1 @chapter Filtering Introduction
2 @c man begin FILTERING INTRODUCTION
4 Filtering in FFmpeg is enabled through the libavfilter library.
6 In libavfilter, a filter can have multiple inputs and multiple
8 To illustrate the sorts of things that are possible, we consider the
13 input --> split ---------------------> overlay --> output
16 +-----> crop --> vflip -------+
19 This filtergraph splits the input stream in two streams, then sends one
20 stream through the crop filter and the vflip filter, before merging it
21 back with the other stream by overlaying it on top. You can use the
22 following command to achieve this:
25 ffmpeg -i INPUT -vf "split [main][tmp]; [tmp] crop=iw:ih/2:0:0, vflip [flip]; [main][flip] overlay=0:H/2" OUTPUT
28 The result will be that the top half of the video is mirrored
29 onto the bottom half of the output video.
31 Filters in the same linear chain are separated by commas, and distinct
32 linear chains of filters are separated by semicolons. In our example,
33 @var{crop,vflip} are in one linear chain, @var{split} and
34 @var{overlay} are separately in another. The points where the linear
35 chains join are labelled by names enclosed in square brackets. In the
36 example, the split filter generates two outputs that are associated to
37 the labels @var{[main]} and @var{[tmp]}.
39 The stream sent to the second output of @var{split}, labelled as
40 @var{[tmp]}, is processed through the @var{crop} filter, which crops
41 away the lower half part of the video, and then vertically flipped. The
42 @var{overlay} filter takes in input the first unchanged output of the
43 split filter (which was labelled as @var{[main]}), and overlay on its
44 lower half the output generated by the @var{crop,vflip} filterchain.
46 Some filters take in input a list of parameters: they are specified
47 after the filter name and an equal sign, and are separated from each other
50 There exist so-called @var{source filters} that do not have an
51 audio/video input, and @var{sink filters} that will not have audio/video
54 @c man end FILTERING INTRODUCTION
57 @c man begin GRAPH2DOT
59 The @file{graph2dot} program included in the FFmpeg @file{tools}
60 directory can be used to parse a filtergraph description and issue a
61 corresponding textual representation in the dot language.
68 to see how to use @file{graph2dot}.
70 You can then pass the dot description to the @file{dot} program (from
71 the graphviz suite of programs) and obtain a graphical representation
74 For example the sequence of commands:
76 echo @var{GRAPH_DESCRIPTION} | \
77 tools/graph2dot -o graph.tmp && \
78 dot -Tpng graph.tmp -o graph.png && \
82 can be used to create and display an image representing the graph
83 described by the @var{GRAPH_DESCRIPTION} string. Note that this string must be
84 a complete self-contained graph, with its inputs and outputs explicitly defined.
85 For example if your command line is of the form:
87 ffmpeg -i infile -vf scale=640:360 outfile
89 your @var{GRAPH_DESCRIPTION} string will need to be of the form:
91 nullsrc,scale=640:360,nullsink
93 you may also need to set the @var{nullsrc} parameters and add a @var{format}
94 filter in order to simulate a specific input file.
98 @chapter Filtergraph description
99 @c man begin FILTERGRAPH DESCRIPTION
101 A filtergraph is a directed graph of connected filters. It can contain
102 cycles, and there can be multiple links between a pair of
103 filters. Each link has one input pad on one side connecting it to one
104 filter from which it takes its input, and one output pad on the other
105 side connecting it to one filter accepting its output.
107 Each filter in a filtergraph is an instance of a filter class
108 registered in the application, which defines the features and the
109 number of input and output pads of the filter.
111 A filter with no input pads is called a "source", and a filter with no
112 output pads is called a "sink".
114 @anchor{Filtergraph syntax}
115 @section Filtergraph syntax
117 A filtergraph has a textual representation, which is recognized by the
118 @option{-filter}/@option{-vf}/@option{-af} and
119 @option{-filter_complex} options in @command{ffmpeg} and
120 @option{-vf}/@option{-af} in @command{ffplay}, and by the
121 @code{avfilter_graph_parse_ptr()} function defined in
122 @file{libavfilter/avfilter.h}.
124 A filterchain consists of a sequence of connected filters, each one
125 connected to the previous one in the sequence. A filterchain is
126 represented by a list of ","-separated filter descriptions.
128 A filtergraph consists of a sequence of filterchains. A sequence of
129 filterchains is represented by a list of ";"-separated filterchain
132 A filter is represented by a string of the form:
133 [@var{in_link_1}]...[@var{in_link_N}]@var{filter_name}@@@var{id}=@var{arguments}[@var{out_link_1}]...[@var{out_link_M}]
135 @var{filter_name} is the name of the filter class of which the
136 described filter is an instance of, and has to be the name of one of
137 the filter classes registered in the program optionally followed by "@@@var{id}".
138 The name of the filter class is optionally followed by a string
141 @var{arguments} is a string which contains the parameters used to
142 initialize the filter instance. It may have one of two forms:
146 A ':'-separated list of @var{key=value} pairs.
149 A ':'-separated list of @var{value}. In this case, the keys are assumed to be
150 the option names in the order they are declared. E.g. the @code{fade} filter
151 declares three options in this order -- @option{type}, @option{start_frame} and
152 @option{nb_frames}. Then the parameter list @var{in:0:30} means that the value
153 @var{in} is assigned to the option @option{type}, @var{0} to
154 @option{start_frame} and @var{30} to @option{nb_frames}.
157 A ':'-separated list of mixed direct @var{value} and long @var{key=value}
158 pairs. The direct @var{value} must precede the @var{key=value} pairs, and
159 follow the same constraints order of the previous point. The following
160 @var{key=value} pairs can be set in any preferred order.
164 If the option value itself is a list of items (e.g. the @code{format} filter
165 takes a list of pixel formats), the items in the list are usually separated by
168 The list of arguments can be quoted using the character @samp{'} as initial
169 and ending mark, and the character @samp{\} for escaping the characters
170 within the quoted text; otherwise the argument string is considered
171 terminated when the next special character (belonging to the set
172 @samp{[]=;,}) is encountered.
174 The name and arguments of the filter are optionally preceded and
175 followed by a list of link labels.
176 A link label allows one to name a link and associate it to a filter output
177 or input pad. The preceding labels @var{in_link_1}
178 ... @var{in_link_N}, are associated to the filter input pads,
179 the following labels @var{out_link_1} ... @var{out_link_M}, are
180 associated to the output pads.
182 When two link labels with the same name are found in the
183 filtergraph, a link between the corresponding input and output pad is
186 If an output pad is not labelled, it is linked by default to the first
187 unlabelled input pad of the next filter in the filterchain.
188 For example in the filterchain
190 nullsrc, split[L1], [L2]overlay, nullsink
192 the split filter instance has two output pads, and the overlay filter
193 instance two input pads. The first output pad of split is labelled
194 "L1", the first input pad of overlay is labelled "L2", and the second
195 output pad of split is linked to the second input pad of overlay,
196 which are both unlabelled.
198 In a filter description, if the input label of the first filter is not
199 specified, "in" is assumed; if the output label of the last filter is not
200 specified, "out" is assumed.
202 In a complete filterchain all the unlabelled filter input and output
203 pads must be connected. A filtergraph is considered valid if all the
204 filter input and output pads of all the filterchains are connected.
206 Libavfilter will automatically insert @ref{scale} filters where format
207 conversion is required. It is possible to specify swscale flags
208 for those automatically inserted scalers by prepending
209 @code{sws_flags=@var{flags};}
210 to the filtergraph description.
212 Here is a BNF description of the filtergraph syntax:
214 @var{NAME} ::= sequence of alphanumeric characters and '_'
215 @var{FILTER_NAME} ::= @var{NAME}["@@"@var{NAME}]
216 @var{LINKLABEL} ::= "[" @var{NAME} "]"
217 @var{LINKLABELS} ::= @var{LINKLABEL} [@var{LINKLABELS}]
218 @var{FILTER_ARGUMENTS} ::= sequence of chars (possibly quoted)
219 @var{FILTER} ::= [@var{LINKLABELS}] @var{FILTER_NAME} ["=" @var{FILTER_ARGUMENTS}] [@var{LINKLABELS}]
220 @var{FILTERCHAIN} ::= @var{FILTER} [,@var{FILTERCHAIN}]
221 @var{FILTERGRAPH} ::= [sws_flags=@var{flags};] @var{FILTERCHAIN} [;@var{FILTERGRAPH}]
224 @anchor{filtergraph escaping}
225 @section Notes on filtergraph escaping
227 Filtergraph description composition entails several levels of
228 escaping. See @ref{quoting_and_escaping,,the "Quoting and escaping"
229 section in the ffmpeg-utils(1) manual,ffmpeg-utils} for more
230 information about the employed escaping procedure.
232 A first level escaping affects the content of each filter option
233 value, which may contain the special character @code{:} used to
234 separate values, or one of the escaping characters @code{\'}.
236 A second level escaping affects the whole filter description, which
237 may contain the escaping characters @code{\'} or the special
238 characters @code{[],;} used by the filtergraph description.
240 Finally, when you specify a filtergraph on a shell commandline, you
241 need to perform a third level escaping for the shell special
242 characters contained within it.
244 For example, consider the following string to be embedded in
245 the @ref{drawtext} filter description @option{text} value:
247 this is a 'string': may contain one, or more, special characters
250 This string contains the @code{'} special escaping character, and the
251 @code{:} special character, so it needs to be escaped in this way:
253 text=this is a \'string\'\: may contain one, or more, special characters
256 A second level of escaping is required when embedding the filter
257 description in a filtergraph description, in order to escape all the
258 filtergraph special characters. Thus the example above becomes:
260 drawtext=text=this is a \\\'string\\\'\\: may contain one\, or more\, special characters
262 (note that in addition to the @code{\'} escaping special characters,
263 also @code{,} needs to be escaped).
265 Finally an additional level of escaping is needed when writing the
266 filtergraph description in a shell command, which depends on the
267 escaping rules of the adopted shell. For example, assuming that
268 @code{\} is special and needs to be escaped with another @code{\}, the
269 previous string will finally result in:
271 -vf "drawtext=text=this is a \\\\\\'string\\\\\\'\\\\: may contain one\\, or more\\, special characters"
274 @chapter Timeline editing
276 Some filters support a generic @option{enable} option. For the filters
277 supporting timeline editing, this option can be set to an expression which is
278 evaluated before sending a frame to the filter. If the evaluation is non-zero,
279 the filter will be enabled, otherwise the frame will be sent unchanged to the
280 next filter in the filtergraph.
282 The expression accepts the following values:
285 timestamp expressed in seconds, NAN if the input timestamp is unknown
288 sequential number of the input frame, starting from 0
291 the position in the file of the input frame, NAN if unknown
295 width and height of the input frame if video
298 Additionally, these filters support an @option{enable} command that can be used
299 to re-define the expression.
301 Like any other filtering option, the @option{enable} option follows the same
304 For example, to enable a blur filter (@ref{smartblur}) from 10 seconds to 3
305 minutes, and a @ref{curves} filter starting at 3 seconds:
307 smartblur = enable='between(t,10,3*60)',
308 curves = enable='gte(t,3)' : preset=cross_process
311 See @code{ffmpeg -filters} to view which filters have timeline support.
313 @c man end FILTERGRAPH DESCRIPTION
316 @chapter Changing options at runtime with a command
318 Some options can be changed during the operation of the filter using
319 a command. These options are marked 'T' on the output of
320 @command{ffmpeg} @option{-h filter=<name of filter>}.
321 The name of the command is the name of the option and the argument is
325 @chapter Options for filters with several inputs (framesync)
326 @c man begin OPTIONS FOR FILTERS WITH SEVERAL INPUTS
328 Some filters with several inputs support a common set of options.
329 These options can only be set by name, not with the short notation.
333 The action to take when EOF is encountered on the secondary input; it accepts
334 one of the following values:
338 Repeat the last frame (the default).
342 Pass the main input through.
346 If set to 1, force the output to terminate when the shortest input
347 terminates. Default value is 0.
350 If set to 1, force the filter to extend the last frame of secondary streams
351 until the end of the primary stream. A value of 0 disables this behavior.
355 @c man end OPTIONS FOR FILTERS WITH SEVERAL INPUTS
357 @chapter Audio Filters
358 @c man begin AUDIO FILTERS
360 When you configure your FFmpeg build, you can disable any of the
361 existing filters using @code{--disable-filters}.
362 The configure output will show the audio filters included in your
365 Below is a description of the currently available audio filters.
369 A compressor is mainly used to reduce the dynamic range of a signal.
370 Especially modern music is mostly compressed at a high ratio to
371 improve the overall loudness. It's done to get the highest attention
372 of a listener, "fatten" the sound and bring more "power" to the track.
373 If a signal is compressed too much it may sound dull or "dead"
374 afterwards or it may start to "pump" (which could be a powerful effect
375 but can also destroy a track completely).
376 The right compression is the key to reach a professional sound and is
377 the high art of mixing and mastering. Because of its complex settings
378 it may take a long time to get the right feeling for this kind of effect.
380 Compression is done by detecting the volume above a chosen level
381 @code{threshold} and dividing it by the factor set with @code{ratio}.
382 So if you set the threshold to -12dB and your signal reaches -6dB a ratio
383 of 2:1 will result in a signal at -9dB. Because an exact manipulation of
384 the signal would cause distortion of the waveform the reduction can be
385 levelled over the time. This is done by setting "Attack" and "Release".
386 @code{attack} determines how long the signal has to rise above the threshold
387 before any reduction will occur and @code{release} sets the time the signal
388 has to fall below the threshold to reduce the reduction again. Shorter signals
389 than the chosen attack time will be left untouched.
