device must be supplied for the output, either directly or with the
12806 @option{derive_device} option. The input and output devices must be of
12807 different types and compatible - the exact meaning of this is
12808 system-dependent, but typically it means that they must refer to the same
12809 underlying hardware context (for example, refer to the same graphics card).
12811 If the input frames were originally created on the output device, then unmap
12812 to retrieve the original frames.
12814 Otherwise, map the frames to the output device - create new hardware frames
12815 on the output corresponding to the frames on the input.
12818 The following additional parameters are accepted:
12822 Set the frame mapping mode. Some combination of:
12825 The mapped frame should be readable.
12827 The mapped frame should be writeable.
12829 The mapping will always overwrite the entire frame.
12831 This may improve performance in some cases, as the original contents of the
12832 frame need not be loaded.
12834 The mapping must not involve any copying.
12836 Indirect mappings to copies of frames are created in some cases where either
12837 direct mapping is not possible or it would have unexpected properties.
12838 Setting this flag ensures that the mapping is direct and will fail if that is
12841 Defaults to @var{read+write} if not specified.
12843 @item derive_device @var{type}
12844 Rather than using the device supplied at initialisation, instead derive a new
12845 device of type @var{type} from the device the input frames exist on.
12848 In a hardware to hardware mapping, map in reverse - create frames in the sink
12849 and map them back to the source. This may be necessary in some cases where
12850 a mapping in one direction is required but only the opposite direction is
12851 supported by the devices being used.
12853 This option is dangerous - it may break the preceding filter in undefined
12854 ways if there are any additional constraints on that filter's output.
12855 Do not use it without fully understanding the implications of its use.
12861 Upload system memory frames to hardware surfaces.
12863 The device to upload to must be supplied when the filter is initialised. If
12864 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
12865 option or with the @option{derive_device} option. The input and output devices
12866 must be of different types and compatible - the exact meaning of this is
12867 system-dependent, but typically it means that they must refer to the same
12868 underlying hardware context (for example, refer to the same graphics card).
12870 The following additional parameters are accepted:
12873 @item derive_device @var{type}
12874 Rather than using the device supplied at initialisation, instead derive a new
12875 device of type @var{type} from the device the input frames exist on.
12878 @anchor{hwupload_cuda}
12879 @section hwupload_cuda
12881 Upload system memory frames to a CUDA device.
12883 It accepts the following optional parameters:
12887 The number of the CUDA device to use
12892 Apply a high-quality magnification filter designed for pixel art. This filter
12893 was originally created by Maxim Stepin.
12895 It accepts the following option:
12899 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
12900 @code{hq3x} and @code{4} for @code{hq4x}.
12901 Default is @code{3}.
12905 Stack input videos horizontally.
12907 All streams must be of same pixel format and of same height.
12909 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
12910 to create same output.
12912 The filter accepts the following option:
12916 Set number of input streams. Default is 2.
12919 If set to 1, force the output to terminate when the shortest input
12920 terminates. Default value is 0.
12925 Modify the hue and/or the saturation of the input.
12927 It accepts the following parameters:
12931 Specify the hue angle as a number of degrees. It accepts an expression,
12932 and defaults to "0".
12935 Specify the saturation in the [-10,10] range. It accepts an expression and
12939 Specify the hue angle as a number of radians. It accepts an
12940 expression, and defaults to "0".
12943 Specify the brightness in the [-10,10] range. It accepts an expression and
12947 @option{h} and @option{H} are mutually exclusive, and can't be
12948 specified at the same time.
12950 The @option{b}, @option{h}, @option{H} and @option{s} option values are
12951 expressions containing the following constants:
12955 frame count of the input frame starting from 0
12958 presentation timestamp of the input frame expressed in time base units
12961 frame rate of the input video, NAN if the input frame rate is unknown
12964 timestamp expressed in seconds, NAN if the input timestamp is unknown
12967 time base of the input video
12970 @subsection Examples
12974 Set the hue to 90 degrees and the saturation to 1.0:
12980 Same command but expressing the hue in radians:
12986 Rotate hue and make the saturation swing between 0
12987 and 2 over a period of 1 second:
12989 hue="H=2*PI*t: s=sin(2*PI*t)+1"
12993 Apply a 3 seconds saturation fade-in effect starting at 0:
12995 hue="s=min(t/3\,1)"
12998 The general fade-in expression can be written as:
13000 hue="s=min(0\, max((t-START)/DURATION\, 1))"
13004 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
13006 hue="s=max(0\, min(1\, (8-t)/3))"
13009 The general fade-out expression can be written as:
13011 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
13016 @subsection Commands
13018 This filter supports the following commands:
13024 Modify the hue and/or the saturation and/or brightness of the input video.
13025 The command accepts the same syntax of the corresponding option.
13027 If the specified expression is not valid, it is kept at its current
13031 @section hysteresis
13033 Grow first stream into second stream by connecting components.
13034 This makes it possible to build more robust edge masks.
13036 This filter accepts the following options:
13040 Set which planes will be processed as bitmap, unprocessed planes will be
13041 copied from first stream.
13042 By default value 0xf, all planes will be processed.
13045 Set threshold which is used in filtering. If pixel component value is higher than
13046 this value filter algorithm for connecting components is activated.
13047 By default value is 0.
13050 The @code{hysteresis} filter also supports the @ref{framesync} options.
13054 Detect video interlacing type.
13056 This filter tries to detect if the input frames are interlaced, progressive,
13057 top or bottom field first. It will also try to detect fields that are
13058 repeated between adjacent frames (a sign of telecine).
13060 Single frame detection considers only immediately adjacent frames when classifying each frame.
13061 Multiple frame detection incorporates the classification history of previous frames.
13063 The filter will log these metadata values:
13066 @item single.current_frame
13067 Detected type of current frame using single-frame detection. One of:
13068 ``tff'' (top field first), ``bff'' (bottom field first),
13069 ``progressive'', or ``undetermined''
13072 Cumulative number of frames detected as top field first using single-frame detection.
13075 Cumulative number of frames detected as top field first using multiple-frame detection.
13078 Cumulative number of frames detected as bottom field first using single-frame detection.
13080 @item multiple.current_frame
13081 Detected type of current frame using multiple-frame detection. One of:
13082 ``tff'' (top field first), ``bff'' (bottom field first),
13083 ``progressive'', or ``undetermined''
13086 Cumulative number of frames detected as bottom field first using multiple-frame detection.
13088 @item single.progressive
13089 Cumulative number of frames detected as progressive using single-frame detection.
13091 @item multiple.progressive
13092 Cumulative number of frames detected as progressive using multiple-frame detection.
13094 @item single.undetermined
13095 Cumulative number of frames that could not be classified using single-frame detection.
13097 @item multiple.undetermined
13098 Cumulative number of frames that could not be classified using multiple-frame detection.
13100 @item repeated.current_frame
13101 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
13103 @item repeated.neither
13104 Cumulative number of frames with no repeated field.
13107 Cumulative number of frames with the top field repeated from the previous frame's top field.
13109 @item repeated.bottom
13110 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
13113 The filter accepts the following options:
13117 Set interlacing threshold.
13119 Set progressive threshold.
13121 Threshold for repeated field detection.
13123 Number of frames after which a given frame's contribution to the
13124 statistics is halved (i.e., it contributes only 0.5 to its
13125 classification). The default of 0 means that all frames seen are given
13126 full weight of 1.0 forever.
13127 @item analyze_interlaced_flag
13128 When this is not 0 then idet will use the specified number of frames to determine
13129 if the interlaced flag is accurate, it will not count undetermined frames.
13130 If the flag is found to be accurate it will be used without any further
13131 computations, if it is found to be inaccurate it will be cleared without any
13132 further computations. This allows inserting the idet filter as a low computational
13133 method to clean up the interlaced flag
13138 Deinterleave or interleave fields.
13140 This filter allows one to process interlaced images fields without
13141 deinterlacing them. Deinterleaving splits the input frame into 2
13142 fields (so called half pictures). Odd lines are moved to the top
13143 half of the output image, even lines to the bottom half.
13144 You can process (filter) them independently and then re-interleave them.
13146 The filter accepts the following options:
13150 @item chroma_mode, c
13151 @item alpha_mode, a
13152 Available values for @var{luma_mode}, @var{chroma_mode} and
13153 @var{alpha_mode} are:
13159 @item deinterleave, d
13160 Deinterleave fields, placing one above the other.
13162 @item interleave, i
13163 Interleave fields. Reverse the effect of deinterleaving.
13165 Default value is @code{none}.
13167 @item luma_swap, ls
13168 @item chroma_swap, cs
13169 @item alpha_swap, as
13170 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
13173 @subsection Commands
13175 This filter supports the all above options as @ref{commands}.
13179 Apply inflate effect to the video.
13181 This filter replaces the pixel by the local(3x3) average by taking into account
13182 only values higher than the pixel.
13184 It accepts the following options:
13191 Limit the maximum change for each plane, default is 65535.
13192 If 0, plane will remain unchanged.
13195 @subsection Commands
13197 This filter supports the all above options as @ref{commands}.
13201 Simple interlacing filter from progressive contents. This interleaves upper (or
13202 lower) lines from odd frames with lower (or upper) lines from even frames,
13203 halving the frame rate and preserving image height.
13206 Original Original New Frame
13207 Frame 'j' Frame 'j+1' (tff)
13208 ========== =========== ==================
13209 Line 0 --------------------> Frame 'j' Line 0
13210 Line 1 Line 1 ----> Frame 'j+1' Line 1
13211 Line 2 ---------------------> Frame 'j' Line 2
13212 Line 3 Line 3 ----> Frame 'j+1' Line 3
13214 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
13217 It accepts the following optional parameters:
13221 This determines whether the interlaced frame is taken from the even
13222 (tff - default) or odd (bff) lines of the progressive frame.
13225 Vertical lowpass filter to avoid twitter interlacing and
13226 reduce moire patterns.
13230 Disable vertical lowpass filter
13233 Enable linear filter (default)
13236 Enable complex filter. This will slightly less reduce twitter and moire
13237 but better retain detail and subjective sharpness impression.
13244 Deinterlace input video by applying Donald Graft's adaptive kernel
13245 deinterling. Work on interlaced parts of a video to produce
13246 progressive frames.
13248 The description of the accepted parameters follows.
13252 Set the threshold which affects the filter's tolerance when
13253 determining if a pixel line must be processed. It must be an integer
13254 in the range [0,255] and defaults to 10. A value of 0 will result in
13255 applying the process on every pixels.
13258 Paint pixels exceeding the threshold value to white if set to 1.
13262 Set the fields order. Swap fields if set to 1, leave fields alone if
13266 Enable additional sharpening if set to 1. Default is 0.
13269 Enable twoway sharpening if set to 1. Default is 0.
13272 @subsection Examples
13276 Apply default values:
13278 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
13282 Enable additional sharpening:
13288 Paint processed pixels in white:
13296 Slowly update darker pixels.
13298 This filter makes short flashes of light appear longer.
13299 This filter accepts the following options:
13303 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
13306 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
13309 @section lenscorrection
13311 Correct radial lens distortion
13313 This filter can be used to correct for radial distortion as can result from the use
13314 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
13315 one can use tools available for example as part of opencv or simply trial-and-error.
13316 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
13317 and extract the k1 and k2 coefficients from the resulting matrix.
13319 Note that effectively the same filter is available in the open-source tools Krita and
13320 Digikam from the KDE project.
13322 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
13323 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
13324 brightness distribution, so you may want to use both filters together in certain
13325 cases, though you will have to take care of ordering, i.e. whether vignetting should
13326 be applied before or after lens correction.
13328 @subsection Options
13330 The filter accepts the following options:
13334 Relative x-coordinate of the focal point of the image, and thereby the center of the
13335 distortion. This value has a range [0,1] and is expressed as fractions of the image
13336 width. Default is 0.5.
13338 Relative y-coordinate of the focal point of the image, and thereby the center of the
13339 distortion. This value has a range [0,1] and is expressed as fractions of the image
13340 height. Default is 0.5.
13342 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
13343 no correction. Default is 0.
13345 Coefficient of the double quadratic correction term. This value has a range [-1,1].
13346 0 means no correction. Default is 0.
13349 The formula that generates the correction is:
13351 @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)
13353 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
13354 distances from the focal point in the source and target images, respectively.
13358 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
13360 The @code{lensfun} filter requires the camera make, camera model, and lens model
13361 to apply the lens correction. The filter will load the lensfun database and
13362 query it to find the corresponding camera and lens entries in the database. As
13363 long as these entries can be found with the given options, the filter can
13364 perform corrections on frames. Note that incomplete strings will result in the
13365 filter choosing the best match with the given options, and the filter will
13366 output the chosen camera and lens models (logged with level "info"). You must
13367 provide the make, camera model, and lens model as they are required.
13369 The filter accepts the following options:
13373 The make of the camera (for example, "Canon"). This option is required.
13376 The model of the camera (for example, "Canon EOS 100D"). This option is
13380 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
13381 option is required.
13384 The type of correction to apply. The following values are valid options:
13388 Enables fixing lens vignetting.
13391 Enables fixing lens geometry. This is the default.
13394 Enables fixing chromatic aberrations.
13397 Enables fixing lens vignetting and lens geometry.
13400 Enables fixing lens vignetting and chromatic aberrations.
13403 Enables fixing both lens geometry and chromatic aberrations.
13406 Enables all possible corrections.
13410 The focal length of the image/video (zoom; expected constant for video). For
13411 example, a 18--55mm lens has focal length range of [18--55], so a value in that
13412 range should be chosen when using that lens. Default 18.
13415 The aperture of the image/video (expected constant for video). Note that
13416 aperture is only used for vignetting correction. Default 3.5.
13418 @item focus_distance
13419 The focus distance of the image/video (expected constant for video). Note that
13420 focus distance is only used for vignetting and only slightly affects the
13421 vignetting correction process. If unknown, leave it at the default value (which
13425 The scale factor which is applied after transformation. After correction the
13426 video is no longer necessarily rectangular. This parameter controls how much of
13427 the resulting image is visible. The value 0 means that a value will be chosen
13428 automatically such that there is little or no unmapped area in the output
13429 image. 1.0 means that no additional scaling is done. Lower values may result
13430 in more of the corrected image being visible, while higher values may avoid
13431 unmapped areas in the output.
13433 @item target_geometry
13434 The target geometry of the output image/video. The following values are valid
13438 @item rectilinear (default)
13441 @item equirectangular
13442 @item fisheye_orthographic
13443 @item fisheye_stereographic
13444 @item fisheye_equisolid
13445 @item fisheye_thoby
13448 Apply the reverse of image correction (instead of correcting distortion, apply
13451 @item interpolation
13452 The type of interpolation used when correcting distortion. The following values
13457 @item linear (default)
13462 @subsection Examples
13466 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
13467 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
13471 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
13475 Apply the same as before, but only for the first 5 seconds of video.
13478 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
13485 Obtain the VMAF (Video Multi-Method Assessment Fusion)
13486 score between two input videos.
13488 The obtained VMAF score is printed through the logging system.
13490 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
13491 After installing the library it can be enabled using:
13492 @code{./configure --enable-libvmaf}.
13493 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
13495 The filter has following options:
13499 Set the model path which is to be used for SVM.
13500 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
13503 Set the file path to be used to store logs.
13506 Set the format of the log file (csv, json or xml).
13508 @item enable_transform
13509 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
13510 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
13511 Default value: @code{false}
13514 Invokes the phone model which will generate VMAF scores higher than in the
13515 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
13516 Default value: @code{false}
13519 Enables computing psnr along with vmaf.
13520 Default value: @code{false}
13523 Enables computing ssim along with vmaf.
13524 Default value: @code{false}
13527 Enables computing ms_ssim along with vmaf.
13528 Default value: @code{false}
13531 Set the pool method to be used for computing vmaf.
13532 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
13535 Set number of threads to be used when computing vmaf.
13536 Default value: @code{0}, which makes use of all available logical processors.
13539 Set interval for frame subsampling used when computing vmaf.
13540 Default value: @code{1}
13542 @item enable_conf_interval
13543 Enables confidence interval.
13544 Default value: @code{false}
13547 This filter also supports the @ref{framesync} options.
13549 @subsection Examples
13552 On the below examples the input file @file{main.mpg} being processed is
13553 compared with the reference file @file{ref.mpg}.
13556 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
13560 Example with options:
13562 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
13566 Example with options and different containers:
13568 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 -
13574 Limits the pixel components values to the specified range [min, max].
13576 The filter accepts the following options:
13580 Lower bound. Defaults to the lowest allowed value for the input.
13583 Upper bound. Defaults to the highest allowed value for the input.
13586 Specify which planes will be processed. Defaults to all available.
13589 @subsection Commands
13591 This filter supports the all above options as @ref{commands}.
13597 The filter accepts the following options:
13601 Set the number of loops. Setting this value to -1 will result in infinite loops.
13605 Set maximal size in number of frames. Default is 0.
13608 Set first frame of loop. Default is 0.
13611 @subsection Examples
13615 Loop single first frame infinitely:
13617 loop=loop=-1:size=1:start=0
13621 Loop single first frame 10 times:
13623 loop=loop=10:size=1:start=0
13627 Loop 10 first frames 5 times:
13629 loop=loop=5:size=10:start=0
13635 Apply a 1D LUT to an input video.
13637 The filter accepts the following options:
13641 Set the 1D LUT file name.
13643 Currently supported formats:
13652 Select interpolation mode.
13654 Available values are:
13658 Use values from the nearest defined point.
13660 Interpolate values using the linear interpolation.
13662 Interpolate values using the cosine interpolation.
13664 Interpolate values using the cubic interpolation.
13666 Interpolate values using the spline interpolation.
13673 Apply a 3D LUT to an input video.
13675 The filter accepts the following options:
13679 Set the 3D LUT file name.
13681 Currently supported formats:
13695 Select interpolation mode.
13697 Available values are:
13701 Use values from the nearest defined point.
13703 Interpolate values using the 8 points defining a cube.
13705 Interpolate values using a tetrahedron.
13711 Turn certain luma values into transparency.
13713 The filter accepts the following options:
13717 Set the luma which will be used as base for transparency.
13718 Default value is @code{0}.
13721 Set the range of luma values to be keyed out.
13722 Default value is @code{0.01}.
13725 Set the range of softness. Default value is @code{0}.
13726 Use this to control gradual transition from zero to full transparency.
13729 @subsection Commands
13730 This filter supports same @ref{commands} as options.
13731 The command accepts the same syntax of the corresponding option.
13733 If the specified expression is not valid, it is kept at its current
13736 @section lut, lutrgb, lutyuv
13738 Compute a look-up table for binding each pixel component input value
13739 to an output value, and apply it to the input video.
13741 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
13742 to an RGB input video.
13744 These filters accept the following parameters:
13747 set first pixel component expression
13749 set second pixel component expression
13751 set third pixel component expression
13753 set fourth pixel component expression, corresponds to the alpha component
13756 set red component expression
13758 set green component expression
13760 set blue component expression
13762 alpha component expression
13765 set Y/luminance component expression
13767 set U/Cb component expression
13769 set V/Cr component expression
13772 Each of them specifies the expression to use for computing the lookup table for
13773 the corresponding pixel component values.
13775 The exact component associated to each of the @var{c*} options depends on the
13778 The @var{lut} filter requires either YUV or RGB pixel formats in input,
13779 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
13781 The expressions can contain the following constants and functions:
13786 The input width and height.
13789 The input value for the pixel component.
13792 The input value, clipped to the @var{minval}-@var{maxval} range.
13795 The maximum value for the pixel component.
13798 The minimum value for the pixel component.
13801 The negated value for the pixel component value, clipped to the
13802 @var{minval}-@var{maxval} range; it corresponds to the expression
13803 "maxval-clipval+minval".
13806 The computed value in @var{val}, clipped to the
13807 @var{minval}-@var{maxval} range.
13809 @item gammaval(gamma)
13810 The computed gamma correction value of the pixel component value,
13811 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
13813 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
13817 All expressions default to "val".
13819 @subsection Examples
13823 Negate input video:
13825 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
13826 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
13829 The above is the same as:
13831 lutrgb="r=negval:g=negval:b=negval"
13832 lutyuv="y=negval:u=negval:v=negval"
13842 Remove chroma components, turning the video into a graytone image:
13844 lutyuv="u=128:v=128"
13848 Apply a luma burning effect:
13854 Remove green and blue components:
13860 Set a constant alpha channel value on input:
13862 format=rgba,lutrgb=a="maxval-minval/2"
13866 Correct luminance gamma by a factor of 0.5:
13868 lutyuv=y=gammaval(0.5)
13872 Discard least significant bits of luma:
13874 lutyuv=y='bitand(val, 128+64+32)'
13878 Technicolor like effect:
13880 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
13884 @section lut2, tlut2
13886 The @code{lut2} filter takes two input streams and outputs one
13889 The @code{tlut2} (time lut2) filter takes two consecutive frames
13890 from one single stream.
13892 This filter accepts the following parameters:
13895 set first pixel component expression
13897 set second pixel component expression
13899 set third pixel component expression
13901 set fourth pixel component expression, corresponds to the alpha component
13904 set output bit depth, only available for @code{lut2} filter. By default is 0,
13905 which means bit depth is automatically picked from first input format.
13908 The @code{lut2} filter also supports the @ref{framesync} options.
13910 Each of them specifies the expression to use for computing the lookup table for
13911 the corresponding pixel component values.
13913 The exact component associated to each of the @var{c*} options depends on the
13916 The expressions can contain the following constants:
13921 The input width and height.
13924 The first input value for the pixel component.
13927 The second input value for the pixel component.
13930 The first input video bit depth.
13933 The second input video bit depth.
13936 All expressions default to "x".
13938 @subsection Examples
13942 Highlight differences between two RGB video streams:
13944 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)'
13948 Highlight differences between two YUV video streams:
13950 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)'
13954 Show max difference between two video streams:
13956 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)))'
13960 @section maskedclamp
13962 Clamp the first input stream with the second input and third input stream.
13964 Returns the value of first stream to be between second input
13965 stream - @code{undershoot} and third input stream + @code{overshoot}.
13967 This filter accepts the following options:
13970 Default value is @code{0}.
13973 Default value is @code{0}.
13976 Set which planes will be processed as bitmap, unprocessed planes will be
13977 copied from first stream.
13978 By default value 0xf, all planes will be processed.
13981 @subsection Commands
13983 This filter supports the all above options as @ref{commands}.
13987 Merge the second and third input stream into output stream using absolute differences
13988 between second input stream and first input stream and absolute difference between
13989 third input stream and first input stream. The picked value will be from second input
13990 stream if second absolute difference is greater than first one or from third input stream
13993 This filter accepts the following options:
13996 Set which planes will be processed as bitmap, unprocessed planes will be
13997 copied from first stream.
13998 By default value 0xf, all planes will be processed.
14001 @subsection Commands
14003 This filter supports the all above options as @ref{commands}.
14005 @section maskedmerge
14007 Merge the first input stream with the second input stream using per pixel
14008 weights in the third input stream.
14010 A value of 0 in the third stream pixel component means that pixel component
14011 from first stream is returned unchanged, while maximum value (eg. 255 for
14012 8-bit videos) means that pixel component from second stream is returned
14013 unchanged. Intermediate values define the amount of merging between both
14014 input stream's pixel components.
14016 This filter accepts the following options:
14019 Set which planes will be processed as bitmap, unprocessed planes will be
14020 copied from first stream.
14021 By default value 0xf, all planes will be processed.
14026 Merge the second and third input stream into output stream using absolute differences
14027 between second input stream and first input stream and absolute difference between
14028 third input stream and first input stream. The picked value will be from second input
14029 stream if second absolute difference is less than first one or from third input stream
14032 This filter accepts the following options:
14035 Set which planes will be processed as bitmap, unprocessed planes will be
14036 copied from first stream.
14037 By default value 0xf, all planes will be processed.
14040 @subsection Commands
14042 This filter supports the all above options as @ref{commands}.
14044 @section maskedthreshold
14045 Pick pixels comparing absolute difference of two video streams with fixed
14048 If absolute difference between pixel component of first and second video
14049 stream is equal or lower than user supplied threshold than pixel component
14050 from first video stream is picked, otherwise pixel component from second
14051 video stream is picked.
14053 This filter accepts the following options:
14056 Set threshold used when picking pixels from absolute difference from two input
14060 Set which planes will be processed as bitmap, unprocessed planes will be
14061 copied from second stream.
14062 By default value 0xf, all planes will be processed.
14065 @subsection Commands
14067 This filter supports the all above options as @ref{commands}.
14070 Create mask from input video.
14072 For example it is useful to create motion masks after @code{tblend} filter.
14074 This filter accepts the following options:
14078 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
14081 Set high threshold. Any pixel component higher than this value will be set to max value
14082 allowed for current pixel format.
14085 Set planes to filter, by default all available planes are filtered.
14088 Fill all frame pixels with this value.
14091 Set max average pixel value for frame. If sum of all pixel components is higher that this
14092 average, output frame will be completely filled with value set by @var{fill} option.
14093 Typically useful for scene changes when used in combination with @code{tblend} filter.
14098 Apply motion-compensation deinterlacing.
14100 It needs one field per frame as input and must thus be used together
14101 with yadif=1/3 or equivalent.
14103 This filter accepts the following options:
14106 Set the deinterlacing mode.
14108 It accepts one of the following values:
14113 use iterative motion estimation
14115 like @samp{slow}, but use multiple reference frames.
14117 Default value is @samp{fast}.
14120 Set the picture field parity assumed for the input video. It must be
14121 one of the following values:
14125 assume top field first
14127 assume bottom field first
14130 Default value is @samp{bff}.
14133 Set per-block quantization parameter (QP) used by the internal
14136 Higher values should result in a smoother motion vector field but less
14137 optimal individual vectors. Default value is 1.
14142 Pick median pixel from certain rectangle defined by radius.
14144 This filter accepts the following options:
14148 Set horizontal radius size. Default value is @code{1}.
14149 Allowed range is integer from 1 to 127.
14152 Set which planes to process. Default is @code{15}, which is all available planes.
14155 Set vertical radius size. Default value is @code{0}.
14156 Allowed range is integer from 0 to 127.
14157 If it is 0, value will be picked from horizontal @code{radius} option.
14160 Set median percentile. Default value is @code{0.5}.
14161 Default value of @code{0.5} will pick always median values, while @code{0} will pick
14162 minimum values, and @code{1} maximum values.
14165 @subsection Commands
14166 This filter supports same @ref{commands} as options.
14167 The command accepts the same syntax of the corresponding option.
14169 If the specified expression is not valid, it is kept at its current
14172 @section mergeplanes
14174 Merge color channel components from several video streams.
14176 The filter accepts up to 4 input streams, and merge selected input
14177 planes to the output video.
14179 This filter accepts the following options:
14182 Set input to output plane mapping. Default is @code{0}.
14184 The mappings is specified as a bitmap. It should be specified as a
14185 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
14186 mapping for the first plane of the output stream. 'A' sets the number of
14187 the input stream to use (from 0 to 3), and 'a' the plane number of the
14188 corresponding input to use (from 0 to 3). The rest of the mappings is
14189 similar, 'Bb' describes the mapping for the output stream second
14190 plane, 'Cc' describes the mapping for the output stream third plane and
14191 'Dd' describes the mapping for the output stream fourth plane.
14194 Set output pixel format. Default is @code{yuva444p}.
14197 @subsection Examples
14201 Merge three gray video streams of same width and height into single video stream:
14203 [a0][a1][a2]mergeplanes=0x001020:yuv444p
14207 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
14209 [a0][a1]mergeplanes=0x00010210:yuva444p
14213 Swap Y and A plane in yuva444p stream:
14215 format=yuva444p,mergeplanes=0x03010200:yuva444p
14219 Swap U and V plane in yuv420p stream:
14221 format=yuv420p,mergeplanes=0x000201:yuv420p
14225 Cast a rgb24 clip to yuv444p:
14227 format=rgb24,mergeplanes=0x000102:yuv444p
14233 Estimate and export motion vectors using block matching algorithms.
14234 Motion vectors are stored in frame side data to be used by other filters.
14236 This filter accepts the following options:
14239 Specify the motion estimation method. Accepts one of the following values:
14243 Exhaustive search algorithm.
14245 Three step search algorithm.
14247 Two dimensional logarithmic search algorithm.
14249 New three step search algorithm.
14251 Four step search algorithm.
14253 Diamond search algorithm.
14255 Hexagon-based search algorithm.
14257 Enhanced predictive zonal search algorithm.
14259 Uneven multi-hexagon search algorithm.
14261 Default value is @samp{esa}.
14264 Macroblock size. Default @code{16}.
14267 Search parameter. Default @code{7}.
14270 @section midequalizer
14272 Apply Midway Image Equalization effect using two video streams.
14274 Midway Image Equalization adjusts a pair of images to have the same
14275 histogram, while maintaining their dynamics as much as possible. It's
14276 useful for e.g. matching exposures from a pair of stereo cameras.
14278 This filter has two inputs and one output, which must be of same pixel format, but
14279 may be of different sizes. The output of filter is first input adjusted with
14280 midway histogram of both inputs.
14282 This filter accepts the following option:
14286 Set which planes to process. Default is @code{15}, which is all available planes.
14289 @section minterpolate
14291 Convert the video to specified frame rate using motion interpolation.
14293 This filter accepts the following options:
14296 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}.
14299 Motion interpolation mode. Following values are accepted:
14302 Duplicate previous or next frame for interpolating new ones.
14304 Blend source frames. Interpolated frame is mean of previous and next frames.
14306 Motion compensated interpolation. Following options are effective when this mode is selected:
14310 Motion compensation mode. Following values are accepted:
14313 Overlapped block motion compensation.
14315 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
14317 Default mode is @samp{obmc}.
14320 Motion estimation mode. Following values are accepted:
14323 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
14325 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
14327 Default mode is @samp{bilat}.
14330 The algorithm to be used for motion estimation. Following values are accepted:
14333 Exhaustive search algorithm.
14335 Three step search algorithm.
14337 Two dimensional logarithmic search algorithm.
14339 New three step search algorithm.
14341 Four step search algorithm.
14343 Diamond search algorithm.
14345 Hexagon-based search algorithm.
14347 Enhanced predictive zonal search algorithm.
14349 Uneven multi-hexagon search algorithm.
14351 Default algorithm is @samp{epzs}.
14354 Macroblock size. Default @code{16}.
14357 Motion estimation search parameter. Default @code{32}.
14360 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).
14365 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:
14368 Disable scene change detection.
14370 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
14372 Default method is @samp{fdiff}.
14374 @item scd_threshold
14375 Scene change detection threshold. Default is @code{10.}.
14380 Mix several video input streams into one video stream.
14382 A description of the accepted options follows.
14386 The number of inputs. If unspecified, it defaults to 2.
14389 Specify weight of each input video stream as sequence.
14390 Each weight is separated by space. If number of weights
14391 is smaller than number of @var{frames} last specified
14392 weight will be used for all remaining unset weights.
14395 Specify scale, if it is set it will be multiplied with sum
14396 of each weight multiplied with pixel values to give final destination
14397 pixel value. By default @var{scale} is auto scaled to sum of weights.
14400 Specify how end of stream is determined.
14403 The duration of the longest input. (default)
14406 The duration of the shortest input.
14409 The duration of the first input.
14413 @section mpdecimate
14415 Drop frames that do not differ greatly from the previous frame in
14416 order to reduce frame rate.
14418 The main use of this filter is for very-low-bitrate encoding
14419 (e.g. streaming over dialup modem), but it could in theory be used for
14420 fixing movies that were inverse-telecined incorrectly.
14422 A description of the accepted options follows.
14426 Set the maximum number of consecutive frames which can be dropped (if
14427 positive), or the minimum interval between dropped frames (if
14428 negative). If the value is 0, the frame is dropped disregarding the
14429 number of previous sequentially dropped frames.
14431 Default value is 0.
14436 Set the dropping threshold values.
14438 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
14439 represent actual pixel value differences, so a threshold of 64
14440 corresponds to 1 unit of difference for each pixel, or the same spread
14441 out differently over the block.
14443 A frame is a candidate for dropping if no 8x8 blocks differ by more
14444 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
14445 meaning the whole image) differ by more than a threshold of @option{lo}.
14447 Default value for @option{hi} is 64*12, default value for @option{lo} is
14448 64*5, and default value for @option{frac} is 0.33.
14454 Negate (invert) the input video.
14456 It accepts the following option:
14461 With value 1, it negates the alpha component, if present. Default value is 0.
14467 Denoise frames using Non-Local Means algorithm.
14469 Each pixel is adjusted by looking for other pixels with similar contexts. This
14470 context similarity is defined by comparing their surrounding patches of size
14471 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
14474 Note that the research area defines centers for patches, which means some
14475 patches will be made of pixels outside that research area.
14477 The filter accepts the following options.
14481 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
14484 Set patch size. Default is 7. Must be odd number in range [0, 99].
14487 Same as @option{p} but for chroma planes.
14489 The default value is @var{0} and means automatic.
14492 Set research size. Default is 15. Must be odd number in range [0, 99].
14495 Same as @option{r} but for chroma planes.
14497 The default value is @var{0} and means automatic.
14502 Deinterlace video using neural network edge directed interpolation.
14504 This filter accepts the following options:
14508 Mandatory option, without binary file filter can not work.
14509 Currently file can be found here:
14510 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
14513 Set which frames to deinterlace, by default it is @code{all}.
14514 Can be @code{all} or @code{interlaced}.
14517 Set mode of operation.
14519 Can be one of the following:
14523 Use frame flags, both fields.
14525 Use frame flags, single field.
14527 Use top field only.
14529 Use bottom field only.
14531 Use both fields, top first.
14533 Use both fields, bottom first.
14537 Set which planes to process, by default filter process all frames.
14540 Set size of local neighborhood around each pixel, used by the predictor neural
14543 Can be one of the following:
14556 Set the number of neurons in predictor neural network.
14557 Can be one of the following:
14568 Controls the number of different neural network predictions that are blended
14569 together to compute the final output value. Can be @code{fast}, default or
14573 Set which set of weights to use in the predictor.
14574 Can be one of the following:
14578 weights trained to minimize absolute error
14580 weights trained to minimize squared error
14584 Controls whether or not the prescreener neural network is used to decide
14585 which pixels should be processed by the predictor neural network and which
14586 can be handled by simple cubic interpolation.
14587 The prescreener is trained to know whether cubic interpolation will be
14588 sufficient for a pixel or whether it should be predicted by the predictor nn.
14589 The computational complexity of the prescreener nn is much less than that of
14590 the predictor nn. Since most pixels can be handled by cubic interpolation,
14591 using the prescreener generally results in much faster processing.
14592 The prescreener is pretty accurate, so the difference between using it and not
14593 using it is almost always unnoticeable.
14595 Can be one of the following:
14603 Default is @code{new}.
14606 Set various debugging flags.
14611 Force libavfilter not to use any of the specified pixel formats for the
14612 input to the next filter.
14614 It accepts the following parameters:
14618 A '|'-separated list of pixel format names, such as
14619 pix_fmts=yuv420p|monow|rgb24".
14623 @subsection Examples
14627 Force libavfilter to use a format different from @var{yuv420p} for the
14628 input to the vflip filter:
14630 noformat=pix_fmts=yuv420p,vflip
14634 Convert the input video to any of the formats not contained in the list:
14636 noformat=yuv420p|yuv444p|yuv410p
14642 Add noise on video input frame.
14644 The filter accepts the following options:
14652 Set noise seed for specific pixel component or all pixel components in case
14653 of @var{all_seed}. Default value is @code{123457}.
14655 @item all_strength, alls
14656 @item c0_strength, c0s
14657 @item c1_strength, c1s
14658 @item c2_strength, c2s
14659 @item c3_strength, c3s
14660 Set noise strength for specific pixel component or all pixel components in case
14661 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
14663 @item all_flags, allf
14664 @item c0_flags, c0f
14665 @item c1_flags, c1f
14666 @item c2_flags, c2f
14667 @item c3_flags, c3f
14668 Set pixel component flags or set flags for all components if @var{all_flags}.
14669 Available values for component flags are:
14672 averaged temporal noise (smoother)
14674 mix random noise with a (semi)regular pattern
14676 temporal noise (noise pattern changes between frames)
14678 uniform noise (gaussian otherwise)
14682 @subsection Examples
14684 Add temporal and uniform noise to input video:
14686 noise=alls=20:allf=t+u
14691 Normalize RGB video (aka histogram stretching, contrast stretching).
14692 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
14694 For each channel of each frame, the filter computes the input range and maps
14695 it linearly to the user-specified output range. The output range defaults
14696 to the full dynamic range from pure black to pure white.
14698 Temporal smoothing can be used on the input range to reduce flickering (rapid
14699 changes in brightness) caused when small dark or bright objects enter or leave
14700 the scene. This is similar to the auto-exposure (automatic gain control) on a
14701 video camera, and, like a video camera, it may cause a period of over- or
14702 under-exposure of the video.
14704 The R,G,B channels can be normalized independently, which may cause some
14705 color shifting, or linked together as a single channel, which prevents
14706 color shifting. Linked normalization preserves hue. Independent normalization
14707 does not, so it can be used to remove some color casts. Independent and linked
14708 normalization can be combined in any ratio.
14710 The normalize filter accepts the following options:
14715 Colors which define the output range. The minimum input value is mapped to
14716 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
14717 The defaults are black and white respectively. Specifying white for
14718 @var{blackpt} and black for @var{whitept} will give color-inverted,
14719 normalized video. Shades of grey can be used to reduce the dynamic range
14720 (contrast). Specifying saturated colors here can create some interesting
14724 The number of previous frames to use for temporal smoothing. The input range
14725 of each channel is smoothed using a rolling average over the current frame
14726 and the @var{smoothing} previous frames. The default is 0 (no temporal
14730 Controls the ratio of independent (color shifting) channel normalization to
14731 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
14732 independent. Defaults to 1.0 (fully independent).
14735 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
14736 expensive no-op. Defaults to 1.0 (full strength).
14740 @subsection Commands
14741 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
14742 The command accepts the same syntax of the corresponding option.
14744 If the specified expression is not valid, it is kept at its current
14747 @subsection Examples
14749 Stretch video contrast to use the full dynamic range, with no temporal
14750 smoothing; may flicker depending on the source content:
14752 normalize=blackpt=black:whitept=white:smoothing=0
14755 As above, but with 50 frames of temporal smoothing; flicker should be
14756 reduced, depending on the source content:
14758 normalize=blackpt=black:whitept=white:smoothing=50
14761 As above, but with hue-preserving linked channel normalization:
14763 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
14766 As above, but with half strength:
14768 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
14771 Map the darkest input color to red, the brightest input color to cyan:
14773 normalize=blackpt=red:whitept=cyan
14778 Pass the video source unchanged to the output.
14781 Optical Character Recognition
14783 This filter uses Tesseract for optical character recognition. To enable
14784 compilation of this filter, you need to configure FFmpeg with
14785 @code{--enable-libtesseract}.
14787 It accepts the following options:
14791 Set datapath to tesseract data. Default is to use whatever was
14792 set at installation.
14795 Set language, default is "eng".
14798 Set character whitelist.
14801 Set character blacklist.
14804 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
14805 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
14809 Apply a video transform using libopencv.
14811 To enable this filter, install the libopencv library and headers and
14812 configure FFmpeg with @code{--enable-libopencv}.
14814 It accepts the following parameters:
14819 The name of the libopencv filter to apply.
14821 @item filter_params
14822 The parameters to pass to the libopencv filter. If not specified, the default
14823 values are assumed.
14827 Refer to the official libopencv documentation for more precise
14829 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
14831 Several libopencv filters are supported; see the following subsections.
14836 Dilate an image by using a specific structuring element.
14837 It corresponds to the libopencv function @code{cvDilate}.
14839 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
14841 @var{struct_el} represents a structuring element, and has the syntax:
14842 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
14844 @var{cols} and @var{rows} represent the number of columns and rows of
14845 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
14846 point, and @var{shape} the shape for the structuring element. @var{shape}
14847 must be "rect", "cross", "ellipse", or "custom".
14849 If the value for @var{shape} is "custom", it must be followed by a
14850 string of the form "=@var{filename}". The file with name
14851 @var{filename} is assumed to represent a binary image, with each
14852 printable character corresponding to a bright pixel. When a custom
14853 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
14854 or columns and rows of the read file are assumed instead.
14856 The default value for @var{struct_el} is "3x3+0x0/rect".
14858 @var{nb_iterations} specifies the number of times the transform is
14859 applied to the image, and defaults to 1.
14863 # Use the default values
14866 # Dilate using a structuring element with a 5x5 cross, iterating two times
14867 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
14869 # Read the shape from the file diamond.shape, iterating two times.
14870 # The file diamond.shape may contain a pattern of characters like this
14876 # The specified columns and rows are ignored
14877 # but the anchor point coordinates are not
14878 ocv=dilate:0x0+2x2/custom=diamond.shape|2
14883 Erode an image by using a specific structuring element.
14884 It corresponds to the libopencv function @code{cvErode}.
14886 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
14887 with the same syntax and semantics as the @ref{dilate} filter.
14891 Smooth the input video.
14893 The filter takes the following parameters:
14894 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
14896 @var{type} is the type of smooth filter to apply, and must be one of
14897 the following values: "blur", "blur_no_scale", "median", "gaussian",
14898 or "bilateral". The default value is "gaussian".
14900 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
14901 depends on the smooth type. @var{param1} and
14902 @var{param2} accept integer positive values or 0. @var{param3} and
14903 @var{param4} accept floating point values.
14905 The default value for @var{param1} is 3. The default value for the
14906 other parameters is 0.
14908 These parameters correspond to the parameters assigned to the
14909 libopencv function @code{cvSmooth}.
14911 @section oscilloscope
14913 2D Video Oscilloscope.
14915 Useful to measure spatial impulse, step responses, chroma delays, etc.
14917 It accepts the following parameters:
14921 Set scope center x position.
14924 Set scope center y position.
14927 Set scope size, relative to frame diagonal.
14930 Set scope tilt/rotation.
14936 Set trace center x position.
14939 Set trace center y position.
14942 Set trace width, relative to width of frame.
14945 Set trace height, relative to height of frame.
14948 Set which components to trace. By default it traces first three components.
14951 Draw trace grid. By default is enabled.
14954 Draw some statistics. By default is enabled.
14957 Draw scope. By default is enabled.
14960 @subsection Commands
14961 This filter supports same @ref{commands} as options.
14962 The command accepts the same syntax of the corresponding option.
14964 If the specified expression is not valid, it is kept at its current
14967 @subsection Examples
14971 Inspect full first row of video frame.
14973 oscilloscope=x=0.5:y=0:s=1
14977 Inspect full last row of video frame.
14979 oscilloscope=x=0.5:y=1:s=1
14983 Inspect full 5th line of video frame of height 1080.
14985 oscilloscope=x=0.5:y=5/1080:s=1
14989 Inspect full last column of video frame.
14991 oscilloscope=x=1:y=0.5:s=1:t=1
14999 Overlay one video on top of another.
15001 It takes two inputs and has one output. The first input is the "main"
15002 video on which the second input is overlaid.
15004 It accepts the following parameters:
15006 A description of the accepted options follows.
15011 Set the expression for the x and y coordinates of the overlaid video
15012 on the main video. Default value is "0" for both expressions. In case
15013 the expression is invalid, it is set to a huge value (meaning that the
15014 overlay will not be displayed within the output visible area).
15017 See @ref{framesync}.
15020 Set when the expressions for @option{x}, and @option{y} are evaluated.
15022 It accepts the following values:
15025 only evaluate expressions once during the filter initialization or
15026 when a command is processed
15029 evaluate expressions for each incoming frame
15032 Default value is @samp{frame}.
15035 See @ref{framesync}.
15038 Set the format for the output video.
15040 It accepts the following values:
15043 force YUV420 output
15046 force YUV420p10 output
15049 force YUV422 output
15052 force YUV422p10 output
15055 force YUV444 output
15058 force packed RGB output
15061 force planar RGB output
15064 automatically pick format
15067 Default value is @samp{yuv420}.
15070 See @ref{framesync}.
15073 Set format of alpha of the overlaid video, it can be @var{straight} or
15074 @var{premultiplied}. Default is @var{straight}.
15077 The @option{x}, and @option{y} expressions can contain the following
15083 The main input width and height.
15087 The overlay input width and height.
15091 The computed values for @var{x} and @var{y}. They are evaluated for
15096 horizontal and vertical chroma subsample values of the output
15097 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
15101 the number of input frame, starting from 0
15104 the position in the file of the input frame, NAN if unknown
15107 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
15111 This filter also supports the @ref{framesync} options.
15113 Note that the @var{n}, @var{pos}, @var{t} variables are available only
15114 when evaluation is done @emph{per frame}, and will evaluate to NAN
15115 when @option{eval} is set to @samp{init}.
15117 Be aware that frames are taken from each input video in timestamp
15118 order, hence, if their initial timestamps differ, it is a good idea
15119 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
15120 have them begin in the same zero timestamp, as the example for
15121 the @var{movie} filter does.
15123 You can chain together more overlays but you should test the
15124 efficiency of such approach.
15126 @subsection Commands
15128 This filter supports the following commands:
15132 Modify the x and y of the overlay input.
15133 The command accepts the same syntax of the corresponding option.
15135 If the specified expression is not valid, it is kept at its current
15139 @subsection Examples
15143 Draw the overlay at 10 pixels from the bottom right corner of the main
15146 overlay=main_w-overlay_w-10:main_h-overlay_h-10
15149 Using named options the example above becomes:
15151 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
15155 Insert a transparent PNG logo in the bottom left corner of the input,
15156 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
15158 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
15162 Insert 2 different transparent PNG logos (second logo on bottom
15163 right corner) using the @command{ffmpeg} tool:
15165 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
15169 Add a transparent color layer on top of the main video; @code{WxH}
15170 must specify the size of the main input to the overlay filter:
15172 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
15176 Play an original video and a filtered version (here with the deshake
15177 filter) side by side using the @command{ffplay} tool:
15179 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
15182 The above command is the same as:
15184 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
15188 Make a sliding overlay appearing from the left to the right top part of the
15189 screen starting since time 2:
15191 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
15195 Compose output by putting two input videos side to side:
15197 ffmpeg -i left.avi -i right.avi -filter_complex "
15198 nullsrc=size=200x100 [background];
15199 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
15200 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
15201 [background][left] overlay=shortest=1 [background+left];
15202 [background+left][right] overlay=shortest=1:x=100 [left+right]
15207 Mask 10-20 seconds of a video by applying the delogo filter to a section
15209 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
15210 -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]'
15215 Chain several overlays in cascade:
15217 nullsrc=s=200x200 [bg];
15218 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
15219 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
15220 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
15221 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
15222 [in3] null, [mid2] overlay=100:100 [out0]
15227 @anchor{overlay_cuda}
15228 @section overlay_cuda
15230 Overlay one video on top of another.
15232 This is the CUDA variant of the @ref{overlay} filter.
15233 It only accepts CUDA frames. The underlying input pixel formats have to match.
15235 It takes two inputs and has one output. The first input is the "main"
15236 video on which the second input is overlaid.
15238 It accepts the following parameters:
15243 Set the x and y coordinates of the overlaid video on the main video.
15244 Default value is "0" for both expressions.
15247 See @ref{framesync}.
15250 See @ref{framesync}.
15253 See @ref{framesync}.
15257 This filter also supports the @ref{framesync} options.
15261 Apply Overcomplete Wavelet denoiser.
15263 The filter accepts the following options:
15269 Larger depth values will denoise lower frequency components more, but
15270 slow down filtering.
15272 Must be an int in the range 8-16, default is @code{8}.
15274 @item luma_strength, ls
15277 Must be a double value in the range 0-1000, default is @code{1.0}.
15279 @item chroma_strength, cs
15280 Set chroma strength.
15282 Must be a double value in the range 0-1000, default is @code{1.0}.
15288 Add paddings to the input image, and place the original input at the
15289 provided @var{x}, @var{y} coordinates.
15291 It accepts the following parameters:
15296 Specify an expression for the size of the output image with the
15297 paddings added. If the value for @var{width} or @var{height} is 0, the
15298 corresponding input size is used for the output.
15300 The @var{width} expression can reference the value set by the
15301 @var{height} expression, and vice versa.
15303 The default value of @var{width} and @var{height} is 0.
15307 Specify the offsets to place the input image at within the padded area,
15308 with respect to the top/left border of the output image.
15310 The @var{x} expression can reference the value set by the @var{y}
15311 expression, and vice versa.
15313 The default value of @var{x} and @var{y} is 0.
15315 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
15316 so the input image is centered on the padded area.
15319 Specify the color of the padded area. For the syntax of this option,
15320 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
15321 manual,ffmpeg-utils}.
15323 The default value of @var{color} is "black".
15326 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
15328 It accepts the following values:
15332 Only evaluate expressions once during the filter initialization or when
15333 a command is processed.
15336 Evaluate expressions for each incoming frame.
15340 Default value is @samp{init}.
15343 Pad to aspect instead to a resolution.
15347 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
15348 options are expressions containing the following constants:
15353 The input video width and height.
15357 These are the same as @var{in_w} and @var{in_h}.
15361 The output width and height (the size of the padded area), as
15362 specified by the @var{width} and @var{height} expressions.
15366 These are the same as @var{out_w} and @var{out_h}.
15370 The x and y offsets as specified by the @var{x} and @var{y}
15371 expressions, or NAN if not yet specified.
15374 same as @var{iw} / @var{ih}
15377 input sample aspect ratio
15380 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
15384 The horizontal and vertical chroma subsample values. For example for the
15385 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15388 @subsection Examples
15392 Add paddings with the color "violet" to the input video. The output video
15393 size is 640x480, and the top-left corner of the input video is placed at
15396 pad=640:480:0:40:violet
15399 The example above is equivalent to the following command:
15401 pad=width=640:height=480:x=0:y=40:color=violet
15405 Pad the input to get an output with dimensions increased by 3/2,
15406 and put the input video at the center of the padded area:
15408 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
15412 Pad the input to get a squared output with size equal to the maximum
15413 value between the input width and height, and put the input video at
15414 the center of the padded area:
15416 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
15420 Pad the input to get a final w/h ratio of 16:9:
15422 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
15426 In case of anamorphic video, in order to set the output display aspect
15427 correctly, it is necessary to use @var{sar} in the expression,
15428 according to the relation:
15430 (ih * X / ih) * sar = output_dar
15431 X = output_dar / sar
15434 Thus the previous example needs to be modified to:
15436 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
15440 Double the output size and put the input video in the bottom-right
15441 corner of the output padded area:
15443 pad="2*iw:2*ih:ow-iw:oh-ih"
15447 @anchor{palettegen}
15448 @section palettegen
15450 Generate one palette for a whole video stream.
15452 It accepts the following options:
15456 Set the maximum number of colors to quantize in the palette.
15457 Note: the palette will still contain 256 colors; the unused palette entries
15460 @item reserve_transparent
15461 Create a palette of 255 colors maximum and reserve the last one for
15462 transparency. Reserving the transparency color is useful for GIF optimization.
15463 If not set, the maximum of colors in the palette will be 256. You probably want
15464 to disable this option for a standalone image.
15467 @item transparency_color
15468 Set the color that will be used as background for transparency.
15471 Set statistics mode.
15473 It accepts the following values:
15476 Compute full frame histograms.
15478 Compute histograms only for the part that differs from previous frame. This
15479 might be relevant to give more importance to the moving part of your input if
15480 the background is static.
15482 Compute new histogram for each frame.
15485 Default value is @var{full}.
15488 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
15489 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
15490 color quantization of the palette. This information is also visible at
15491 @var{info} logging level.
15493 @subsection Examples
15497 Generate a representative palette of a given video using @command{ffmpeg}:
15499 ffmpeg -i input.mkv -vf palettegen palette.png
15503 @section paletteuse
15505 Use a palette to downsample an input video stream.
15507 The filter takes two inputs: one video stream and a palette. The palette must
15508 be a 256 pixels image.
15510 It accepts the following options:
15514 Select dithering mode. Available algorithms are:
15517 Ordered 8x8 bayer dithering (deterministic)
15519 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
15520 Note: this dithering is sometimes considered "wrong" and is included as a
15522 @item floyd_steinberg
15523 Floyd and Steingberg dithering (error diffusion)
15525 Frankie Sierra dithering v2 (error diffusion)
15527 Frankie Sierra dithering v2 "Lite" (error diffusion)
15530 Default is @var{sierra2_4a}.
15533 When @var{bayer} dithering is selected, this option defines the scale of the
15534 pattern (how much the crosshatch pattern is visible). A low value means more
15535 visible pattern for less banding, and higher value means less visible pattern
15536 at the cost of more banding.
15538 The option must be an integer value in the range [0,5]. Default is @var{2}.
15541 If set, define the zone to process
15545 Only the changing rectangle will be reprocessed. This is similar to GIF
15546 cropping/offsetting compression mechanism. This option can be useful for speed
15547 if only a part of the image is changing, and has use cases such as limiting the
15548 scope of the error diffusal @option{dither} to the rectangle that bounds the
15549 moving scene (it leads to more deterministic output if the scene doesn't change
15550 much, and as a result less moving noise and better GIF compression).
15553 Default is @var{none}.
15556 Take new palette for each output frame.
15558 @item alpha_threshold
15559 Sets the alpha threshold for transparency. Alpha values above this threshold
15560 will be treated as completely opaque, and values below this threshold will be
15561 treated as completely transparent.
15563 The option must be an integer value in the range [0,255]. Default is @var{128}.
15566 @subsection Examples
15570 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
15571 using @command{ffmpeg}:
15573 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
15577 @section perspective
15579 Correct perspective of video not recorded perpendicular to the screen.
15581 A description of the accepted parameters follows.
15592 Set coordinates expression for top left, top right, bottom left and bottom right corners.
15593 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
15594 If the @code{sense} option is set to @code{source}, then the specified points will be sent
15595 to the corners of the destination. If the @code{sense} option is set to @code{destination},
15596 then the corners of the source will be sent to the specified coordinates.
15598 The expressions can use the following variables:
15603 the width and height of video frame.
15607 Output frame count.
15610 @item interpolation
15611 Set interpolation for perspective correction.
15613 It accepts the following values:
15619 Default value is @samp{linear}.
15622 Set interpretation of coordinate options.
15624 It accepts the following values:
15628 Send point in the source specified by the given coordinates to
15629 the corners of the destination.
15631 @item 1, destination
15633 Send the corners of the source to the point in the destination specified
15634 by the given coordinates.
15636 Default value is @samp{source}.
15640 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
15642 It accepts the following values:
15645 only evaluate expressions once during the filter initialization or
15646 when a command is processed
15649 evaluate expressions for each incoming frame
15652 Default value is @samp{init}.
15657 Delay interlaced video by one field time so that the field order changes.
15659 The intended use is to fix PAL movies that have been captured with the
15660 opposite field order to the film-to-video transfer.
15662 A description of the accepted parameters follows.
15668 It accepts the following values:
15671 Capture field order top-first, transfer bottom-first.
15672 Filter will delay the bottom field.
15675 Capture field order bottom-first, transfer top-first.
15676 Filter will delay the top field.
15679 Capture and transfer with the same field order. This mode only exists
15680 for the documentation of the other options to refer to, but if you
15681 actually select it, the filter will faithfully do nothing.
15684 Capture field order determined automatically by field flags, transfer
15686 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
15687 basis using field flags. If no field information is available,
15688 then this works just like @samp{u}.
15691 Capture unknown or varying, transfer opposite.
15692 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
15693 analyzing the images and selecting the alternative that produces best
15694 match between the fields.
15697 Capture top-first, transfer unknown or varying.
15698 Filter selects among @samp{t} and @samp{p} using image analysis.
15701 Capture bottom-first, transfer unknown or varying.
15702 Filter selects among @samp{b} and @samp{p} using image analysis.
15705 Capture determined by field flags, transfer unknown or varying.
15706 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
15707 image analysis. If no field information is available, then this works just
15708 like @samp{U}. This is the default mode.
15711 Both capture and transfer unknown or varying.
15712 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
15716 @section photosensitivity
15717 Reduce various flashes in video, so to help users with epilepsy.
15719 It accepts the following options:
15722 Set how many frames to use when filtering. Default is 30.
15725 Set detection threshold factor. Default is 1.
15729 Set how many pixels to skip when sampling frames. Default is 1.
15730 Allowed range is from 1 to 1024.
15733 Leave frames unchanged. Default is disabled.
15736 @section pixdesctest
15738 Pixel format descriptor test filter, mainly useful for internal
15739 testing. The output video should be equal to the input video.
15743 format=monow, pixdesctest
15746 can be used to test the monowhite pixel format descriptor definition.
15750 Display sample values of color channels. Mainly useful for checking color
15751 and levels. Minimum supported resolution is 640x480.
15753 The filters accept the following options:
15757 Set scope X position, relative offset on X axis.
15760 Set scope Y position, relative offset on Y axis.
15769 Set window opacity. This window also holds statistics about pixel area.
15772 Set window X position, relative offset on X axis.
15775 Set window Y position, relative offset on Y axis.
15780 Enable the specified chain of postprocessing subfilters using libpostproc. This
15781 library should be automatically selected with a GPL build (@code{--enable-gpl}).
15782 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
15783 Each subfilter and some options have a short and a long name that can be used
15784 interchangeably, i.e. dr/dering are the same.
15786 The filters accept the following options:
15790 Set postprocessing subfilters string.
15793 All subfilters share common options to determine their scope:
15797 Honor the quality commands for this subfilter.
15800 Do chrominance filtering, too (default).
15803 Do luminance filtering only (no chrominance).
15806 Do chrominance filtering only (no luminance).
15809 These options can be appended after the subfilter name, separated by a '|'.
15811 Available subfilters are:
15814 @item hb/hdeblock[|difference[|flatness]]
15815 Horizontal deblocking filter
15818 Difference factor where higher values mean more deblocking (default: @code{32}).
15820 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15823 @item vb/vdeblock[|difference[|flatness]]
15824 Vertical deblocking filter
15827 Difference factor where higher values mean more deblocking (default: @code{32}).
15829 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15832 @item ha/hadeblock[|difference[|flatness]]
15833 Accurate horizontal deblocking filter
15836 Difference factor where higher values mean more deblocking (default: @code{32}).
15838 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15841 @item va/vadeblock[|difference[|flatness]]
15842 Accurate vertical deblocking filter
15845 Difference factor where higher values mean more deblocking (default: @code{32}).
15847 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15851 The horizontal and vertical deblocking filters share the difference and
15852 flatness values so you cannot set different horizontal and vertical
15856 @item h1/x1hdeblock
15857 Experimental horizontal deblocking filter
15859 @item v1/x1vdeblock
15860 Experimental vertical deblocking filter
15865 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
15868 larger -> stronger filtering
15870 larger -> stronger filtering
15872 larger -> stronger filtering
15875 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
15878 Stretch luminance to @code{0-255}.
15881 @item lb/linblenddeint
15882 Linear blend deinterlacing filter that deinterlaces the given block by
15883 filtering all lines with a @code{(1 2 1)} filter.
15885 @item li/linipoldeint
15886 Linear interpolating deinterlacing filter that deinterlaces the given block by
15887 linearly interpolating every second line.
15889 @item ci/cubicipoldeint
15890 Cubic interpolating deinterlacing filter deinterlaces the given block by
15891 cubically interpolating every second line.
15893 @item md/mediandeint
15894 Median deinterlacing filter that deinterlaces the given block by applying a
15895 median filter to every second line.
15897 @item fd/ffmpegdeint
15898 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
15899 second line with a @code{(-1 4 2 4 -1)} filter.
15902 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
15903 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
15905 @item fq/forceQuant[|quantizer]
15906 Overrides the quantizer table from the input with the constant quantizer you
15914 Default pp filter combination (@code{hb|a,vb|a,dr|a})
15917 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
15920 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
15923 @subsection Examples
15927 Apply horizontal and vertical deblocking, deringing and automatic
15928 brightness/contrast:
15934 Apply default filters without brightness/contrast correction:
15940 Apply default filters and temporal denoiser:
15942 pp=default/tmpnoise|1|2|3
15946 Apply deblocking on luminance only, and switch vertical deblocking on or off
15947 automatically depending on available CPU time:
15954 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
15955 similar to spp = 6 with 7 point DCT, where only the center sample is
15958 The filter accepts the following options:
15962 Force a constant quantization parameter. It accepts an integer in range
15963 0 to 63. If not set, the filter will use the QP from the video stream
15967 Set thresholding mode. Available modes are:
15971 Set hard thresholding.
15973 Set soft thresholding (better de-ringing effect, but likely blurrier).
15975 Set medium thresholding (good results, default).
15979 @section premultiply
15980 Apply alpha premultiply effect to input video stream using first plane
15981 of second stream as alpha.
15983 Both streams must have same dimensions and same pixel format.
15985 The filter accepts the following option:
15989 Set which planes will be processed, unprocessed planes will be copied.
15990 By default value 0xf, all planes will be processed.
15993 Do not require 2nd input for processing, instead use alpha plane from input stream.
15997 Apply prewitt operator to input video stream.
15999 The filter accepts the following option:
16003 Set which planes will be processed, unprocessed planes will be copied.
16004 By default value 0xf, all planes will be processed.
16007 Set value which will be multiplied with filtered result.
16010 Set value which will be added to filtered result.
16013 @subsection Commands
16015 This filter supports the all above options as @ref{commands}.
16017 @section pseudocolor
16019 Alter frame colors in video with pseudocolors.
16021 This filter accepts the following options:
16025 set pixel first component expression
16028 set pixel second component expression
16031 set pixel third component expression
16034 set pixel fourth component expression, corresponds to the alpha component
16037 set component to use as base for altering colors
16040 Each of them specifies the expression to use for computing the lookup table for
16041 the corresponding pixel component values.
16043 The expressions can contain the following constants and functions:
16048 The input width and height.
16051 The input value for the pixel component.
16053 @item ymin, umin, vmin, amin
16054 The minimum allowed component value.
16056 @item ymax, umax, vmax, amax
16057 The maximum allowed component value.
16060 All expressions default to "val".
16062 @subsection Examples
16066 Change too high luma values to gradient:
16068 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'"
16074 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
16075 Ratio) between two input videos.
16077 This filter takes in input two input videos, the first input is
16078 considered the "main" source and is passed unchanged to the
16079 output. The second input is used as a "reference" video for computing
16082 Both video inputs must have the same resolution and pixel format for
16083 this filter to work correctly. Also it assumes that both inputs
16084 have the same number of frames, which are compared one by one.
16086 The obtained average PSNR is printed through the logging system.
16088 The filter stores the accumulated MSE (mean squared error) of each
16089 frame, and at the end of the processing it is averaged across all frames
16090 equally, and the following formula is applied to obtain the PSNR:
16093 PSNR = 10*log10(MAX^2/MSE)
16096 Where MAX is the average of the maximum values of each component of the
16099 The description of the accepted parameters follows.
16102 @item stats_file, f
16103 If specified the filter will use the named file to save the PSNR of
16104 each individual frame. When filename equals "-" the data is sent to
16107 @item stats_version
16108 Specifies which version of the stats file format to use. Details of
16109 each format are written below.