390 The overall reduction of the signal can be made up afterwards with the
391 @code{makeup} setting. So compressing the peaks of a signal about 6dB and
392 raising the makeup to this level results in a signal twice as loud than the
393 source. To gain a softer entry in the compression the @code{knee} flattens the
394 hard edge at the threshold in the range of the chosen decibels.
396 The filter accepts the following options:
400 Set input gain. Default is 1. Range is between 0.015625 and 64.
403 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
404 Default is @code{downward}.
407 If a signal of stream rises above this level it will affect the gain
409 By default it is 0.125. Range is between 0.00097563 and 1.
412 Set a ratio by which the signal is reduced. 1:2 means that if the level
413 rose 4dB above the threshold, it will be only 2dB above after the reduction.
414 Default is 2. Range is between 1 and 20.
417 Amount of milliseconds the signal has to rise above the threshold before gain
418 reduction starts. Default is 20. Range is between 0.01 and 2000.
421 Amount of milliseconds the signal has to fall below the threshold before
422 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
425 Set the amount by how much signal will be amplified after processing.
426 Default is 1. Range is from 1 to 64.
429 Curve the sharp knee around the threshold to enter gain reduction more softly.
430 Default is 2.82843. Range is between 1 and 8.
433 Choose if the @code{average} level between all channels of input stream
434 or the louder(@code{maximum}) channel of input stream affects the
435 reduction. Default is @code{average}.
438 Should the exact signal be taken in case of @code{peak} or an RMS one in case
439 of @code{rms}. Default is @code{rms} which is mostly smoother.
442 How much to use compressed signal in output. Default is 1.
443 Range is between 0 and 1.
448 This filter supports the all above options as @ref{commands}.
451 Simple audio dynamic range compression/expansion filter.
453 The filter accepts the following options:
457 Set contrast. Default is 33. Allowed range is between 0 and 100.
462 Copy the input audio source unchanged to the output. This is mainly useful for
467 Apply cross fade from one input audio stream to another input audio stream.
468 The cross fade is applied for specified duration near the end of first stream.
470 The filter accepts the following options:
474 Specify the number of samples for which the cross fade effect has to last.
475 At the end of the cross fade effect the first input audio will be completely
476 silent. Default is 44100.
479 Specify the duration of the cross fade effect. See
480 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
481 for the accepted syntax.
482 By default the duration is determined by @var{nb_samples}.
483 If set this option is used instead of @var{nb_samples}.
486 Should first stream end overlap with second stream start. Default is enabled.
489 Set curve for cross fade transition for first stream.
492 Set curve for cross fade transition for second stream.
494 For description of available curve types see @ref{afade} filter description.
501 Cross fade from one input to another:
503 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
507 Cross fade from one input to another but without overlapping:
509 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
514 Split audio stream into several bands.
516 This filter splits audio stream into two or more frequency ranges.
517 Summing all streams back will give flat output.
519 The filter accepts the following options:
523 Set split frequencies. Those must be positive and increasing.
526 Set filter order for each band split. This controls filter roll-off or steepness
527 of filter transfer function.
528 Available values are:
553 Default is @var{4th}.
556 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
559 Set output gain for each band. Default value is 1 for all bands.
566 Split input audio stream into two bands (low and high) with split frequency of 1500 Hz,
567 each band will be in separate stream:
569 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
573 Same as above, but with higher filter order:
575 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500:order=8th[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
579 Same as above, but also with additional middle band (frequencies between 1500 and 8000):
581 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500 8000:order=8th[LOW][MID][HIGH]' -map '[LOW]' low.wav -map '[MID]' mid.wav -map '[HIGH]' high.wav
587 Reduce audio bit resolution.
589 This filter is bit crusher with enhanced functionality. A bit crusher
590 is used to audibly reduce number of bits an audio signal is sampled
591 with. This doesn't change the bit depth at all, it just produces the
592 effect. Material reduced in bit depth sounds more harsh and "digital".
593 This filter is able to even round to continuous values instead of discrete
595 Additionally it has a D/C offset which results in different crushing of
596 the lower and the upper half of the signal.
597 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
599 Another feature of this filter is the logarithmic mode.
600 This setting switches from linear distances between bits to logarithmic ones.
601 The result is a much more "natural" sounding crusher which doesn't gate low
602 signals for example. The human ear has a logarithmic perception,
603 so this kind of crushing is much more pleasant.
604 Logarithmic crushing is also able to get anti-aliased.
606 The filter accepts the following options:
622 Can be linear: @code{lin} or logarithmic: @code{log}.
631 Set sample reduction.
634 Enable LFO. By default disabled.
645 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
649 Remove impulsive noise from input audio.
651 Samples detected as impulsive noise are replaced by interpolated samples using
652 autoregressive modelling.
656 Set window size, in milliseconds. Allowed range is from @code{10} to
657 @code{100}. Default value is @code{55} milliseconds.
658 This sets size of window which will be processed at once.
661 Set window overlap, in percentage of window size. Allowed range is from
662 @code{50} to @code{95}. Default value is @code{75} percent.
663 Setting this to a very high value increases impulsive noise removal but makes
664 whole process much slower.
667 Set autoregression order, in percentage of window size. Allowed range is from
668 @code{0} to @code{25}. Default value is @code{2} percent. This option also
669 controls quality of interpolated samples using neighbour good samples.
672 Set threshold value. Allowed range is from @code{1} to @code{100}.
673 Default value is @code{2}.
674 This controls the strength of impulsive noise which is going to be removed.
675 The lower value, the more samples will be detected as impulsive noise.
678 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
679 @code{10}. Default value is @code{2}.
680 If any two samples detected as noise are spaced less than this value then any
681 sample between those two samples will be also detected as noise.
686 It accepts the following values:
689 Select overlap-add method. Even not interpolated samples are slightly
690 changed with this method.
693 Select overlap-save method. Not interpolated samples remain unchanged.
696 Default value is @code{a}.
700 Remove clipped samples from input audio.
702 Samples detected as clipped are replaced by interpolated samples using
703 autoregressive modelling.
707 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
708 Default value is @code{55} milliseconds.
709 This sets size of window which will be processed at once.
712 Set window overlap, in percentage of window size. Allowed range is from @code{50}
713 to @code{95}. Default value is @code{75} percent.
716 Set autoregression order, in percentage of window size. Allowed range is from
717 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
718 quality of interpolated samples using neighbour good samples.
721 Set threshold value. Allowed range is from @code{1} to @code{100}.
722 Default value is @code{10}. Higher values make clip detection less aggressive.
725 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
726 Default value is @code{1000}. Higher values make clip detection less aggressive.
731 It accepts the following values:
734 Select overlap-add method. Even not interpolated samples are slightly changed
738 Select overlap-save method. Not interpolated samples remain unchanged.
741 Default value is @code{a}.
746 Delay one or more audio channels.
748 Samples in delayed channel are filled with silence.
750 The filter accepts the following option:
754 Set list of delays in milliseconds for each channel separated by '|'.
755 Unused delays will be silently ignored. If number of given delays is
756 smaller than number of channels all remaining channels will not be delayed.
757 If you want to delay exact number of samples, append 'S' to number.
758 If you want instead to delay in seconds, append 's' to number.
761 Use last set delay for all remaining channels. By default is disabled.
762 This option if enabled changes how option @code{delays} is interpreted.
769 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
770 the second channel (and any other channels that may be present) unchanged.
776 Delay second channel by 500 samples, the third channel by 700 samples and leave
777 the first channel (and any other channels that may be present) unchanged.
783 Delay all channels by same number of samples:
785 adelay=delays=64S:all=1
790 Remedy denormals in audio by adding extremely low-level noise.
792 This filter shall be placed before any filter that can produce denormals.
794 A description of the accepted parameters follows.
798 Set level of added noise in dB. Default is @code{-351}.
799 Allowed range is from -451 to -90.
802 Set type of added noise.
815 Default is @code{dc}.
820 This filter supports the all above options as @ref{commands}.
822 @section aderivative, aintegral
824 Compute derivative/integral of audio stream.
826 Applying both filters one after another produces original audio.
830 Apply echoing to the input audio.
832 Echoes are reflected sound and can occur naturally amongst mountains
833 (and sometimes large buildings) when talking or shouting; digital echo
834 effects emulate this behaviour and are often used to help fill out the
835 sound of a single instrument or vocal. The time difference between the
836 original signal and the reflection is the @code{delay}, and the
837 loudness of the reflected signal is the @code{decay}.
838 Multiple echoes can have different delays and decays.
840 A description of the accepted parameters follows.
844 Set input gain of reflected signal. Default is @code{0.6}.
847 Set output gain of reflected signal. Default is @code{0.3}.
850 Set list of time intervals in milliseconds between original signal and reflections
851 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
852 Default is @code{1000}.
855 Set list of loudness of reflected signals separated by '|'.
856 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
857 Default is @code{0.5}.
864 Make it sound as if there are twice as many instruments as are actually playing:
866 aecho=0.8:0.88:60:0.4
870 If delay is very short, then it sounds like a (metallic) robot playing music:
876 A longer delay will sound like an open air concert in the mountains:
878 aecho=0.8:0.9:1000:0.3
882 Same as above but with one more mountain:
884 aecho=0.8:0.9:1000|1800:0.3|0.25
889 Audio emphasis filter creates or restores material directly taken from LPs or
890 emphased CDs with different filter curves. E.g. to store music on vinyl the
891 signal has to be altered by a filter first to even out the disadvantages of
892 this recording medium.
893 Once the material is played back the inverse filter has to be applied to
894 restore the distortion of the frequency response.
896 The filter accepts the following options:
906 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
907 use @code{production} mode. Default is @code{reproduction} mode.
910 Set filter type. Selects medium. Can be one of the following:
922 select Compact Disc (CD).
928 select 50µs (FM-KF).
930 select 75µs (FM-KF).
936 This filter supports the all above options as @ref{commands}.
940 Modify an audio signal according to the specified expressions.
942 This filter accepts one or more expressions (one for each channel),
943 which are evaluated and used to modify a corresponding audio signal.
945 It accepts the following parameters:
949 Set the '|'-separated expressions list for each separate channel. If
950 the number of input channels is greater than the number of
951 expressions, the last specified expression is used for the remaining
954 @item channel_layout, c
955 Set output channel layout. If not specified, the channel layout is
956 specified by the number of expressions. If set to @samp{same}, it will
957 use by default the same input channel layout.
960 Each expression in @var{exprs} can contain the following constants and functions:
964 channel number of the current expression
967 number of the evaluated sample, starting from 0
973 time of the evaluated sample expressed in seconds
976 @item nb_out_channels
977 input and output number of channels
980 the value of input channel with number @var{CH}
983 Note: this filter is slow. For faster processing you should use a
992 aeval=val(ch)/2:c=same
996 Invert phase of the second channel:
1005 Apply fade-in/out effect to input audio.
1007 A description of the accepted parameters follows.
1011 Specify the effect type, can be either @code{in} for fade-in, or
1012 @code{out} for a fade-out effect. Default is @code{in}.
1014 @item start_sample, ss
1015 Specify the number of the start sample for starting to apply the fade
1016 effect. Default is 0.
1018 @item nb_samples, ns
1019 Specify the number of samples for which the fade effect has to last. At
1020 the end of the fade-in effect the output audio will have the same
1021 volume as the input audio, at the end of the fade-out transition
1022 the output audio will be silence. Default is 44100.
1024 @item start_time, st
1025 Specify the start time of the fade effect. Default is 0.
1026 The value must be specified as a time duration; see
1027 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1028 for the accepted syntax.
1029 If set this option is used instead of @var{start_sample}.
1032 Specify the duration of the fade effect. See
1033 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1034 for the accepted syntax.
1035 At the end of the fade-in effect the output audio will have the same
1036 volume as the input audio, at the end of the fade-out transition
1037 the output audio will be silence.
1038 By default the duration is determined by @var{nb_samples}.
1039 If set this option is used instead of @var{nb_samples}.
1042 Set curve for fade transition.
1044 It accepts the following values:
1047 select triangular, linear slope (default)
1049 select quarter of sine wave
1051 select half of sine wave
1053 select exponential sine wave
1057 select inverted parabola
1071 select inverted quarter of sine wave
1073 select inverted half of sine wave
1075 select double-exponential seat
1077 select double-exponential sigmoid
1079 select logistic sigmoid
1081 select sine cardinal function
1083 select inverted sine cardinal function
1089 @subsection Commands
1091 This filter supports the all above options as @ref{commands}.
1093 @subsection Examples
1097 Fade in first 15 seconds of audio:
1099 afade=t=in:ss=0:d=15
1103 Fade out last 25 seconds of a 900 seconds audio:
1105 afade=t=out:st=875:d=25
1110 Denoise audio samples with FFT.
1112 A description of the accepted parameters follows.
1116 Set the noise reduction in dB, allowed range is 0.01 to 97.
1117 Default value is 12 dB.
1120 Set the noise floor in dB, allowed range is -80 to -20.
1121 Default value is -50 dB.
1126 It accepts the following values:
1135 Select shellac noise.
1138 Select custom noise, defined in @code{bn} option.
1140 Default value is white noise.
1144 Set custom band noise for every one of 15 bands.
1145 Bands are separated by ' ' or '|'.
1148 Set the residual floor in dB, allowed range is -80 to -20.
1149 Default value is -38 dB.
1152 Enable noise tracking. By default is disabled.
1153 With this enabled, noise floor is automatically adjusted.
1156 Enable residual tracking. By default is disabled.
1159 Set the output mode.
1161 It accepts the following values:
1164 Pass input unchanged.
1167 Pass noise filtered out.
1172 Default value is @var{o}.
1176 @subsection Commands
1178 This filter supports the following commands:
1180 @item sample_noise, sn
1181 Start or stop measuring noise profile.
1182 Syntax for the command is : "start" or "stop" string.
1183 After measuring noise profile is stopped it will be
1184 automatically applied in filtering.
1186 @item noise_reduction, nr
1187 Change noise reduction. Argument is single float number.
1188 Syntax for the command is : "@var{noise_reduction}"
1190 @item noise_floor, nf
1191 Change noise floor. Argument is single float number.
1192 Syntax for the command is : "@var{noise_floor}"
1194 @item output_mode, om
1195 Change output mode operation.
1196 Syntax for the command is : "i", "o" or "n" string.
1200 Apply arbitrary expressions to samples in frequency domain.
1204 Set frequency domain real expression for each separate channel separated
1205 by '|'. Default is "re".
1206 If the number of input channels is greater than the number of
1207 expressions, the last specified expression is used for the remaining
1211 Set frequency domain imaginary expression for each separate channel
1212 separated by '|'. Default is "im".
1214 Each expression in @var{real} and @var{imag} can contain the following
1215 constants and functions:
1222 current frequency bin number
1225 number of available bins
1228 channel number of the current expression
1237 current real part of frequency bin of current channel
1240 current imaginary part of frequency bin of current channel
1243 Return the value of real part of frequency bin at location (@var{bin},@var{channel})
1246 Return the value of imaginary part of frequency bin at location (@var{bin},@var{channel})
1250 Set window size. Allowed range is from 16 to 131072.
1251 Default is @code{4096}
1254 Set window function. Default is @code{hann}.
1257 Set window overlap. If set to 1, the recommended overlap for selected
1258 window function will be picked. Default is @code{0.75}.
1261 @subsection Examples
1265 Leave almost only low frequencies in audio:
1267 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1271 Apply robotize effect:
1273 afftfilt="real='hypot(re,im)*sin(0)':imag='hypot(re,im)*cos(0)':win_size=512:overlap=0.75"
1277 Apply whisper effect:
1279 afftfilt="real='hypot(re,im)*cos((random(0)*2-1)*2*3.14)':imag='hypot(re,im)*sin((random(1)*2-1)*2*3.14)':win_size=128:overlap=0.8"
1286 Apply an arbitrary Finite Impulse Response filter.
1288 This filter is designed for applying long FIR filters,
1289 up to 60 seconds long.
1291 It can be used as component for digital crossover filters,
1292 room equalization, cross talk cancellation, wavefield synthesis,
1293 auralization, ambiophonics, ambisonics and spatialization.
1295 This filter uses the streams higher than first one as FIR coefficients.
1296 If the non-first stream holds a single channel, it will be used
1297 for all input channels in the first stream, otherwise
1298 the number of channels in the non-first stream must be same as
1299 the number of channels in the first stream.
1301 It accepts the following parameters:
1305 Set dry gain. This sets input gain.
1308 Set wet gain. This sets final output gain.
1311 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1314 Enable applying gain measured from power of IR.
1316 Set which approach to use for auto gain measurement.
1320 Do not apply any gain.
1323 select peak gain, very conservative approach. This is default value.
1326 select DC gain, limited application.
1329 select gain to noise approach, this is most popular one.
1333 Set gain to be applied to IR coefficients before filtering.
1334 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1337 Set format of IR stream. Can be @code{mono} or @code{input}.
1338 Default is @code{input}.
1341 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1342 Allowed range is 0.1 to 60 seconds.
1345 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1346 By default it is disabled.
1349 Set for which IR channel to display frequency response. By default is first channel
1350 displayed. This option is used only when @var{response} is enabled.
1353 Set video stream size. This option is used only when @var{response} is enabled.
1356 Set video stream frame rate. This option is used only when @var{response} is enabled.
1359 Set minimal partition size used for convolution. Default is @var{8192}.
1360 Allowed range is from @var{1} to @var{32768}.
1361 Lower values decreases latency at cost of higher CPU usage.
1364 Set maximal partition size used for convolution. Default is @var{8192}.
1365 Allowed range is from @var{8} to @var{32768}.
1366 Lower values may increase CPU usage.
1369 Set number of input impulse responses streams which will be switchable at runtime.
1370 Allowed range is from @var{1} to @var{32}. Default is @var{1}.
1373 Set IR stream which will be used for convolution, starting from @var{0}, should always be
1374 lower than supplied value by @code{nbirs} option. Default is @var{0}.
1375 This option can be changed at runtime via @ref{commands}.
1378 @subsection Examples
1382 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1384 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1391 Set output format constraints for the input audio. The framework will
1392 negotiate the most appropriate format to minimize conversions.
1394 It accepts the following parameters:
1397 @item sample_fmts, f
1398 A '|'-separated list of requested sample formats.
1400 @item sample_rates, r
1401 A '|'-separated list of requested sample rates.
1403 @item channel_layouts, cl
1404 A '|'-separated list of requested channel layouts.
1406 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1407 for the required syntax.
1410 If a parameter is omitted, all values are allowed.
1412 Force the output to either unsigned 8-bit or signed 16-bit stereo
1414 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1418 Apply frequency shift to input audio samples.
1420 The filter accepts the following options:
1424 Specify frequency shift. Allowed range is -INT_MAX to INT_MAX.
1425 Default value is 0.0.
1428 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
1429 Default value is 1.0.
1432 @subsection Commands
1434 This filter supports the all above options as @ref{commands}.
1438 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1439 processing reduces disturbing noise between useful signals.
1441 Gating is done by detecting the volume below a chosen level @var{threshold}
1442 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1443 floor is set via @var{range}. Because an exact manipulation of the signal
1444 would cause distortion of the waveform the reduction can be levelled over
1445 time. This is done by setting @var{attack} and @var{release}.
1447 @var{attack} determines how long the signal has to fall below the threshold
1448 before any reduction will occur and @var{release} sets the time the signal
1449 has to rise above the threshold to reduce the reduction again.
1450 Shorter signals than the chosen attack time will be left untouched.
1454 Set input level before filtering.
1455 Default is 1. Allowed range is from 0.015625 to 64.
1458 Set the mode of operation. Can be @code{upward} or @code{downward}.
1459 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
1460 will be amplified, expanding dynamic range in upward direction.
1461 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
1464 Set the level of gain reduction when the signal is below the threshold.
1465 Default is 0.06125. Allowed range is from 0 to 1.
1466 Setting this to 0 disables reduction and then filter behaves like expander.
1469 If a signal rises above this level the gain reduction is released.
1470 Default is 0.125. Allowed range is from 0 to 1.
1473 Set a ratio by which the signal is reduced.
1474 Default is 2. Allowed range is from 1 to 9000.
1477 Amount of milliseconds the signal has to rise above the threshold before gain
1479 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1482 Amount of milliseconds the signal has to fall below the threshold before the
1483 reduction is increased again. Default is 250 milliseconds.
1484 Allowed range is from 0.01 to 9000.
1487 Set amount of amplification of signal after processing.
1488 Default is 1. Allowed range is from 1 to 64.
1491 Curve the sharp knee around the threshold to enter gain reduction more softly.
1492 Default is 2.828427125. Allowed range is from 1 to 8.
1495 Choose if exact signal should be taken for detection or an RMS like one.
1496 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1499 Choose if the average level between all channels or the louder channel affects
1501 Default is @code{average}. Can be @code{average} or @code{maximum}.
1504 @subsection Commands
1506 This filter supports the all above options as @ref{commands}.
1510 Apply an arbitrary Infinite Impulse Response filter.
1512 It accepts the following parameters:
1516 Set B/numerator/zeros/reflection coefficients.
1519 Set A/denominator/poles/ladder coefficients.
1531 Set coefficients format.
1535 lattice-ladder function
1537 analog transfer function
1539 digital transfer function
1541 Z-plane zeros/poles, cartesian (default)
1543 Z-plane zeros/poles, polar radians
1545 Z-plane zeros/poles, polar degrees
1551 Set type of processing.
1563 Set filtering precision.
1567 double-precision floating-point (default)
1569 single-precision floating-point
1577 Normalize filter coefficients, by default is enabled.
1578 Enabling it will normalize magnitude response at DC to 0dB.
1581 How much to use filtered signal in output. Default is 1.
1582 Range is between 0 and 1.
1585 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1586 By default it is disabled.
1589 Set for which IR channel to display frequency response. By default is first channel
1590 displayed. This option is used only when @var{response} is enabled.
1593 Set video stream size. This option is used only when @var{response} is enabled.
1596 Coefficients in @code{tf} and @code{sf} format are separated by spaces and are in ascending
1599 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1600 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1603 Different coefficients and gains can be provided for every channel, in such case
1604 use '|' to separate coefficients or gains. Last provided coefficients will be
1605 used for all remaining channels.
1607 @subsection Examples
1611 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1613 aiir=k=1:z=7.957584807809675810E-1 -2.575128568908332300 3.674839853930788710 -2.57512875289799137 7.957586296317130880E-1:p=1 -2.86950072432325953 3.63022088054647218 -2.28075678147272232 6.361362326477423500E-1:f=tf:r=d
1617 Same as above but in @code{zp} format:
1619 aiir=k=0.79575848078096756:z=0.80918701+0.58773007i 0.80918701-0.58773007i 0.80884700+0.58784055i 0.80884700-0.58784055i:p=0.63892345+0.59951235i 0.63892345-0.59951235i 0.79582691+0.44198673i 0.79582691-0.44198673i:f=zp:r=s
1623 Apply 3-rd order analog normalized Butterworth low-pass filter, using analog transfer function format:
1625 aiir=z=1.3057 0 0 0:p=1.3057 2.3892 2.1860 1:f=sf:r=d
1631 The limiter prevents an input signal from rising over a desired threshold.
1632 This limiter uses lookahead technology to prevent your signal from distorting.
1633 It means that there is a small delay after the signal is processed. Keep in mind
1634 that the delay it produces is the attack time you set.
1636 The filter accepts the following options:
1640 Set input gain. Default is 1.
1643 Set output gain. Default is 1.
1646 Don't let signals above this level pass the limiter. Default is 1.
1649 The limiter will reach its attenuation level in this amount of time in
1650 milliseconds. Default is 5 milliseconds.
1653 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1654 Default is 50 milliseconds.
1657 When gain reduction is always needed ASC takes care of releasing to an
1658 average reduction level rather than reaching a reduction of 0 in the release
1662 Select how much the release time is affected by ASC, 0 means nearly no changes
1663 in release time while 1 produces higher release times.
1666 Auto level output signal. Default is enabled.
1667 This normalizes audio back to 0dB if enabled.
1670 Depending on picked setting it is recommended to upsample input 2x or 4x times
1671 with @ref{aresample} before applying this filter.
1675 Apply a two-pole all-pass filter with central frequency (in Hz)
1676 @var{frequency}, and filter-width @var{width}.
1677 An all-pass filter changes the audio's frequency to phase relationship
1678 without changing its frequency to amplitude relationship.
1680 The filter accepts the following options:
1684 Set frequency in Hz.
1687 Set method to specify band-width of filter.
1702 Specify the band-width of a filter in width_type units.
1705 How much to use filtered signal in output. Default is 1.
1706 Range is between 0 and 1.
1709 Specify which channels to filter, by default all available are filtered.
1712 Normalize biquad coefficients, by default is disabled.
1713 Enabling it will normalize magnitude response at DC to 0dB.
1716 Set the filter order, can be 1 or 2. Default is 2.
1719 Set transform type of IIR filter.
1728 Set precison of filtering.
1731 Pick automatic sample format depending on surround filters.
1733 Always use signed 16-bit.
1735 Always use signed 32-bit.
1737 Always use float 32-bit.
1739 Always use float 64-bit.
1743 @subsection Commands
1745 This filter supports the following commands:
1748 Change allpass frequency.
1749 Syntax for the command is : "@var{frequency}"
1752 Change allpass width_type.
1753 Syntax for the command is : "@var{width_type}"
1756 Change allpass width.
1757 Syntax for the command is : "@var{width}"
1761 Syntax for the command is : "@var{mix}"
1768 The filter accepts the following options:
1772 Set the number of loops. Setting this value to -1 will result in infinite loops.
1776 Set maximal number of samples. Default is 0.
1779 Set first sample of loop. Default is 0.
1785 Merge two or more audio streams into a single multi-channel stream.
1787 The filter accepts the following options:
1792 Set the number of inputs. Default is 2.
1796 If the channel layouts of the inputs are disjoint, and therefore compatible,
1797 the channel layout of the output will be set accordingly and the channels
1798 will be reordered as necessary. If the channel layouts of the inputs are not
1799 disjoint, the output will have all the channels of the first input then all
1800 the channels of the second input, in that order, and the channel layout of
1801 the output will be the default value corresponding to the total number of
1804 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1805 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1806 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1807 first input, b1 is the first channel of the second input).
1809 On the other hand, if both input are in stereo, the output channels will be
1810 in the default order: a1, a2, b1, b2, and the channel layout will be
1811 arbitrarily set to 4.0, which may or may not be the expected value.
1813 All inputs must have the same sample rate, and format.
1815 If inputs do not have the same duration, the output will stop with the
1818 @subsection Examples
1822 Merge two mono files into a stereo stream:
1824 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1828 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1830 ffmpeg -i input.mkv -filter_complex "[0:1][0:2][0:3][0:4][0:5][0:6] amerge=inputs=6" -c:a pcm_s16le output.mkv
1836 Mixes multiple audio inputs into a single output.
1838 Note that this filter only supports float samples (the @var{amerge}
1839 and @var{pan} audio filters support many formats). If the @var{amix}
1840 input has integer samples then @ref{aresample} will be automatically
1841 inserted to perform the conversion to float samples.
1845 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1847 will mix 3 input audio streams to a single output with the same duration as the
1848 first input and a dropout transition time of 3 seconds.
1850 It accepts the following parameters:
1854 The number of inputs. If unspecified, it defaults to 2.
1857 How to determine the end-of-stream.
1861 The duration of the longest input. (default)
1864 The duration of the shortest input.
1867 The duration of the first input.
1871 @item dropout_transition
1872 The transition time, in seconds, for volume renormalization when an input
1873 stream ends. The default value is 2 seconds.