16110 Default value is 1.
16112 @item stats_add_max
16113 Determines whether the max value is output to the stats log.
16114 Default value is 0.
16115 Requires stats_version >= 2. If this is set and stats_version < 2,
16116 the filter will return an error.
16119 This filter also supports the @ref{framesync} options.
16121 The file printed if @var{stats_file} is selected, contains a sequence of
16122 key/value pairs of the form @var{key}:@var{value} for each compared
16125 If a @var{stats_version} greater than 1 is specified, a header line precedes
16126 the list of per-frame-pair stats, with key value pairs following the frame
16127 format with the following parameters:
16130 @item psnr_log_version
16131 The version of the log file format. Will match @var{stats_version}.
16134 A comma separated list of the per-frame-pair parameters included in
16138 A description of each shown per-frame-pair parameter follows:
16142 sequential number of the input frame, starting from 1
16145 Mean Square Error pixel-by-pixel average difference of the compared
16146 frames, averaged over all the image components.
16148 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
16149 Mean Square Error pixel-by-pixel average difference of the compared
16150 frames for the component specified by the suffix.
16152 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
16153 Peak Signal to Noise ratio of the compared frames for the component
16154 specified by the suffix.
16156 @item max_avg, max_y, max_u, max_v
16157 Maximum allowed value for each channel, and average over all
16161 @subsection Examples
16166 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
16167 [main][ref] psnr="stats_file=stats.log" [out]
16170 On this example the input file being processed is compared with the
16171 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
16172 is stored in @file{stats.log}.
16175 Another example with different containers:
16177 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 -
16184 Pulldown reversal (inverse telecine) filter, capable of handling mixed
16185 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
16188 The pullup filter is designed to take advantage of future context in making
16189 its decisions. This filter is stateless in the sense that it does not lock
16190 onto a pattern to follow, but it instead looks forward to the following
16191 fields in order to identify matches and rebuild progressive frames.
16193 To produce content with an even framerate, insert the fps filter after
16194 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
16195 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
16197 The filter accepts the following options:
16204 These options set the amount of "junk" to ignore at the left, right, top, and
16205 bottom of the image, respectively. Left and right are in units of 8 pixels,
16206 while top and bottom are in units of 2 lines.
16207 The default is 8 pixels on each side.
16210 Set the strict breaks. Setting this option to 1 will reduce the chances of
16211 filter generating an occasional mismatched frame, but it may also cause an
16212 excessive number of frames to be dropped during high motion sequences.
16213 Conversely, setting it to -1 will make filter match fields more easily.
16214 This may help processing of video where there is slight blurring between
16215 the fields, but may also cause there to be interlaced frames in the output.
16216 Default value is @code{0}.
16219 Set the metric plane to use. It accepts the following values:
16225 Use chroma blue plane.
16228 Use chroma red plane.
16231 This option may be set to use chroma plane instead of the default luma plane
16232 for doing filter's computations. This may improve accuracy on very clean
16233 source material, but more likely will decrease accuracy, especially if there
16234 is chroma noise (rainbow effect) or any grayscale video.
16235 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
16236 load and make pullup usable in realtime on slow machines.
16239 For best results (without duplicated frames in the output file) it is
16240 necessary to change the output frame rate. For example, to inverse
16241 telecine NTSC input:
16243 ffmpeg -i input -vf pullup -r 24000/1001 ...
16248 Change video quantization parameters (QP).
16250 The filter accepts the following option:
16254 Set expression for quantization parameter.
16257 The expression is evaluated through the eval API and can contain, among others,
16258 the following constants:
16262 1 if index is not 129, 0 otherwise.
16265 Sequential index starting from -129 to 128.
16268 @subsection Examples
16272 Some equation like:
16280 Flush video frames from internal cache of frames into a random order.
16281 No frame is discarded.
16282 Inspired by @ref{frei0r} nervous filter.
16286 Set size in number of frames of internal cache, in range from @code{2} to
16287 @code{512}. Default is @code{30}.
16290 Set seed for random number generator, must be an integer included between
16291 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
16292 less than @code{0}, the filter will try to use a good random seed on a
16296 @section readeia608
16298 Read closed captioning (EIA-608) information from the top lines of a video frame.
16300 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
16301 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
16302 with EIA-608 data (starting from 0). A description of each metadata value follows:
16305 @item lavfi.readeia608.X.cc
16306 The two bytes stored as EIA-608 data (printed in hexadecimal).
16308 @item lavfi.readeia608.X.line
16309 The number of the line on which the EIA-608 data was identified and read.
16312 This filter accepts the following options:
16316 Set the line to start scanning for EIA-608 data. Default is @code{0}.
16319 Set the line to end scanning for EIA-608 data. Default is @code{29}.
16322 Set the ratio of width reserved for sync code detection.
16323 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
16326 Enable checking the parity bit. In the event of a parity error, the filter will output
16327 @code{0x00} for that character. Default is false.
16330 Lowpass lines prior to further processing. Default is enabled.
16333 @subsection Commands
16335 This filter supports the all above options as @ref{commands}.
16337 @subsection Examples
16341 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
16343 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
16349 Read vertical interval timecode (VITC) information from the top lines of a
16352 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
16353 timecode value, if a valid timecode has been detected. Further metadata key
16354 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
16355 timecode data has been found or not.
16357 This filter accepts the following options:
16361 Set the maximum number of lines to scan for VITC data. If the value is set to
16362 @code{-1} the full video frame is scanned. Default is @code{45}.
16365 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
16366 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
16369 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
16370 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
16373 @subsection Examples
16377 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
16378 draw @code{--:--:--:--} as a placeholder:
16380 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
16386 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
16388 Destination pixel at position (X, Y) will be picked from source (x, y) position
16389 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
16390 value for pixel will be used for destination pixel.
16392 Xmap and Ymap input video streams must be of same dimensions. Output video stream
16393 will have Xmap/Ymap video stream dimensions.
16394 Xmap and Ymap input video streams are 16bit depth, single channel.
16398 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
16399 Default is @code{color}.
16402 Specify the color of the unmapped pixels. For the syntax of this option,
16403 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
16404 manual,ffmpeg-utils}. Default color is @code{black}.
16407 @section removegrain
16409 The removegrain filter is a spatial denoiser for progressive video.
16413 Set mode for the first plane.
16416 Set mode for the second plane.
16419 Set mode for the third plane.
16422 Set mode for the fourth plane.
16425 Range of mode is from 0 to 24. Description of each mode follows:
16429 Leave input plane unchanged. Default.
16432 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
16435 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
16438 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
16441 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
16442 This is equivalent to a median filter.
16445 Line-sensitive clipping giving the minimal change.
16448 Line-sensitive clipping, intermediate.
16451 Line-sensitive clipping, intermediate.
16454 Line-sensitive clipping, intermediate.
16457 Line-sensitive clipping on a line where the neighbours pixels are the closest.
16460 Replaces the target pixel with the closest neighbour.
16463 [1 2 1] horizontal and vertical kernel blur.
16469 Bob mode, interpolates top field from the line where the neighbours
16470 pixels are the closest.
16473 Bob mode, interpolates bottom field from the line where the neighbours
16474 pixels are the closest.
16477 Bob mode, interpolates top field. Same as 13 but with a more complicated
16478 interpolation formula.
16481 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
16482 interpolation formula.
16485 Clips the pixel with the minimum and maximum of respectively the maximum and
16486 minimum of each pair of opposite neighbour pixels.
16489 Line-sensitive clipping using opposite neighbours whose greatest distance from
16490 the current pixel is minimal.
16493 Replaces the pixel with the average of its 8 neighbours.
16496 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
16499 Clips pixels using the averages of opposite neighbour.
16502 Same as mode 21 but simpler and faster.
16505 Small edge and halo removal, but reputed useless.
16511 @section removelogo
16513 Suppress a TV station logo, using an image file to determine which
16514 pixels comprise the logo. It works by filling in the pixels that
16515 comprise the logo with neighboring pixels.
16517 The filter accepts the following options:
16521 Set the filter bitmap file, which can be any image format supported by
16522 libavformat. The width and height of the image file must match those of the
16523 video stream being processed.
16526 Pixels in the provided bitmap image with a value of zero are not
16527 considered part of the logo, non-zero pixels are considered part of
16528 the logo. If you use white (255) for the logo and black (0) for the
16529 rest, you will be safe. For making the filter bitmap, it is
16530 recommended to take a screen capture of a black frame with the logo
16531 visible, and then using a threshold filter followed by the erode
16532 filter once or twice.
16534 If needed, little splotches can be fixed manually. Remember that if
16535 logo pixels are not covered, the filter quality will be much
16536 reduced. Marking too many pixels as part of the logo does not hurt as
16537 much, but it will increase the amount of blurring needed to cover over
16538 the image and will destroy more information than necessary, and extra
16539 pixels will slow things down on a large logo.
16541 @section repeatfields
16543 This filter uses the repeat_field flag from the Video ES headers and hard repeats
16544 fields based on its value.
16548 Reverse a video clip.
16550 Warning: This filter requires memory to buffer the entire clip, so trimming
16553 @subsection Examples
16557 Take the first 5 seconds of a clip, and reverse it.
16564 Shift R/G/B/A pixels horizontally and/or vertically.
16566 The filter accepts the following options:
16569 Set amount to shift red horizontally.
16571 Set amount to shift red vertically.
16573 Set amount to shift green horizontally.
16575 Set amount to shift green vertically.
16577 Set amount to shift blue horizontally.
16579 Set amount to shift blue vertically.
16581 Set amount to shift alpha horizontally.
16583 Set amount to shift alpha vertically.
16585 Set edge mode, can be @var{smear}, default, or @var{warp}.
16588 @subsection Commands
16590 This filter supports the all above options as @ref{commands}.
16593 Apply roberts cross operator to input video stream.
16595 The filter accepts the following option:
16599 Set which planes will be processed, unprocessed planes will be copied.
16600 By default value 0xf, all planes will be processed.
16603 Set value which will be multiplied with filtered result.
16606 Set value which will be added to filtered result.
16609 @subsection Commands
16611 This filter supports the all above options as @ref{commands}.
16615 Rotate video by an arbitrary angle expressed in radians.
16617 The filter accepts the following options:
16619 A description of the optional parameters follows.
16622 Set an expression for the angle by which to rotate the input video
16623 clockwise, expressed as a number of radians. A negative value will
16624 result in a counter-clockwise rotation. By default it is set to "0".
16626 This expression is evaluated for each frame.
16629 Set the output width expression, default value is "iw".
16630 This expression is evaluated just once during configuration.
16633 Set the output height expression, default value is "ih".
16634 This expression is evaluated just once during configuration.
16637 Enable bilinear interpolation if set to 1, a value of 0 disables
16638 it. Default value is 1.
16641 Set the color used to fill the output area not covered by the rotated
16642 image. For the general syntax of this option, check the
16643 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
16644 If the special value "none" is selected then no
16645 background is printed (useful for example if the background is never shown).
16647 Default value is "black".
16650 The expressions for the angle and the output size can contain the
16651 following constants and functions:
16655 sequential number of the input frame, starting from 0. It is always NAN
16656 before the first frame is filtered.
16659 time in seconds of the input frame, it is set to 0 when the filter is
16660 configured. It is always NAN before the first frame is filtered.
16664 horizontal and vertical chroma subsample values. For example for the
16665 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16669 the input video width and height
16673 the output width and height, that is the size of the padded area as
16674 specified by the @var{width} and @var{height} expressions
16678 the minimal width/height required for completely containing the input
16679 video rotated by @var{a} radians.
16681 These are only available when computing the @option{out_w} and
16682 @option{out_h} expressions.
16685 @subsection Examples
16689 Rotate the input by PI/6 radians clockwise:
16695 Rotate the input by PI/6 radians counter-clockwise:
16701 Rotate the input by 45 degrees clockwise:
16707 Apply a constant rotation with period T, starting from an angle of PI/3:
16709 rotate=PI/3+2*PI*t/T
16713 Make the input video rotation oscillating with a period of T
16714 seconds and an amplitude of A radians:
16716 rotate=A*sin(2*PI/T*t)
16720 Rotate the video, output size is chosen so that the whole rotating
16721 input video is always completely contained in the output:
16723 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
16727 Rotate the video, reduce the output size so that no background is ever
16730 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
16734 @subsection Commands
16736 The filter supports the following commands:
16740 Set the angle expression.
16741 The command accepts the same syntax of the corresponding option.
16743 If the specified expression is not valid, it is kept at its current
16749 Apply Shape Adaptive Blur.
16751 The filter accepts the following options:
16754 @item luma_radius, lr
16755 Set luma blur filter strength, must be a value in range 0.1-4.0, default
16756 value is 1.0. A greater value will result in a more blurred image, and
16757 in slower processing.
16759 @item luma_pre_filter_radius, lpfr
16760 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
16763 @item luma_strength, ls
16764 Set luma maximum difference between pixels to still be considered, must
16765 be a value in the 0.1-100.0 range, default value is 1.0.
16767 @item chroma_radius, cr
16768 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
16769 greater value will result in a more blurred image, and in slower
16772 @item chroma_pre_filter_radius, cpfr
16773 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
16775 @item chroma_strength, cs
16776 Set chroma maximum difference between pixels to still be considered,
16777 must be a value in the -0.9-100.0 range.
16780 Each chroma option value, if not explicitly specified, is set to the
16781 corresponding luma option value.
16786 Scale (resize) the input video, using the libswscale library.
16788 The scale filter forces the output display aspect ratio to be the same
16789 of the input, by changing the output sample aspect ratio.
16791 If the input image format is different from the format requested by
16792 the next filter, the scale filter will convert the input to the
16795 @subsection Options
16796 The filter accepts the following options, or any of the options
16797 supported by the libswscale scaler.
16799 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
16800 the complete list of scaler options.
16805 Set the output video dimension expression. Default value is the input
16808 If the @var{width} or @var{w} value is 0, the input width is used for
16809 the output. If the @var{height} or @var{h} value is 0, the input height
16810 is used for the output.
16812 If one and only one of the values is -n with n >= 1, the scale filter
16813 will use a value that maintains the aspect ratio of the input image,
16814 calculated from the other specified dimension. After that it will,
16815 however, make sure that the calculated dimension is divisible by n and
16816 adjust the value if necessary.
16818 If both values are -n with n >= 1, the behavior will be identical to
16819 both values being set to 0 as previously detailed.
16821 See below for the list of accepted constants for use in the dimension
16825 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
16829 Only evaluate expressions once during the filter initialization or when a command is processed.
16832 Evaluate expressions for each incoming frame.
16836 Default value is @samp{init}.
16840 Set the interlacing mode. It accepts the following values:
16844 Force interlaced aware scaling.
16847 Do not apply interlaced scaling.
16850 Select interlaced aware scaling depending on whether the source frames
16851 are flagged as interlaced or not.
16854 Default value is @samp{0}.
16857 Set libswscale scaling flags. See
16858 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16859 complete list of values. If not explicitly specified the filter applies
16863 @item param0, param1
16864 Set libswscale input parameters for scaling algorithms that need them. See
16865 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16866 complete documentation. If not explicitly specified the filter applies
16872 Set the video size. For the syntax of this option, check the
16873 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16875 @item in_color_matrix
16876 @item out_color_matrix
16877 Set in/output YCbCr color space type.
16879 This allows the autodetected value to be overridden as well as allows forcing
16880 a specific value used for the output and encoder.
16882 If not specified, the color space type depends on the pixel format.
16888 Choose automatically.
16891 Format conforming to International Telecommunication Union (ITU)
16892 Recommendation BT.709.
16895 Set color space conforming to the United States Federal Communications
16896 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
16901 Set color space conforming to:
16905 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
16908 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
16911 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
16916 Set color space conforming to SMPTE ST 240:1999.
16919 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
16924 Set in/output YCbCr sample range.
16926 This allows the autodetected value to be overridden as well as allows forcing
16927 a specific value used for the output and encoder. If not specified, the
16928 range depends on the pixel format. Possible values:
16932 Choose automatically.
16935 Set full range (0-255 in case of 8-bit luma).
16937 @item mpeg/limited/tv
16938 Set "MPEG" range (16-235 in case of 8-bit luma).
16941 @item force_original_aspect_ratio
16942 Enable decreasing or increasing output video width or height if necessary to
16943 keep the original aspect ratio. Possible values:
16947 Scale the video as specified and disable this feature.
16950 The output video dimensions will automatically be decreased if needed.
16953 The output video dimensions will automatically be increased if needed.
16957 One useful instance of this option is that when you know a specific device's
16958 maximum allowed resolution, you can use this to limit the output video to
16959 that, while retaining the aspect ratio. For example, device A allows
16960 1280x720 playback, and your video is 1920x800. Using this option (set it to
16961 decrease) and specifying 1280x720 to the command line makes the output
16964 Please note that this is a different thing than specifying -1 for @option{w}
16965 or @option{h}, you still need to specify the output resolution for this option
16968 @item force_divisible_by
16969 Ensures that both the output dimensions, width and height, are divisible by the
16970 given integer when used together with @option{force_original_aspect_ratio}. This
16971 works similar to using @code{-n} in the @option{w} and @option{h} options.
16973 This option respects the value set for @option{force_original_aspect_ratio},
16974 increasing or decreasing the resolution accordingly. The video's aspect ratio
16975 may be slightly modified.
16977 This option can be handy if you need to have a video fit within or exceed
16978 a defined resolution using @option{force_original_aspect_ratio} but also have
16979 encoder restrictions on width or height divisibility.
16983 The values of the @option{w} and @option{h} options are expressions
16984 containing the following constants:
16989 The input width and height
16993 These are the same as @var{in_w} and @var{in_h}.
16997 The output (scaled) width and height
17001 These are the same as @var{out_w} and @var{out_h}
17004 The same as @var{iw} / @var{ih}
17007 input sample aspect ratio
17010 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
17014 horizontal and vertical input chroma subsample values. For example for the
17015 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17019 horizontal and vertical output chroma subsample values. For example for the
17020 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17023 The (sequential) number of the input frame, starting from 0.
17024 Only available with @code{eval=frame}.
17027 The presentation timestamp of the input frame, expressed as a number of
17028 seconds. Only available with @code{eval=frame}.
17031 The position (byte offset) of the frame in the input stream, or NaN if
17032 this information is unavailable and/or meaningless (for example in case of synthetic video).
17033 Only available with @code{eval=frame}.
17036 @subsection Examples
17040 Scale the input video to a size of 200x100
17045 This is equivalent to:
17056 Specify a size abbreviation for the output size:
17061 which can also be written as:
17067 Scale the input to 2x:
17069 scale=w=2*iw:h=2*ih
17073 The above is the same as:
17075 scale=2*in_w:2*in_h
17079 Scale the input to 2x with forced interlaced scaling:
17081 scale=2*iw:2*ih:interl=1
17085 Scale the input to half size:
17087 scale=w=iw/2:h=ih/2
17091 Increase the width, and set the height to the same size:
17097 Seek Greek harmony:
17104 Increase the height, and set the width to 3/2 of the height:
17106 scale=w=3/2*oh:h=3/5*ih
17110 Increase the size, making the size a multiple of the chroma
17113 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
17117 Increase the width to a maximum of 500 pixels,
17118 keeping the same aspect ratio as the input:
17120 scale=w='min(500\, iw*3/2):h=-1'
17124 Make pixels square by combining scale and setsar:
17126 scale='trunc(ih*dar):ih',setsar=1/1
17130 Make pixels square by combining scale and setsar,
17131 making sure the resulting resolution is even (required by some codecs):
17133 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
17137 @subsection Commands
17139 This filter supports the following commands:
17143 Set the output video dimension expression.
17144 The command accepts the same syntax of the corresponding option.
17146 If the specified expression is not valid, it is kept at its current
17152 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
17153 format conversion on CUDA video frames. Setting the output width and height
17154 works in the same way as for the @var{scale} filter.
17156 The following additional options are accepted:
17159 The pixel format of the output CUDA frames. If set to the string "same" (the
17160 default), the input format will be kept. Note that automatic format negotiation
17161 and conversion is not yet supported for hardware frames
17164 The interpolation algorithm used for resizing. One of the following:
17171 @item cubic2p_bspline
17172 2-parameter cubic (B=1, C=0)
17174 @item cubic2p_catmullrom
17175 2-parameter cubic (B=0, C=1/2)
17177 @item cubic2p_b05c03
17178 2-parameter cubic (B=1/2, C=3/10)
17186 @item force_original_aspect_ratio
17187 Enable decreasing or increasing output video width or height if necessary to
17188 keep the original aspect ratio. Possible values:
17192 Scale the video as specified and disable this feature.
17195 The output video dimensions will automatically be decreased if needed.
17198 The output video dimensions will automatically be increased if needed.
17202 One useful instance of this option is that when you know a specific device's
17203 maximum allowed resolution, you can use this to limit the output video to
17204 that, while retaining the aspect ratio. For example, device A allows
17205 1280x720 playback, and your video is 1920x800. Using this option (set it to
17206 decrease) and specifying 1280x720 to the command line makes the output
17209 Please note that this is a different thing than specifying -1 for @option{w}
17210 or @option{h}, you still need to specify the output resolution for this option
17213 @item force_divisible_by
17214 Ensures that both the output dimensions, width and height, are divisible by the
17215 given integer when used together with @option{force_original_aspect_ratio}. This
17216 works similar to using @code{-n} in the @option{w} and @option{h} options.
17218 This option respects the value set for @option{force_original_aspect_ratio},
17219 increasing or decreasing the resolution accordingly. The video's aspect ratio
17220 may be slightly modified.
17222 This option can be handy if you need to have a video fit within or exceed
17223 a defined resolution using @option{force_original_aspect_ratio} but also have
17224 encoder restrictions on width or height divisibility.
17230 Scale (resize) the input video, based on a reference video.
17232 See the scale filter for available options, scale2ref supports the same but
17233 uses the reference video instead of the main input as basis. scale2ref also
17234 supports the following additional constants for the @option{w} and
17235 @option{h} options:
17240 The main input video's width and height
17243 The same as @var{main_w} / @var{main_h}
17246 The main input video's sample aspect ratio
17248 @item main_dar, mdar
17249 The main input video's display aspect ratio. Calculated from
17250 @code{(main_w / main_h) * main_sar}.
17254 The main input video's horizontal and vertical chroma subsample values.
17255 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
17259 The (sequential) number of the main input frame, starting from 0.
17260 Only available with @code{eval=frame}.
17263 The presentation timestamp of the main input frame, expressed as a number of
17264 seconds. Only available with @code{eval=frame}.
17267 The position (byte offset) of the frame in the main input stream, or NaN if
17268 this information is unavailable and/or meaningless (for example in case of synthetic video).
17269 Only available with @code{eval=frame}.
17272 @subsection Examples
17276 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
17278 'scale2ref[b][a];[a][b]overlay'
17282 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
17284 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
17288 @subsection Commands
17290 This filter supports the following commands:
17294 Set the output video dimension expression.
17295 The command accepts the same syntax of the corresponding option.
17297 If the specified expression is not valid, it is kept at its current
17302 Scroll input video horizontally and/or vertically by constant speed.
17304 The filter accepts the following options:
17306 @item horizontal, h
17307 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
17308 Negative values changes scrolling direction.
17311 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
17312 Negative values changes scrolling direction.
17315 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
17318 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
17321 @subsection Commands
17323 This filter supports the following @ref{commands}:
17325 @item horizontal, h
17326 Set the horizontal scrolling speed.
17328 Set the vertical scrolling speed.
17334 Detect video scene change.
17336 This filter sets frame metadata with mafd between frame, the scene score, and
17337 forward the frame to the next filter, so they can use these metadata to detect
17338 scene change or others.
17340 In addition, this filter logs a message and sets frame metadata when it detects
17341 a scene change by @option{threshold}.
17343 @code{lavfi.scd.mafd} metadata keys are set with mafd for every frame.
17345 @code{lavfi.scd.score} metadata keys are set with scene change score for every frame
17346 to detect scene change.
17348 @code{lavfi.scd.time} metadata keys are set with current filtered frame time which
17349 detect scene change with @option{threshold}.
17351 The filter accepts the following options:
17355 Set the scene change detection threshold as a percentage of maximum change. Good
17356 values are in the @code{[8.0, 14.0]} range. The range for @option{threshold} is
17359 Default value is @code{10.}.
17362 Set the flag to pass scene change frames to the next filter. Default value is @code{0}
17363 You can enable it if you want to get snapshot of scene change frames only.
17366 @anchor{selectivecolor}
17367 @section selectivecolor
17369 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
17370 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
17371 by the "purity" of the color (that is, how saturated it already is).
17373 This filter is similar to the Adobe Photoshop Selective Color tool.
17375 The filter accepts the following options:
17378 @item correction_method
17379 Select color correction method.
17381 Available values are:
17384 Specified adjustments are applied "as-is" (added/subtracted to original pixel
17387 Specified adjustments are relative to the original component value.
17389 Default is @code{absolute}.
17391 Adjustments for red pixels (pixels where the red component is the maximum)
17393 Adjustments for yellow pixels (pixels where the blue component is the minimum)
17395 Adjustments for green pixels (pixels where the green component is the maximum)
17397 Adjustments for cyan pixels (pixels where the red component is the minimum)
17399 Adjustments for blue pixels (pixels where the blue component is the maximum)
17401 Adjustments for magenta pixels (pixels where the green component is the minimum)
17403 Adjustments for white pixels (pixels where all components are greater than 128)
17405 Adjustments for all pixels except pure black and pure white
17407 Adjustments for black pixels (pixels where all components are lesser than 128)
17409 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
17412 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
17413 4 space separated floating point adjustment values in the [-1,1] range,
17414 respectively to adjust the amount of cyan, magenta, yellow and black for the
17415 pixels of its range.
17417 @subsection Examples
17421 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
17422 increase magenta by 27% in blue areas:
17424 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
17428 Use a Photoshop selective color preset:
17430 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
17434 @anchor{separatefields}
17435 @section separatefields
17437 The @code{separatefields} takes a frame-based video input and splits
17438 each frame into its components fields, producing a new half height clip
17439 with twice the frame rate and twice the frame count.
17441 This filter use field-dominance information in frame to decide which
17442 of each pair of fields to place first in the output.
17443 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
17445 @section setdar, setsar
17447 The @code{setdar} filter sets the Display Aspect Ratio for the filter
17450 This is done by changing the specified Sample (aka Pixel) Aspect
17451 Ratio, according to the following equation:
17453 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
17456 Keep in mind that the @code{setdar} filter does not modify the pixel
17457 dimensions of the video frame. Also, the display aspect ratio set by
17458 this filter may be changed by later filters in the filterchain,
17459 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
17462 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
17463 the filter output video.
17465 Note that as a consequence of the application of this filter, the
17466 output display aspect ratio will change according to the equation
17469 Keep in mind that the sample aspect ratio set by the @code{setsar}
17470 filter may be changed by later filters in the filterchain, e.g. if
17471 another "setsar" or a "setdar" filter is applied.
17473 It accepts the following parameters:
17476 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
17477 Set the aspect ratio used by the filter.
17479 The parameter can be a floating point number string, an expression, or
17480 a string of the form @var{num}:@var{den}, where @var{num} and
17481 @var{den} are the numerator and denominator of the aspect ratio. If
17482 the parameter is not specified, it is assumed the value "0".
17483 In case the form "@var{num}:@var{den}" is used, the @code{:} character
17487 Set the maximum integer value to use for expressing numerator and
17488 denominator when reducing the expressed aspect ratio to a rational.
17489 Default value is @code{100}.
17493 The parameter @var{sar} is an expression containing
17494 the following constants:
17498 These are approximated values for the mathematical constants e
17499 (Euler's number), pi (Greek pi), and phi (the golden ratio).
17502 The input width and height.
17505 These are the same as @var{w} / @var{h}.
17508 The input sample aspect ratio.
17511 The input display aspect ratio. It is the same as
17512 (@var{w} / @var{h}) * @var{sar}.
17515 Horizontal and vertical chroma subsample values. For example, for the
17516 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17519 @subsection Examples
17524 To change the display aspect ratio to 16:9, specify one of the following:
17531 To change the sample aspect ratio to 10:11, specify:
17537 To set a display aspect ratio of 16:9, and specify a maximum integer value of
17538 1000 in the aspect ratio reduction, use the command:
17540 setdar=ratio=16/9:max=1000
17548 Force field for the output video frame.
17550 The @code{setfield} filter marks the interlace type field for the
17551 output frames. It does not change the input frame, but only sets the
17552 corresponding property, which affects how the frame is treated by
17553 following filters (e.g. @code{fieldorder} or @code{yadif}).
17555 The filter accepts the following options:
17560 Available values are:
17564 Keep the same field property.
17567 Mark the frame as bottom-field-first.
17570 Mark the frame as top-field-first.
17573 Mark the frame as progressive.
17580 Force frame parameter for the output video frame.
17582 The @code{setparams} filter marks interlace and color range for the
17583 output frames. It does not change the input frame, but only sets the
17584 corresponding property, which affects how the frame is treated by
17589 Available values are:
17593 Keep the same field property (default).
17596 Mark the frame as bottom-field-first.
17599 Mark the frame as top-field-first.
17602 Mark the frame as progressive.
17606 Available values are:
17610 Keep the same color range property (default).
17612 @item unspecified, unknown
17613 Mark the frame as unspecified color range.
17615 @item limited, tv, mpeg
17616 Mark the frame as limited range.
17618 @item full, pc, jpeg
17619 Mark the frame as full range.
17622 @item color_primaries
17623 Set the color primaries.
17624 Available values are:
17628 Keep the same color primaries property (default).
17645 Set the color transfer.
17646 Available values are:
17650 Keep the same color trc property (default).
17672 Set the colorspace.
17673 Available values are:
17677 Keep the same colorspace property (default).
17690 @item chroma-derived-nc
17691 @item chroma-derived-c
17698 Show a line containing various information for each input video frame.
17699 The input video is not modified.
17701 This filter supports the following options:
17705 Calculate checksums of each plane. By default enabled.
17708 The shown line contains a sequence of key/value pairs of the form
17709 @var{key}:@var{value}.
17711 The following values are shown in the output:
17715 The (sequential) number of the input frame, starting from 0.
17718 The Presentation TimeStamp of the input frame, expressed as a number of
17719 time base units. The time base unit depends on the filter input pad.
17722 The Presentation TimeStamp of the input frame, expressed as a number of
17726 The position of the frame in the input stream, or -1 if this information is
17727 unavailable and/or meaningless (for example in case of synthetic video).
17730 The pixel format name.
17733 The sample aspect ratio of the input frame, expressed in the form
17734 @var{num}/@var{den}.
17737 The size of the input frame. For the syntax of this option, check the
17738 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17741 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
17742 for bottom field first).
17745 This is 1 if the frame is a key frame, 0 otherwise.
17748 The picture type of the input frame ("I" for an I-frame, "P" for a
17749 P-frame, "B" for a B-frame, or "?" for an unknown type).
17750 Also refer to the documentation of the @code{AVPictureType} enum and of
17751 the @code{av_get_picture_type_char} function defined in
17752 @file{libavutil/avutil.h}.
17755 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
17757 @item plane_checksum
17758 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
17759 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
17762 The mean value of pixels in each plane of the input frame, expressed in the form
17763 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
17766 The standard deviation of pixel values in each plane of the input frame, expressed
17767 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
17771 @section showpalette
17773 Displays the 256 colors palette of each frame. This filter is only relevant for
17774 @var{pal8} pixel format frames.