1876 Specify weight of each input audio stream as sequence.
1877 Each weight is separated by space. By default all inputs have same weight.
1880 @subsection Commands
1882 This filter supports the following commands:
1885 Syntax is same as option with same name.
1890 Multiply first audio stream with second audio stream and store result
1891 in output audio stream. Multiplication is done by multiplying each
1892 sample from first stream with sample at same position from second stream.
1894 With this element-wise multiplication one can create amplitude fades and
1895 amplitude modulations.
1897 @section anequalizer
1899 High-order parametric multiband equalizer for each channel.
1901 It accepts the following parameters:
1905 This option string is in format:
1906 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1907 Each equalizer band is separated by '|'.
1911 Set channel number to which equalization will be applied.
1912 If input doesn't have that channel the entry is ignored.
1915 Set central frequency for band.
1916 If input doesn't have that frequency the entry is ignored.
1919 Set band width in Hertz.
1922 Set band gain in dB.
1925 Set filter type for band, optional, can be:
1929 Butterworth, this is default.
1940 With this option activated frequency response of anequalizer is displayed
1944 Set video stream size. Only useful if curves option is activated.
1947 Set max gain that will be displayed. Only useful if curves option is activated.
1948 Setting this to a reasonable value makes it possible to display gain which is derived from
1949 neighbour bands which are too close to each other and thus produce higher gain
1950 when both are activated.
1953 Set frequency scale used to draw frequency response in video output.
1954 Can be linear or logarithmic. Default is logarithmic.
1957 Set color for each channel curve which is going to be displayed in video stream.
1958 This is list of color names separated by space or by '|'.
1959 Unrecognised or missing colors will be replaced by white color.
1962 @subsection Examples
1966 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1967 for first 2 channels using Chebyshev type 1 filter:
1969 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1973 @subsection Commands
1975 This filter supports the following commands:
1978 Alter existing filter parameters.
1979 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1981 @var{fN} is existing filter number, starting from 0, if no such filter is available
1983 @var{freq} set new frequency parameter.
1984 @var{width} set new width parameter in Hertz.
1985 @var{gain} set new gain parameter in dB.
1987 Full filter invocation with asendcmd may look like this:
1988 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1993 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1995 Each sample is adjusted by looking for other samples with similar contexts. This
1996 context similarity is defined by comparing their surrounding patches of size
1997 @option{p}. Patches are searched in an area of @option{r} around the sample.
1999 The filter accepts the following options:
2003 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
2006 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
2007 Default value is 2 milliseconds.
2010 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
2011 Default value is 6 milliseconds.
2014 Set the output mode.
2016 It accepts the following values:
2019 Pass input unchanged.
2022 Pass noise filtered out.
2027 Default value is @var{o}.
2031 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
2034 @subsection Commands
2036 This filter supports the all above options as @ref{commands}.
2039 Apply Normalized Least-Mean-Squares algorithm to the first audio stream using the second audio stream.
2041 This adaptive filter is used to mimic a desired filter by finding the filter coefficients that
2042 relate to producing the least mean square of the error signal (difference between the desired,
2043 2nd input audio stream and the actual signal, the 1st input audio stream).
2045 A description of the accepted options follows.
2058 Set the filter leakage.
2061 It accepts the following values:
2070 Pass filtered samples.
2073 Pass difference between desired and filtered samples.
2075 Default value is @var{o}.
2079 @subsection Examples
2083 One of many usages of this filter is noise reduction, input audio is filtered
2084 with same samples that are delayed by fixed amount, one such example for stereo audio is:
2086 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
2090 @subsection Commands
2092 This filter supports the same commands as options, excluding option @code{order}.
2096 Pass the audio source unchanged to the output.
2100 Pad the end of an audio stream with silence.
2102 This can be used together with @command{ffmpeg} @option{-shortest} to
2103 extend audio streams to the same length as the video stream.
2105 A description of the accepted options follows.
2109 Set silence packet size. Default value is 4096.
2112 Set the number of samples of silence to add to the end. After the
2113 value is reached, the stream is terminated. This option is mutually
2114 exclusive with @option{whole_len}.
2117 Set the minimum total number of samples in the output audio stream. If
2118 the value is longer than the input audio length, silence is added to
2119 the end, until the value is reached. This option is mutually exclusive
2120 with @option{pad_len}.
2123 Specify the duration of samples of silence to add. See
2124 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2125 for the accepted syntax. Used only if set to non-zero value.
2128 Specify the minimum total duration in the output audio stream. See
2129 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2130 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
2131 the input audio length, silence is added to the end, until the value is reached.
2132 This option is mutually exclusive with @option{pad_dur}
2135 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
2136 nor @option{whole_dur} option is set, the filter will add silence to the end of
2137 the input stream indefinitely.
2139 @subsection Examples
2143 Add 1024 samples of silence to the end of the input:
2149 Make sure the audio output will contain at least 10000 samples, pad
2150 the input with silence if required:
2152 apad=whole_len=10000
2156 Use @command{ffmpeg} to pad the audio input with silence, so that the
2157 video stream will always result the shortest and will be converted
2158 until the end in the output file when using the @option{shortest}
2161 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
2166 Add a phasing effect to the input audio.
2168 A phaser filter creates series of peaks and troughs in the frequency spectrum.
2169 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
2171 A description of the accepted parameters follows.
2175 Set input gain. Default is 0.4.
2178 Set output gain. Default is 0.74
2181 Set delay in milliseconds. Default is 3.0.
2184 Set decay. Default is 0.4.
2187 Set modulation speed in Hz. Default is 0.5.
2190 Set modulation type. Default is triangular.
2192 It accepts the following values:
2199 @section aphaseshift
2200 Apply phase shift to input audio samples.
2202 The filter accepts the following options:
2206 Specify phase shift. Allowed range is from -1.0 to 1.0.
2207 Default value is 0.0.
2210 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
2211 Default value is 1.0.
2214 @subsection Commands
2216 This filter supports the all above options as @ref{commands}.
2220 Audio pulsator is something between an autopanner and a tremolo.
2221 But it can produce funny stereo effects as well. Pulsator changes the volume
2222 of the left and right channel based on a LFO (low frequency oscillator) with
2223 different waveforms and shifted phases.
2224 This filter have the ability to define an offset between left and right
2225 channel. An offset of 0 means that both LFO shapes match each other.
2226 The left and right channel are altered equally - a conventional tremolo.
2227 An offset of 50% means that the shape of the right channel is exactly shifted
2228 in phase (or moved backwards about half of the frequency) - pulsator acts as
2229 an autopanner. At 1 both curves match again. Every setting in between moves the
2230 phase shift gapless between all stages and produces some "bypassing" sounds with
2231 sine and triangle waveforms. The more you set the offset near 1 (starting from
2232 the 0.5) the faster the signal passes from the left to the right speaker.
2234 The filter accepts the following options:
2238 Set input gain. By default it is 1. Range is [0.015625 - 64].
2241 Set output gain. By default it is 1. Range is [0.015625 - 64].
2244 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
2245 sawup or sawdown. Default is sine.
2248 Set modulation. Define how much of original signal is affected by the LFO.
2251 Set left channel offset. Default is 0. Allowed range is [0 - 1].
2254 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
2257 Set pulse width. Default is 1. Allowed range is [0 - 2].
2260 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
2263 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
2267 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
2271 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
2272 if timing is set to hz.
2278 Resample the input audio to the specified parameters, using the
2279 libswresample library. If none are specified then the filter will
2280 automatically convert between its input and output.
2282 This filter is also able to stretch/squeeze the audio data to make it match
2283 the timestamps or to inject silence / cut out audio to make it match the
2284 timestamps, do a combination of both or do neither.
2286 The filter accepts the syntax
2287 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
2288 expresses a sample rate and @var{resampler_options} is a list of
2289 @var{key}=@var{value} pairs, separated by ":". See the
2290 @ref{Resampler Options,,"Resampler Options" section in the
2291 ffmpeg-resampler(1) manual,ffmpeg-resampler}
2292 for the complete list of supported options.
2294 @subsection Examples
2298 Resample the input audio to 44100Hz:
2304 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2305 samples per second compensation:
2307 aresample=async=1000
2313 Reverse an audio clip.
2315 Warning: This filter requires memory to buffer the entire clip, so trimming
2318 @subsection Examples
2322 Take the first 5 seconds of a clip, and reverse it.
2324 atrim=end=5,areverse
2330 Reduce noise from speech using Recurrent Neural Networks.
2332 This filter accepts the following options:
2336 Set train model file to load. This option is always required.
2339 Set how much to mix filtered samples into final output.
2340 Allowed range is from -1 to 1. Default value is 1.
2341 Negative values are special, they set how much to keep filtered noise
2342 in the final filter output. Set this option to -1 to hear actual
2343 noise removed from input signal.
2346 @section asetnsamples
2348 Set the number of samples per each output audio frame.
2350 The last output packet may contain a different number of samples, as
2351 the filter will flush all the remaining samples when the input audio
2354 The filter accepts the following options:
2358 @item nb_out_samples, n
2359 Set the number of frames per each output audio frame. The number is
2360 intended as the number of samples @emph{per each channel}.
2361 Default value is 1024.
2364 If set to 1, the filter will pad the last audio frame with zeroes, so
2365 that the last frame will contain the same number of samples as the
2366 previous ones. Default value is 1.
2369 For example, to set the number of per-frame samples to 1234 and
2370 disable padding for the last frame, use:
2372 asetnsamples=n=1234:p=0
2377 Set the sample rate without altering the PCM data.
2378 This will result in a change of speed and pitch.
2380 The filter accepts the following options:
2383 @item sample_rate, r
2384 Set the output sample rate. Default is 44100 Hz.
2389 Show a line containing various information for each input audio frame.
2390 The input audio is not modified.
2392 The shown line contains a sequence of key/value pairs of the form
2393 @var{key}:@var{value}.
2395 The following values are shown in the output:
2399 The (sequential) number of the input frame, starting from 0.
2402 The presentation timestamp of the input frame, in time base units; the time base
2403 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2406 The presentation timestamp of the input frame in seconds.
2409 position of the frame in the input stream, -1 if this information in
2410 unavailable and/or meaningless (for example in case of synthetic audio)
2419 The sample rate for the audio frame.
2422 The number of samples (per channel) in the frame.
2425 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2426 audio, the data is treated as if all the planes were concatenated.
2428 @item plane_checksums
2429 A list of Adler-32 checksums for each data plane.
2433 Apply audio soft clipping.
2435 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2436 along a smooth curve, rather than the abrupt shape of hard-clipping.
2438 This filter accepts the following options:
2442 Set type of soft-clipping.
2444 It accepts the following values:
2458 Set threshold from where to start clipping. Default value is 0dB or 1.
2461 Set gain applied to output. Default value is 0dB or 1.
2464 Set additional parameter which controls sigmoid function.
2467 Set oversampling factor.
2470 @subsection Commands
2472 This filter supports the all above options as @ref{commands}.
2475 Automatic Speech Recognition
2477 This filter uses PocketSphinx for speech recognition. To enable
2478 compilation of this filter, you need to configure FFmpeg with
2479 @code{--enable-pocketsphinx}.
2481 It accepts the following options:
2485 Set sampling rate of input audio. Defaults is @code{16000}.
2486 This need to match speech models, otherwise one will get poor results.
2489 Set dictionary containing acoustic model files.
2492 Set pronunciation dictionary.
2495 Set language model file.
2498 Set language model set.
2501 Set which language model to use.
2504 Set output for log messages.
2507 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2512 Display time domain statistical information about the audio channels.
2513 Statistics are calculated and displayed for each audio channel and,
2514 where applicable, an overall figure is also given.
2516 It accepts the following option:
2519 Short window length in seconds, used for peak and trough RMS measurement.
2520 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2524 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2525 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2528 Available keys for each channel are:
2574 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2575 this @code{lavfi.astats.Overall.Peak_count}.
2577 For description what each key means read below.
2580 Set number of frame after which stats are going to be recalculated.
2581 Default is disabled.
2583 @item measure_perchannel
2584 Select the entries which need to be measured per channel. The metadata keys can
2585 be used as flags, default is @option{all} which measures everything.
2586 @option{none} disables all per channel measurement.
2588 @item measure_overall
2589 Select the entries which need to be measured overall. The metadata keys can
2590 be used as flags, default is @option{all} which measures everything.
2591 @option{none} disables all overall measurement.
2595 A description of each shown parameter follows:
2599 Mean amplitude displacement from zero.
2602 Minimal sample level.
2605 Maximal sample level.
2607 @item Min difference
2608 Minimal difference between two consecutive samples.
2610 @item Max difference
2611 Maximal difference between two consecutive samples.
2613 @item Mean difference
2614 Mean difference between two consecutive samples.
2615 The average of each difference between two consecutive samples.
2617 @item RMS difference
2618 Root Mean Square difference between two consecutive samples.
2622 Standard peak and RMS level measured in dBFS.
2626 Peak and trough values for RMS level measured over a short window.
2629 Standard ratio of peak to RMS level (note: not in dB).
2632 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2633 (i.e. either @var{Min level} or @var{Max level}).
2636 Number of occasions (not the number of samples) that the signal attained either
2637 @var{Min level} or @var{Max level}.