17776 It accepts the following option:
17780 Set the size of the box used to represent one palette color entry. Default is
17781 @code{30} (for a @code{30x30} pixel box).
17784 @section shuffleframes
17786 Reorder and/or duplicate and/or drop video frames.
17788 It accepts the following parameters:
17792 Set the destination indexes of input frames.
17793 This is space or '|' separated list of indexes that maps input frames to output
17794 frames. Number of indexes also sets maximal value that each index may have.
17795 '-1' index have special meaning and that is to drop frame.
17798 The first frame has the index 0. The default is to keep the input unchanged.
17800 @subsection Examples
17804 Swap second and third frame of every three frames of the input:
17806 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
17810 Swap 10th and 1st frame of every ten frames of the input:
17812 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
17816 @section shufflepixels
17818 Reorder pixels in video frames.
17820 This filter accepts the following options:
17824 Set shuffle direction. Can be forward or inverse direction.
17825 Default direction is forward.
17828 Set shuffle mode. Can be horizontal, vertical or block mode.
17832 Set shuffle block_size. In case of horizontal shuffle mode only width
17833 part of size is used, and in case of vertical shuffle mode only height
17834 part of size is used.
17837 Set random seed used with shuffling pixels. Mainly useful to set to be able
17838 to reverse filtering process to get original input.
17839 For example, to reverse forward shuffle you need to use same parameters
17840 and exact same seed and to set direction to inverse.
17843 @section shuffleplanes
17845 Reorder and/or duplicate video planes.
17847 It accepts the following parameters:
17852 The index of the input plane to be used as the first output plane.
17855 The index of the input plane to be used as the second output plane.
17858 The index of the input plane to be used as the third output plane.
17861 The index of the input plane to be used as the fourth output plane.
17865 The first plane has the index 0. The default is to keep the input unchanged.
17867 @subsection Examples
17871 Swap the second and third planes of the input:
17873 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
17877 @anchor{signalstats}
17878 @section signalstats
17879 Evaluate various visual metrics that assist in determining issues associated
17880 with the digitization of analog video media.
17882 By default the filter will log these metadata values:
17886 Display the minimal Y value contained within the input frame. Expressed in
17890 Display the Y value at the 10% percentile within the input frame. Expressed in
17894 Display the average Y value within the input frame. Expressed in range of
17898 Display the Y value at the 90% percentile within the input frame. Expressed in
17902 Display the maximum Y value contained within the input frame. Expressed in
17906 Display the minimal U value contained within the input frame. Expressed in
17910 Display the U value at the 10% percentile within the input frame. Expressed in
17914 Display the average U value within the input frame. Expressed in range of
17918 Display the U value at the 90% percentile within the input frame. Expressed in
17922 Display the maximum U value contained within the input frame. Expressed in
17926 Display the minimal V value contained within the input frame. Expressed in
17930 Display the V value at the 10% percentile within the input frame. Expressed in
17934 Display the average V value within the input frame. Expressed in range of
17938 Display the V value at the 90% percentile within the input frame. Expressed in
17942 Display the maximum V value contained within the input frame. Expressed in
17946 Display the minimal saturation value contained within the input frame.
17947 Expressed in range of [0-~181.02].
17950 Display the saturation value at the 10% percentile within the input frame.
17951 Expressed in range of [0-~181.02].
17954 Display the average saturation value within the input frame. Expressed in range
17958 Display the saturation value at the 90% percentile within the input frame.
17959 Expressed in range of [0-~181.02].
17962 Display the maximum saturation value contained within the input frame.
17963 Expressed in range of [0-~181.02].
17966 Display the median value for hue within the input frame. Expressed in range of
17970 Display the average value for hue within the input frame. Expressed in range of
17974 Display the average of sample value difference between all values of the Y
17975 plane in the current frame and corresponding values of the previous input frame.
17976 Expressed in range of [0-255].
17979 Display the average of sample value difference between all values of the U
17980 plane in the current frame and corresponding values of the previous input frame.
17981 Expressed in range of [0-255].
17984 Display the average of sample value difference between all values of the V
17985 plane in the current frame and corresponding values of the previous input frame.
17986 Expressed in range of [0-255].
17989 Display bit depth of Y plane in current frame.
17990 Expressed in range of [0-16].
17993 Display bit depth of U plane in current frame.
17994 Expressed in range of [0-16].
17997 Display bit depth of V plane in current frame.
17998 Expressed in range of [0-16].
18001 The filter accepts the following options:
18007 @option{stat} specify an additional form of image analysis.
18008 @option{out} output video with the specified type of pixel highlighted.
18010 Both options accept the following values:
18014 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
18015 unlike the neighboring pixels of the same field. Examples of temporal outliers
18016 include the results of video dropouts, head clogs, or tape tracking issues.
18019 Identify @var{vertical line repetition}. Vertical line repetition includes
18020 similar rows of pixels within a frame. In born-digital video vertical line
18021 repetition is common, but this pattern is uncommon in video digitized from an
18022 analog source. When it occurs in video that results from the digitization of an
18023 analog source it can indicate concealment from a dropout compensator.
18026 Identify pixels that fall outside of legal broadcast range.
18030 Set the highlight color for the @option{out} option. The default color is
18034 @subsection Examples
18038 Output data of various video metrics:
18040 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
18044 Output specific data about the minimum and maximum values of the Y plane per frame:
18046 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
18050 Playback video while highlighting pixels that are outside of broadcast range in red.
18052 ffplay example.mov -vf signalstats="out=brng:color=red"
18056 Playback video with signalstats metadata drawn over the frame.
18058 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
18061 The contents of signalstat_drawtext.txt used in the command are:
18064 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
18065 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
18066 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
18067 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
18075 Calculates the MPEG-7 Video Signature. The filter can handle more than one
18076 input. In this case the matching between the inputs can be calculated additionally.
18077 The filter always passes through the first input. The signature of each stream can
18078 be written into a file.
18080 It accepts the following options:
18084 Enable or disable the matching process.
18086 Available values are:
18090 Disable the calculation of a matching (default).
18092 Calculate the matching for the whole video and output whether the whole video
18093 matches or only parts.
18095 Calculate only until a matching is found or the video ends. Should be faster in
18100 Set the number of inputs. The option value must be a non negative integer.
18101 Default value is 1.
18104 Set the path to which the output is written. If there is more than one input,
18105 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
18106 integer), that will be replaced with the input number. If no filename is
18107 specified, no output will be written. This is the default.
18110 Choose the output format.
18112 Available values are:
18116 Use the specified binary representation (default).
18118 Use the specified xml representation.
18122 Set threshold to detect one word as similar. The option value must be an integer
18123 greater than zero. The default value is 9000.
18126 Set threshold to detect all words as similar. The option value must be an integer
18127 greater than zero. The default value is 60000.
18130 Set threshold to detect frames as similar. The option value must be an integer
18131 greater than zero. The default value is 116.
18134 Set the minimum length of a sequence in frames to recognize it as matching
18135 sequence. The option value must be a non negative integer value.
18136 The default value is 0.
18139 Set the minimum relation, that matching frames to all frames must have.
18140 The option value must be a double value between 0 and 1. The default value is 0.5.
18143 @subsection Examples
18147 To calculate the signature of an input video and store it in signature.bin:
18149 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
18153 To detect whether two videos match and store the signatures in XML format in
18154 signature0.xml and signature1.xml:
18156 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 -
18164 Blur the input video without impacting the outlines.
18166 It accepts the following options:
18169 @item luma_radius, lr
18170 Set the luma radius. The option value must be a float number in
18171 the range [0.1,5.0] that specifies the variance of the gaussian filter
18172 used to blur the image (slower if larger). Default value is 1.0.
18174 @item luma_strength, ls
18175 Set the luma strength. The option value must be a float number
18176 in the range [-1.0,1.0] that configures the blurring. A value included
18177 in [0.0,1.0] will blur the image whereas a value included in
18178 [-1.0,0.0] will sharpen the image. Default value is 1.0.
18180 @item luma_threshold, lt
18181 Set the luma threshold used as a coefficient to determine
18182 whether a pixel should be blurred or not. The option value must be an
18183 integer in the range [-30,30]. A value of 0 will filter all the image,
18184 a value included in [0,30] will filter flat areas and a value included
18185 in [-30,0] will filter edges. Default value is 0.
18187 @item chroma_radius, cr
18188 Set the chroma radius. The option value must be a float number in
18189 the range [0.1,5.0] that specifies the variance of the gaussian filter
18190 used to blur the image (slower if larger). Default value is @option{luma_radius}.
18192 @item chroma_strength, cs
18193 Set the chroma strength. The option value must be a float number
18194 in the range [-1.0,1.0] that configures the blurring. A value included
18195 in [0.0,1.0] will blur the image whereas a value included in
18196 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
18198 @item chroma_threshold, ct
18199 Set the chroma threshold used as a coefficient to determine
18200 whether a pixel should be blurred or not. The option value must be an
18201 integer in the range [-30,30]. A value of 0 will filter all the image,
18202 a value included in [0,30] will filter flat areas and a value included
18203 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
18206 If a chroma option is not explicitly set, the corresponding luma value
18210 Apply sobel operator to input video stream.
18212 The filter accepts the following option:
18216 Set which planes will be processed, unprocessed planes will be copied.
18217 By default value 0xf, all planes will be processed.
18220 Set value which will be multiplied with filtered result.
18223 Set value which will be added to filtered result.
18226 @subsection Commands
18228 This filter supports the all above options as @ref{commands}.
18233 Apply a simple postprocessing filter that compresses and decompresses the image
18234 at several (or - in the case of @option{quality} level @code{6} - all) shifts
18235 and average the results.
18237 The filter accepts the following options:
18241 Set quality. This option defines the number of levels for averaging. It accepts
18242 an integer in the range 0-6. If set to @code{0}, the filter will have no
18243 effect. A value of @code{6} means the higher quality. For each increment of
18244 that value the speed drops by a factor of approximately 2. Default value is
18248 Force a constant quantization parameter. If not set, the filter will use the QP
18249 from the video stream (if available).
18252 Set thresholding mode. Available modes are:
18256 Set hard thresholding (default).
18258 Set soft thresholding (better de-ringing effect, but likely blurrier).
18261 @item use_bframe_qp
18262 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
18263 option may cause flicker since the B-Frames have often larger QP. Default is
18264 @code{0} (not enabled).
18267 @subsection Commands
18269 This filter supports the following commands:
18271 @item quality, level
18272 Set quality level. The value @code{max} can be used to set the maximum level,
18273 currently @code{6}.
18279 Scale the input by applying one of the super-resolution methods based on
18280 convolutional neural networks. Supported models:
18284 Super-Resolution Convolutional Neural Network model (SRCNN).
18285 See @url{https://arxiv.org/abs/1501.00092}.
18288 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
18289 See @url{https://arxiv.org/abs/1609.05158}.
18292 Training scripts as well as scripts for model file (.pb) saving can be found at
18293 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
18294 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
18296 Native model files (.model) can be generated from TensorFlow model
18297 files (.pb) by using tools/python/convert.py
18299 The filter accepts the following options:
18303 Specify which DNN backend to use for model loading and execution. This option accepts
18304 the following values:
18308 Native implementation of DNN loading and execution.
18311 TensorFlow backend. To enable this backend you
18312 need to install the TensorFlow for C library (see
18313 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
18314 @code{--enable-libtensorflow}
18317 Default value is @samp{native}.
18320 Set path to model file specifying network architecture and its parameters.
18321 Note that different backends use different file formats. TensorFlow backend
18322 can load files for both formats, while native backend can load files for only
18326 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
18327 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
18328 input upscaled using bicubic upscaling with proper scale factor.
18331 This feature can also be finished with @ref{dnn_processing} filter.
18335 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
18337 This filter takes in input two input videos, the first input is
18338 considered the "main" source and is passed unchanged to the
18339 output. The second input is used as a "reference" video for computing
18342 Both video inputs must have the same resolution and pixel format for
18343 this filter to work correctly. Also it assumes that both inputs
18344 have the same number of frames, which are compared one by one.
18346 The filter stores the calculated SSIM of each frame.
18348 The description of the accepted parameters follows.
18351 @item stats_file, f
18352 If specified the filter will use the named file to save the SSIM of
18353 each individual frame. When filename equals "-" the data is sent to
18357 The file printed if @var{stats_file} is selected, contains a sequence of
18358 key/value pairs of the form @var{key}:@var{value} for each compared
18361 A description of each shown parameter follows:
18365 sequential number of the input frame, starting from 1
18367 @item Y, U, V, R, G, B
18368 SSIM of the compared frames for the component specified by the suffix.
18371 SSIM of the compared frames for the whole frame.
18374 Same as above but in dB representation.
18377 This filter also supports the @ref{framesync} options.
18379 @subsection Examples
18384 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
18385 [main][ref] ssim="stats_file=stats.log" [out]
18388 On this example the input file being processed is compared with the
18389 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
18390 is stored in @file{stats.log}.
18393 Another example with both psnr and ssim at same time:
18395 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
18399 Another example with different containers:
18401 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 -
18407 Convert between different stereoscopic image formats.
18409 The filters accept the following options:
18413 Set stereoscopic image format of input.
18415 Available values for input image formats are:
18418 side by side parallel (left eye left, right eye right)
18421 side by side crosseye (right eye left, left eye right)
18424 side by side parallel with half width resolution
18425 (left eye left, right eye right)
18428 side by side crosseye with half width resolution
18429 (right eye left, left eye right)
18433 above-below (left eye above, right eye below)
18437 above-below (right eye above, left eye below)
18441 above-below with half height resolution
18442 (left eye above, right eye below)
18446 above-below with half height resolution
18447 (right eye above, left eye below)
18450 alternating frames (left eye first, right eye second)
18453 alternating frames (right eye first, left eye second)
18456 interleaved rows (left eye has top row, right eye starts on next row)
18459 interleaved rows (right eye has top row, left eye starts on next row)
18462 interleaved columns, left eye first
18465 interleaved columns, right eye first
18467 Default value is @samp{sbsl}.
18471 Set stereoscopic image format of output.
18475 side by side parallel (left eye left, right eye right)
18478 side by side crosseye (right eye left, left eye right)
18481 side by side parallel with half width resolution
18482 (left eye left, right eye right)
18485 side by side crosseye with half width resolution
18486 (right eye left, left eye right)
18490 above-below (left eye above, right eye below)
18494 above-below (right eye above, left eye below)
18498 above-below with half height resolution
18499 (left eye above, right eye below)
18503 above-below with half height resolution
18504 (right eye above, left eye below)
18507 alternating frames (left eye first, right eye second)
18510 alternating frames (right eye first, left eye second)
18513 interleaved rows (left eye has top row, right eye starts on next row)
18516 interleaved rows (right eye has top row, left eye starts on next row)
18519 anaglyph red/blue gray
18520 (red filter on left eye, blue filter on right eye)
18523 anaglyph red/green gray
18524 (red filter on left eye, green filter on right eye)
18527 anaglyph red/cyan gray
18528 (red filter on left eye, cyan filter on right eye)
18531 anaglyph red/cyan half colored
18532 (red filter on left eye, cyan filter on right eye)
18535 anaglyph red/cyan color
18536 (red filter on left eye, cyan filter on right eye)
18539 anaglyph red/cyan color optimized with the least squares projection of dubois
18540 (red filter on left eye, cyan filter on right eye)
18543 anaglyph green/magenta gray
18544 (green filter on left eye, magenta filter on right eye)
18547 anaglyph green/magenta half colored
18548 (green filter on left eye, magenta filter on right eye)
18551 anaglyph green/magenta colored
18552 (green filter on left eye, magenta filter on right eye)
18555 anaglyph green/magenta color optimized with the least squares projection of dubois
18556 (green filter on left eye, magenta filter on right eye)
18559 anaglyph yellow/blue gray
18560 (yellow filter on left eye, blue filter on right eye)
18563 anaglyph yellow/blue half colored
18564 (yellow filter on left eye, blue filter on right eye)
18567 anaglyph yellow/blue colored
18568 (yellow filter on left eye, blue filter on right eye)
18571 anaglyph yellow/blue color optimized with the least squares projection of dubois
18572 (yellow filter on left eye, blue filter on right eye)
18575 mono output (left eye only)
18578 mono output (right eye only)
18581 checkerboard, left eye first
18584 checkerboard, right eye first
18587 interleaved columns, left eye first
18590 interleaved columns, right eye first
18596 Default value is @samp{arcd}.
18599 @subsection Examples
18603 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
18609 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
18615 @section streamselect, astreamselect
18616 Select video or audio streams.
18618 The filter accepts the following options:
18622 Set number of inputs. Default is 2.
18625 Set input indexes to remap to outputs.
18628 @subsection Commands
18630 The @code{streamselect} and @code{astreamselect} filter supports the following
18635 Set input indexes to remap to outputs.
18638 @subsection Examples
18642 Select first 5 seconds 1st stream and rest of time 2nd stream:
18644 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
18648 Same as above, but for audio:
18650 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
18657 Draw subtitles on top of input video using the libass library.
18659 To enable compilation of this filter you need to configure FFmpeg with
18660 @code{--enable-libass}. This filter also requires a build with libavcodec and
18661 libavformat to convert the passed subtitles file to ASS (Advanced Substation
18662 Alpha) subtitles format.
18664 The filter accepts the following options:
18668 Set the filename of the subtitle file to read. It must be specified.
18670 @item original_size
18671 Specify the size of the original video, the video for which the ASS file
18672 was composed. For the syntax of this option, check the
18673 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18674 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
18675 correctly scale the fonts if the aspect ratio has been changed.
18678 Set a directory path containing fonts that can be used by the filter.
18679 These fonts will be used in addition to whatever the font provider uses.
18682 Process alpha channel, by default alpha channel is untouched.
18685 Set subtitles input character encoding. @code{subtitles} filter only. Only
18686 useful if not UTF-8.
18688 @item stream_index, si
18689 Set subtitles stream index. @code{subtitles} filter only.
18692 Override default style or script info parameters of the subtitles. It accepts a
18693 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
18696 If the first key is not specified, it is assumed that the first value
18697 specifies the @option{filename}.
18699 For example, to render the file @file{sub.srt} on top of the input
18700 video, use the command:
18705 which is equivalent to:
18707 subtitles=filename=sub.srt
18710 To render the default subtitles stream from file @file{video.mkv}, use:
18712 subtitles=video.mkv
18715 To render the second subtitles stream from that file, use:
18717 subtitles=video.mkv:si=1
18720 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
18721 @code{DejaVu Serif}, use:
18723 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
18726 @section super2xsai
18728 Scale the input by 2x and smooth using the Super2xSaI (Scale and
18729 Interpolate) pixel art scaling algorithm.
18731 Useful for enlarging pixel art images without reducing sharpness.
18735 Swap two rectangular objects in video.
18737 This filter accepts the following options:
18747 Set 1st rect x coordinate.
18750 Set 1st rect y coordinate.
18753 Set 2nd rect x coordinate.
18756 Set 2nd rect y coordinate.
18758 All expressions are evaluated once for each frame.
18761 The all options are expressions containing the following constants:
18766 The input width and height.
18769 same as @var{w} / @var{h}
18772 input sample aspect ratio
18775 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
18778 The number of the input frame, starting from 0.
18781 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
18784 the position in the file of the input frame, NAN if unknown
18791 Blend successive video frames.
18797 Apply telecine process to the video.
18799 This filter accepts the following options:
18808 The default value is @code{top}.
18812 A string of numbers representing the pulldown pattern you wish to apply.
18813 The default value is @code{23}.
18817 Some typical patterns:
18822 24p: 2332 (preferred)
18829 24p: 222222222223 ("Euro pulldown")
18834 @section thistogram
18836 Compute and draw a color distribution histogram for the input video across time.
18838 Unlike @ref{histogram} video filter which only shows histogram of single input frame
18839 at certain time, this filter shows also past histograms of number of frames defined
18840 by @code{width} option.
18842 The computed histogram is a representation of the color component
18843 distribution in an image.
18845 The filter accepts the following options:
18849 Set width of single color component output. Default value is @code{0}.
18850 Value of @code{0} means width will be picked from input video.
18851 This also set number of passed histograms to keep.
18852 Allowed range is [0, 8192].
18854 @item display_mode, d
18856 It accepts the following values:
18859 Per color component graphs are placed below each other.
18862 Per color component graphs are placed side by side.
18865 Presents information identical to that in the @code{parade}, except
18866 that the graphs representing color components are superimposed directly
18869 Default is @code{stack}.
18871 @item levels_mode, m
18872 Set mode. Can be either @code{linear}, or @code{logarithmic}.
18873 Default is @code{linear}.
18875 @item components, c
18876 Set what color components to display.
18877 Default is @code{7}.
18880 Set background opacity. Default is @code{0.9}.
18883 Show envelope. Default is disabled.
18886 Set envelope color. Default is @code{gold}.
18891 Available values for slide is:
18894 Draw new frame when right border is reached.
18897 Replace old columns with new ones.
18900 Scroll from right to left.
18903 Scroll from left to right.
18906 Draw single picture.
18909 Default is @code{replace}.
18914 Apply threshold effect to video stream.
18916 This filter needs four video streams to perform thresholding.
18917 First stream is stream we are filtering.
18918 Second stream is holding threshold values, third stream is holding min values,
18919 and last, fourth stream is holding max values.
18921 The filter accepts the following option:
18925 Set which planes will be processed, unprocessed planes will be copied.
18926 By default value 0xf, all planes will be processed.
18929 For example if first stream pixel's component value is less then threshold value
18930 of pixel component from 2nd threshold stream, third stream value will picked,
18931 otherwise fourth stream pixel component value will be picked.
18933 Using color source filter one can perform various types of thresholding:
18935 @subsection Examples
18939 Binary threshold, using gray color as threshold:
18941 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
18945 Inverted binary threshold, using gray color as threshold:
18947 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
18951 Truncate binary threshold, using gray color as threshold:
18953 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
18957 Threshold to zero, using gray color as threshold:
18959 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
18963 Inverted threshold to zero, using gray color as threshold:
18965 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
18970 Select the most representative frame in a given sequence of consecutive frames.
18972 The filter accepts the following options:
18976 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
18977 will pick one of them, and then handle the next batch of @var{n} frames until
18978 the end. Default is @code{100}.
18981 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
18982 value will result in a higher memory usage, so a high value is not recommended.
18984 @subsection Examples
18988 Extract one picture each 50 frames:
18994 Complete example of a thumbnail creation with @command{ffmpeg}:
18996 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
19003 Tile several successive frames together.
19005 The @ref{untile} filter can do the reverse.
19007 The filter accepts the following options:
19012 Set the grid size (i.e. the number of lines and columns). For the syntax of
19013 this option, check the
19014 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19017 Set the maximum number of frames to render in the given area. It must be less
19018 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
19019 the area will be used.
19022 Set the outer border margin in pixels.
19025 Set the inner border thickness (i.e. the number of pixels between frames). For
19026 more advanced padding options (such as having different values for the edges),
19027 refer to the pad video filter.
19030 Specify the color of the unused area. For the syntax of this option, check the
19031 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
19032 The default value of @var{color} is "black".
19035 Set the number of frames to overlap when tiling several successive frames together.
19036 The value must be between @code{0} and @var{nb_frames - 1}.
19039 Set the number of frames to initially be empty before displaying first output frame.
19040 This controls how soon will one get first output frame.
19041 The value must be between @code{0} and @var{nb_frames - 1}.
19044 @subsection Examples
19048 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
19050 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
19052 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
19053 duplicating each output frame to accommodate the originally detected frame
19057 Display @code{5} pictures in an area of @code{3x2} frames,
19058 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
19059 mixed flat and named options:
19061 tile=3x2:nb_frames=5:padding=7:margin=2
19065 @section tinterlace
19067 Perform various types of temporal field interlacing.
19069 Frames are counted starting from 1, so the first input frame is
19072 The filter accepts the following options:
19077 Specify the mode of the interlacing. This option can also be specified
19078 as a value alone. See below for a list of values for this option.
19080 Available values are:
19084 Move odd frames into the upper field, even into the lower field,
19085 generating a double height frame at half frame rate.
19089 Frame 1 Frame 2 Frame 3 Frame 4
19091 11111 22222 33333 44444
19092 11111 22222 33333 44444
19093 11111 22222 33333 44444
19094 11111 22222 33333 44444
19108 Only output odd frames, even frames are dropped, generating a frame with
19109 unchanged height at half frame rate.
19114 Frame 1 Frame 2 Frame 3 Frame 4
19116 11111 22222 33333 44444
19117 11111 22222 33333 44444
19118 11111 22222 33333 44444
19119 11111 22222 33333 44444
19129 Only output even frames, odd frames are dropped, generating a frame with
19130 unchanged height at half frame rate.
19135 Frame 1 Frame 2 Frame 3 Frame 4
19137 11111 22222 33333 44444
19138 11111 22222 33333 44444
19139 11111 22222 33333 44444
19140 11111 22222 33333 44444
19150 Expand each frame to full height, but pad alternate lines with black,
19151 generating a frame with double height at the same input frame rate.
19156 Frame 1 Frame 2 Frame 3 Frame 4
19158 11111 22222 33333 44444
19159 11111 22222 33333 44444
19160 11111 22222 33333 44444
19161 11111 22222 33333 44444
19164 11111 ..... 33333 .....
19165 ..... 22222 ..... 44444
19166 11111 ..... 33333 .....
19167 ..... 22222 ..... 44444
19168 11111 ..... 33333 .....
19169 ..... 22222 ..... 44444
19170 11111 ..... 33333 .....
19171 ..... 22222 ..... 44444
19175 @item interleave_top, 4
19176 Interleave the upper field from odd frames with the lower field from
19177 even frames, generating a frame with unchanged height at half frame rate.
19182 Frame 1 Frame 2 Frame 3 Frame 4
19184 11111<- 22222 33333<- 44444
19185 11111 22222<- 33333 44444<-
19186 11111<- 22222 33333<- 44444
19187 11111 22222<- 33333 44444<-
19197 @item interleave_bottom, 5
19198 Interleave the lower field from odd frames with the upper field from
19199 even frames, generating a frame with unchanged height at half frame rate.
19204 Frame 1 Frame 2 Frame 3 Frame 4
19206 11111 22222<- 33333 44444<-
19207 11111<- 22222 33333<- 44444
19208 11111 22222<- 33333 44444<-
19209 11111<- 22222 33333<- 44444
19219 @item interlacex2, 6
19220 Double frame rate with unchanged height. Frames are inserted each
19221 containing the second temporal field from the previous input frame and
19222 the first temporal field from the next input frame. This mode relies on
19223 the top_field_first flag. Useful for interlaced video displays with no
19224 field synchronisation.
19229 Frame 1 Frame 2 Frame 3 Frame 4
19231 11111 22222 33333 44444
19232 11111 22222 33333 44444
19233 11111 22222 33333 44444
19234 11111 22222 33333 44444
19237 11111 22222 22222 33333 33333 44444 44444
19238 11111 11111 22222 22222 33333 33333 44444
19239 11111 22222 22222 33333 33333 44444 44444
19240 11111 11111 22222 22222 33333 33333 44444
19245 Move odd frames into the upper field, even into the lower field,
19246 generating a double height frame at same frame rate.
19251 Frame 1 Frame 2 Frame 3 Frame 4
19253 11111 22222 33333 44444
19254 11111 22222 33333 44444
19255 11111 22222 33333 44444
19256 11111 22222 33333 44444
19259 11111 33333 33333 55555
19260 22222 22222 44444 44444
19261 11111 33333 33333 55555
19262 22222 22222 44444 44444
19263 11111 33333 33333 55555
19264 22222 22222 44444 44444
19265 11111 33333 33333 55555
19266 22222 22222 44444 44444
19271 Numeric values are deprecated but are accepted for backward
19272 compatibility reasons.
19274 Default mode is @code{merge}.
19277 Specify flags influencing the filter process.
19279 Available value for @var{flags} is:
19282 @item low_pass_filter, vlpf
19283 Enable linear vertical low-pass filtering in the filter.
19284 Vertical low-pass filtering is required when creating an interlaced
19285 destination from a progressive source which contains high-frequency
19286 vertical detail. Filtering will reduce interlace 'twitter' and Moire
19289 @item complex_filter, cvlpf
19290 Enable complex vertical low-pass filtering.
19291 This will slightly less reduce interlace 'twitter' and Moire
19292 patterning but better retain detail and subjective sharpness impression.
19295 Bypass already interlaced frames, only adjust the frame rate.
19298 Vertical low-pass filtering and bypassing already interlaced frames can only be
19299 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
19304 Pick median pixels from several successive input video frames.
19306 The filter accepts the following options:
19310 Set radius of median filter.
19311 Default is 1. Allowed range is from 1 to 127.
19314 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
19317 Set median percentile. Default value is @code{0.5}.
19318 Default value of @code{0.5} will pick always median values, while @code{0} will pick
19319 minimum values, and @code{1} maximum values.
19322 @subsection Commands
19324 This filter supports all above options as @ref{commands}, excluding option @code{radius}.
19326 @section tmidequalizer
19328 Apply Temporal Midway Video Equalization effect.
19330 Midway Video Equalization adjusts a sequence of video frames to have the same
19331 histograms, while maintaining their dynamics as much as possible. It's
19332 useful for e.g. matching exposures from a video frames sequence.
19334 This filter accepts the following option:
19338 Set filtering radius. Default is @code{5}. Allowed range is from 1 to 127.
19341 Set filtering sigma. Default is @code{0.5}. This controls strength of filtering.
19342 Setting this option to 0 effectively does nothing.
19345 Set which planes to process. Default is @code{15}, which is all available planes.
19350 Mix successive video frames.
19352 A description of the accepted options follows.
19356 The number of successive frames to mix. If unspecified, it defaults to 3.
19359 Specify weight of each input video frame.
19360 Each weight is separated by space. If number of weights is smaller than
19361 number of @var{frames} last specified weight will be used for all remaining
19365 Specify scale, if it is set it will be multiplied with sum
19366 of each weight multiplied with pixel values to give final destination
19367 pixel value. By default @var{scale} is auto scaled to sum of weights.
19370 @subsection Examples
19374 Average 7 successive frames:
19376 tmix=frames=7:weights="1 1 1 1 1 1 1"
19380 Apply simple temporal convolution:
19382 tmix=frames=3:weights="-1 3 -1"
19386 Similar as above but only showing temporal differences:
19388 tmix=frames=3:weights="-1 2 -1":scale=1
19394 Tone map colors from different dynamic ranges.
19396 This filter expects data in single precision floating point, as it needs to
19397 operate on (and can output) out-of-range values. Another filter, such as
19398 @ref{zscale}, is needed to convert the resulting frame to a usable format.
19400 The tonemapping algorithms implemented only work on linear light, so input
19401 data should be linearized beforehand (and possibly correctly tagged).
19404 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
19407 @subsection Options
19408 The filter accepts the following options.
19412 Set the tone map algorithm to use.
19414 Possible values are:
19417 Do not apply any tone map, only desaturate overbright pixels.
19420 Hard-clip any out-of-range values. Use it for perfect color accuracy for
19421 in-range values, while distorting out-of-range values.
19424 Stretch the entire reference gamut to a linear multiple of the display.
19427 Fit a logarithmic transfer between the tone curves.
19430 Preserve overall image brightness with a simple curve, using nonlinear
19431 contrast, which results in flattening details and degrading color accuracy.
19434 Preserve both dark and bright details better than @var{reinhard}, at the cost
19435 of slightly darkening everything. Use it when detail preservation is more
19436 important than color and brightness accuracy.
19439 Smoothly map out-of-range values, while retaining contrast and colors for
19440 in-range material as much as possible. Use it when color accuracy is more
19441 important than detail preservation.
19447 Tune the tone mapping algorithm.
19449 This affects the following algorithms:
19455 Specifies the scale factor to use while stretching.
19459 Specifies the exponent of the function.
19463 Specify an extra linear coefficient to multiply into the signal before clipping.
19467 Specify the local contrast coefficient at the display peak.
19468 Default to 0.5, which means that in-gamut values will be about half as bright
19475 Specify the transition point from linear to mobius transform. Every value
19476 below this point is guaranteed to be mapped 1:1. The higher the value, the
19477 more accurate the result will be, at the cost of losing bright details.
19478 Default to 0.3, which due to the steep initial slope still preserves in-range
19479 colors fairly accurately.
19483 Apply desaturation for highlights that exceed this level of brightness. The
19484 higher the parameter, the more color information will be preserved. This
19485 setting helps prevent unnaturally blown-out colors for super-highlights, by
19486 (smoothly) turning into white instead. This makes images feel more natural,
19487 at the cost of reducing information about out-of-range colors.
19489 The default of 2.0 is somewhat conservative and will mostly just apply to
19490 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
19492 This option works only if the input frame has a supported color tag.
19495 Override signal/nominal/reference peak with this value. Useful when the
19496 embedded peak information in display metadata is not reliable or when tone
19497 mapping from a lower range to a higher range.
19502 Temporarily pad video frames.
19504 The filter accepts the following options:
19508 Specify number of delay frames before input video stream. Default is 0.
19511 Specify number of padding frames after input video stream.
19512 Set to -1 to pad indefinitely. Default is 0.
19515 Set kind of frames added to beginning of stream.
19516 Can be either @var{add} or @var{clone}.
19517 With @var{add} frames of solid-color are added.
19518 With @var{clone} frames are clones of first frame.
19519 Default is @var{add}.
19522 Set kind of frames added to end of stream.
19523 Can be either @var{add} or @var{clone}.
19524 With @var{add} frames of solid-color are added.
19525 With @var{clone} frames are clones of last frame.
19526 Default is @var{add}.
19528 @item start_duration, stop_duration
19529 Specify the duration of the start/stop delay. See
19530 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19531 for the accepted syntax.
19532 These options override @var{start} and @var{stop}. Default is 0.
19535 Specify the color of the padded area. For the syntax of this option,
19536 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
19537 manual,ffmpeg-utils}.
19539 The default value of @var{color} is "black".
19545 Transpose rows with columns in the input video and optionally flip it.
19547 It accepts the following parameters:
19552 Specify the transposition direction.
19554 Can assume the following values:
19556 @item 0, 4, cclock_flip
19557 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
19565 Rotate by 90 degrees clockwise, that is:
19573 Rotate by 90 degrees counterclockwise, that is:
19580 @item 3, 7, clock_flip
19581 Rotate by 90 degrees clockwise and vertically flip, that is:
19589 For values between 4-7, the transposition is only done if the input
19590 video geometry is portrait and not landscape. These values are
19591 deprecated, the @code{passthrough} option should be used instead.
19593 Numerical values are deprecated, and should be dropped in favor of
19594 symbolic constants.
19597 Do not apply the transposition if the input geometry matches the one
19598 specified by the specified value. It accepts the following values:
19601 Always apply transposition.
19603 Preserve portrait geometry (when @var{height} >= @var{width}).
19605 Preserve landscape geometry (when @var{width} >= @var{height}).
19608 Default value is @code{none}.
19611 For example to rotate by 90 degrees clockwise and preserve portrait
19614 transpose=dir=1:passthrough=portrait
19617 The command above can also be specified as:
19619 transpose=1:portrait
19622 @section transpose_npp
19624 Transpose rows with columns in the input video and optionally flip it.
19625 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
19627 It accepts the following parameters:
19632 Specify the transposition direction.
19634 Can assume the following values:
19637 Rotate by 90 degrees counterclockwise and vertically flip. (default)
19640 Rotate by 90 degrees clockwise.
19643 Rotate by 90 degrees counterclockwise.
19646 Rotate by 90 degrees clockwise and vertically flip.
19650 Do not apply the transposition if the input geometry matches the one
19651 specified by the specified value. It accepts the following values:
19654 Always apply transposition. (default)
19656 Preserve portrait geometry (when @var{height} >= @var{width}).
19658 Preserve landscape geometry (when @var{width} >= @var{height}).
19664 Trim the input so that the output contains one continuous subpart of the input.
19666 It accepts the following parameters:
19669 Specify the time of the start of the kept section, i.e. the frame with the
19670 timestamp @var{start} will be the first frame in the output.
19673 Specify the time of the first frame that will be dropped, i.e. the frame
19674 immediately preceding the one with the timestamp @var{end} will be the last
19675 frame in the output.
19678 This is the same as @var{start}, except this option sets the start timestamp
19679 in timebase units instead of seconds.
19682 This is the same as @var{end}, except this option sets the end timestamp
19683 in timebase units instead of seconds.
19686 The maximum duration of the output in seconds.
19689 The number of the first frame that should be passed to the output.
19692 The number of the first frame that should be dropped.
19695 @option{start}, @option{end}, and @option{duration} are expressed as time
19696 duration specifications; see
19697 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19698 for the accepted syntax.
19700 Note that the first two sets of the start/end options and the @option{duration}
19701 option look at the frame timestamp, while the _frame variants simply count the
19702 frames that pass through the filter. Also note that this filter does not modify
19703 the timestamps. If you wish for the output timestamps to start at zero, insert a
19704 setpts filter after the trim filter.
19706 If multiple start or end options are set, this filter tries to be greedy and
19707 keep all the frames that match at least one of the specified constraints. To keep
19708 only the part that matches all the constraints at once, chain multiple trim
19711 The defaults are such that all the input is kept. So it is possible to set e.g.
19712 just the end values to keep everything before the specified time.
19717 Drop everything except the second minute of input:
19719 ffmpeg -i INPUT -vf trim=60:120
19723 Keep only the first second:
19725 ffmpeg -i INPUT -vf trim=duration=1
19730 @section unpremultiply
19731 Apply alpha unpremultiply effect to input video stream using first plane
19732 of second stream as alpha.
19734 Both streams must have same dimensions and same pixel format.
19736 The filter accepts the following option:
19740 Set which planes will be processed, unprocessed planes will be copied.
19741 By default value 0xf, all planes will be processed.
19743 If the format has 1 or 2 components, then luma is bit 0.
19744 If the format has 3 or 4 components:
19745 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
19746 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
19747 If present, the alpha channel is always the last bit.
19750 Do not require 2nd input for processing, instead use alpha plane from input stream.
19756 Sharpen or blur the input video.
19758 It accepts the following parameters:
19761 @item luma_msize_x, lx
19762 Set the luma matrix horizontal size. It must be an odd integer between
19763 3 and 23. The default value is 5.
19765 @item luma_msize_y, ly
19766 Set the luma matrix vertical size. It must be an odd integer between 3
19767 and 23. The default value is 5.
19769 @item luma_amount, la
19770 Set the luma effect strength. It must be a floating point number, reasonable
19771 values lay between -1.5 and 1.5.
19773 Negative values will blur the input video, while positive values will
19774 sharpen it, a value of zero will disable the effect.
19776 Default value is 1.0.
19778 @item chroma_msize_x, cx
19779 Set the chroma matrix horizontal size. It must be an odd integer
19780 between 3 and 23. The default value is 5.
19782 @item chroma_msize_y, cy
19783 Set the chroma matrix vertical size. It must be an odd integer
19784 between 3 and 23. The default value is 5.
19786 @item chroma_amount, ca
19787 Set the chroma effect strength. It must be a floating point number, reasonable
19788 values lay between -1.5 and 1.5.
19790 Negative values will blur the input video, while positive values will
19791 sharpen it, a value of zero will disable the effect.
19793 Default value is 0.0.
19797 All parameters are optional and default to the equivalent of the
19798 string '5:5:1.0:5:5:0.0'.
19800 @subsection Examples
19804 Apply strong luma sharpen effect:
19806 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
19810 Apply a strong blur of both luma and chroma parameters:
19812 unsharp=7:7:-2:7:7:-2
19819 Decompose a video made of tiled images into the individual images.
19821 The frame rate of the output video is the frame rate of the input video
19822 multiplied by the number of tiles.
19824 This filter does the reverse of @ref{tile}.
19826 The filter accepts the following options:
19831 Set the grid size (i.e. the number of lines and columns). For the syntax of
19832 this option, check the
19833 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19836 @subsection Examples
19840 Produce a 1-second video from a still image file made of 25 frames stacked
19841 vertically, like an analogic film reel:
19843 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
19849 Apply ultra slow/simple postprocessing filter that compresses and decompresses
19850 the image at several (or - in the case of @option{quality} level @code{8} - all)
19851 shifts and average the results.
19853 The way this differs from the behavior of spp is that uspp actually encodes &
19854 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
19855 DCT similar to MJPEG.
19857 The filter accepts the following options:
19861 Set quality. This option defines the number of levels for averaging. It accepts
19862 an integer in the range 0-8. If set to @code{0}, the filter will have no
19863 effect. A value of @code{8} means the higher quality. For each increment of
19864 that value the speed drops by a factor of approximately 2. Default value is
19868 Force a constant quantization parameter. If not set, the filter will use the QP
19869 from the video stream (if available).
19874 Convert 360 videos between various formats.
19876 The filter accepts the following options:
19882 Set format of the input/output video.
19890 Equirectangular projection.
19895 Cubemap with 3x2/6x1/1x6 layout.
19897 Format specific options:
19902 Set padding proportion for the input/output cubemap. Values in decimals.
19909 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)
19912 Default value is @b{@samp{0}}.
19913 Maximum value is @b{@samp{0.1}}.
19917 Set fixed padding for the input/output cubemap. Values in pixels.
19919 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
19923 Set order of faces for the input/output cubemap. Choose one direction for each position.
19925 Designation of directions:
19941 Default value is @b{@samp{rludfb}}.
19945 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
19947 Designation of angles:
19950 0 degrees clockwise
19952 90 degrees clockwise
19954 180 degrees clockwise
19956 270 degrees clockwise
19959 Default value is @b{@samp{000000}}.
19963 Equi-Angular Cubemap.
19970 Format specific options:
19975 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19977 If diagonal field of view is set it overrides horizontal and vertical field of view.
19982 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19984 If diagonal field of view is set it overrides horizontal and vertical field of view.
19990 Format specific options:
19995 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19997 If diagonal field of view is set it overrides horizontal and vertical field of view.
20002 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20004 If diagonal field of view is set it overrides horizontal and vertical field of view.
20010 Facebook's 360 formats.
20013 Stereographic format.
20015 Format specific options:
20020 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20022 If diagonal field of view is set it overrides horizontal and vertical field of view.
20027 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20029 If diagonal field of view is set it overrides horizontal and vertical field of view.
20036 Ball format, gives significant distortion toward the back.
20039 Hammer-Aitoff map projection format.
20042 Sinusoidal map projection format.
20045 Fisheye projection.
20047 Format specific options:
20052 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20054 If diagonal field of view is set it overrides horizontal and vertical field of view.
20059 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20061 If diagonal field of view is set it overrides horizontal and vertical field of view.
20065 Pannini projection.
20067 Format specific options:
20070 Set output pannini parameter.
20073 Set input pannini parameter.
20077 Cylindrical projection.
20079 Format specific options:
20084 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20086 If diagonal field of view is set it overrides horizontal and vertical field of view.
20091 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20093 If diagonal field of view is set it overrides horizontal and vertical field of view.
20097 Perspective projection. @i{(output only)}
20099 Format specific options:
20102 Set perspective parameter.
20106 Tetrahedron projection.
20109 Truncated square pyramid projection.
20113 Half equirectangular projection.
20118 Format specific options:
20123 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20125 If diagonal field of view is set it overrides horizontal and vertical field of view.
20130 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20132 If diagonal field of view is set it overrides horizontal and vertical field of view.
20136 Orthographic format.
20138 Format specific options:
20143 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20145 If diagonal field of view is set it overrides horizontal and vertical field of view.
20150 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20152 If diagonal field of view is set it overrides horizontal and vertical field of view.
20156 Octahedron projection.
20160 Set interpolation method.@*
20161 @i{Note: more complex interpolation methods require much more memory to run.}
20171 Bilinear interpolation.
20173 Lagrange9 interpolation.
20176 Bicubic interpolation.
20179 Lanczos interpolation.
20182 Spline16 interpolation.
20185 Gaussian interpolation.
20187 Mitchell interpolation.
20190 Default value is @b{@samp{line}}.
20194 Set the output video resolution.
20196 Default resolution depends on formats.
20200 Set the input/output stereo format.
20211 Default value is @b{@samp{2d}} for input and output format.
20216 Set rotation for the output video. Values in degrees.
20219 Set rotation order for the output video. Choose one item for each position.
20230 Default value is @b{@samp{ypr}}.
20235 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
20239 Set if input video is flipped horizontally/vertically. Boolean values.
20242 Set if input video is transposed. Boolean value, by default disabled.
20245 Set if output video needs to be transposed. Boolean value, by default disabled.
20248 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
20251 @subsection Examples
20255 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
20257 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
20260 Extract back view of Equi-Angular Cubemap:
20262 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
20265 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
20267 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
20271 @subsection Commands
20273 This filter supports subset of above options as @ref{commands}.
20275 @section vaguedenoiser
20277 Apply a wavelet based denoiser.
20279 It transforms each frame from the video input into the wavelet domain,
20280 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
20281 the obtained coefficients. It does an inverse wavelet transform after.
20282 Due to wavelet properties, it should give a nice smoothed result, and
20283 reduced noise, without blurring picture features.
20285 This filter accepts the following options:
20289 The filtering strength. The higher, the more filtered the video will be.
20290 Hard thresholding can use a higher threshold than soft thresholding
20291 before the video looks overfiltered. Default value is 2.
20294 The filtering method the filter will use.
20296 It accepts the following values:
20299 All values under the threshold will be zeroed.
20302 All values under the threshold will be zeroed. All values above will be
20303 reduced by the threshold.
20306 Scales or nullifies coefficients - intermediary between (more) soft and
20307 (less) hard thresholding.
20310 Default is garrote.
20313 Number of times, the wavelet will decompose the picture. Picture can't
20314 be decomposed beyond a particular point (typically, 8 for a 640x480
20315 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
20318 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
20321 A list of the planes to process. By default all planes are processed.
20324 The threshold type the filter will use.
20326 It accepts the following values:
20329 Threshold used is same for all decompositions.
20332 Threshold used depends also on each decomposition coefficients.
20335 Default is universal.
20338 @section vectorscope
20340 Display 2 color component values in the two dimensional graph (which is called
20343 This filter accepts the following options:
20347 Set vectorscope mode.
20349 It accepts the following values:
20353 Gray values are displayed on graph, higher brightness means more pixels have
20354 same component color value on location in graph. This is the default mode.
20357 Gray values are displayed on graph. Surrounding pixels values which are not
20358 present in video frame are drawn in gradient of 2 color components which are
20359 set by option @code{x} and @code{y}. The 3rd color component is static.
20362 Actual color components values present in video frame are displayed on graph.
20365 Similar as color2 but higher frequency of same values @code{x} and @code{y}
20366 on graph increases value of another color component, which is luminance by
20367 default values of @code{x} and @code{y}.
20370 Actual colors present in video frame are displayed on graph. If two different
20371 colors map to same position on graph then color with higher value of component
20372 not present in graph is picked.
20375 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
20376 component picked from radial gradient.
20380 Set which color component will be represented on X-axis. Default is @code{1}.
20383 Set which color component will be represented on Y-axis. Default is @code{2}.
20386 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
20387 of color component which represents frequency of (X, Y) location in graph.
20392 No envelope, this is default.
20395 Instant envelope, even darkest single pixel will be clearly highlighted.
20398 Hold maximum and minimum values presented in graph over time. This way you
20399 can still spot out of range values without constantly looking at vectorscope.
20402 Peak and instant envelope combined together.
20406 Set what kind of graticule to draw.
20415 Set graticule opacity.
20418 Set graticule flags.
20422 Draw graticule for white point.
20425 Draw graticule for black point.
20428 Draw color points short names.
20432 Set background opacity.
20434 @item lthreshold, l
20435 Set low threshold for color component not represented on X or Y axis.
20436 Values lower than this value will be ignored. Default is 0.
20437 Note this value is multiplied with actual max possible value one pixel component
20438 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
20441 @item hthreshold, h
20442 Set high threshold for color component not represented on X or Y axis.
20443 Values higher than this value will be ignored. Default is 1.
20444 Note this value is multiplied with actual max possible value one pixel component
20445 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
20446 is 0.9 * 255 = 230.
20448 @item colorspace, c
20449 Set what kind of colorspace to use when drawing graticule.
20459 Set color tint for gray/tint vectorscope mode. By default both options are zero.
20460 This means no tint, and output will remain gray.
20463 @anchor{vidstabdetect}
20464 @section vidstabdetect
20466 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
20467 @ref{vidstabtransform} for pass 2.
20469 This filter generates a file with relative translation and rotation
20470 transform information about subsequent frames, which is then used by
20471 the @ref{vidstabtransform} filter.
20473 To enable compilation of this filter you need to configure FFmpeg with
20474 @code{--enable-libvidstab}.
20476 This filter accepts the following options:
20480 Set the path to the file used to write the transforms information.
20481 Default value is @file{transforms.trf}.
20484 Set how shaky the video is and how quick the camera is. It accepts an
20485 integer in the range 1-10, a value of 1 means little shakiness, a
20486 value of 10 means strong shakiness. Default value is 5.
20489 Set the accuracy of the detection process. It must be a value in the
20490 range 1-15. A value of 1 means low accuracy, a value of 15 means high
20491 accuracy. Default value is 15.
20494 Set stepsize of the search process. The region around minimum is
20495 scanned with 1 pixel resolution. Default value is 6.
20498 Set minimum contrast. Below this value a local measurement field is
20499 discarded. Must be a floating point value in the range 0-1. Default
20503 Set reference frame number for tripod mode.
20505 If enabled, the motion of the frames is compared to a reference frame
20506 in the filtered stream, identified by the specified number. The idea
20507 is to compensate all movements in a more-or-less static scene and keep
20508 the camera view absolutely still.
20510 If set to 0, it is disabled. The frames are counted starting from 1.
20513 Show fields and transforms in the resulting frames. It accepts an
20514 integer in the range 0-2. Default value is 0, which disables any
20518 @subsection Examples
20522 Use default values:
20528 Analyze strongly shaky movie and put the results in file
20529 @file{mytransforms.trf}:
20531 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
20535 Visualize the result of internal transformations in the resulting
20538 vidstabdetect=show=1
20542 Analyze a video with medium shakiness using @command{ffmpeg}:
20544 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
20548 @anchor{vidstabtransform}
20549 @section vidstabtransform
20551 Video stabilization/deshaking: pass 2 of 2,
20552 see @ref{vidstabdetect} for pass 1.
20554 Read a file with transform information for each frame and
20555 apply/compensate them. Together with the @ref{vidstabdetect}
20556 filter this can be used to deshake videos. See also
20557 @url{http://public.hronopik.de/vid.stab}. It is important to also use
20558 the @ref{unsharp} filter, see below.
20560 To enable compilation of this filter you need to configure FFmpeg with
20561 @code{--enable-libvidstab}.
20563 @subsection Options
20567 Set path to the file used to read the transforms. Default value is
20568 @file{transforms.trf}.
20571 Set the number of frames (value*2 + 1) used for lowpass filtering the
20572 camera movements. Default value is 10.
20574 For example a number of 10 means that 21 frames are used (10 in the
20575 past and 10 in the future) to smoothen the motion in the video. A
20576 larger value leads to a smoother video, but limits the acceleration of
20577 the camera (pan/tilt movements). 0 is a special case where a static
20578 camera is simulated.
20581 Set the camera path optimization algorithm.
20583 Accepted values are:
20586 gaussian kernel low-pass filter on camera motion (default)
20588 averaging on transformations
20592 Set maximal number of pixels to translate frames. Default value is -1,
20596 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
20597 value is -1, meaning no limit.
20600 Specify how to deal with borders that may be visible due to movement
20603 Available values are:
20606 keep image information from previous frame (default)
20608 fill the border black
20612 Invert transforms if set to 1. Default value is 0.
20615 Consider transforms as relative to previous frame if set to 1,
20616 absolute if set to 0. Default value is 0.
20619 Set percentage to zoom. A positive value will result in a zoom-in
20620 effect, a negative value in a zoom-out effect. Default value is 0 (no
20624 Set optimal zooming to avoid borders.
20626 Accepted values are:
20631 optimal static zoom value is determined (only very strong movements
20632 will lead to visible borders) (default)
20634 optimal adaptive zoom value is determined (no borders will be
20635 visible), see @option{zoomspeed}
20638 Note that the value given at zoom is added to the one calculated here.
20641 Set percent to zoom maximally each frame (enabled when
20642 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
20646 Specify type of interpolation.
20648 Available values are:
20653 linear only horizontal
20655 linear in both directions (default)
20657 cubic in both directions (slow)
20661 Enable virtual tripod mode if set to 1, which is equivalent to
20662 @code{relative=0:smoothing=0}. Default value is 0.
20664 Use also @code{tripod} option of @ref{vidstabdetect}.
20667 Increase log verbosity if set to 1. Also the detected global motions
20668 are written to the temporary file @file{global_motions.trf}. Default
20672 @subsection Examples
20676 Use @command{ffmpeg} for a typical stabilization with default values:
20678 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
20681 Note the use of the @ref{unsharp} filter which is always recommended.
20684 Zoom in a bit more and load transform data from a given file:
20686 vidstabtransform=zoom=5:input="mytransforms.trf"
20690 Smoothen the video even more:
20692 vidstabtransform=smoothing=30
20698 Flip the input video vertically.
20700 For example, to vertically flip a video with @command{ffmpeg}:
20702 ffmpeg -i in.avi -vf "vflip" out.avi
20707 Detect variable frame rate video.
20709 This filter tries to detect if the input is variable or constant frame rate.
20711 At end it will output number of frames detected as having variable delta pts,
20712 and ones with constant delta pts.
20713 If there was frames with variable delta, than it will also show min, max and
20714 average delta encountered.
20718 Boost or alter saturation.
20720 The filter accepts the following options:
20723 Set strength of boost if positive value or strength of alter if negative value.
20724 Default is 0. Allowed range is from -2 to 2.
20727 Set the red balance. Default is 1. Allowed range is from -10 to 10.
20730 Set the green balance. Default is 1. Allowed range is from -10 to 10.
20733 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
20736 Set the red luma coefficient.
20739 Set the green luma coefficient.
20742 Set the blue luma coefficient.
20745 If @code{intensity} is negative and this is set to 1, colors will change,
20746 otherwise colors will be less saturated, more towards gray.
20749 @subsection Commands
20751 This filter supports the all above options as @ref{commands}.
20756 Make or reverse a natural vignetting effect.
20758 The filter accepts the following options:
20762 Set lens angle expression as a number of radians.
20764 The value is clipped in the @code{[0,PI/2]} range.
20766 Default value: @code{"PI/5"}
20770 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
20774 Set forward/backward mode.
20776 Available modes are:
20779 The larger the distance from the central point, the darker the image becomes.
20782 The larger the distance from the central point, the brighter the image becomes.
20783 This can be used to reverse a vignette effect, though there is no automatic
20784 detection to extract the lens @option{angle} and other settings (yet). It can
20785 also be used to create a burning effect.
20788 Default value is @samp{forward}.
20791 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
20793 It accepts the following values:
20796 Evaluate expressions only once during the filter initialization.
20799 Evaluate expressions for each incoming frame. This is way slower than the
20800 @samp{init} mode since it requires all the scalers to be re-computed, but it
20801 allows advanced dynamic expressions.
20804 Default value is @samp{init}.
20807 Set dithering to reduce the circular banding effects. Default is @code{1}
20811 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
20812 Setting this value to the SAR of the input will make a rectangular vignetting
20813 following the dimensions of the video.
20815 Default is @code{1/1}.
20818 @subsection Expressions
20820 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
20821 following parameters.
20826 input width and height
20829 the number of input frame, starting from 0
20832 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
20833 @var{TB} units, NAN if undefined
20836 frame rate of the input video, NAN if the input frame rate is unknown
20839 the PTS (Presentation TimeStamp) of the filtered video frame,
20840 expressed in seconds, NAN if undefined
20843 time base of the input video
20847 @subsection Examples
20851 Apply simple strong vignetting effect:
20857 Make a flickering vignetting:
20859 vignette='PI/4+random(1)*PI/50':eval=frame
20864 @section vmafmotion
20866 Obtain the average VMAF motion score of a video.
20867 It is one of the component metrics of VMAF.
20869 The obtained average motion score is printed through the logging system.
20871 The filter accepts the following options:
20875 If specified, the filter will use the named file to save the motion score of
20876 each frame with respect to the previous frame.
20877 When filename equals "-" the data is sent to standard output.
20882 ffmpeg -i ref.mpg -vf vmafmotion -f null -
20886 Stack input videos vertically.
20888 All streams must be of same pixel format and of same width.
20890 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
20891 to create same output.
20893 The filter accepts the following options:
20897 Set number of input streams. Default is 2.
20900 If set to 1, force the output to terminate when the shortest input
20901 terminates. Default value is 0.
20906 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
20907 Deinterlacing Filter").
20909 Based on the process described by Martin Weston for BBC R&D, and
20910 implemented based on the de-interlace algorithm written by Jim
20911 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
20912 uses filter coefficients calculated by BBC R&D.
20914 This filter uses field-dominance information in frame to decide which
20915 of each pair of fields to place first in the output.
20916 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
20918 There are two sets of filter coefficients, so called "simple"
20919 and "complex". Which set of filter coefficients is used can
20920 be set by passing an optional parameter:
20924 Set the interlacing filter coefficients. Accepts one of the following values:
20928 Simple filter coefficient set.
20930 More-complex filter coefficient set.
20932 Default value is @samp{complex}.
20935 Specify which frames to deinterlace. Accepts one of the following values:
20939 Deinterlace all frames,
20941 Only deinterlace frames marked as interlaced.
20944 Default value is @samp{all}.
20947 @subsection Commands
20948 This filter supports same @ref{commands} as options.
20951 Video waveform monitor.
20953 The waveform monitor plots color component intensity. By default luminance
20954 only. Each column of the waveform corresponds to a column of pixels in the
20957 It accepts the following options:
20961 Can be either @code{row}, or @code{column}. Default is @code{column}.
20962 In row mode, the graph on the left side represents color component value 0 and
20963 the right side represents value = 255. In column mode, the top side represents
20964 color component value = 0 and bottom side represents value = 255.
20967 Set intensity. Smaller values are useful to find out how many values of the same
20968 luminance are distributed across input rows/columns.
20969 Default value is @code{0.04}. Allowed range is [0, 1].
20972 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
20973 In mirrored mode, higher values will be represented on the left
20974 side for @code{row} mode and at the top for @code{column} mode. Default is
20975 @code{1} (mirrored).
20979 It accepts the following values:
20982 Presents information identical to that in the @code{parade}, except
20983 that the graphs representing color components are superimposed directly
20986 This display mode makes it easier to spot relative differences or similarities
20987 in overlapping areas of the color components that are supposed to be identical,
20988 such as neutral whites, grays, or blacks.
20991 Display separate graph for the color components side by side in
20992 @code{row} mode or one below the other in @code{column} mode.
20995 Display separate graph for the color components side by side in
20996 @code{column} mode or one below the other in @code{row} mode.
20998 Using this display mode makes it easy to spot color casts in the highlights
20999 and shadows of an image, by comparing the contours of the top and the bottom
21000 graphs of each waveform. Since whites, grays, and blacks are characterized
21001 by exactly equal amounts of red, green, and blue, neutral areas of the picture
21002 should display three waveforms of roughly equal width/height. If not, the
21003 correction is easy to perform by making level adjustments the three waveforms.
21005 Default is @code{stack}.
21007 @item components, c
21008 Set which color components to display. Default is 1, which means only luminance
21009 or red color component if input is in RGB colorspace. If is set for example to
21010 7 it will display all 3 (if) available color components.
21015 No envelope, this is default.
21018 Instant envelope, minimum and maximum values presented in graph will be easily
21019 visible even with small @code{step} value.
21022 Hold minimum and maximum values presented in graph across time. This way you
21023 can still spot out of range values without constantly looking at waveforms.
21026 Peak and instant envelope combined together.
21032 No filtering, this is default.
21035 Luma and chroma combined together.
21038 Similar as above, but shows difference between blue and red chroma.
21041 Similar as above, but use different colors.
21044 Similar as above, but again with different colors.
21047 Displays only chroma.
21050 Displays actual color value on waveform.
21053 Similar as above, but with luma showing frequency of chroma values.
21057 Set which graticule to display.
21061 Do not display graticule.
21064 Display green graticule showing legal broadcast ranges.
21067 Display orange graticule showing legal broadcast ranges.
21070 Display invert graticule showing legal broadcast ranges.
21074 Set graticule opacity.
21077 Set graticule flags.
21081 Draw numbers above lines. By default enabled.
21084 Draw dots instead of lines.
21088 Set scale used for displaying graticule.
21095 Default is digital.
21098 Set background opacity.
21102 Set tint for output.
21103 Only used with lowpass filter and when display is not overlay and input
21104 pixel formats are not RGB.
21107 @section weave, doubleweave
21109 The @code{weave} takes a field-based video input and join
21110 each two sequential fields into single frame, producing a new double
21111 height clip with half the frame rate and half the frame count.
21113 The @code{doubleweave} works same as @code{weave} but without
21114 halving frame rate and frame count.
21116 It accepts the following option:
21120 Set first field. Available values are:
21124 Set the frame as top-field-first.
21127 Set the frame as bottom-field-first.
21131 @subsection Examples
21135 Interlace video using @ref{select} and @ref{separatefields} filter:
21137 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
21142 Apply the xBR high-quality magnification filter which is designed for pixel
21143 art. It follows a set of edge-detection rules, see
21144 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
21146 It accepts the following option:
21150 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
21151 @code{3xBR} and @code{4} for @code{4xBR}.
21152 Default is @code{3}.
21157 Apply cross fade from one input video stream to another input video stream.
21158 The cross fade is applied for specified duration.
21160 The filter accepts the following options:
21164 Set one of available transition effects:
21212 Default transition effect is fade.
21215 Set cross fade duration in seconds.
21216 Default duration is 1 second.
21219 Set cross fade start relative to first input stream in seconds.
21220 Default offset is 0.
21223 Set expression for custom transition effect.
21225 The expressions can use the following variables and functions:
21230 The coordinates of the current sample.
21234 The width and height of the image.
21237 Progress of transition effect.
21240 Currently processed plane.
21243 Return value of first input at current location and plane.
21246 Return value of second input at current location and plane.
21252 Return the value of the pixel at location (@var{x},@var{y}) of the
21253 first/second/third/fourth component of first input.
21259 Return the value of the pixel at location (@var{x},@var{y}) of the
21260 first/second/third/fourth component of second input.
21264 @subsection Examples
21268 Cross fade from one input video to another input video, with fade transition and duration of transition
21269 of 2 seconds starting at offset of 5 seconds:
21271 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
21276 Pick median pixels from several input videos.
21278 The filter accepts the following options:
21282 Set number of inputs.
21283 Default is 3. Allowed range is from 3 to 255.
21284 If number of inputs is even number, than result will be mean value between two median values.
21287 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
21290 Set median percentile. Default value is @code{0.5}.
21291 Default value of @code{0.5} will pick always median values, while @code{0} will pick
21292 minimum values, and @code{1} maximum values.