2639 @item Noise floor dB
2640 Minimum local peak measured in dBFS over a short window.
2642 @item Noise floor count
2643 Number of occasions (not the number of samples) that the signal attained
2647 Overall bit depth of audio. Number of bits used for each sample.
2650 Measured dynamic range of audio in dB.
2652 @item Zero crossings
2653 Number of points where the waveform crosses the zero level axis.
2655 @item Zero crossings rate
2656 Rate of Zero crossings and number of audio samples.
2660 Boost subwoofer frequencies.
2662 The filter accepts the following options:
2666 Set dry gain, how much of original signal is kept. Allowed range is from 0 to 1.
2667 Default value is 0.7.
2670 Set wet gain, how much of filtered signal is kept. Allowed range is from 0 to 1.
2671 Default value is 0.7.
2674 Set delay line decay gain value. Allowed range is from 0 to 1.
2675 Default value is 0.7.
2678 Set delay line feedback gain value. Allowed range is from 0 to 1.
2679 Default value is 0.9.
2682 Set cutoff frequency in Hertz. Allowed range is 50 to 900.
2683 Default value is 100.
2686 Set slope amount for cutoff frequency. Allowed range is 0.0001 to 1.
2687 Default value is 0.5.
2690 Set delay. Allowed range is from 1 to 100.
2691 Default value is 20.
2694 @subsection Commands
2696 This filter supports the all above options as @ref{commands}.
2699 Cut subwoofer frequencies.
2701 This filter allows to set custom, steeper
2702 roll off than highpass filter, and thus is able to more attenuate
2703 frequency content in stop-band.
2705 The filter accepts the following options:
2709 Set cutoff frequency in Hertz. Allowed range is 2 to 200.
2710 Default value is 20.
2713 Set filter order. Available values are from 3 to 20.
2714 Default value is 10.
2717 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2720 @subsection Commands
2722 This filter supports the all above options as @ref{commands}.
2725 Cut super frequencies.
2727 The filter accepts the following options:
2731 Set cutoff frequency in Hertz. Allowed range is 20000 to 192000.
2732 Default value is 20000.
2735 Set filter order. Available values are from 3 to 20.
2736 Default value is 10.
2739 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2742 @subsection Commands
2744 This filter supports the all above options as @ref{commands}.
2747 Apply high order Butterworth band-pass filter.
2749 The filter accepts the following options:
2753 Set center frequency in Hertz. Allowed range is 2 to 999999.
2754 Default value is 1000.
2757 Set filter order. Available values are from 4 to 20.
2761 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2764 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2767 @subsection Commands
2769 This filter supports the all above options as @ref{commands}.
2772 Apply high order Butterworth band-stop filter.
2774 The filter accepts the following options:
2778 Set center frequency in Hertz. Allowed range is 2 to 999999.
2779 Default value is 1000.
2782 Set filter order. Available values are from 4 to 20.
2786 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2789 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2792 @subsection Commands
2794 This filter supports the all above options as @ref{commands}.
2800 The filter accepts exactly one parameter, the audio tempo. If not
2801 specified then the filter will assume nominal 1.0 tempo. Tempo must
2802 be in the [0.5, 100.0] range.
2804 Note that tempo greater than 2 will skip some samples rather than
2805 blend them in. If for any reason this is a concern it is always
2806 possible to daisy-chain several instances of atempo to achieve the
2807 desired product tempo.
2809 @subsection Examples
2813 Slow down audio to 80% tempo:
2819 To speed up audio to 300% tempo:
2825 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2827 atempo=sqrt(3),atempo=sqrt(3)
2831 @subsection Commands
2833 This filter supports the following commands:
2836 Change filter tempo scale factor.
2837 Syntax for the command is : "@var{tempo}"
2842 Trim the input so that the output contains one continuous subpart of the input.
2844 It accepts the following parameters:
2847 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2848 sample with the timestamp @var{start} will be the first sample in the output.
2851 Specify time of the first audio sample that will be dropped, i.e. the
2852 audio sample immediately preceding the one with the timestamp @var{end} will be
2853 the last sample in the output.
2856 Same as @var{start}, except this option sets the start timestamp in samples
2860 Same as @var{end}, except this option sets the end timestamp in samples instead
2864 The maximum duration of the output in seconds.
2867 The number of the first sample that should be output.
2870 The number of the first sample that should be dropped.
2873 @option{start}, @option{end}, and @option{duration} are expressed as time
2874 duration specifications; see
2875 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2877 Note that the first two sets of the start/end options and the @option{duration}
2878 option look at the frame timestamp, while the _sample options simply count the
2879 samples that pass through the filter. So start/end_pts and start/end_sample will
2880 give different results when the timestamps are wrong, inexact or do not start at
2881 zero. Also note that this filter does not modify the timestamps. If you wish
2882 to have the output timestamps start at zero, insert the asetpts filter after the
2885 If multiple start or end options are set, this filter tries to be greedy and
2886 keep all samples that match at least one of the specified constraints. To keep
2887 only the part that matches all the constraints at once, chain multiple atrim
2890 The defaults are such that all the input is kept. So it is possible to set e.g.
2891 just the end values to keep everything before the specified time.
2896 Drop everything except the second minute of input:
2898 ffmpeg -i INPUT -af atrim=60:120
2902 Keep only the first 1000 samples:
2904 ffmpeg -i INPUT -af atrim=end_sample=1000
2909 @section axcorrelate
2910 Calculate normalized cross-correlation between two input audio streams.
2912 Resulted samples are always between -1 and 1 inclusive.
2913 If result is 1 it means two input samples are highly correlated in that selected segment.
2914 Result 0 means they are not correlated at all.
2915 If result is -1 it means two input samples are out of phase, which means they cancel each
2918 The filter accepts the following options:
2922 Set size of segment over which cross-correlation is calculated.
2923 Default is 256. Allowed range is from 2 to 131072.
2926 Set algorithm for cross-correlation. Can be @code{slow} or @code{fast}.
2927 Default is @code{slow}. Fast algorithm assumes mean values over any given segment
2928 are always zero and thus need much less calculations to make.
2929 This is generally not true, but is valid for typical audio streams.
2932 @subsection Examples
2936 Calculate correlation between channels in stereo audio stream:
2938 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
2944 Apply a two-pole Butterworth band-pass filter with central
2945 frequency @var{frequency}, and (3dB-point) band-width width.
2946 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2947 instead of the default: constant 0dB peak gain.
2948 The filter roll off at 6dB per octave (20dB per decade).
2950 The filter accepts the following options:
2954 Set the filter's central frequency. Default is @code{3000}.
2957 Constant skirt gain if set to 1. Defaults to 0.
2960 Set method to specify band-width of filter.
2975 Specify the band-width of a filter in width_type units.
2978 How much to use filtered signal in output. Default is 1.
2979 Range is between 0 and 1.
2982 Specify which channels to filter, by default all available are filtered.
2985 Normalize biquad coefficients, by default is disabled.
2986 Enabling it will normalize magnitude response at DC to 0dB.
2989 Set transform type of IIR filter.
2998 Set precison of filtering.
3001 Pick automatic sample format depending on surround filters.
3003 Always use signed 16-bit.
3005 Always use signed 32-bit.
3007 Always use float 32-bit.
3009 Always use float 64-bit.
3013 @subsection Commands
3015 This filter supports the following commands:
3018 Change bandpass frequency.
3019 Syntax for the command is : "@var{frequency}"
3022 Change bandpass width_type.
3023 Syntax for the command is : "@var{width_type}"
3026 Change bandpass width.
3027 Syntax for the command is : "@var{width}"
3030 Change bandpass mix.
3031 Syntax for the command is : "@var{mix}"
3036 Apply a two-pole Butterworth band-reject filter with central
3037 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
3038 The filter roll off at 6dB per octave (20dB per decade).
3040 The filter accepts the following options:
3044 Set the filter's central frequency. Default is @code{3000}.
3047 Set method to specify band-width of filter.
3062 Specify the band-width of a filter in width_type units.
3065 How much to use filtered signal in output. Default is 1.
3066 Range is between 0 and 1.
3069 Specify which channels to filter, by default all available are filtered.
3072 Normalize biquad coefficients, by default is disabled.
3073 Enabling it will normalize magnitude response at DC to 0dB.
3076 Set transform type of IIR filter.
3085 Set precison of filtering.
3088 Pick automatic sample format depending on surround filters.
3090 Always use signed 16-bit.
3092 Always use signed 32-bit.
3094 Always use float 32-bit.
3096 Always use float 64-bit.
3100 @subsection Commands
3102 This filter supports the following commands:
3105 Change bandreject frequency.
3106 Syntax for the command is : "@var{frequency}"
3109 Change bandreject width_type.
3110 Syntax for the command is : "@var{width_type}"
3113 Change bandreject width.
3114 Syntax for the command is : "@var{width}"
3117 Change bandreject mix.
3118 Syntax for the command is : "@var{mix}"
3121 @section bass, lowshelf
3123 Boost or cut the bass (lower) frequencies of the audio using a two-pole
3124 shelving filter with a response similar to that of a standard
3125 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
3127 The filter accepts the following options:
3131 Give the gain at 0 Hz. Its useful range is about -20
3132 (for a large cut) to +20 (for a large boost).
3133 Beware of clipping when using a positive gain.
3136 Set the filter's central frequency and so can be used
3137 to extend or reduce the frequency range to be boosted or cut.
3138 The default value is @code{100} Hz.
3141 Set method to specify band-width of filter.
3156 Determine how steep is the filter's shelf transition.
3159 Set number of poles. Default is 2.
3162 How much to use filtered signal in output. Default is 1.
3163 Range is between 0 and 1.
3166 Specify which channels to filter, by default all available are filtered.
3169 Normalize biquad coefficients, by default is disabled.
3170 Enabling it will normalize magnitude response at DC to 0dB.
3173 Set transform type of IIR filter.
3182 Set precison of filtering.
3185 Pick automatic sample format depending on surround filters.
3187 Always use signed 16-bit.
3189 Always use signed 32-bit.
3191 Always use float 32-bit.
3193 Always use float 64-bit.
3197 @subsection Commands
3199 This filter supports the following commands:
3202 Change bass frequency.
3203 Syntax for the command is : "@var{frequency}"
3206 Change bass width_type.
3207 Syntax for the command is : "@var{width_type}"
3211 Syntax for the command is : "@var{width}"
3215 Syntax for the command is : "@var{gain}"
3219 Syntax for the command is : "@var{mix}"
3224 Apply a biquad IIR filter with the given coefficients.
3225 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
3226 are the numerator and denominator coefficients respectively.
3227 and @var{channels}, @var{c} specify which channels to filter, by default all
3228 available are filtered.
3230 @subsection Commands
3232 This filter supports the following commands:
3240 Change biquad parameter.
3241 Syntax for the command is : "@var{value}"
3244 How much to use filtered signal in output. Default is 1.
3245 Range is between 0 and 1.
3248 Specify which channels to filter, by default all available are filtered.
3251 Normalize biquad coefficients, by default is disabled.
3252 Enabling it will normalize magnitude response at DC to 0dB.
3255 Set transform type of IIR filter.
3264 Set precison of filtering.
3267 Pick automatic sample format depending on surround filters.
3269 Always use signed 16-bit.
3271 Always use signed 32-bit.
3273 Always use float 32-bit.
3275 Always use float 64-bit.
3280 Bauer stereo to binaural transformation, which improves headphone listening of
3281 stereo audio records.
3283 To enable compilation of this filter you need to configure FFmpeg with
3284 @code{--enable-libbs2b}.
3286 It accepts the following parameters:
3290 Pre-defined crossfeed level.
3294 Default level (fcut=700, feed=50).
3297 Chu Moy circuit (fcut=700, feed=60).
3300 Jan Meier circuit (fcut=650, feed=95).
3305 Cut frequency (in Hz).
3314 Remap input channels to new locations.
3316 It accepts the following parameters:
3319 Map channels from input to output. The argument is a '|'-separated list of
3320 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
3321 @var{in_channel} form. @var{in_channel} can be either the name of the input
3322 channel (e.g. FL for front left) or its index in the input channel layout.
3323 @var{out_channel} is the name of the output channel or its index in the output
3324 channel layout. If @var{out_channel} is not given then it is implicitly an
3325 index, starting with zero and increasing by one for each mapping.
3327 @item channel_layout
3328 The channel layout of the output stream.
3331 If no mapping is present, the filter will implicitly map input channels to
3332 output channels, preserving indices.
3334 @subsection Examples
3338 For example, assuming a 5.1+downmix input MOV file,
3340 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
3342 will create an output WAV file tagged as stereo from the downmix channels of
3346 To fix a 5.1 WAV improperly encoded in AAC's native channel order
3348 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
3352 @section channelsplit
3354 Split each channel from an input audio stream into a separate output stream.
3356 It accepts the following parameters:
3358 @item channel_layout
3359 The channel layout of the input stream. The default is "stereo".
3361 A channel layout describing the channels to be extracted as separate output streams
3362 or "all" to extract each input channel as a separate stream. The default is "all".
3364 Choosing channels not present in channel layout in the input will result in an error.
3367 @subsection Examples
3371 For example, assuming a stereo input MP3 file,
3373 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
3375 will create an output Matroska file with two audio streams, one containing only
3376 the left channel and the other the right channel.
3379 Split a 5.1 WAV file into per-channel files:
3381 ffmpeg -i in.wav -filter_complex
3382 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
3383 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
3384 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
3389 Extract only LFE from a 5.1 WAV file:
3391 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
3392 -map '[LFE]' lfe.wav
3397 Add a chorus effect to the audio.
3399 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
3401 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
3402 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
3403 The modulation depth defines the range the modulated delay is played before or after
3404 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
3405 sound tuned around the original one, like in a chorus where some vocals are slightly
3408 It accepts the following parameters:
3411 Set input gain. Default is 0.4.
3414 Set output gain. Default is 0.4.
3417 Set delays. A typical delay is around 40ms to 60ms.
3429 @subsection Examples
3435 chorus=0.7:0.9:55:0.4:0.25:2
3441 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
3445 Fuller sounding chorus with three delays:
3447 chorus=0.5:0.9:50|60|40:0.4|0.32|0.3:0.25|0.4|0.3:2|2.3|1.3
3452 Compress or expand the audio's dynamic range.
3454 It accepts the following parameters:
3460 A list of times in seconds for each channel over which the instantaneous level
3461 of the input signal is averaged to determine its volume. @var{attacks} refers to
3462 increase of volume and @var{decays} refers to decrease of volume. For most
3463 situations, the attack time (response to the audio getting louder) should be
3464 shorter than the decay time, because the human ear is more sensitive to sudden
3465 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
3466 a typical value for decay is 0.8 seconds.
3467 If specified number of attacks & decays is lower than number of channels, the last
3468 set attack/decay will be used for all remaining channels.
3471 A list of points for the transfer function, specified in dB relative to the
3472 maximum possible signal amplitude. Each key points list must be defined using
3473 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
3474 @code{x0/y0 x1/y1 x2/y2 ....}
3476 The input values must be in strictly increasing order but the transfer function
3477 does not have to be monotonically rising. The point @code{0/0} is assumed but
3478 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
3479 function are @code{-70/-70|-60/-20|1/0}.
3482 Set the curve radius in dB for all joints. It defaults to 0.01.
3485 Set the additional gain in dB to be applied at all points on the transfer
3486 function. This allows for easy adjustment of the overall gain.
3490 Set an initial volume, in dB, to be assumed for each channel when filtering
3491 starts. This permits the user to supply a nominal level initially, so that, for
3492 example, a very large gain is not applied to initial signal levels before the
3493 companding has begun to operate. A typical value for audio which is initially
3494 quiet is -90 dB. It defaults to 0.
3497 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
3498 delayed before being fed to the volume adjuster. Specifying a delay
3499 approximately equal to the attack/decay times allows the filter to effectively
3500 operate in predictive rather than reactive mode. It defaults to 0.
3504 @subsection Examples
3508 Make music with both quiet and loud passages suitable for listening to in a
3511 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
3514 Another example for audio with whisper and explosion parts:
3516 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
3520 A noise gate for when the noise is at a lower level than the signal:
3522 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
3526 Here is another noise gate, this time for when the noise is at a higher level
3527 than the signal (making it, in some ways, similar to squelch):
3529 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3533 2:1 compression starting at -6dB:
3535 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3539 2:1 compression starting at -9dB:
3541 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3545 2:1 compression starting at -12dB:
3547 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3551 2:1 compression starting at -18dB:
3553 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3557 3:1 compression starting at -15dB:
3559 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3565 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3571 compand=attacks=0:points=-80/-169|-54/-80|-49.5/-64.6|-41.1/-41.1|-25.8/-15|-10.8/-4.5|0/0|20/8.3
3575 Hard limiter at -6dB:
3577 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3581 Hard limiter at -12dB:
3583 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3587 Hard noise gate at -35 dB:
3589 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3595 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3599 @section compensationdelay
3601 Compensation Delay Line is a metric based delay to compensate differing
3602 positions of microphones or speakers.
3604 For example, you have recorded guitar with two microphones placed in
3605 different locations. Because the front of sound wave has fixed speed in
3606 normal conditions, the phasing of microphones can vary and depends on
3607 their location and interposition. The best sound mix can be achieved when
3608 these microphones are in phase (synchronized). Note that a distance of
3609 ~30 cm between microphones makes one microphone capture the signal in
3610 antiphase to the other microphone. That makes the final mix sound moody.
3611 This filter helps to solve phasing problems by adding different delays
3612 to each microphone track and make them synchronized.
3614 The best result can be reached when you take one track as base and
3615 synchronize other tracks one by one with it.
3616 Remember that synchronization/delay tolerance depends on sample rate, too.
3617 Higher sample rates will give more tolerance.
3619 The filter accepts the following parameters:
3623 Set millimeters distance. This is compensation distance for fine tuning.
3627 Set cm distance. This is compensation distance for tightening distance setup.
3631 Set meters distance. This is compensation distance for hard distance setup.
3635 Set dry amount. Amount of unprocessed (dry) signal.
3639 Set wet amount. Amount of processed (wet) signal.
3643 Set temperature in degrees Celsius. This is the temperature of the environment.
3648 Apply headphone crossfeed filter.
3650 Crossfeed is the process of blending the left and right channels of stereo
3652 It is mainly used to reduce extreme stereo separation of low frequencies.
3654 The intent is to produce more speaker like sound to the listener.
3656 The filter accepts the following options:
3660 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3661 This sets gain of low shelf filter for side part of stereo image.
3662 Default is -6dB. Max allowed is -30db when strength is set to 1.
3665 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3666 This sets cut off frequency of low shelf filter. Default is cut off near
3667 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3670 Set curve slope of low shelf filter. Default is 0.5.
3671 Allowed range is from 0.01 to 1.
3674 Set input gain. Default is 0.9.
3677 Set output gain. Default is 1.
3680 @subsection Commands
3682 This filter supports the all above options as @ref{commands}.
3684 @section crystalizer
3685 Simple algorithm for audio noise sharpening.
3687 This filter linearly increases differences betweeen each audio sample.
3689 The filter accepts the following options:
3693 Sets the intensity of effect (default: 2.0). Must be in range between -10.0 to 0
3694 (unchanged sound) to 10.0 (maximum effect).
3695 To inverse filtering use negative value.
3698 Enable clipping. By default is enabled.
3701 @subsection Commands
3703 This filter supports the all above options as @ref{commands}.
3706 Apply a DC shift to the audio.
3708 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3709 in the recording chain) from the audio. The effect of a DC offset is reduced
3710 headroom and hence volume. The @ref{astats} filter can be used to determine if
3711 a signal has a DC offset.
3715 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3719 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3720 used to prevent clipping.
3725 Apply de-essing to the audio samples.
3729 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3733 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3737 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3741 Set the output mode.
3743 It accepts the following values:
3746 Pass input unchanged.
3749 Pass ess filtered out.
3754 Default value is @var{o}.
3760 Measure audio dynamic range.
3762 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3763 is found in transition material. And anything less that 8 have very poor dynamics
3764 and is very compressed.
3766 The filter accepts the following options:
3770 Set window length in seconds used to split audio into segments of equal length.
3771 Default is 3 seconds.
3775 Dynamic Audio Normalizer.
3777 This filter applies a certain amount of gain to the input audio in order
3778 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3779 contrast to more "simple" normalization algorithms, the Dynamic Audio
3780 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3781 This allows for applying extra gain to the "quiet" sections of the audio
3782 while avoiding distortions or clipping the "loud" sections. In other words:
3783 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3784 sections, in the sense that the volume of each section is brought to the
3785 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3786 this goal *without* applying "dynamic range compressing". It will retain 100%
3787 of the dynamic range *within* each section of the audio file.
3791 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3792 Default is 500 milliseconds.
3793 The Dynamic Audio Normalizer processes the input audio in small chunks,
3794 referred to as frames. This is required, because a peak magnitude has no
3795 meaning for just a single sample value. Instead, we need to determine the
3796 peak magnitude for a contiguous sequence of sample values. While a "standard"
3797 normalizer would simply use the peak magnitude of the complete file, the
3798 Dynamic Audio Normalizer determines the peak magnitude individually for each
3799 frame. The length of a frame is specified in milliseconds. By default, the
3800 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3801 been found to give good results with most files.
3802 Note that the exact frame length, in number of samples, will be determined
3803 automatically, based on the sampling rate of the individual input audio file.
3806 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3807 number. Default is 31.
3808 Probably the most important parameter of the Dynamic Audio Normalizer is the
3809 @code{window size} of the Gaussian smoothing filter. The filter's window size
3810 is specified in frames, centered around the current frame. For the sake of
3811 simplicity, this must be an odd number. Consequently, the default value of 31
3812 takes into account the current frame, as well as the 15 preceding frames and
3813 the 15 subsequent frames. Using a larger window results in a stronger
3814 smoothing effect and thus in less gain variation, i.e. slower gain
3815 adaptation. Conversely, using a smaller window results in a weaker smoothing
3816 effect and thus in more gain variation, i.e. faster gain adaptation.
3817 In other words, the more you increase this value, the more the Dynamic Audio
3818 Normalizer will behave like a "traditional" normalization filter. On the
3819 contrary, the more you decrease this value, the more the Dynamic Audio
3820 Normalizer will behave like a dynamic range compressor.
3823 Set the target peak value. This specifies the highest permissible magnitude
3824 level for the normalized audio input. This filter will try to approach the
3825 target peak magnitude as closely as possible, but at the same time it also
3826 makes sure that the normalized signal will never exceed the peak magnitude.
3827 A frame's maximum local gain factor is imposed directly by the target peak
3828 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3829 It is not recommended to go above this value.
3832 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3833 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3834 factor for each input frame, i.e. the maximum gain factor that does not
3835 result in clipping or distortion. The maximum gain factor is determined by
3836 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3837 additionally bounds the frame's maximum gain factor by a predetermined
3838 (global) maximum gain factor. This is done in order to avoid excessive gain
3839 factors in "silent" or almost silent frames. By default, the maximum gain
3840 factor is 10.0, For most inputs the default value should be sufficient and
3841 it usually is not recommended to increase this value. Though, for input
3842 with an extremely low overall volume level, it may be necessary to allow even
3843 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3844 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3845 Instead, a "sigmoid" threshold function will be applied. This way, the
3846 gain factors will smoothly approach the threshold value, but never exceed that
3850 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3851 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3852 This means that the maximum local gain factor for each frame is defined
3853 (only) by the frame's highest magnitude sample. This way, the samples can
3854 be amplified as much as possible without exceeding the maximum signal
3855 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3856 Normalizer can also take into account the frame's root mean square,
3857 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3858 determine the power of a time-varying signal. It is therefore considered
3859 that the RMS is a better approximation of the "perceived loudness" than
3860 just looking at the signal's peak magnitude. Consequently, by adjusting all
3861 frames to a constant RMS value, a uniform "perceived loudness" can be
3862 established. If a target RMS value has been specified, a frame's local gain
3863 factor is defined as the factor that would result in exactly that RMS value.
3864 Note, however, that the maximum local gain factor is still restricted by the
3865 frame's highest magnitude sample, in order to prevent clipping.
3868 Enable channels coupling. By default is enabled.
3869 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3870 amount. This means the same gain factor will be applied to all channels, i.e.
3871 the maximum possible gain factor is determined by the "loudest" channel.
3872 However, in some recordings, it may happen that the volume of the different
3873 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3874 In this case, this option can be used to disable the channel coupling. This way,
3875 the gain factor will be determined independently for each channel, depending
3876 only on the individual channel's highest magnitude sample. This allows for
3877 harmonizing the volume of the different channels.
3880 Enable DC bias correction. By default is disabled.
3881 An audio signal (in the time domain) is a sequence of sample values.
3882 In the Dynamic Audio Normalizer these sample values are represented in the
3883 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3884 audio signal, or "waveform", should be centered around the zero point.
3885 That means if we calculate the mean value of all samples in a file, or in a
3886 single frame, then the result should be 0.0 or at least very close to that
3887 value. If, however, there is a significant deviation of the mean value from
3888 0.0, in either positive or negative direction, this is referred to as a
3889 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3890 Audio Normalizer provides optional DC bias correction.
3891 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3892 the mean value, or "DC correction" offset, of each input frame and subtract
3893 that value from all of the frame's sample values which ensures those samples
3894 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3895 boundaries, the DC correction offset values will be interpolated smoothly
3896 between neighbouring frames.
3898 @item altboundary, b
3899 Enable alternative boundary mode. By default is disabled.
3900 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3901 around each frame. This includes the preceding frames as well as the
3902 subsequent frames. However, for the "boundary" frames, located at the very
3903 beginning and at the very end of the audio file, not all neighbouring
3904 frames are available. In particular, for the first few frames in the audio
3905 file, the preceding frames are not known. And, similarly, for the last few
3906 frames in the audio file, the subsequent frames are not known. Thus, the
3907 question arises which gain factors should be assumed for the missing frames
3908 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3909 to deal with this situation. The default boundary mode assumes a gain factor
3910 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3911 "fade out" at the beginning and at the end of the input, respectively.
3914 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3915 By default, the Dynamic Audio Normalizer does not apply "traditional"
3916 compression. This means that signal peaks will not be pruned and thus the
3917 full dynamic range will be retained within each local neighbourhood. However,
3918 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3919 normalization algorithm with a more "traditional" compression.
3920 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3921 (thresholding) function. If (and only if) the compression feature is enabled,
3922 all input frames will be processed by a soft knee thresholding function prior
3923 to the actual normalization process. Put simply, the thresholding function is
3924 going to prune all samples whose magnitude exceeds a certain threshold value.
3925 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3926 value. Instead, the threshold value will be adjusted for each individual
3928 In general, smaller parameters result in stronger compression, and vice versa.
3929 Values below 3.0 are not recommended, because audible distortion may appear.
3932 Set the target threshold value. This specifies the lowest permissible
3933 magnitude level for the audio input which will be normalized.
3934 If input frame volume is above this value frame will be normalized.
3935 Otherwise frame may not be normalized at all. The default value is set
3936 to 0, which means all input frames will be normalized.
3937 This option is mostly useful if digital noise is not wanted to be amplified.
3940 @subsection Commands
3942 This filter supports the all above options as @ref{commands}.
3946 Make audio easier to listen to on headphones.
3948 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3949 so that when listened to on headphones the stereo image is moved from
3950 inside your head (standard for headphones) to outside and in front of
3951 the listener (standard for speakers).
3957 Apply a two-pole peaking equalisation (EQ) filter. With this
3958 filter, the signal-level at and around a selected frequency can
3959 be increased or decreased, whilst (unlike bandpass and bandreject
3960 filters) that at all other frequencies is unchanged.
3962 In order to produce complex equalisation curves, this filter can
3963 be given several times, each with a different central frequency.