21295 @subsection Commands
21297 This filter supports all above options as @ref{commands}, excluding option @code{inputs}.
21300 Stack video inputs into custom layout.
21302 All streams must be of same pixel format.
21304 The filter accepts the following options:
21308 Set number of input streams. Default is 2.
21311 Specify layout of inputs.
21312 This option requires the desired layout configuration to be explicitly set by the user.
21313 This sets position of each video input in output. Each input
21314 is separated by '|'.
21315 The first number represents the column, and the second number represents the row.
21316 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
21317 where X is video input from which to take width or height.
21318 Multiple values can be used when separated by '+'. In such
21319 case values are summed together.
21321 Note that if inputs are of different sizes gaps may appear, as not all of
21322 the output video frame will be filled. Similarly, videos can overlap each
21323 other if their position doesn't leave enough space for the full frame of
21326 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
21327 a layout must be set by the user.
21330 If set to 1, force the output to terminate when the shortest input
21331 terminates. Default value is 0.
21334 If set to valid color, all unused pixels will be filled with that color.
21335 By default fill is set to none, so it is disabled.
21338 @subsection Examples
21342 Display 4 inputs into 2x2 grid.
21346 input1(0, 0) | input3(w0, 0)
21347 input2(0, h0) | input4(w0, h0)
21351 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
21354 Note that if inputs are of different sizes, gaps or overlaps may occur.
21357 Display 4 inputs into 1x4 grid.
21364 input4(0, h0+h1+h2)
21368 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
21371 Note that if inputs are of different widths, unused space will appear.
21374 Display 9 inputs into 3x3 grid.
21378 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
21379 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
21380 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
21384 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
21387 Note that if inputs are of different sizes, gaps or overlaps may occur.
21390 Display 16 inputs into 4x4 grid.
21394 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
21395 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
21396 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
21397 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
21401 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|
21402 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
21405 Note that if inputs are of different sizes, gaps or overlaps may occur.
21412 Deinterlace the input video ("yadif" means "yet another deinterlacing
21415 It accepts the following parameters:
21421 The interlacing mode to adopt. It accepts one of the following values:
21424 @item 0, send_frame
21425 Output one frame for each frame.
21426 @item 1, send_field
21427 Output one frame for each field.
21428 @item 2, send_frame_nospatial
21429 Like @code{send_frame}, but it skips the spatial interlacing check.
21430 @item 3, send_field_nospatial
21431 Like @code{send_field}, but it skips the spatial interlacing check.
21434 The default value is @code{send_frame}.
21437 The picture field parity assumed for the input interlaced video. It accepts one
21438 of the following values:
21442 Assume the top field is first.
21444 Assume the bottom field is first.
21446 Enable automatic detection of field parity.
21449 The default value is @code{auto}.
21450 If the interlacing is unknown or the decoder does not export this information,
21451 top field first will be assumed.
21454 Specify which frames to deinterlace. Accepts one of the following
21459 Deinterlace all frames.
21460 @item 1, interlaced
21461 Only deinterlace frames marked as interlaced.
21464 The default value is @code{all}.
21467 @section yadif_cuda
21469 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
21470 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
21473 It accepts the following parameters:
21479 The interlacing mode to adopt. It accepts one of the following values:
21482 @item 0, send_frame
21483 Output one frame for each frame.
21484 @item 1, send_field
21485 Output one frame for each field.
21486 @item 2, send_frame_nospatial
21487 Like @code{send_frame}, but it skips the spatial interlacing check.
21488 @item 3, send_field_nospatial
21489 Like @code{send_field}, but it skips the spatial interlacing check.
21492 The default value is @code{send_frame}.
21495 The picture field parity assumed for the input interlaced video. It accepts one
21496 of the following values:
21500 Assume the top field is first.
21502 Assume the bottom field is first.
21504 Enable automatic detection of field parity.
21507 The default value is @code{auto}.
21508 If the interlacing is unknown or the decoder does not export this information,
21509 top field first will be assumed.
21512 Specify which frames to deinterlace. Accepts one of the following
21517 Deinterlace all frames.
21518 @item 1, interlaced
21519 Only deinterlace frames marked as interlaced.
21522 The default value is @code{all}.
21527 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
21528 The algorithm is described in
21529 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
21531 It accepts the following parameters:
21535 Set the window radius. Default value is 3.
21538 Set which planes to filter. Default is only the first plane.
21541 Set blur strength. Default value is 128.
21544 @subsection Commands
21545 This filter supports same @ref{commands} as options.
21549 Apply Zoom & Pan effect.
21551 This filter accepts the following options:
21555 Set the zoom expression. Range is 1-10. Default is 1.
21559 Set the x and y expression. Default is 0.
21562 Set the duration expression in number of frames.
21563 This sets for how many number of frames effect will last for
21564 single input image.
21567 Set the output image size, default is 'hd720'.
21570 Set the output frame rate, default is '25'.
21573 Each expression can contain the following constants:
21592 Output frame count.
21595 The input timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
21597 @item out_time, time, ot
21598 The output timestamp expressed in seconds.
21602 Last calculated 'x' and 'y' position from 'x' and 'y' expression
21603 for current input frame.
21607 'x' and 'y' of last output frame of previous input frame or 0 when there was
21608 not yet such frame (first input frame).
21611 Last calculated zoom from 'z' expression for current input frame.
21614 Last calculated zoom of last output frame of previous input frame.
21617 Number of output frames for current input frame. Calculated from 'd' expression
21618 for each input frame.
21621 number of output frames created for previous input frame
21624 Rational number: input width / input height
21627 sample aspect ratio
21630 display aspect ratio
21634 @subsection Examples
21638 Zoom in up to 1.5x and pan at same time to some spot near center of picture:
21640 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
21644 Zoom in up to 1.5x and pan always at center of picture:
21646 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21650 Same as above but without pausing:
21652 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21656 Zoom in 2x into center of picture only for the first second of the input video:
21658 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21665 Scale (resize) the input video, using the z.lib library:
21666 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
21667 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
21669 The zscale filter forces the output display aspect ratio to be the same
21670 as the input, by changing the output sample aspect ratio.
21672 If the input image format is different from the format requested by
21673 the next filter, the zscale filter will convert the input to the
21676 @subsection Options
21677 The filter accepts the following options.
21682 Set the output video dimension expression. Default value is the input
21685 If the @var{width} or @var{w} value is 0, the input width is used for
21686 the output. If the @var{height} or @var{h} value is 0, the input height
21687 is used for the output.
21689 If one and only one of the values is -n with n >= 1, the zscale filter
21690 will use a value that maintains the aspect ratio of the input image,
21691 calculated from the other specified dimension. After that it will,
21692 however, make sure that the calculated dimension is divisible by n and
21693 adjust the value if necessary.
21695 If both values are -n with n >= 1, the behavior will be identical to
21696 both values being set to 0 as previously detailed.
21698 See below for the list of accepted constants for use in the dimension
21702 Set the video size. For the syntax of this option, check the
21703 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21706 Set the dither type.
21708 Possible values are:
21713 @item error_diffusion
21719 Set the resize filter type.
21721 Possible values are:
21731 Default is bilinear.
21734 Set the color range.
21736 Possible values are:
21743 Default is same as input.
21746 Set the color primaries.
21748 Possible values are:
21758 Default is same as input.
21761 Set the transfer characteristics.
21763 Possible values are:
21777 Default is same as input.
21780 Set the colorspace matrix.
21782 Possible value are:
21793 Default is same as input.
21796 Set the input color range.
21798 Possible values are:
21805 Default is same as input.
21807 @item primariesin, pin
21808 Set the input color primaries.
21810 Possible values are:
21820 Default is same as input.
21822 @item transferin, tin
21823 Set the input transfer characteristics.
21825 Possible values are:
21836 Default is same as input.
21838 @item matrixin, min
21839 Set the input colorspace matrix.
21841 Possible value are:
21853 Set the output chroma location.
21855 Possible values are:
21866 @item chromalin, cin
21867 Set the input chroma location.
21869 Possible values are:
21881 Set the nominal peak luminance.
21884 The values of the @option{w} and @option{h} options are expressions
21885 containing the following constants:
21890 The input width and height
21894 These are the same as @var{in_w} and @var{in_h}.
21898 The output (scaled) width and height
21902 These are the same as @var{out_w} and @var{out_h}
21905 The same as @var{iw} / @var{ih}
21908 input sample aspect ratio
21911 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
21915 horizontal and vertical input chroma subsample values. For example for the
21916 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21920 horizontal and vertical output chroma subsample values. For example for the
21921 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21924 @subsection Commands
21926 This filter supports the following commands:
21930 Set the output video dimension expression.
21931 The command accepts the same syntax of the corresponding option.
21933 If the specified expression is not valid, it is kept at its current
21937 @c man end VIDEO FILTERS
21939 @chapter OpenCL Video Filters
21940 @c man begin OPENCL VIDEO FILTERS
21942 Below is a description of the currently available OpenCL video filters.
21944 To enable compilation of these filters you need to configure FFmpeg with
21945 @code{--enable-opencl}.
21947 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
21950 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
21951 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
21952 given device parameters.
21954 @item -filter_hw_device @var{name}
21955 Pass the hardware device called @var{name} to all filters in any filter graph.
21959 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
21963 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
21965 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
21969 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.
21971 @section avgblur_opencl
21973 Apply average blur filter.
21975 The filter accepts the following options:
21979 Set horizontal radius size.
21980 Range is @code{[1, 1024]} and default value is @code{1}.
21983 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21986 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
21989 @subsection Example
21993 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.
21995 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
21999 @section boxblur_opencl
22001 Apply a boxblur algorithm to the input video.
22003 It accepts the following parameters:
22007 @item luma_radius, lr
22008 @item luma_power, lp
22009 @item chroma_radius, cr
22010 @item chroma_power, cp
22011 @item alpha_radius, ar
22012 @item alpha_power, ap
22016 A description of the accepted options follows.
22019 @item luma_radius, lr
22020 @item chroma_radius, cr
22021 @item alpha_radius, ar
22022 Set an expression for the box radius in pixels used for blurring the
22023 corresponding input plane.
22025 The radius value must be a non-negative number, and must not be
22026 greater than the value of the expression @code{min(w,h)/2} for the
22027 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
22030 Default value for @option{luma_radius} is "2". If not specified,
22031 @option{chroma_radius} and @option{alpha_radius} default to the
22032 corresponding value set for @option{luma_radius}.
22034 The expressions can contain the following constants:
22038 The input width and height in pixels.
22042 The input chroma image width and height in pixels.
22046 The horizontal and vertical chroma subsample values. For example, for the
22047 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
22050 @item luma_power, lp
22051 @item chroma_power, cp
22052 @item alpha_power, ap
22053 Specify how many times the boxblur filter is applied to the
22054 corresponding plane.
22056 Default value for @option{luma_power} is 2. If not specified,
22057 @option{chroma_power} and @option{alpha_power} default to the
22058 corresponding value set for @option{luma_power}.
22060 A value of 0 will disable the effect.
22063 @subsection Examples
22065 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.
22069 Apply a boxblur filter with the luma, chroma, and alpha radius
22070 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.
22072 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
22073 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
22077 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.
22079 For the luma plane, a 2x2 box radius will be run once.
22081 For the chroma plane, a 4x4 box radius will be run 5 times.
22083 For the alpha plane, a 3x3 box radius will be run 7 times.
22085 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
22089 @section colorkey_opencl
22090 RGB colorspace color keying.
22092 The filter accepts the following options:
22096 The color which will be replaced with transparency.
22099 Similarity percentage with the key color.
22101 0.01 matches only the exact key color, while 1.0 matches everything.
22106 0.0 makes pixels either fully transparent, or not transparent at all.
22108 Higher values result in semi-transparent pixels, with a higher transparency
22109 the more similar the pixels color is to the key color.
22112 @subsection Examples
22116 Make every semi-green pixel in the input transparent with some slight blending:
22118 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
22122 @section convolution_opencl
22124 Apply convolution of 3x3, 5x5, 7x7 matrix.
22126 The filter accepts the following options:
22133 Set matrix for each plane.
22134 Matrix is sequence of 9, 25 or 49 signed numbers.
22135 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
22141 Set multiplier for calculated value for each plane.
22142 If unset or 0, it will be sum of all matrix elements.
22143 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
22149 Set bias for each plane. This value is added to the result of the multiplication.
22150 Useful for making the overall image brighter or darker.
22151 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
22155 @subsection Examples
22161 -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
22167 -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
22171 Apply edge enhance:
22173 -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
22179 -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
22183 Apply laplacian edge detector which includes diagonals:
22185 -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
22191 -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
22195 @section erosion_opencl
22197 Apply erosion effect to the video.
22199 This filter replaces the pixel by the local(3x3) minimum.
22201 It accepts the following options:
22208 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22209 If @code{0}, plane will remain unchanged.
22212 Flag which specifies the pixel to refer to.
22213 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22215 Flags to local 3x3 coordinates region centered on @code{x}:
22224 @subsection Example
22228 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.
22230 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22234 @section deshake_opencl
22235 Feature-point based video stabilization filter.
22237 The filter accepts the following options:
22241 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
22244 Whether or not additional debug info should be displayed, both in the processed output and in the console.
22246 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
22248 Viewing point matches in the output video is only supported for RGB input.
22250 Defaults to @code{0}.
22252 @item adaptive_crop
22253 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
22255 Defaults to @code{1}.
22257 @item refine_features
22258 Whether or not feature points should be refined at a sub-pixel level.
22260 This can be turned off for a slight performance gain at the cost of precision.
22262 Defaults to @code{1}.
22264 @item smooth_strength
22265 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
22267 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
22269 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
22271 Defaults to @code{0.0}.
22273 @item smooth_window_multiplier
22274 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
22276 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
22278 Acceptable values range from @code{0.1} to @code{10.0}.
22280 Larger values increase the amount of motion data available for determining how to smooth the camera path,
22281 potentially improving smoothness, but also increase latency and memory usage.
22283 Defaults to @code{2.0}.
22287 @subsection Examples
22291 Stabilize a video with a fixed, medium smoothing strength:
22293 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
22297 Stabilize a video with debugging (both in console and in rendered video):
22299 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
22303 @section dilation_opencl
22305 Apply dilation effect to the video.
22307 This filter replaces the pixel by the local(3x3) maximum.
22309 It accepts the following options:
22316 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22317 If @code{0}, plane will remain unchanged.
22320 Flag which specifies the pixel to refer to.
22321 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22323 Flags to local 3x3 coordinates region centered on @code{x}:
22332 @subsection Example
22336 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.
22338 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22342 @section nlmeans_opencl
22344 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
22346 @section overlay_opencl
22348 Overlay one video on top of another.
22350 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
22351 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
22353 The filter accepts the following options:
22358 Set the x coordinate of the overlaid video on the main video.
22359 Default value is @code{0}.
22362 Set the y coordinate of the overlaid video on the main video.
22363 Default value is @code{0}.
22367 @subsection Examples
22371 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
22373 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22376 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
22378 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22383 @section pad_opencl
22385 Add paddings to the input image, and place the original input at the
22386 provided @var{x}, @var{y} coordinates.
22388 It accepts the following options:
22393 Specify an expression for the size of the output image with the
22394 paddings added. If the value for @var{width} or @var{height} is 0, the
22395 corresponding input size is used for the output.
22397 The @var{width} expression can reference the value set by the
22398 @var{height} expression, and vice versa.
22400 The default value of @var{width} and @var{height} is 0.
22404 Specify the offsets to place the input image at within the padded area,
22405 with respect to the top/left border of the output image.
22407 The @var{x} expression can reference the value set by the @var{y}
22408 expression, and vice versa.
22410 The default value of @var{x} and @var{y} is 0.
22412 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
22413 so the input image is centered on the padded area.
22416 Specify the color of the padded area. For the syntax of this option,
22417 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
22418 manual,ffmpeg-utils}.
22421 Pad to an aspect instead to a resolution.
22424 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
22425 options are expressions containing the following constants:
22430 The input video width and height.
22434 These are the same as @var{in_w} and @var{in_h}.
22438 The output width and height (the size of the padded area), as
22439 specified by the @var{width} and @var{height} expressions.
22443 These are the same as @var{out_w} and @var{out_h}.
22447 The x and y offsets as specified by the @var{x} and @var{y}
22448 expressions, or NAN if not yet specified.
22451 same as @var{iw} / @var{ih}
22454 input sample aspect ratio
22457 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
22460 @section prewitt_opencl
22462 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
22464 The filter accepts the following option:
22468 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22471 Set value which will be multiplied with filtered result.
22472 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22475 Set value which will be added to filtered result.
22476 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22479 @subsection Example
22483 Apply the Prewitt operator with scale set to 2 and delta set to 10.
22485 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
22489 @anchor{program_opencl}
22490 @section program_opencl
22492 Filter video using an OpenCL program.
22497 OpenCL program source file.
22500 Kernel name in program.
22503 Number of inputs to the filter. Defaults to 1.
22506 Size of output frames. Defaults to the same as the first input.
22510 The @code{program_opencl} filter also supports the @ref{framesync} options.
22512 The program source file must contain a kernel function with the given name,
22513 which will be run once for each plane of the output. Each run on a plane
22514 gets enqueued as a separate 2D global NDRange with one work-item for each
22515 pixel to be generated. The global ID offset for each work-item is therefore
22516 the coordinates of a pixel in the destination image.
22518 The kernel function needs to take the following arguments:
22521 Destination image, @var{__write_only image2d_t}.
22523 This image will become the output; the kernel should write all of it.
22525 Frame index, @var{unsigned int}.
22527 This is a counter starting from zero and increasing by one for each frame.
22529 Source images, @var{__read_only image2d_t}.
22531 These are the most recent images on each input. The kernel may read from
22532 them to generate the output, but they can't be written to.
22539 Copy the input to the output (output must be the same size as the input).
22541 __kernel void copy(__write_only image2d_t destination,
22542 unsigned int index,
22543 __read_only image2d_t source)
22545 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
22547 int2 location = (int2)(get_global_id(0), get_global_id(1));
22549 float4 value = read_imagef(source, sampler, location);
22551 write_imagef(destination, location, value);
22556 Apply a simple transformation, rotating the input by an amount increasing
22557 with the index counter. Pixel values are linearly interpolated by the
22558 sampler, and the output need not have the same dimensions as the input.
22560 __kernel void rotate_image(__write_only image2d_t dst,
22561 unsigned int index,
22562 __read_only image2d_t src)
22564 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22565 CLK_FILTER_LINEAR);
22567 float angle = (float)index / 100.0f;
22569 float2 dst_dim = convert_float2(get_image_dim(dst));
22570 float2 src_dim = convert_float2(get_image_dim(src));
22572 float2 dst_cen = dst_dim / 2.0f;
22573 float2 src_cen = src_dim / 2.0f;
22575 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22577 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
22579 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
22580 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
22582 src_pos = src_pos * src_dim / dst_dim;
22584 float2 src_loc = src_pos + src_cen;
22586 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
22587 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
22588 write_imagef(dst, dst_loc, 0.5f);
22590 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
22595 Blend two inputs together, with the amount of each input used varying
22596 with the index counter.
22598 __kernel void blend_images(__write_only image2d_t dst,
22599 unsigned int index,
22600 __read_only image2d_t src1,
22601 __read_only image2d_t src2)
22603 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22604 CLK_FILTER_LINEAR);
22606 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
22608 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22609 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
22610 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
22612 float4 val1 = read_imagef(src1, sampler, src1_loc);
22613 float4 val2 = read_imagef(src2, sampler, src2_loc);
22615 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
22621 @section roberts_opencl
22622 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
22624 The filter accepts the following option:
22628 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22631 Set value which will be multiplied with filtered result.
22632 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22635 Set value which will be added to filtered result.
22636 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22639 @subsection Example
22643 Apply the Roberts cross operator with scale set to 2 and delta set to 10
22645 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
22649 @section sobel_opencl
22651 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
22653 The filter accepts the following option:
22657 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22660 Set value which will be multiplied with filtered result.
22661 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22664 Set value which will be added to filtered result.
22665 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22668 @subsection Example
22672 Apply sobel operator with scale set to 2 and delta set to 10
22674 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
22678 @section tonemap_opencl
22680 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
22682 It accepts the following parameters:
22686 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
22689 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
22692 Apply desaturation for highlights that exceed this level of brightness. The
22693 higher the parameter, the more color information will be preserved. This
22694 setting helps prevent unnaturally blown-out colors for super-highlights, by
22695 (smoothly) turning into white instead. This makes images feel more natural,
22696 at the cost of reducing information about out-of-range colors.
22698 The default value is 0.5, and the algorithm here is a little different from
22699 the cpu version tonemap currently. A setting of 0.0 disables this option.
22702 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
22703 is used to detect whether the scene has changed or not. If the distance between
22704 the current frame average brightness and the current running average exceeds
22705 a threshold value, we would re-calculate scene average and peak brightness.
22706 The default value is 0.2.
22709 Specify the output pixel format.
22711 Currently supported formats are:
22718 Set the output color range.
22720 Possible values are:
22726 Default is same as input.
22729 Set the output color primaries.
22731 Possible values are:
22737 Default is same as input.
22740 Set the output transfer characteristics.
22742 Possible values are:
22751 Set the output colorspace matrix.
22753 Possible value are:
22759 Default is same as input.
22763 @subsection Example
22767 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
22769 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
22773 @section unsharp_opencl
22775 Sharpen or blur the input video.
22777 It accepts the following parameters:
22780 @item luma_msize_x, lx
22781 Set the luma matrix horizontal size.
22782 Range is @code{[1, 23]} and default value is @code{5}.
22784 @item luma_msize_y, ly
22785 Set the luma matrix vertical size.
22786 Range is @code{[1, 23]} and default value is @code{5}.
22788 @item luma_amount, la
22789 Set the luma effect strength.
22790 Range is @code{[-10, 10]} and default value is @code{1.0}.
22792 Negative values will blur the input video, while positive values will
22793 sharpen it, a value of zero will disable the effect.
22795 @item chroma_msize_x, cx
22796 Set the chroma matrix horizontal size.
22797 Range is @code{[1, 23]} and default value is @code{5}.
22799 @item chroma_msize_y, cy
22800 Set the chroma matrix vertical size.
22801 Range is @code{[1, 23]} and default value is @code{5}.
22803 @item chroma_amount, ca
22804 Set the chroma effect strength.
22805 Range is @code{[-10, 10]} and default value is @code{0.0}.
22807 Negative values will blur the input video, while positive values will
22808 sharpen it, a value of zero will disable the effect.
22812 All parameters are optional and default to the equivalent of the
22813 string '5:5:1.0:5:5:0.0'.
22815 @subsection Examples
22819 Apply strong luma sharpen effect:
22821 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
22825 Apply a strong blur of both luma and chroma parameters:
22827 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
22831 @section xfade_opencl
22833 Cross fade two videos with custom transition effect by using OpenCL.
22835 It accepts the following options:
22839 Set one of possible transition effects.
22843 Select custom transition effect, the actual transition description
22844 will be picked from source and kernel options.
22856 Default transition is fade.
22860 OpenCL program source file for custom transition.
22863 Set name of kernel to use for custom transition from program source file.
22866 Set duration of video transition.
22869 Set time of start of transition relative to first video.
22872 The program source file must contain a kernel function with the given name,
22873 which will be run once for each plane of the output. Each run on a plane
22874 gets enqueued as a separate 2D global NDRange with one work-item for each
22875 pixel to be generated. The global ID offset for each work-item is therefore
22876 the coordinates of a pixel in the destination image.
22878 The kernel function needs to take the following arguments:
22881 Destination image, @var{__write_only image2d_t}.
22883 This image will become the output; the kernel should write all of it.
22886 First Source image, @var{__read_only image2d_t}.
22887 Second Source image, @var{__read_only image2d_t}.
22889 These are the most recent images on each input. The kernel may read from
22890 them to generate the output, but they can't be written to.
22893 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
22900 Apply dots curtain transition effect:
22902 __kernel void blend_images(__write_only image2d_t dst,
22903 __read_only image2d_t src1,
22904 __read_only image2d_t src2,
22907 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22908 CLK_FILTER_LINEAR);
22909 int2 p = (int2)(get_global_id(0), get_global_id(1));
22910 float2 rp = (float2)(get_global_id(0), get_global_id(1));
22911 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
22914 float2 dots = (float2)(20.0, 20.0);
22915 float2 center = (float2)(0,0);
22918 float4 val1 = read_imagef(src1, sampler, p);
22919 float4 val2 = read_imagef(src2, sampler, p);
22920 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
22922 write_imagef(dst, p, next ? val1 : val2);
22928 @c man end OPENCL VIDEO FILTERS
22930 @chapter VAAPI Video Filters
22931 @c man begin VAAPI VIDEO FILTERS
22933 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
22935 To enable compilation of these filters you need to configure FFmpeg with
22936 @code{--enable-vaapi}.
22938 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}
22940 @section tonemap_vaapi
22942 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
22943 It maps the dynamic range of HDR10 content to the SDR content.
22944 It currently only accepts HDR10 as input.
22946 It accepts the following parameters:
22950 Specify the output pixel format.
22952 Currently supported formats are:
22961 Set the output color primaries.
22963 Default is same as input.
22966 Set the output transfer characteristics.
22971 Set the output colorspace matrix.
22973 Default is same as input.
22977 @subsection Example
22981 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
22983 tonemap_vaapi=format=p010:t=bt2020-10
22987 @c man end VAAPI VIDEO FILTERS
22989 @chapter Video Sources
22990 @c man begin VIDEO SOURCES
22992 Below is a description of the currently available video sources.
22996 Buffer video frames, and make them available to the filter chain.
22998 This source is mainly intended for a programmatic use, in particular
22999 through the interface defined in @file{libavfilter/buffersrc.h}.
23001 It accepts the following parameters:
23006 Specify the size (width and height) of the buffered video frames. For the
23007 syntax of this option, check the
23008 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23011 The input video width.
23014 The input video height.
23017 A string representing the pixel format of the buffered video frames.
23018 It may be a number corresponding to a pixel format, or a pixel format
23022 Specify the timebase assumed by the timestamps of the buffered frames.
23025 Specify the frame rate expected for the video stream.
23027 @item pixel_aspect, sar
23028 The sample (pixel) aspect ratio of the input video.
23031 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
23032 to the filtergraph description to specify swscale flags for automatically
23033 inserted scalers. See @ref{Filtergraph syntax}.
23035 @item hw_frames_ctx
23036 When using a hardware pixel format, this should be a reference to an
23037 AVHWFramesContext describing input frames.
23042 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
23045 will instruct the source to accept video frames with size 320x240 and
23046 with format "yuv410p", assuming 1/24 as the timestamps timebase and
23047 square pixels (1:1 sample aspect ratio).
23048 Since the pixel format with name "yuv410p" corresponds to the number 6
23049 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
23050 this example corresponds to:
23052 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
23055 Alternatively, the options can be specified as a flat string, but this
23056 syntax is deprecated:
23058 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
23062 Create a pattern generated by an elementary cellular automaton.
23064 The initial state of the cellular automaton can be defined through the
23065 @option{filename} and @option{pattern} options. If such options are
23066 not specified an initial state is created randomly.
23068 At each new frame a new row in the video is filled with the result of
23069 the cellular automaton next generation. The behavior when the whole
23070 frame is filled is defined by the @option{scroll} option.
23072 This source accepts the following options:
23076 Read the initial cellular automaton state, i.e. the starting row, from
23077 the specified file.
23078 In the file, each non-whitespace character is considered an alive
23079 cell, a newline will terminate the row, and further characters in the
23080 file will be ignored.
23083 Read the initial cellular automaton state, i.e. the starting row, from
23084 the specified string.
23086 Each non-whitespace character in the string is considered an alive
23087 cell, a newline will terminate the row, and further characters in the
23088 string will be ignored.
23091 Set the video rate, that is the number of frames generated per second.
23094 @item random_fill_ratio, ratio
23095 Set the random fill ratio for the initial cellular automaton row. It
23096 is a floating point number value ranging from 0 to 1, defaults to
23099 This option is ignored when a file or a pattern is specified.
23101 @item random_seed, seed
23102 Set the seed for filling randomly the initial row, must be an integer
23103 included between 0 and UINT32_MAX. If not specified, or if explicitly
23104 set to -1, the filter will try to use a good random seed on a best
23108 Set the cellular automaton rule, it is a number ranging from 0 to 255.
23109 Default value is 110.
23112 Set the size of the output video. For the syntax of this option, check the
23113 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23115 If @option{filename} or @option{pattern} is specified, the size is set
23116 by default to the width of the specified initial state row, and the
23117 height is set to @var{width} * PHI.
23119 If @option{size} is set, it must contain the width of the specified
23120 pattern string, and the specified pattern will be centered in the
23123 If a filename or a pattern string is not specified, the size value
23124 defaults to "320x518" (used for a randomly generated initial state).
23127 If set to 1, scroll the output upward when all the rows in the output
23128 have been already filled. If set to 0, the new generated row will be
23129 written over the top row just after the bottom row is filled.
23132 @item start_full, full
23133 If set to 1, completely fill the output with generated rows before
23134 outputting the first frame.
23135 This is the default behavior, for disabling set the value to 0.
23138 If set to 1, stitch the left and right row edges together.
23139 This is the default behavior, for disabling set the value to 0.
23142 @subsection Examples
23146 Read the initial state from @file{pattern}, and specify an output of
23149 cellauto=f=pattern:s=200x400
23153 Generate a random initial row with a width of 200 cells, with a fill
23156 cellauto=ratio=2/3:s=200x200
23160 Create a pattern generated by rule 18 starting by a single alive cell
23161 centered on an initial row with width 100:
23163 cellauto=p=@@:s=100x400:full=0:rule=18
23167 Specify a more elaborated initial pattern:
23169 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
23174 @anchor{coreimagesrc}
23175 @section coreimagesrc
23176 Video source generated on GPU using Apple's CoreImage API on OSX.
23178 This video source is a specialized version of the @ref{coreimage} video filter.
23179 Use a core image generator at the beginning of the applied filterchain to
23180 generate the content.
23182 The coreimagesrc video source accepts the following options:
23184 @item list_generators
23185 List all available generators along with all their respective options as well as
23186 possible minimum and maximum values along with the default values.
23188 list_generators=true
23192 Specify the size of the sourced video. For the syntax of this option, check the
23193 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23194 The default value is @code{320x240}.
23197 Specify the frame rate of the sourced video, as the number of frames
23198 generated per second. It has to be a string in the format
23199 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23200 number or a valid video frame rate abbreviation. The default value is
23204 Set the sample aspect ratio of the sourced video.
23207 Set the duration of the sourced video. See
23208 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23209 for the accepted syntax.
23211 If not specified, or the expressed duration is negative, the video is
23212 supposed to be generated forever.
23215 Additionally, all options of the @ref{coreimage} video filter are accepted.
23216 A complete filterchain can be used for further processing of the
23217 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
23218 and examples for details.
23220 @subsection Examples
23225 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
23226 given as complete and escaped command-line for Apple's standard bash shell:
23228 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
23230 This example is equivalent to the QRCode example of @ref{coreimage} without the
23231 need for a nullsrc video source.