3965 The filter accepts the following options:
3969 Set the filter's central frequency in Hz.
3972 Set method to specify band-width of filter.
3987 Specify the band-width of a filter in width_type units.
3990 Set the required gain or attenuation in dB.
3991 Beware of clipping when using a positive gain.
3994 How much to use filtered signal in output. Default is 1.
3995 Range is between 0 and 1.
3998 Specify which channels to filter, by default all available are filtered.
4001 Normalize biquad coefficients, by default is disabled.
4002 Enabling it will normalize magnitude response at DC to 0dB.
4005 Set transform type of IIR filter.
4014 Set precison of filtering.
4017 Pick automatic sample format depending on surround filters.
4019 Always use signed 16-bit.
4021 Always use signed 32-bit.
4023 Always use float 32-bit.
4025 Always use float 64-bit.
4029 @subsection Examples
4032 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
4034 equalizer=f=1000:t=h:width=200:g=-10
4038 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
4040 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
4044 @subsection Commands
4046 This filter supports the following commands:
4049 Change equalizer frequency.
4050 Syntax for the command is : "@var{frequency}"
4053 Change equalizer width_type.
4054 Syntax for the command is : "@var{width_type}"
4057 Change equalizer width.
4058 Syntax for the command is : "@var{width}"
4061 Change equalizer gain.
4062 Syntax for the command is : "@var{gain}"
4065 Change equalizer mix.
4066 Syntax for the command is : "@var{mix}"
4069 @section extrastereo
4071 Linearly increases the difference between left and right channels which
4072 adds some sort of "live" effect to playback.
4074 The filter accepts the following options:
4078 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
4079 (average of both channels), with 1.0 sound will be unchanged, with
4080 -1.0 left and right channels will be swapped.
4083 Enable clipping. By default is enabled.
4086 @subsection Commands
4088 This filter supports the all above options as @ref{commands}.
4090 @section firequalizer
4091 Apply FIR Equalization using arbitrary frequency response.
4093 The filter accepts the following option:
4097 Set gain curve equation (in dB). The expression can contain variables:
4100 the evaluated frequency
4104 channel number, set to 0 when multichannels evaluation is disabled
4106 channel id, see libavutil/channel_layout.h, set to the first channel id when
4107 multichannels evaluation is disabled
4111 channel_layout, see libavutil/channel_layout.h
4116 @item gain_interpolate(f)
4117 interpolate gain on frequency f based on gain_entry
4118 @item cubic_interpolate(f)
4119 same as gain_interpolate, but smoother
4121 This option is also available as command. Default is @code{gain_interpolate(f)}.
4124 Set gain entry for gain_interpolate function. The expression can
4128 store gain entry at frequency f with value g
4130 This option is also available as command.
4133 Set filter delay in seconds. Higher value means more accurate.
4134 Default is @code{0.01}.
4137 Set filter accuracy in Hz. Lower value means more accurate.
4138 Default is @code{5}.
4141 Set window function. Acceptable values are:
4144 rectangular window, useful when gain curve is already smooth
4146 hann window (default)
4152 3-terms continuous 1st derivative nuttall window
4154 minimum 3-terms discontinuous nuttall window
4156 4-terms continuous 1st derivative nuttall window
4158 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
4160 blackman-harris window
4166 If enabled, use fixed number of audio samples. This improves speed when
4167 filtering with large delay. Default is disabled.
4170 Enable multichannels evaluation on gain. Default is disabled.
4173 Enable zero phase mode by subtracting timestamp to compensate delay.
4174 Default is disabled.
4177 Set scale used by gain. Acceptable values are:
4180 linear frequency, linear gain
4182 linear frequency, logarithmic (in dB) gain (default)
4184 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
4186 logarithmic frequency, logarithmic gain
4190 Set file for dumping, suitable for gnuplot.
4193 Set scale for dumpfile. Acceptable values are same with scale option.
4197 Enable 2-channel convolution using complex FFT. This improves speed significantly.
4198 Default is disabled.
4201 Enable minimum phase impulse response. Default is disabled.
4204 @subsection Examples
4209 firequalizer=gain='if(lt(f,1000), 0, -INF)'
4212 lowpass at 1000 Hz with gain_entry:
4214 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
4217 custom equalization:
4219 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
4222 higher delay with zero phase to compensate delay:
4224 firequalizer=delay=0.1:fixed=on:zero_phase=on
4227 lowpass on left channel, highpass on right channel:
4229 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
4230 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
4235 Apply a flanging effect to the audio.
4237 The filter accepts the following options:
4241 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
4244 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
4247 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
4251 Set percentage of delayed signal mixed with original. Range from 0 to 100.
4252 Default value is 71.
4255 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
4258 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
4259 Default value is @var{sinusoidal}.
4262 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
4263 Default value is 25.
4266 Set delay-line interpolation, @var{linear} or @var{quadratic}.
4267 Default is @var{linear}.
4271 Apply Haas effect to audio.
4273 Note that this makes most sense to apply on mono signals.
4274 With this filter applied to mono signals it give some directionality and
4275 stretches its stereo image.
4277 The filter accepts the following options:
4281 Set input level. By default is @var{1}, or 0dB
4284 Set output level. By default is @var{1}, or 0dB.
4287 Set gain applied to side part of signal. By default is @var{1}.
4290 Set kind of middle source. Can be one of the following:
4300 Pick middle part signal of stereo image.
4303 Pick side part signal of stereo image.
4307 Change middle phase. By default is disabled.
4310 Set left channel delay. By default is @var{2.05} milliseconds.
4313 Set left channel balance. By default is @var{-1}.
4316 Set left channel gain. By default is @var{1}.
4319 Change left phase. By default is disabled.
4322 Set right channel delay. By defaults is @var{2.12} milliseconds.
4325 Set right channel balance. By default is @var{1}.
4328 Set right channel gain. By default is @var{1}.
4331 Change right phase. By default is enabled.
4336 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
4337 embedded HDCD codes is expanded into a 20-bit PCM stream.
4339 The filter supports the Peak Extend and Low-level Gain Adjustment features
4340 of HDCD, and detects the Transient Filter flag.
4343 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
4346 When using the filter with wav, note the default encoding for wav is 16-bit,
4347 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
4348 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
4350 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
4351 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
4354 The filter accepts the following options:
4357 @item disable_autoconvert
4358 Disable any automatic format conversion or resampling in the filter graph.
4360 @item process_stereo
4361 Process the stereo channels together. If target_gain does not match between
4362 channels, consider it invalid and use the last valid target_gain.
4365 Set the code detect timer period in ms.
4368 Always extend peaks above -3dBFS even if PE isn't signaled.
4371 Replace audio with a solid tone and adjust the amplitude to signal some
4372 specific aspect of the decoding process. The output file can be loaded in
4373 an audio editor alongside the original to aid analysis.
4375 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
4382 Gain adjustment level at each sample
4384 Samples where peak extend occurs
4386 Samples where the code detect timer is active
4388 Samples where the target gain does not match between channels
4394 Apply head-related transfer functions (HRTFs) to create virtual
4395 loudspeakers around the user for binaural listening via headphones.
4396 The HRIRs are provided via additional streams, for each channel
4397 one stereo input stream is needed.
4399 The filter accepts the following options:
4403 Set mapping of input streams for convolution.
4404 The argument is a '|'-separated list of channel names in order as they
4405 are given as additional stream inputs for filter.
4406 This also specify number of input streams. Number of input streams
4407 must be not less than number of channels in first stream plus one.
4410 Set gain applied to audio. Value is in dB. Default is 0.
4413 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4414 processing audio in time domain which is slow.
4415 @var{freq} is processing audio in frequency domain which is fast.
4416 Default is @var{freq}.
4419 Set custom gain for LFE channels. Value is in dB. Default is 0.
4422 Set size of frame in number of samples which will be processed at once.
4423 Default value is @var{1024}. Allowed range is from 1024 to 96000.
4426 Set format of hrir stream.
4427 Default value is @var{stereo}. Alternative value is @var{multich}.
4428 If value is set to @var{stereo}, number of additional streams should
4429 be greater or equal to number of input channels in first input stream.
4430 Also each additional stream should have stereo number of channels.
4431 If value is set to @var{multich}, number of additional streams should
4432 be exactly one. Also number of input channels of additional stream
4433 should be equal or greater than twice number of channels of first input
4437 @subsection Examples
4441 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4442 each amovie filter use stereo file with IR coefficients as input.
4443 The files give coefficients for each position of virtual loudspeaker:
4446 -filter_complex "amovie=azi_270_ele_0_DFC.wav[sr];amovie=azi_90_ele_0_DFC.wav[sl];amovie=azi_225_ele_0_DFC.wav[br];amovie=azi_135_ele_0_DFC.wav[bl];amovie=azi_0_ele_0_DFC.wav,asplit[fc][lfe];amovie=azi_35_ele_0_DFC.wav[fl];amovie=azi_325_ele_0_DFC.wav[fr];[0:a][fl][fr][fc][lfe][bl][br][sl][sr]headphone=FL|FR|FC|LFE|BL|BR|SL|SR"
4451 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4452 but now in @var{multich} @var{hrir} format.
4454 ffmpeg -i input.wav -filter_complex "amovie=minp.wav[hrirs];[0:a][hrirs]headphone=map=FL|FR|FC|LFE|BL|BR|SL|SR:hrir=multich"
4461 Apply a high-pass filter with 3dB point frequency.
4462 The filter can be either single-pole, or double-pole (the default).
4463 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4465 The filter accepts the following options:
4469 Set frequency in Hz. Default is 3000.
4472 Set number of poles. Default is 2.
4475 Set method to specify band-width of filter.
4490 Specify the band-width of a filter in width_type units.
4491 Applies only to double-pole filter.
4492 The default is 0.707q and gives a Butterworth response.
4495 How much to use filtered signal in output. Default is 1.
4496 Range is between 0 and 1.
4499 Specify which channels to filter, by default all available are filtered.
4502 Normalize biquad coefficients, by default is disabled.
4503 Enabling it will normalize magnitude response at DC to 0dB.
4506 Set transform type of IIR filter.
4515 Set precison of filtering.
4518 Pick automatic sample format depending on surround filters.
4520 Always use signed 16-bit.
4522 Always use signed 32-bit.
4524 Always use float 32-bit.
4526 Always use float 64-bit.
4530 @subsection Commands
4532 This filter supports the following commands:
4535 Change highpass frequency.
4536 Syntax for the command is : "@var{frequency}"
4539 Change highpass width_type.
4540 Syntax for the command is : "@var{width_type}"
4543 Change highpass width.
4544 Syntax for the command is : "@var{width}"
4547 Change highpass mix.
4548 Syntax for the command is : "@var{mix}"
4553 Join multiple input streams into one multi-channel stream.
4555 It accepts the following parameters:
4559 The number of input streams. It defaults to 2.
4561 @item channel_layout
4562 The desired output channel layout. It defaults to stereo.
4565 Map channels from inputs to output. The argument is a '|'-separated list of
4566 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4567 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4568 can be either the name of the input channel (e.g. FL for front left) or its
4569 index in the specified input stream. @var{out_channel} is the name of the output
4573 The filter will attempt to guess the mappings when they are not specified
4574 explicitly. It does so by first trying to find an unused matching input channel
4575 and if that fails it picks the first unused input channel.
4577 Join 3 inputs (with properly set channel layouts):
4579 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4582 Build a 5.1 output from 6 single-channel streams:
4584 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4585 'join=inputs=6:channel_layout=5.1:map=0.0-FL|1.0-FR|2.0-FC|3.0-SL|4.0-SR|5.0-LFE'
4591 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4593 To enable compilation of this filter you need to configure FFmpeg with
4594 @code{--enable-ladspa}.
4598 Specifies the name of LADSPA plugin library to load. If the environment
4599 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4600 each one of the directories specified by the colon separated list in
4601 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4602 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4603 @file{/usr/lib/ladspa/}.
4606 Specifies the plugin within the library. Some libraries contain only
4607 one plugin, but others contain many of them. If this is not set filter
4608 will list all available plugins within the specified library.
4611 Set the '|' separated list of controls which are zero or more floating point
4612 values that determine the behavior of the loaded plugin (for example delay,
4614 Controls need to be defined using the following syntax:
4615 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4616 @var{valuei} is the value set on the @var{i}-th control.
4617 Alternatively they can be also defined using the following syntax:
4618 @var{value0}|@var{value1}|@var{value2}|..., where
4619 @var{valuei} is the value set on the @var{i}-th control.
4620 If @option{controls} is set to @code{help}, all available controls and
4621 their valid ranges are printed.
4623 @item sample_rate, s
4624 Specify the sample rate, default to 44100. Only used if plugin have
4628 Set the number of samples per channel per each output frame, default
4629 is 1024. Only used if plugin have zero inputs.
4632 Set the minimum duration of the sourced audio. See
4633 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4634 for the accepted syntax.
4635 Note that the resulting duration may be greater than the specified duration,
4636 as the generated audio is always cut at the end of a complete frame.
4637 If not specified, or the expressed duration is negative, the audio is
4638 supposed to be generated forever.