23236 Generate several gradients.
23240 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23241 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23244 Set frame rate, expressed as number of frames per second. Default
23247 @item c0, c1, c2, c3, c4, c5, c6, c7
23248 Set 8 colors. Default values for colors is to pick random one.
23250 @item x0, y0, y0, y1
23251 Set gradient line source and destination points. If negative or out of range, random ones
23255 Set number of colors to use at once. Allowed range is from 2 to 8. Default value is 2.
23258 Set seed for picking gradient line points.
23261 Set the duration of the sourced video. See
23262 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23263 for the accepted syntax.
23265 If not specified, or the expressed duration is negative, the video is
23266 supposed to be generated forever.
23269 Set speed of gradients rotation.
23273 @section mandelbrot
23275 Generate a Mandelbrot set fractal, and progressively zoom towards the
23276 point specified with @var{start_x} and @var{start_y}.
23278 This source accepts the following options:
23283 Set the terminal pts value. Default value is 400.
23286 Set the terminal scale value.
23287 Must be a floating point value. Default value is 0.3.
23290 Set the inner coloring mode, that is the algorithm used to draw the
23291 Mandelbrot fractal internal region.
23293 It shall assume one of the following values:
23298 Show time until convergence.
23300 Set color based on point closest to the origin of the iterations.
23305 Default value is @var{mincol}.
23308 Set the bailout value. Default value is 10.0.
23311 Set the maximum of iterations performed by the rendering
23312 algorithm. Default value is 7189.
23315 Set outer coloring mode.
23316 It shall assume one of following values:
23318 @item iteration_count
23319 Set iteration count mode.
23320 @item normalized_iteration_count
23321 set normalized iteration count mode.
23323 Default value is @var{normalized_iteration_count}.
23326 Set frame rate, expressed as number of frames per second. Default
23330 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23331 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23334 Set the initial scale value. Default value is 3.0.
23337 Set the initial x position. Must be a floating point value between
23338 -100 and 100. Default value is -0.743643887037158704752191506114774.
23341 Set the initial y position. Must be a floating point value between
23342 -100 and 100. Default value is -0.131825904205311970493132056385139.
23347 Generate various test patterns, as generated by the MPlayer test filter.
23349 The size of the generated video is fixed, and is 256x256.
23350 This source is useful in particular for testing encoding features.
23352 This source accepts the following options:
23357 Specify the frame rate of the sourced video, as the number of frames
23358 generated per second. It has to be a string in the format
23359 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23360 number or a valid video frame rate abbreviation. The default value is
23364 Set the duration of the sourced video. See
23365 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23366 for the accepted syntax.
23368 If not specified, or the expressed duration is negative, the video is
23369 supposed to be generated forever.
23373 Set the number or the name of the test to perform. Supported tests are:
23387 @item max_frames, m
23388 Set the maximum number of frames generated for each test, default value is 30.
23392 Default value is "all", which will cycle through the list of all tests.
23397 mptestsrc=t=dc_luma
23400 will generate a "dc_luma" test pattern.
23402 @section frei0r_src
23404 Provide a frei0r source.
23406 To enable compilation of this filter you need to install the frei0r
23407 header and configure FFmpeg with @code{--enable-frei0r}.
23409 This source accepts the following parameters:
23414 The size of the video to generate. For the syntax of this option, check the
23415 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23418 The framerate of the generated video. It may be a string of the form
23419 @var{num}/@var{den} or a frame rate abbreviation.
23422 The name to the frei0r source to load. For more information regarding frei0r and
23423 how to set the parameters, read the @ref{frei0r} section in the video filters
23426 @item filter_params
23427 A '|'-separated list of parameters to pass to the frei0r source.
23431 For example, to generate a frei0r partik0l source with size 200x200
23432 and frame rate 10 which is overlaid on the overlay filter main input:
23434 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
23439 Generate a life pattern.
23441 This source is based on a generalization of John Conway's life game.
23443 The sourced input represents a life grid, each pixel represents a cell
23444 which can be in one of two possible states, alive or dead. Every cell
23445 interacts with its eight neighbours, which are the cells that are
23446 horizontally, vertically, or diagonally adjacent.
23448 At each interaction the grid evolves according to the adopted rule,
23449 which specifies the number of neighbor alive cells which will make a
23450 cell stay alive or born. The @option{rule} option allows one to specify
23453 This source accepts the following options:
23457 Set the file from which to read the initial grid state. In the file,
23458 each non-whitespace character is considered an alive cell, and newline
23459 is used to delimit the end of each row.
23461 If this option is not specified, the initial grid is generated
23465 Set the video rate, that is the number of frames generated per second.
23468 @item random_fill_ratio, ratio
23469 Set the random fill ratio for the initial random grid. It is a
23470 floating point number value ranging from 0 to 1, defaults to 1/PHI.
23471 It is ignored when a file is specified.
23473 @item random_seed, seed
23474 Set the seed for filling the initial random grid, must be an integer
23475 included between 0 and UINT32_MAX. If not specified, or if explicitly
23476 set to -1, the filter will try to use a good random seed on a best
23482 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
23483 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
23484 @var{NS} specifies the number of alive neighbor cells which make a
23485 live cell stay alive, and @var{NB} the number of alive neighbor cells
23486 which make a dead cell to become alive (i.e. to "born").
23487 "s" and "b" can be used in place of "S" and "B", respectively.
23489 Alternatively a rule can be specified by an 18-bits integer. The 9
23490 high order bits are used to encode the next cell state if it is alive
23491 for each number of neighbor alive cells, the low order bits specify
23492 the rule for "borning" new cells. Higher order bits encode for an
23493 higher number of neighbor cells.
23494 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
23495 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
23497 Default value is "S23/B3", which is the original Conway's game of life
23498 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
23499 cells, and will born a new cell if there are three alive cells around
23503 Set the size of the output video. For the syntax of this option, check the
23504 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23506 If @option{filename} is specified, the size is set by default to the
23507 same size of the input file. If @option{size} is set, it must contain
23508 the size specified in the input file, and the initial grid defined in
23509 that file is centered in the larger resulting area.
23511 If a filename is not specified, the size value defaults to "320x240"
23512 (used for a randomly generated initial grid).
23515 If set to 1, stitch the left and right grid edges together, and the
23516 top and bottom edges also. Defaults to 1.
23519 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
23520 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
23521 value from 0 to 255.
23524 Set the color of living (or new born) cells.
23527 Set the color of dead cells. If @option{mold} is set, this is the first color
23528 used to represent a dead cell.
23531 Set mold color, for definitely dead and moldy cells.
23533 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
23534 ffmpeg-utils manual,ffmpeg-utils}.
23537 @subsection Examples
23541 Read a grid from @file{pattern}, and center it on a grid of size
23544 life=f=pattern:s=300x300
23548 Generate a random grid of size 200x200, with a fill ratio of 2/3:
23550 life=ratio=2/3:s=200x200
23554 Specify a custom rule for evolving a randomly generated grid:
23560 Full example with slow death effect (mold) using @command{ffplay}:
23562 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
23569 @anchor{haldclutsrc}
23572 @anchor{pal100bars}
23573 @anchor{rgbtestsrc}
23575 @anchor{smptehdbars}
23578 @anchor{yuvtestsrc}
23579 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
23581 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
23583 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
23585 The @code{color} source provides an uniformly colored input.
23587 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
23588 @ref{haldclut} filter.
23590 The @code{nullsrc} source returns unprocessed video frames. It is
23591 mainly useful to be employed in analysis / debugging tools, or as the
23592 source for filters which ignore the input data.
23594 The @code{pal75bars} source generates a color bars pattern, based on
23595 EBU PAL recommendations with 75% color levels.
23597 The @code{pal100bars} source generates a color bars pattern, based on
23598 EBU PAL recommendations with 100% color levels.
23600 The @code{rgbtestsrc} source generates an RGB test pattern useful for
23601 detecting RGB vs BGR issues. You should see a red, green and blue
23602 stripe from top to bottom.
23604 The @code{smptebars} source generates a color bars pattern, based on
23605 the SMPTE Engineering Guideline EG 1-1990.
23607 The @code{smptehdbars} source generates a color bars pattern, based on
23608 the SMPTE RP 219-2002.
23610 The @code{testsrc} source generates a test video pattern, showing a
23611 color pattern, a scrolling gradient and a timestamp. This is mainly
23612 intended for testing purposes.
23614 The @code{testsrc2} source is similar to testsrc, but supports more
23615 pixel formats instead of just @code{rgb24}. This allows using it as an
23616 input for other tests without requiring a format conversion.
23618 The @code{yuvtestsrc} source generates an YUV test pattern. You should
23619 see a y, cb and cr stripe from top to bottom.
23621 The sources accept the following parameters:
23626 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
23627 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
23628 pixels to be used as identity matrix for 3D lookup tables. Each component is
23629 coded on a @code{1/(N*N)} scale.
23632 Specify the color of the source, only available in the @code{color}
23633 source. For the syntax of this option, check the
23634 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
23637 Specify the size of the sourced video. For the syntax of this option, check the
23638 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23639 The default value is @code{320x240}.
23641 This option is not available with the @code{allrgb}, @code{allyuv}, and
23642 @code{haldclutsrc} filters.
23645 Specify the frame rate of the sourced video, as the number of frames
23646 generated per second. It has to be a string in the format
23647 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23648 number or a valid video frame rate abbreviation. The default value is
23652 Set the duration of the sourced video. See
23653 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23654 for the accepted syntax.
23656 If not specified, or the expressed duration is negative, the video is
23657 supposed to be generated forever.
23659 Since the frame rate is used as time base, all frames including the last one
23660 will have their full duration. If the specified duration is not a multiple
23661 of the frame duration, it will be rounded up.
23664 Set the sample aspect ratio of the sourced video.
23667 Specify the alpha (opacity) of the background, only available in the
23668 @code{testsrc2} source. The value must be between 0 (fully transparent) and
23669 255 (fully opaque, the default).
23672 Set the number of decimals to show in the timestamp, only available in the
23673 @code{testsrc} source.
23675 The displayed timestamp value will correspond to the original
23676 timestamp value multiplied by the power of 10 of the specified
23677 value. Default value is 0.
23680 @subsection Examples
23684 Generate a video with a duration of 5.3 seconds, with size
23685 176x144 and a frame rate of 10 frames per second:
23687 testsrc=duration=5.3:size=qcif:rate=10
23691 The following graph description will generate a red source
23692 with an opacity of 0.2, with size "qcif" and a frame rate of 10
23695 color=c=red@@0.2:s=qcif:r=10
23699 If the input content is to be ignored, @code{nullsrc} can be used. The
23700 following command generates noise in the luminance plane by employing
23701 the @code{geq} filter:
23703 nullsrc=s=256x256, geq=random(1)*255:128:128
23707 @subsection Commands
23709 The @code{color} source supports the following commands:
23713 Set the color of the created image. Accepts the same syntax of the
23714 corresponding @option{color} option.
23719 Generate video using an OpenCL program.
23724 OpenCL program source file.
23727 Kernel name in program.
23730 Size of frames to generate. This must be set.
23733 Pixel format to use for the generated frames. This must be set.
23736 Number of frames generated every second. Default value is '25'.
23740 For details of how the program loading works, see the @ref{program_opencl}
23747 Generate a colour ramp by setting pixel values from the position of the pixel
23748 in the output image. (Note that this will work with all pixel formats, but
23749 the generated output will not be the same.)
23751 __kernel void ramp(__write_only image2d_t dst,
23752 unsigned int index)
23754 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23757 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
23759 write_imagef(dst, loc, val);
23764 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
23766 __kernel void sierpinski_carpet(__write_only image2d_t dst,
23767 unsigned int index)
23769 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23771 float4 value = 0.0f;
23772 int x = loc.x + index;
23773 int y = loc.y + index;
23774 while (x > 0 || y > 0) {
23775 if (x % 3 == 1 && y % 3 == 1) {
23783 write_imagef(dst, loc, value);
23789 @section sierpinski
23791 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
23793 This source accepts the following options:
23797 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23798 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23801 Set frame rate, expressed as number of frames per second. Default
23805 Set seed which is used for random panning.
23808 Set max jump for single pan destination. Allowed range is from 1 to 10000.
23811 Set fractal type, can be default @code{carpet} or @code{triangle}.
23814 @c man end VIDEO SOURCES
23816 @chapter Video Sinks
23817 @c man begin VIDEO SINKS
23819 Below is a description of the currently available video sinks.
23821 @section buffersink
23823 Buffer video frames, and make them available to the end of the filter
23826 This sink is mainly intended for programmatic use, in particular
23827 through the interface defined in @file{libavfilter/buffersink.h}
23828 or the options system.
23830 It accepts a pointer to an AVBufferSinkContext structure, which
23831 defines the incoming buffers' formats, to be passed as the opaque
23832 parameter to @code{avfilter_init_filter} for initialization.
23836 Null video sink: do absolutely nothing with the input video. It is
23837 mainly useful as a template and for use in analysis / debugging
23840 @c man end VIDEO SINKS
23842 @chapter Multimedia Filters
23843 @c man begin MULTIMEDIA FILTERS
23845 Below is a description of the currently available multimedia filters.
23849 Convert input audio to a video output, displaying the audio bit scope.
23851 The filter accepts the following options:
23855 Set frame rate, expressed as number of frames per second. Default
23859 Specify the video size for the output. For the syntax of this option, check the
23860 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23861 Default value is @code{1024x256}.
23864 Specify list of colors separated by space or by '|' which will be used to
23865 draw channels. Unrecognized or missing colors will be replaced
23869 @section adrawgraph
23870 Draw a graph using input audio metadata.
23872 See @ref{drawgraph}
23874 @section agraphmonitor
23876 See @ref{graphmonitor}.
23878 @section ahistogram
23880 Convert input audio to a video output, displaying the volume histogram.
23882 The filter accepts the following options:
23886 Specify how histogram is calculated.
23888 It accepts the following values:
23891 Use single histogram for all channels.
23893 Use separate histogram for each channel.
23895 Default is @code{single}.
23898 Set frame rate, expressed as number of frames per second. Default
23902 Specify the video size for the output. For the syntax of this option, check the
23903 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23904 Default value is @code{hd720}.
23909 It accepts the following values:
23920 reverse logarithmic
23922 Default is @code{log}.
23925 Set amplitude scale.
23927 It accepts the following values:
23934 Default is @code{log}.
23937 Set how much frames to accumulate in histogram.
23938 Default is 1. Setting this to -1 accumulates all frames.
23941 Set histogram ratio of window height.
23944 Set sonogram sliding.
23946 It accepts the following values:
23949 replace old rows with new ones.
23951 scroll from top to bottom.
23953 Default is @code{replace}.
23956 @section aphasemeter
23958 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
23959 representing mean phase of current audio frame. A video output can also be produced and is
23960 enabled by default. The audio is passed through as first output.
23962 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
23963 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
23964 and @code{1} means channels are in phase.
23966 The filter accepts the following options, all related to its video output:
23970 Set the output frame rate. Default value is @code{25}.
23973 Set the video size for the output. For the syntax of this option, check the
23974 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23975 Default value is @code{800x400}.
23980 Specify the red, green, blue contrast. Default values are @code{2},
23981 @code{7} and @code{1}.
23982 Allowed range is @code{[0, 255]}.
23985 Set color which will be used for drawing median phase. If color is
23986 @code{none} which is default, no median phase value will be drawn.
23989 Enable video output. Default is enabled.
23992 @subsection phasing detection
23994 The filter also detects out of phase and mono sequences in stereo streams.
23995 It logs the sequence start, end and duration when it lasts longer or as long as the minimum set.
23997 The filter accepts the following options for this detection:
24001 Enable mono and out of phase detection. Default is disabled.
24004 Set phase tolerance for mono detection, in amplitude ratio. Default is @code{0}.
24005 Allowed range is @code{[0, 1]}.
24008 Set angle threshold for out of phase detection, in degree. Default is @code{170}.
24009 Allowed range is @code{[90, 180]}.
24012 Set mono or out of phase duration until notification, expressed in seconds. Default is @code{2}.
24015 @subsection Examples
24019 Complete example with @command{ffmpeg} to detect 1 second of mono with 0.001 phase tolerance:
24021 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
24025 @section avectorscope
24027 Convert input audio to a video output, representing the audio vector
24030 The filter is used to measure the difference between channels of stereo
24031 audio stream. A monaural signal, consisting of identical left and right
24032 signal, results in straight vertical line. Any stereo separation is visible
24033 as a deviation from this line, creating a Lissajous figure.
24034 If the straight (or deviation from it) but horizontal line appears this
24035 indicates that the left and right channels are out of phase.
24037 The filter accepts the following options:
24041 Set the vectorscope mode.
24043 Available values are:
24046 Lissajous rotated by 45 degrees.
24049 Same as above but not rotated.
24052 Shape resembling half of circle.
24055 Default value is @samp{lissajous}.
24058 Set the video size for the output. For the syntax of this option, check the
24059 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24060 Default value is @code{400x400}.
24063 Set the output frame rate. Default value is @code{25}.
24069 Specify the red, green, blue and alpha contrast. Default values are @code{40},
24070 @code{160}, @code{80} and @code{255}.
24071 Allowed range is @code{[0, 255]}.
24077 Specify the red, green, blue and alpha fade. Default values are @code{15},
24078 @code{10}, @code{5} and @code{5}.
24079 Allowed range is @code{[0, 255]}.
24082 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
24083 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
24086 Set the vectorscope drawing mode.
24088 Available values are:
24091 Draw dot for each sample.
24094 Draw line between previous and current sample.
24097 Default value is @samp{dot}.
24100 Specify amplitude scale of audio samples.
24102 Available values are:
24118 Swap left channel axis with right channel axis.
24128 Mirror only x axis.
24131 Mirror only y axis.
24139 @subsection Examples
24143 Complete example using @command{ffplay}:
24145 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
24146 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
24150 @section bench, abench
24152 Benchmark part of a filtergraph.
24154 The filter accepts the following options:
24158 Start or stop a timer.
24160 Available values are:
24163 Get the current time, set it as frame metadata (using the key
24164 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
24167 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
24168 the input frame metadata to get the time difference. Time difference, average,
24169 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
24170 @code{min}) are then printed. The timestamps are expressed in seconds.
24174 @subsection Examples
24178 Benchmark @ref{selectivecolor} filter:
24180 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
24186 Concatenate audio and video streams, joining them together one after the
24189 The filter works on segments of synchronized video and audio streams. All
24190 segments must have the same number of streams of each type, and that will
24191 also be the number of streams at output.
24193 The filter accepts the following options:
24198 Set the number of segments. Default is 2.
24201 Set the number of output video streams, that is also the number of video
24202 streams in each segment. Default is 1.
24205 Set the number of output audio streams, that is also the number of audio
24206 streams in each segment. Default is 0.
24209 Activate unsafe mode: do not fail if segments have a different format.
24213 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
24214 @var{a} audio outputs.
24216 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
24217 segment, in the same order as the outputs, then the inputs for the second
24220 Related streams do not always have exactly the same duration, for various
24221 reasons including codec frame size or sloppy authoring. For that reason,
24222 related synchronized streams (e.g. a video and its audio track) should be
24223 concatenated at once. The concat filter will use the duration of the longest
24224 stream in each segment (except the last one), and if necessary pad shorter
24225 audio streams with silence.
24227 For this filter to work correctly, all segments must start at timestamp 0.
24229 All corresponding streams must have the same parameters in all segments; the
24230 filtering system will automatically select a common pixel format for video
24231 streams, and a common sample format, sample rate and channel layout for
24232 audio streams, but other settings, such as resolution, must be converted
24233 explicitly by the user.
24235 Different frame rates are acceptable but will result in variable frame rate
24236 at output; be sure to configure the output file to handle it.
24238 @subsection Examples
24242 Concatenate an opening, an episode and an ending, all in bilingual version
24243 (video in stream 0, audio in streams 1 and 2):
24245 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
24246 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
24247 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
24248 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
24252 Concatenate two parts, handling audio and video separately, using the
24253 (a)movie sources, and adjusting the resolution:
24255 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
24256 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
24257 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
24259 Note that a desync will happen at the stitch if the audio and video streams
24260 do not have exactly the same duration in the first file.
24264 @subsection Commands
24266 This filter supports the following commands:
24269 Close the current segment and step to the next one
24275 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
24276 level. By default, it logs a message at a frequency of 10Hz with the
24277 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
24278 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
24280 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
24281 sample format is double-precision floating point. The input stream will be converted to
24282 this specification, if needed. Users may need to insert aformat and/or aresample filters
24283 after this filter to obtain the original parameters.
24285 The filter also has a video output (see the @var{video} option) with a real
24286 time graph to observe the loudness evolution. The graphic contains the logged
24287 message mentioned above, so it is not printed anymore when this option is set,
24288 unless the verbose logging is set. The main graphing area contains the
24289 short-term loudness (3 seconds of analysis), and the gauge on the right is for
24290 the momentary loudness (400 milliseconds), but can optionally be configured
24291 to instead display short-term loudness (see @var{gauge}).
24293 The green area marks a +/- 1LU target range around the target loudness
24294 (-23LUFS by default, unless modified through @var{target}).
24296 More information about the Loudness Recommendation EBU R128 on
24297 @url{http://tech.ebu.ch/loudness}.
24299 The filter accepts the following options:
24304 Activate the video output. The audio stream is passed unchanged whether this
24305 option is set or no. The video stream will be the first output stream if
24306 activated. Default is @code{0}.
24309 Set the video size. This option is for video only. For the syntax of this
24311 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24312 Default and minimum resolution is @code{640x480}.
24315 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
24316 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
24317 other integer value between this range is allowed.
24320 Set metadata injection. If set to @code{1}, the audio input will be segmented
24321 into 100ms output frames, each of them containing various loudness information
24322 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
24324 Default is @code{0}.
24327 Force the frame logging level.
24329 Available values are:
24332 information logging level
24334 verbose logging level
24337 By default, the logging level is set to @var{info}. If the @option{video} or
24338 the @option{metadata} options are set, it switches to @var{verbose}.
24343 Available modes can be cumulated (the option is a @code{flag} type). Possible
24347 Disable any peak mode (default).
24349 Enable sample-peak mode.
24351 Simple peak mode looking for the higher sample value. It logs a message
24352 for sample-peak (identified by @code{SPK}).
24354 Enable true-peak mode.
24356 If enabled, the peak lookup is done on an over-sampled version of the input
24357 stream for better peak accuracy. It logs a message for true-peak.
24358 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
24359 This mode requires a build with @code{libswresample}.
24363 Treat mono input files as "dual mono". If a mono file is intended for playback
24364 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
24365 If set to @code{true}, this option will compensate for this effect.
24366 Multi-channel input files are not affected by this option.
24369 Set a specific pan law to be used for the measurement of dual mono files.
24370 This parameter is optional, and has a default value of -3.01dB.
24373 Set a specific target level (in LUFS) used as relative zero in the visualization.
24374 This parameter is optional and has a default value of -23LUFS as specified
24375 by EBU R128. However, material published online may prefer a level of -16LUFS
24376 (e.g. for use with podcasts or video platforms).
24379 Set the value displayed by the gauge. Valid values are @code{momentary} and s
24380 @code{shortterm}. By default the momentary value will be used, but in certain
24381 scenarios it may be more useful to observe the short term value instead (e.g.
24385 Sets the display scale for the loudness. Valid parameters are @code{absolute}
24386 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
24387 video output, not the summary or continuous log output.
24390 @subsection Examples
24394 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
24396 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
24400 Run an analysis with @command{ffmpeg}:
24402 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
24406 @section interleave, ainterleave
24408 Temporally interleave frames from several inputs.
24410 @code{interleave} works with video inputs, @code{ainterleave} with audio.
24412 These filters read frames from several inputs and send the oldest
24413 queued frame to the output.
24415 Input streams must have well defined, monotonically increasing frame
24418 In order to submit one frame to output, these filters need to enqueue
24419 at least one frame for each input, so they cannot work in case one
24420 input is not yet terminated and will not receive incoming frames.
24422 For example consider the case when one input is a @code{select} filter
24423 which always drops input frames. The @code{interleave} filter will keep
24424 reading from that input, but it will never be able to send new frames
24425 to output until the input sends an end-of-stream signal.
24427 Also, depending on inputs synchronization, the filters will drop
24428 frames in case one input receives more frames than the other ones, and
24429 the queue is already filled.
24431 These filters accept the following options:
24435 Set the number of different inputs, it is 2 by default.
24438 How to determine the end-of-stream.
24442 The duration of the longest input. (default)
24445 The duration of the shortest input.
24448 The duration of the first input.
24453 @subsection Examples
24457 Interleave frames belonging to different streams using @command{ffmpeg}:
24459 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
24463 Add flickering blur effect:
24465 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
24469 @section metadata, ametadata
24471 Manipulate frame metadata.
24473 This filter accepts the following options:
24477 Set mode of operation of the filter.
24479 Can be one of the following:
24483 If both @code{value} and @code{key} is set, select frames
24484 which have such metadata. If only @code{key} is set, select
24485 every frame that has such key in metadata.
24488 Add new metadata @code{key} and @code{value}. If key is already available
24492 Modify value of already present key.
24495 If @code{value} is set, delete only keys that have such value.
24496 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
24500 Print key and its value if metadata was found. If @code{key} is not set print all
24501 metadata values available in frame.
24505 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
24508 Set metadata value which will be used. This option is mandatory for
24509 @code{modify} and @code{add} mode.
24512 Which function to use when comparing metadata value and @code{value}.
24514 Can be one of following:
24518 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
24521 Values are interpreted as strings, returns true if metadata value starts with
24522 the @code{value} option string.
24525 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
24528 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
24531 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
24534 Values are interpreted as floats, returns true if expression from option @code{expr}
24538 Values are interpreted as strings, returns true if metadata value ends with
24539 the @code{value} option string.
24543 Set expression which is used when @code{function} is set to @code{expr}.
24544 The expression is evaluated through the eval API and can contain the following
24549 Float representation of @code{value} from metadata key.
24552 Float representation of @code{value} as supplied by user in @code{value} option.
24556 If specified in @code{print} mode, output is written to the named file. Instead of
24557 plain filename any writable url can be specified. Filename ``-'' is a shorthand
24558 for standard output. If @code{file} option is not set, output is written to the log
24559 with AV_LOG_INFO loglevel.
24562 Reduces buffering in print mode when output is written to a URL set using @var{file}.
24566 @subsection Examples
24570 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
24573 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
24576 Print silencedetect output to file @file{metadata.txt}.
24578 silencedetect,ametadata=mode=print:file=metadata.txt
24581 Direct all metadata to a pipe with file descriptor 4.
24583 metadata=mode=print:file='pipe\:4'
24587 @section perms, aperms
24589 Set read/write permissions for the output frames.
24591 These filters are mainly aimed at developers to test direct path in the
24592 following filter in the filtergraph.
24594 The filters accept the following options:
24598 Select the permissions mode.
24600 It accepts the following values:
24603 Do nothing. This is the default.
24605 Set all the output frames read-only.
24607 Set all the output frames directly writable.
24609 Make the frame read-only if writable, and writable if read-only.
24611 Set each output frame read-only or writable randomly.
24615 Set the seed for the @var{random} mode, must be an integer included between
24616 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
24617 @code{-1}, the filter will try to use a good random seed on a best effort
24621 Note: in case of auto-inserted filter between the permission filter and the
24622 following one, the permission might not be received as expected in that
24623 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
24624 perms/aperms filter can avoid this problem.
24626 @section realtime, arealtime
24628 Slow down filtering to match real time approximately.
24630 These filters will pause the filtering for a variable amount of time to
24631 match the output rate with the input timestamps.
24632 They are similar to the @option{re} option to @code{ffmpeg}.
24634 They accept the following options:
24638 Time limit for the pauses. Any pause longer than that will be considered
24639 a timestamp discontinuity and reset the timer. Default is 2 seconds.
24641 Speed factor for processing. The value must be a float larger than zero.
24642 Values larger than 1.0 will result in faster than realtime processing,
24643 smaller will slow processing down. The @var{limit} is automatically adapted
24644 accordingly. Default is 1.0.
24646 A processing speed faster than what is possible without these filters cannot
24651 @section select, aselect
24653 Select frames to pass in output.
24655 This filter accepts the following options:
24660 Set expression, which is evaluated for each input frame.
24662 If the expression is evaluated to zero, the frame is discarded.
24664 If the evaluation result is negative or NaN, the frame is sent to the
24665 first output; otherwise it is sent to the output with index
24666 @code{ceil(val)-1}, assuming that the input index starts from 0.
24668 For example a value of @code{1.2} corresponds to the output with index
24669 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
24672 Set the number of outputs. The output to which to send the selected
24673 frame is based on the result of the evaluation. Default value is 1.
24676 The expression can contain the following constants:
24680 The (sequential) number of the filtered frame, starting from 0.
24683 The (sequential) number of the selected frame, starting from 0.
24685 @item prev_selected_n
24686 The sequential number of the last selected frame. It's NAN if undefined.
24689 The timebase of the input timestamps.
24692 The PTS (Presentation TimeStamp) of the filtered video frame,
24693 expressed in @var{TB} units. It's NAN if undefined.
24696 The PTS of the filtered video frame,
24697 expressed in seconds. It's NAN if undefined.
24700 The PTS of the previously filtered video frame. It's NAN if undefined.
24702 @item prev_selected_pts
24703 The PTS of the last previously filtered video frame. It's NAN if undefined.
24705 @item prev_selected_t
24706 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
24709 The PTS of the first video frame in the video. It's NAN if undefined.
24712 The time of the first video frame in the video. It's NAN if undefined.
24714 @item pict_type @emph{(video only)}
24715 The type of the filtered frame. It can assume one of the following
24727 @item interlace_type @emph{(video only)}
24728 The frame interlace type. It can assume one of the following values:
24731 The frame is progressive (not interlaced).
24733 The frame is top-field-first.
24735 The frame is bottom-field-first.
24738 @item consumed_sample_n @emph{(audio only)}
24739 the number of selected samples before the current frame
24741 @item samples_n @emph{(audio only)}
24742 the number of samples in the current frame
24744 @item sample_rate @emph{(audio only)}
24745 the input sample rate
24748 This is 1 if the filtered frame is a key-frame, 0 otherwise.
24751 the position in the file of the filtered frame, -1 if the information
24752 is not available (e.g. for synthetic video)
24754 @item scene @emph{(video only)}
24755 value between 0 and 1 to indicate a new scene; a low value reflects a low
24756 probability for the current frame to introduce a new scene, while a higher
24757 value means the current frame is more likely to be one (see the example below)
24759 @item concatdec_select
24760 The concat demuxer can select only part of a concat input file by setting an
24761 inpoint and an outpoint, but the output packets may not be entirely contained
24762 in the selected interval. By using this variable, it is possible to skip frames
24763 generated by the concat demuxer which are not exactly contained in the selected
24766 This works by comparing the frame pts against the @var{lavf.concat.start_time}
24767 and the @var{lavf.concat.duration} packet metadata values which are also
24768 present in the decoded frames.
24770 The @var{concatdec_select} variable is -1 if the frame pts is at least
24771 start_time and either the duration metadata is missing or the frame pts is less
24772 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
24775 That basically means that an input frame is selected if its pts is within the
24776 interval set by the concat demuxer.
24780 The default value of the select expression is "1".
24782 @subsection Examples
24786 Select all frames in input:
24791 The example above is the same as:
24803 Select only I-frames:
24805 select='eq(pict_type\,I)'
24809 Select one frame every 100:
24811 select='not(mod(n\,100))'
24815 Select only frames contained in the 10-20 time interval:
24817 select=between(t\,10\,20)
24821 Select only I-frames contained in the 10-20 time interval:
24823 select=between(t\,10\,20)*eq(pict_type\,I)
24827 Select frames with a minimum distance of 10 seconds:
24829 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
24833 Use aselect to select only audio frames with samples number > 100:
24835 aselect='gt(samples_n\,100)'
24839 Create a mosaic of the first scenes:
24841 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
24844 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
24848 Send even and odd frames to separate outputs, and compose them:
24850 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
24854 Select useful frames from an ffconcat file which is using inpoints and
24855 outpoints but where the source files are not intra frame only.
24857 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
24861 @section sendcmd, asendcmd
24863 Send commands to filters in the filtergraph.
24865 These filters read commands to be sent to other filters in the
24868 @code{sendcmd} must be inserted between two video filters,
24869 @code{asendcmd} must be inserted between two audio filters, but apart
24870 from that they act the same way.
24872 The specification of commands can be provided in the filter arguments
24873 with the @var{commands} option, or in a file specified by the
24874 @var{filename} option.
24876 These filters accept the following options:
24879 Set the commands to be read and sent to the other filters.
24881 Set the filename of the commands to be read and sent to the other
24885 @subsection Commands syntax
24887 A commands description consists of a sequence of interval
24888 specifications, comprising a list of commands to be executed when a
24889 particular event related to that interval occurs. The occurring event
24890 is typically the current frame time entering or leaving a given time
24893 An interval is specified by the following syntax:
24895 @var{START}[-@var{END}] @var{COMMANDS};
24898 The time interval is specified by the @var{START} and @var{END} times.
24899 @var{END} is optional and defaults to the maximum time.
24901 The current frame time is considered within the specified interval if
24902 it is included in the interval [@var{START}, @var{END}), that is when
24903 the time is greater or equal to @var{START} and is lesser than
24906 @var{COMMANDS} consists of a sequence of one or more command
24907 specifications, separated by ",", relating to that interval. The
24908 syntax of a command specification is given by:
24910 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
24913 @var{FLAGS} is optional and specifies the type of events relating to
24914 the time interval which enable sending the specified command, and must
24915 be a non-null sequence of identifier flags separated by "+" or "|" and
24916 enclosed between "[" and "]".
24918 The following flags are recognized:
24921 The command is sent when the current frame timestamp enters the
24922 specified interval. In other words, the command is sent when the
24923 previous frame timestamp was not in the given interval, and the
24927 The command is sent when the current frame timestamp leaves the
24928 specified interval. In other words, the command is sent when the
24929 previous frame timestamp was in the given interval, and the
24933 The command @var{ARG} is interpreted as expression and result of
24934 expression is passed as @var{ARG}.
24936 The expression is evaluated through the eval API and can contain the following
24941 Original position in the file of the frame, or undefined if undefined
24942 for the current frame.
24945 The presentation timestamp in input.
24948 The count of the input frame for video or audio, starting from 0.
24951 The time in seconds of the current frame.
24954 The start time in seconds of the current command interval.
24957 The end time in seconds of the current command interval.
24960 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
24965 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
24968 @var{TARGET} specifies the target of the command, usually the name of
24969 the filter class or a specific filter instance name.
24971 @var{COMMAND} specifies the name of the command for the target filter.
24973 @var{ARG} is optional and specifies the optional list of argument for
24974 the given @var{COMMAND}.
24976 Between one interval specification and another, whitespaces, or
24977 sequences of characters starting with @code{#} until the end of line,
24978 are ignored and can be used to annotate comments.
24980 A simplified BNF description of the commands specification syntax
24983 @var{COMMAND_FLAG} ::= "enter" | "leave"
24984 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
24985 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
24986 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
24987 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
24988 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
24991 @subsection Examples
24995 Specify audio tempo change at second 4:
24997 asendcmd=c='4.0 atempo tempo 1.5',atempo
25001 Target a specific filter instance:
25003 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
25007 Specify a list of drawtext and hue commands in a file.
25009 # show text in the interval 5-10
25010 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
25011 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
25013 # desaturate the image in the interval 15-20
25014 15.0-20.0 [enter] hue s 0,
25015 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
25017 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
25019 # apply an exponential saturation fade-out effect, starting from time 25
25020 25 [enter] hue s exp(25-t)
25023 A filtergraph allowing to read and process the above command list
25024 stored in a file @file{test.cmd}, can be specified with:
25026 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
25031 @section setpts, asetpts
25033 Change the PTS (presentation timestamp) of the input frames.
25035 @code{setpts} works on video frames, @code{asetpts} on audio frames.
25037 This filter accepts the following options:
25042 The expression which is evaluated for each frame to construct its timestamp.
25046 The expression is evaluated through the eval API and can contain the following
25050 @item FRAME_RATE, FR
25051 frame rate, only defined for constant frame-rate video
25054 The presentation timestamp in input
25057 The count of the input frame for video or the number of consumed samples,
25058 not including the current frame for audio, starting from 0.
25060 @item NB_CONSUMED_SAMPLES
25061 The number of consumed samples, not including the current frame (only
25064 @item NB_SAMPLES, S
25065 The number of samples in the current frame (only audio)
25067 @item SAMPLE_RATE, SR
25068 The audio sample rate.
25071 The PTS of the first frame.
25074 the time in seconds of the first frame
25077 State whether the current frame is interlaced.
25080 the time in seconds of the current frame
25083 original position in the file of the frame, or undefined if undefined
25084 for the current frame
25087 The previous input PTS.
25090 previous input time in seconds
25093 The previous output PTS.
25096 previous output time in seconds
25099 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
25103 The wallclock (RTC) time at the start of the movie in microseconds.
25106 The timebase of the input timestamps.
25110 @subsection Examples
25114 Start counting PTS from zero
25116 setpts=PTS-STARTPTS
25120 Apply fast motion effect:
25126 Apply slow motion effect:
25132 Set fixed rate of 25 frames per second:
25138 Set fixed rate 25 fps with some jitter:
25140 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
25144 Apply an offset of 10 seconds to the input PTS:
25150 Generate timestamps from a "live source" and rebase onto the current timebase:
25152 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
25156 Generate timestamps by counting samples:
25165 Force color range for the output video frame.
25167 The @code{setrange} filter marks the color range property for the
25168 output frames. It does not change the input frame, but only sets the
25169 corresponding property, which affects how the frame is treated by
25172 The filter accepts the following options:
25177 Available values are:
25181 Keep the same color range property.
25183 @item unspecified, unknown
25184 Set the color range as unspecified.
25186 @item limited, tv, mpeg
25187 Set the color range as limited.
25189 @item full, pc, jpeg
25190 Set the color range as full.
25194 @section settb, asettb
25196 Set the timebase to use for the output frames timestamps.
25197 It is mainly useful for testing timebase configuration.
25199 It accepts the following parameters:
25204 The expression which is evaluated into the output timebase.
25208 The value for @option{tb} is an arithmetic expression representing a
25209 rational. The expression can contain the constants "AVTB" (the default
25210 timebase), "intb" (the input timebase) and "sr" (the sample rate,
25211 audio only). Default value is "intb".
25213 @subsection Examples
25217 Set the timebase to 1/25:
25223 Set the timebase to 1/10:
25229 Set the timebase to 1001/1000:
25235 Set the timebase to 2*intb:
25241 Set the default timebase value:
25248 Convert input audio to a video output representing frequency spectrum
25249 logarithmically using Brown-Puckette constant Q transform algorithm with
25250 direct frequency domain coefficient calculation (but the transform itself
25251 is not really constant Q, instead the Q factor is actually variable/clamped),
25252 with musical tone scale, from E0 to D#10.
25254 The filter accepts the following options:
25258 Specify the video size for the output. It must be even. For the syntax of this option,
25259 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25260 Default value is @code{1920x1080}.
25263 Set the output frame rate. Default value is @code{25}.
25266 Set the bargraph height. It must be even. Default value is @code{-1} which
25267 computes the bargraph height automatically.
25270 Set the axis height. It must be even. Default value is @code{-1} which computes
25271 the axis height automatically.
25274 Set the sonogram height. It must be even. Default value is @code{-1} which
25275 computes the sonogram height automatically.
25278 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
25279 instead. Default value is @code{1}.
25281 @item sono_v, volume
25282 Specify the sonogram volume expression. It can contain variables:
25285 the @var{bar_v} evaluated expression
25286 @item frequency, freq, f
25287 the frequency where it is evaluated
25288 @item timeclamp, tc
25289 the value of @var{timeclamp} option
25293 @item a_weighting(f)
25294 A-weighting of equal loudness
25295 @item b_weighting(f)
25296 B-weighting of equal loudness
25297 @item c_weighting(f)
25298 C-weighting of equal loudness.
25300 Default value is @code{16}.
25302 @item bar_v, volume2
25303 Specify the bargraph volume expression. It can contain variables:
25306 the @var{sono_v} evaluated expression
25307 @item frequency, freq, f
25308 the frequency where it is evaluated
25309 @item timeclamp, tc
25310 the value of @var{timeclamp} option
25314 @item a_weighting(f)
25315 A-weighting of equal loudness
25316 @item b_weighting(f)
25317 B-weighting of equal loudness
25318 @item c_weighting(f)
25319 C-weighting of equal loudness.
25321 Default value is @code{sono_v}.
25323 @item sono_g, gamma
25324 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
25325 higher gamma makes the spectrum having more range. Default value is @code{3}.
25326 Acceptable range is @code{[1, 7]}.
25328 @item bar_g, gamma2
25329 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
25333 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
25334 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
25336 @item timeclamp, tc
25337 Specify the transform timeclamp. At low frequency, there is trade-off between
25338 accuracy in time domain and frequency domain. If timeclamp is lower,
25339 event in time domain is represented more accurately (such as fast bass drum),
25340 otherwise event in frequency domain is represented more accurately
25341 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
25344 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
25345 limits future samples by applying asymmetric windowing in time domain, useful
25346 when low latency is required. Accepted range is @code{[0, 1]}.
25349 Specify the transform base frequency. Default value is @code{20.01523126408007475},
25350 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
25353 Specify the transform end frequency. Default value is @code{20495.59681441799654},
25354 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
25357 This option is deprecated and ignored.
25360 Specify the transform length in time domain. Use this option to control accuracy
25361 trade-off between time domain and frequency domain at every frequency sample.
25362 It can contain variables:
25364 @item frequency, freq, f
25365 the frequency where it is evaluated
25366 @item timeclamp, tc
25367 the value of @var{timeclamp} option.
25369 Default value is @code{384*tc/(384+tc*f)}.
25372 Specify the transform count for every video frame. Default value is @code{6}.
25373 Acceptable range is @code{[1, 30]}.
25376 Specify the transform count for every single pixel. Default value is @code{0},
25377 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
25380 Specify font file for use with freetype to draw the axis. If not specified,
25381 use embedded font. Note that drawing with font file or embedded font is not
25382 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
25386 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
25387 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
25391 Specify font color expression. This is arithmetic expression that should return
25392 integer value 0xRRGGBB. It can contain variables:
25394 @item frequency, freq, f
25395 the frequency where it is evaluated
25396 @item timeclamp, tc
25397 the value of @var{timeclamp} option
25402 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
25403 @item r(x), g(x), b(x)
25404 red, green, and blue value of intensity x.
25406 Default value is @code{st(0, (midi(f)-59.5)/12);
25407 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
25408 r(1-ld(1)) + b(ld(1))}.
25411 Specify image file to draw the axis. This option override @var{fontfile} and
25412 @var{fontcolor} option.
25415 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
25416 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
25417 Default value is @code{1}.
25420 Set colorspace. The accepted values are:
25423 Unspecified (default)
25432 BT.470BG or BT.601-6 625
25435 SMPTE-170M or BT.601-6 525
25441 BT.2020 with non-constant luminance
25446 Set spectrogram color scheme. This is list of floating point values with format
25447 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
25448 The default is @code{1|0.5|0|0|0.5|1}.
25452 @subsection Examples
25456 Playing audio while showing the spectrum:
25458 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
25462 Same as above, but with frame rate 30 fps:
25464 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
25468 Playing at 1280x720:
25470 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
25474 Disable sonogram display:
25480 A1 and its harmonics: A1, A2, (near)E3, A3:
25482 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),
25483 asplit[a][out1]; [a] showcqt [out0]'
25487 Same as above, but with more accuracy in frequency domain:
25489 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),
25490 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
25496 bar_v=10:sono_v=bar_v*a_weighting(f)
25500 Custom gamma, now spectrum is linear to the amplitude.
25506 Custom tlength equation:
25508 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)))'
25512 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
25514 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
25518 Custom font using fontconfig:
25520 font='Courier New,Monospace,mono|bold'
25524 Custom frequency range with custom axis using image file:
25526 axisfile=myaxis.png:basefreq=40:endfreq=10000
25532 Convert input audio to video output representing the audio power spectrum.
25533 Audio amplitude is on Y-axis while frequency is on X-axis.
25535 The filter accepts the following options:
25539 Specify size of video. For the syntax of this option, check the
25540 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25541 Default is @code{1024x512}.
25545 This set how each frequency bin will be represented.
25547 It accepts the following values:
25553 Default is @code{bar}.
25556 Set amplitude scale.
25558 It accepts the following values:
25572 Default is @code{log}.
25575 Set frequency scale.
25577 It accepts the following values:
25586 Reverse logarithmic scale.
25588 Default is @code{lin}.
25591 Set window size. Allowed range is from 16 to 65536.
25593 Default is @code{2048}
25596 Set windowing function.
25598 It accepts the following values:
25621 Default is @code{hanning}.
25624 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25625 which means optimal overlap for selected window function will be picked.
25628 Set time averaging. Setting this to 0 will display current maximal peaks.
25629 Default is @code{1}, which means time averaging is disabled.
25632 Specify list of colors separated by space or by '|' which will be used to
25633 draw channel frequencies. Unrecognized or missing colors will be replaced
25637 Set channel display mode.
25639 It accepts the following values:
25644 Default is @code{combined}.
25647 Set minimum amplitude used in @code{log} amplitude scaler.
25650 Set data display mode.
25652 It accepts the following values:
25658 Default is @code{magnitude}.
25661 @section showspatial
25663 Convert stereo input audio to a video output, representing the spatial relationship
25664 between two channels.
25666 The filter accepts the following options:
25670 Specify the video size for the output. For the syntax of this option, check the
25671 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25672 Default value is @code{512x512}.
25675 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
25678 Set window function.
25680 It accepts the following values:
25705 Default value is @code{hann}.
25708 Set ratio of overlap window. Default value is @code{0.5}.
25709 When value is @code{1} overlap is set to recommended size for specific
25710 window function currently used.
25713 @anchor{showspectrum}
25714 @section showspectrum
25716 Convert input audio to a video output, representing the audio frequency
25719 The filter accepts the following options:
25723 Specify the video size for the output. For the syntax of this option, check the
25724 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25725 Default value is @code{640x512}.
25728 Specify how the spectrum should slide along the window.
25730 It accepts the following values:
25733 the samples start again on the left when they reach the right
25735 the samples scroll from right to left
25737 frames are only produced when the samples reach the right
25739 the samples scroll from left to right
25742 Default value is @code{replace}.
25745 Specify display mode.
25747 It accepts the following values:
25750 all channels are displayed in the same row
25752 all channels are displayed in separate rows
25755 Default value is @samp{combined}.
25758 Specify display color mode.
25760 It accepts the following values:
25763 each channel is displayed in a separate color
25765 each channel is displayed using the same color scheme
25767 each channel is displayed using the rainbow color scheme
25769 each channel is displayed using the moreland color scheme
25771 each channel is displayed using the nebulae color scheme
25773 each channel is displayed using the fire color scheme
25775 each channel is displayed using the fiery color scheme
25777 each channel is displayed using the fruit color scheme
25779 each channel is displayed using the cool color scheme
25781 each channel is displayed using the magma color scheme
25783 each channel is displayed using the green color scheme
25785 each channel is displayed using the viridis color scheme
25787 each channel is displayed using the plasma color scheme
25789 each channel is displayed using the cividis color scheme
25791 each channel is displayed using the terrain color scheme
25794 Default value is @samp{channel}.
25797 Specify scale used for calculating intensity color values.
25799 It accepts the following values:
25804 square root, default
25815 Default value is @samp{sqrt}.
25818 Specify frequency scale.
25820 It accepts the following values:
25828 Default value is @samp{lin}.
25831 Set saturation modifier for displayed colors. Negative values provide
25832 alternative color scheme. @code{0} is no saturation at all.
25833 Saturation must be in [-10.0, 10.0] range.
25834 Default value is @code{1}.
25837 Set window function.
25839 It accepts the following values:
25864 Default value is @code{hann}.
25867 Set orientation of time vs frequency axis. Can be @code{vertical} or
25868 @code{horizontal}. Default is @code{vertical}.
25871 Set ratio of overlap window. Default value is @code{0}.
25872 When value is @code{1} overlap is set to recommended size for specific
25873 window function currently used.
25876 Set scale gain for calculating intensity color values.
25877 Default value is @code{1}.
25880 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
25883 Set color rotation, must be in [-1.0, 1.0] range.
25884 Default value is @code{0}.
25887 Set start frequency from which to display spectrogram. Default is @code{0}.
25890 Set stop frequency to which to display spectrogram. Default is @code{0}.
25893 Set upper frame rate limit. Default is @code{auto}, unlimited.
25896 Draw time and frequency axes and legends. Default is disabled.
25899 The usage is very similar to the showwaves filter; see the examples in that
25902 @subsection Examples
25906 Large window with logarithmic color scaling:
25908 showspectrum=s=1280x480:scale=log
25912 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
25914 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
25915 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
25919 @section showspectrumpic
25921 Convert input audio to a single video frame, representing the audio frequency
25924 The filter accepts the following options:
25928 Specify the video size for the output. For the syntax of this option, check the
25929 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25930 Default value is @code{4096x2048}.
25933 Specify display mode.
25935 It accepts the following values:
25938 all channels are displayed in the same row
25940 all channels are displayed in separate rows
25942 Default value is @samp{combined}.
25945 Specify display color mode.
25947 It accepts the following values:
25950 each channel is displayed in a separate color
25952 each channel is displayed using the same color scheme
25954 each channel is displayed using the rainbow color scheme
25956 each channel is displayed using the moreland color scheme
25958 each channel is displayed using the nebulae color scheme
25960 each channel is displayed using the fire color scheme
25962 each channel is displayed using the fiery color scheme
25964 each channel is displayed using the fruit color scheme
25966 each channel is displayed using the cool color scheme
25968 each channel is displayed using the magma color scheme
25970 each channel is displayed using the green color scheme
25972 each channel is displayed using the viridis color scheme
25974 each channel is displayed using the plasma color scheme
25976 each channel is displayed using the cividis color scheme
25978 each channel is displayed using the terrain color scheme
25980 Default value is @samp{intensity}.
25983 Specify scale used for calculating intensity color values.
25985 It accepts the following values:
25990 square root, default
26000 Default value is @samp{log}.
26003 Specify frequency scale.
26005 It accepts the following values:
26013 Default value is @samp{lin}.
26016 Set saturation modifier for displayed colors. Negative values provide
26017 alternative color scheme. @code{0} is no saturation at all.
26018 Saturation must be in [-10.0, 10.0] range.
26019 Default value is @code{1}.
26022 Set window function.
26024 It accepts the following values:
26048 Default value is @code{hann}.
26051 Set orientation of time vs frequency axis. Can be @code{vertical} or
26052 @code{horizontal}. Default is @code{vertical}.
26055 Set scale gain for calculating intensity color values.
26056 Default value is @code{1}.
26059 Draw time and frequency axes and legends. Default is enabled.
26062 Set color rotation, must be in [-1.0, 1.0] range.
26063 Default value is @code{0}.
26066 Set start frequency from which to display spectrogram. Default is @code{0}.
26069 Set stop frequency to which to display spectrogram. Default is @code{0}.
26072 @subsection Examples
26076 Extract an audio spectrogram of a whole audio track
26077 in a 1024x1024 picture using @command{ffmpeg}:
26079 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
26083 @section showvolume
26085 Convert input audio volume to a video output.
26087 The filter accepts the following options:
26094 Set border width, allowed range is [0, 5]. Default is 1.
26097 Set channel width, allowed range is [80, 8192]. Default is 400.
26100 Set channel height, allowed range is [1, 900]. Default is 20.
26103 Set fade, allowed range is [0, 1]. Default is 0.95.
26106 Set volume color expression.
26108 The expression can use the following variables:
26112 Current max volume of channel in dB.
26118 Current channel number, starting from 0.
26122 If set, displays channel names. Default is enabled.
26125 If set, displays volume values. Default is enabled.
26128 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
26129 default is @code{h}.
26132 Set step size, allowed range is [0, 5]. Default is 0, which means
26136 Set background opacity, allowed range is [0, 1]. Default is 0.
26139 Set metering mode, can be peak: @code{p} or rms: @code{r},
26140 default is @code{p}.
26143 Set display scale, can be linear: @code{lin} or log: @code{log},
26144 default is @code{lin}.
26148 If set to > 0., display a line for the max level
26149 in the previous seconds.
26150 default is disabled: @code{0.}
26153 The color of the max line. Use when @code{dm} option is set to > 0.
26154 default is: @code{orange}
26159 Convert input audio to a video output, representing the samples waves.
26161 The filter accepts the following options:
26165 Specify the video size for the output. For the syntax of this option, check the
26166 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26167 Default value is @code{600x240}.
26172 Available values are:
26175 Draw a point for each sample.
26178 Draw a vertical line for each sample.
26181 Draw a point for each sample and a line between them.
26184 Draw a centered vertical line for each sample.
26187 Default value is @code{point}.
26190 Set the number of samples which are printed on the same column. A
26191 larger value will decrease the frame rate. Must be a positive
26192 integer. This option can be set only if the value for @var{rate}
26193 is not explicitly specified.
26196 Set the (approximate) output frame rate. This is done by setting the
26197 option @var{n}. Default value is "25".
26199 @item split_channels
26200 Set if channels should be drawn separately or overlap. Default value is 0.
26203 Set colors separated by '|' which are going to be used for drawing of each channel.
26206 Set amplitude scale.
26208 Available values are:
26226 Set the draw mode. This is mostly useful to set for high @var{n}.
26228 Available values are:
26231 Scale pixel values for each drawn sample.
26234 Draw every sample directly.
26237 Default value is @code{scale}.
26240 @subsection Examples
26244 Output the input file audio and the corresponding video representation
26247 amovie=a.mp3,asplit[out0],showwaves[out1]
26251 Create a synthetic signal and show it with showwaves, forcing a
26252 frame rate of 30 frames per second:
26254 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
26258 @section showwavespic
26260 Convert input audio to a single video frame, representing the samples waves.
26262 The filter accepts the following options:
26266 Specify the video size for the output. For the syntax of this option, check the
26267 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26268 Default value is @code{600x240}.
26270 @item split_channels
26271 Set if channels should be drawn separately or overlap. Default value is 0.
26274 Set colors separated by '|' which are going to be used for drawing of each channel.
26277 Set amplitude scale.
26279 Available values are:
26299 Available values are:
26302 Scale pixel values for each drawn sample.
26305 Draw every sample directly.
26308 Default value is @code{scale}.
26311 Set the filter mode.
26313 Available values are:
26316 Use average samples values for each drawn sample.
26319 Use peak samples values for each drawn sample.
26322 Default value is @code{average}.
26325 @subsection Examples
26329 Extract a channel split representation of the wave form of a whole audio track
26330 in a 1024x800 picture using @command{ffmpeg}:
26332 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
26336 @section sidedata, asidedata
26338 Delete frame side data, or select frames based on it.
26340 This filter accepts the following options:
26344 Set mode of operation of the filter.
26346 Can be one of the following:
26350 Select every frame with side data of @code{type}.
26353 Delete side data of @code{type}. If @code{type} is not set, delete all side
26359 Set side data type used with all modes. Must be set for @code{select} mode. For
26360 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
26361 in @file{libavutil/frame.h}. For example, to choose
26362 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
26366 @section spectrumsynth
26368 Synthesize audio from 2 input video spectrums, first input stream represents
26369 magnitude across time and second represents phase across time.
26370 The filter will transform from frequency domain as displayed in videos back
26371 to time domain as presented in audio output.
26373 This filter is primarily created for reversing processed @ref{showspectrum}
26374 filter outputs, but can synthesize sound from other spectrograms too.
26375 But in such case results are going to be poor if the phase data is not
26376 available, because in such cases phase data need to be recreated, usually
26377 it's just recreated from random noise.
26378 For best results use gray only output (@code{channel} color mode in
26379 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
26380 @code{lin} scale for phase video. To produce phase, for 2nd video, use
26381 @code{data} option. Inputs videos should generally use @code{fullframe}
26382 slide mode as that saves resources needed for decoding video.
26384 The filter accepts the following options:
26388 Specify sample rate of output audio, the sample rate of audio from which
26389 spectrum was generated may differ.
26392 Set number of channels represented in input video spectrums.
26395 Set scale which was used when generating magnitude input spectrum.
26396 Can be @code{lin} or @code{log}. Default is @code{log}.
26399 Set slide which was used when generating inputs spectrums.
26400 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
26401 Default is @code{fullframe}.
26404 Set window function used for resynthesis.
26407 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
26408 which means optimal overlap for selected window function will be picked.
26411 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
26412 Default is @code{vertical}.
26415 @subsection Examples
26419 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
26420 then resynthesize videos back to audio with spectrumsynth:
26422 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
26423 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
26424 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
26428 @section split, asplit
26430 Split input into several identical outputs.
26432 @code{asplit} works with audio input, @code{split} with video.
26434 The filter accepts a single parameter which specifies the number of outputs. If
26435 unspecified, it defaults to 2.
26437 @subsection Examples
26441 Create two separate outputs from the same input:
26443 [in] split [out0][out1]
26447 To create 3 or more outputs, you need to specify the number of
26450 [in] asplit=3 [out0][out1][out2]
26454 Create two separate outputs from the same input, one cropped and
26457 [in] split [splitout1][splitout2];
26458 [splitout1] crop=100:100:0:0 [cropout];
26459 [splitout2] pad=200:200:100:100 [padout];
26463 Create 5 copies of the input audio with @command{ffmpeg}:
26465 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
26471 Receive commands sent through a libzmq client, and forward them to
26472 filters in the filtergraph.
26474 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
26475 must be inserted between two video filters, @code{azmq} between two
26476 audio filters. Both are capable to send messages to any filter type.
26478 To enable these filters you need to install the libzmq library and
26479 headers and configure FFmpeg with @code{--enable-libzmq}.
26481 For more information about libzmq see:
26482 @url{http://www.zeromq.org/}
26484 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
26485 receives messages sent through a network interface defined by the
26486 @option{bind_address} (or the abbreviation "@option{b}") option.
26487 Default value of this option is @file{tcp://localhost:5555}. You may
26488 want to alter this value to your needs, but do not forget to escape any
26489 ':' signs (see @ref{filtergraph escaping}).
26491 The received message must be in the form:
26493 @var{TARGET} @var{COMMAND} [@var{ARG}]
26496 @var{TARGET} specifies the target of the command, usually the name of
26497 the filter class or a specific filter instance name. The default
26498 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
26499 but you can override this by using the @samp{filter_name@@id} syntax
26500 (see @ref{Filtergraph syntax}).
26502 @var{COMMAND} specifies the name of the command for the target filter.
26504 @var{ARG} is optional and specifies the optional argument list for the
26505 given @var{COMMAND}.
26507 Upon reception, the message is processed and the corresponding command
26508 is injected into the filtergraph. Depending on the result, the filter
26509 will send a reply to the client, adopting the format:
26511 @var{ERROR_CODE} @var{ERROR_REASON}
26515 @var{MESSAGE} is optional.
26517 @subsection Examples
26519 Look at @file{tools/zmqsend} for an example of a zmq client which can
26520 be used to send commands processed by these filters.
26522 Consider the following filtergraph generated by @command{ffplay}.
26523 In this example the last overlay filter has an instance name. All other
26524 filters will have default instance names.
26527 ffplay -dumpgraph 1 -f lavfi "
26528 color=s=100x100:c=red [l];
26529 color=s=100x100:c=blue [r];
26530 nullsrc=s=200x100, zmq [bg];
26531 [bg][l] overlay [bg+l];
26532 [bg+l][r] overlay@@my=x=100 "
26535 To change the color of the left side of the video, the following
26536 command can be used:
26538 echo Parsed_color_0 c yellow | tools/zmqsend
26541 To change the right side:
26543 echo Parsed_color_1 c pink | tools/zmqsend
26546 To change the position of the right side:
26548 echo overlay@@my x 150 | tools/zmqsend
26552 @c man end MULTIMEDIA FILTERS
26554 @chapter Multimedia Sources
26555 @c man begin MULTIMEDIA SOURCES
26557 Below is a description of the currently available multimedia sources.
26561 This is the same as @ref{movie} source, except it selects an audio
26567 Read audio and/or video stream(s) from a movie container.
26569 It accepts the following parameters:
26573 The name of the resource to read (not necessarily a file; it can also be a
26574 device or a stream accessed through some protocol).
26576 @item format_name, f
26577 Specifies the format assumed for the movie to read, and can be either
26578 the name of a container or an input device. If not specified, the
26579 format is guessed from @var{movie_name} or by probing.
26581 @item seek_point, sp
26582 Specifies the seek point in seconds. The frames will be output
26583 starting from this seek point. The parameter is evaluated with
26584 @code{av_strtod}, so the numerical value may be suffixed by an IS
26585 postfix. The default value is "0".
26588 Specifies the streams to read. Several streams can be specified,
26589 separated by "+". The source will then have as many outputs, in the
26590 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
26591 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
26592 respectively the default (best suited) video and audio stream. Default
26593 is "dv", or "da" if the filter is called as "amovie".
26595 @item stream_index, si
26596 Specifies the index of the video stream to read. If the value is -1,
26597 the most suitable video stream will be automatically selected. The default
26598 value is "-1". Deprecated. If the filter is called "amovie", it will select
26599 audio instead of video.
26602 Specifies how many times to read the stream in sequence.
26603 If the value is 0, the stream will be looped infinitely.
26604 Default value is "1".
26606 Note that when the movie is looped the source timestamps are not
26607 changed, so it will generate non monotonically increasing timestamps.
26609 @item discontinuity
26610 Specifies the time difference between frames above which the point is
26611 considered a timestamp discontinuity which is removed by adjusting the later
26615 It allows overlaying a second video on top of the main input of
26616 a filtergraph, as shown in this graph:
26618 input -----------> deltapts0 --> overlay --> output
26621 movie --> scale--> deltapts1 -------+
26623 @subsection Examples
26627 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
26628 on top of the input labelled "in":
26630 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
26631 [in] setpts=PTS-STARTPTS [main];
26632 [main][over] overlay=16:16 [out]
26636 Read from a video4linux2 device, and overlay it on top of the input
26639 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
26640 [in] setpts=PTS-STARTPTS [main];
26641 [main][over] overlay=16:16 [out]
26645 Read the first video stream and the audio stream with id 0x81 from
26646 dvd.vob; the video is connected to the pad named "video" and the audio is
26647 connected to the pad named "audio":
26649 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
26653 @subsection Commands
26655 Both movie and amovie support the following commands:
26658 Perform seek using "av_seek_frame".
26659 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
26662 @var{stream_index}: If stream_index is -1, a default
26663 stream is selected, and @var{timestamp} is automatically converted
26664 from AV_TIME_BASE units to the stream specific time_base.
26666 @var{timestamp}: Timestamp in AVStream.time_base units
26667 or, if no stream is specified, in AV_TIME_BASE units.
26669 @var{flags}: Flags which select direction and seeking mode.
26673 Get movie duration in AV_TIME_BASE units.
26677 @c man end MULTIMEDIA SOURCES