4639 Only used if plugin have zero inputs.
4642 Enable latency compensation, by default is disabled.
4643 Only used if plugin have inputs.
4646 @subsection Examples
4650 List all available plugins within amp (LADSPA example plugin) library:
4656 List all available controls and their valid ranges for @code{vcf_notch}
4657 plugin from @code{VCF} library:
4659 ladspa=f=vcf:p=vcf_notch:c=help
4663 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4666 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4670 Add reverberation to the audio using TAP-plugins
4671 (Tom's Audio Processing plugins):
4673 ladspa=file=tap_reverb:tap_reverb
4677 Generate white noise, with 0.2 amplitude:
4679 ladspa=file=cmt:noise_source_white:c=c0=.2
4683 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4684 @code{C* Audio Plugin Suite} (CAPS) library:
4686 ladspa=file=caps:Click:c=c1=20'
4690 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4692 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4696 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4697 @code{SWH Plugins} collection:
4699 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4703 Attenuate low frequencies using Multiband EQ from Steve Harris
4704 @code{SWH Plugins} collection:
4706 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4710 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4713 ladspa=caps:Narrower
4717 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4719 ladspa=caps:White:.2
4723 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4725 ladspa=caps:Fractal:c=c1=1
4729 Dynamic volume normalization using @code{VLevel} plugin:
4731 ladspa=vlevel-ladspa:vlevel_mono
4735 @subsection Commands
4737 This filter supports the following commands:
4740 Modify the @var{N}-th control value.
4742 If the specified value is not valid, it is ignored and prior one is kept.
4747 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4748 Support for both single pass (livestreams, files) and double pass (files) modes.
4749 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4750 detect true peaks, the audio stream will be upsampled to 192 kHz.
4751 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4753 The filter accepts the following options:
4757 Set integrated loudness target.
4758 Range is -70.0 - -5.0. Default value is -24.0.
4761 Set loudness range target.
4762 Range is 1.0 - 20.0. Default value is 7.0.
4765 Set maximum true peak.
4766 Range is -9.0 - +0.0. Default value is -2.0.
4768 @item measured_I, measured_i
4769 Measured IL of input file.
4770 Range is -99.0 - +0.0.
4772 @item measured_LRA, measured_lra
4773 Measured LRA of input file.
4774 Range is 0.0 - 99.0.
4776 @item measured_TP, measured_tp
4777 Measured true peak of input file.
4778 Range is -99.0 - +99.0.
4780 @item measured_thresh
4781 Measured threshold of input file.
4782 Range is -99.0 - +0.0.
4785 Set offset gain. Gain is applied before the true-peak limiter.
4786 Range is -99.0 - +99.0. Default is +0.0.
4789 Normalize by linearly scaling the source audio.
4790 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4791 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4792 be lower than source LRA and the change in integrated loudness shouldn't
4793 result in a true peak which exceeds the target TP. If any of these
4794 conditions aren't met, normalization mode will revert to @var{dynamic}.
4795 Options are @code{true} or @code{false}. Default is @code{true}.
4798 Treat mono input files as "dual-mono". If a mono file is intended for playback
4799 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4800 If set to @code{true}, this option will compensate for this effect.
4801 Multi-channel input files are not affected by this option.
4802 Options are true or false. Default is false.
4805 Set print format for stats. Options are summary, json, or none.
4806 Default value is none.
4811 Apply a low-pass filter with 3dB point frequency.
4812 The filter can be either single-pole or double-pole (the default).
4813 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4815 The filter accepts the following options:
4819 Set frequency in Hz. Default is 500.
4822 Set number of poles. Default is 2.
4825 Set method to specify band-width of filter.
4840 Specify the band-width of a filter in width_type units.
4841 Applies only to double-pole filter.
4842 The default is 0.707q and gives a Butterworth response.
4845 How much to use filtered signal in output. Default is 1.
4846 Range is between 0 and 1.
4849 Specify which channels to filter, by default all available are filtered.
4852 Normalize biquad coefficients, by default is disabled.
4853 Enabling it will normalize magnitude response at DC to 0dB.
4856 Set transform type of IIR filter.
4865 Set precison of filtering.
4868 Pick automatic sample format depending on surround filters.
4870 Always use signed 16-bit.
4872 Always use signed 32-bit.
4874 Always use float 32-bit.
4876 Always use float 64-bit.
4880 @subsection Examples
4883 Lowpass only LFE channel, it LFE is not present it does nothing:
4889 @subsection Commands
4891 This filter supports the following commands:
4894 Change lowpass frequency.
4895 Syntax for the command is : "@var{frequency}"
4898 Change lowpass width_type.
4899 Syntax for the command is : "@var{width_type}"
4902 Change lowpass width.
4903 Syntax for the command is : "@var{width}"
4907 Syntax for the command is : "@var{mix}"
4912 Load a LV2 (LADSPA Version 2) plugin.
4914 To enable compilation of this filter you need to configure FFmpeg with
4915 @code{--enable-lv2}.
4919 Specifies the plugin URI. You may need to escape ':'.
4922 Set the '|' separated list of controls which are zero or more floating point
4923 values that determine the behavior of the loaded plugin (for example delay,
4925 If @option{controls} is set to @code{help}, all available controls and
4926 their valid ranges are printed.
4928 @item sample_rate, s
4929 Specify the sample rate, default to 44100. Only used if plugin have
4933 Set the number of samples per channel per each output frame, default
4934 is 1024. Only used if plugin have zero inputs.
4937 Set the minimum duration of the sourced audio. See
4938 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4939 for the accepted syntax.
4940 Note that the resulting duration may be greater than the specified duration,
4941 as the generated audio is always cut at the end of a complete frame.
4942 If not specified, or the expressed duration is negative, the audio is
4943 supposed to be generated forever.
4944 Only used if plugin have zero inputs.
4947 @subsection Examples
4951 Apply bass enhancer plugin from Calf:
4953 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4957 Apply vinyl plugin from Calf:
4959 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4963 Apply bit crusher plugin from ArtyFX:
4965 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4970 Multiband Compress or expand the audio's dynamic range.
4972 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4973 This is akin to the crossover of a loudspeaker, and results in flat frequency
4974 response when absent compander action.
4976 It accepts the following parameters:
4980 This option syntax is:
4981 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4982 For explanation of each item refer to compand filter documentation.
4988 Mix channels with specific gain levels. The filter accepts the output
4989 channel layout followed by a set of channels definitions.
4991 This filter is also designed to efficiently remap the channels of an audio
4994 The filter accepts parameters of the form:
4995 "@var{l}|@var{outdef}|@var{outdef}|..."
4999 output channel layout or number of channels
5002 output channel specification, of the form:
5003 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
5006 output channel to define, either a channel name (FL, FR, etc.) or a channel
5007 number (c0, c1, etc.)
5010 multiplicative coefficient for the channel, 1 leaving the volume unchanged
5013 input channel to use, see out_name for details; it is not possible to mix
5014 named and numbered input channels
5017 If the `=' in a channel specification is replaced by `<', then the gains for
5018 that specification will be renormalized so that the total is 1, thus
5019 avoiding clipping noise.
5021 @subsection Mixing examples
5023 For example, if you want to down-mix from stereo to mono, but with a bigger
5024 factor for the left channel:
5026 pan=1c|c0=0.9*c0+0.1*c1
5029 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
5030 7-channels surround:
5032 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
5035 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
5036 that should be preferred (see "-ac" option) unless you have very specific
5039 @subsection Remapping examples
5041 The channel remapping will be effective if, and only if:
5044 @item gain coefficients are zeroes or ones,
5045 @item only one input per channel output,
5048 If all these conditions are satisfied, the filter will notify the user ("Pure
5049 channel mapping detected"), and use an optimized and lossless method to do the
5052 For example, if you have a 5.1 source and want a stereo audio stream by
5053 dropping the extra channels:
5055 pan="stereo| c0=FL | c1=FR"
5058 Given the same source, you can also switch front left and front right channels
5059 and keep the input channel layout:
5061 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
5064 If the input is a stereo audio stream, you can mute the front left channel (and
5065 still keep the stereo channel layout) with:
5070 Still with a stereo audio stream input, you can copy the right channel in both
5071 front left and right:
5073 pan="stereo| c0=FR | c1=FR"
5078 ReplayGain scanner filter. This filter takes an audio stream as an input and
5079 outputs it unchanged.
5080 At end of filtering it displays @code{track_gain} and @code{track_peak}.
5084 Convert the audio sample format, sample rate and channel layout. It is
5085 not meant to be used directly.
5088 Apply time-stretching and pitch-shifting with librubberband.
5090 To enable compilation of this filter, you need to configure FFmpeg with
5091 @code{--enable-librubberband}.
5093 The filter accepts the following options:
5097 Set tempo scale factor.
5100 Set pitch scale factor.
5103 Set transients detector.
5104 Possible values are:
5113 Possible values are:
5122 Possible values are:
5129 Set processing window size.
5130 Possible values are:
5139 Possible values are:
5146 Enable formant preservation when shift pitching.
5147 Possible values are:
5155 Possible values are:
5164 Possible values are:
5171 @subsection Commands
5173 This filter supports the following commands:
5176 Change filter tempo scale factor.
5177 Syntax for the command is : "@var{tempo}"
5180 Change filter pitch scale factor.
5181 Syntax for the command is : "@var{pitch}"
5184 @section sidechaincompress
5186 This filter acts like normal compressor but has the ability to compress
5187 detected signal using second input signal.
5188 It needs two input streams and returns one output stream.
5189 First input stream will be processed depending on second stream signal.
5190 The filtered signal then can be filtered with other filters in later stages of
5191 processing. See @ref{pan} and @ref{amerge} filter.
5193 The filter accepts the following options:
5197 Set input gain. Default is 1. Range is between 0.015625 and 64.
5200 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
5201 Default is @code{downward}.
5204 If a signal of second stream raises above this level it will affect the gain
5205 reduction of first stream.
5206 By default is 0.125. Range is between 0.00097563 and 1.
5209 Set a ratio about which the signal is reduced. 1:2 means that if the level
5210 raised 4dB above the threshold, it will be only 2dB above after the reduction.
5211 Default is 2. Range is between 1 and 20.
5214 Amount of milliseconds the signal has to rise above the threshold before gain
5215 reduction starts. Default is 20. Range is between 0.01 and 2000.
5218 Amount of milliseconds the signal has to fall below the threshold before
5219 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
5222 Set the amount by how much signal will be amplified after processing.
5223 Default is 1. Range is from 1 to 64.
5226 Curve the sharp knee around the threshold to enter gain reduction more softly.
5227 Default is 2.82843. Range is between 1 and 8.
5230 Choose if the @code{average} level between all channels of side-chain stream
5231 or the louder(@code{maximum}) channel of side-chain stream affects the
5232 reduction. Default is @code{average}.
5235 Should the exact signal be taken in case of @code{peak} or an RMS one in case
5236 of @code{rms}. Default is @code{rms} which is mainly smoother.
5239 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
5242 How much to use compressed signal in output. Default is 1.
5243 Range is between 0 and 1.
5246 @subsection Commands
5248 This filter supports the all above options as @ref{commands}.
5250 @subsection Examples
5254 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
5255 depending on the signal of 2nd input and later compressed signal to be
5256 merged with 2nd input:
5258 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
5262 @section sidechaingate
5264 A sidechain gate acts like a normal (wideband) gate but has the ability to
5265 filter the detected signal before sending it to the gain reduction stage.
5266 Normally a gate uses the full range signal to detect a level above the
5268 For example: If you cut all lower frequencies from your sidechain signal
5269 the gate will decrease the volume of your track only if not enough highs
5270 appear. With this technique you are able to reduce the resonation of a
5271 natural drum or remove "rumbling" of muted strokes from a heavily distorted
5273 It needs two input streams and returns one output stream.
5274 First input stream will be processed depending on second stream signal.
5276 The filter accepts the following options:
5280 Set input level before filtering.
5281 Default is 1. Allowed range is from 0.015625 to 64.
5284 Set the mode of operation. Can be @code{upward} or @code{downward}.
5285 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
5286 will be amplified, expanding dynamic range in upward direction.
5287 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
5290 Set the level of gain reduction when the signal is below the threshold.
5291 Default is 0.06125. Allowed range is from 0 to 1.
5292 Setting this to 0 disables reduction and then filter behaves like expander.
5295 If a signal rises above this level the gain reduction is released.
5296 Default is 0.125. Allowed range is from 0 to 1.
5299 Set a ratio about which the signal is reduced.
5300 Default is 2. Allowed range is from 1 to 9000.
5303 Amount of milliseconds the signal has to rise above the threshold before gain
5305 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
5308 Amount of milliseconds the signal has to fall below the threshold before the
5309 reduction is increased again. Default is 250 milliseconds.
5310 Allowed range is from 0.01 to 9000.
5313 Set amount of amplification of signal after processing.
5314 Default is 1. Allowed range is from 1 to 64.
5317 Curve the sharp knee around the threshold to enter gain reduction more softly.
5318 Default is 2.828427125. Allowed range is from 1 to 8.
5321 Choose if exact signal should be taken for detection or an RMS like one.
5322 Default is rms. Can be peak or rms.
5325 Choose if the average level between all channels or the louder channel affects
5327 Default is average. Can be average or maximum.
5330 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
5333 @subsection Commands
5335 This filter supports the all above options as @ref{commands}.
5337 @section silencedetect
5339 Detect silence in an audio stream.
5341 This filter logs a message when it detects that the input audio volume is less
5342 or equal to a noise tolerance value for a duration greater or equal to the
5343 minimum detected noise duration.
5345 The printed times and duration are expressed in seconds. The
5346 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
5347 is set on the first frame whose timestamp equals or exceeds the detection
5348 duration and it contains the timestamp of the first frame of the silence.
5350 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
5351 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
5352 keys are set on the first frame after the silence. If @option{mono} is
5353 enabled, and each channel is evaluated separately, the @code{.X}
5354 suffixed keys are used, and @code{X} corresponds to the channel number.
5356 The filter accepts the following options:
5360 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
5361 specified value) or amplitude ratio. Default is -60dB, or 0.001.
5364 Set silence duration until notification (default is 2 seconds). See
5365 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5366 for the accepted syntax.
5369 Process each channel separately, instead of combined. By default is disabled.
5372 @subsection Examples
5376 Detect 5 seconds of silence with -50dB noise tolerance:
5378 silencedetect=n=-50dB:d=5
5382 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
5383 tolerance in @file{silence.mp3}:
5385 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
5389 @section silenceremove
5391 Remove silence from the beginning, middle or end of the audio.
5393 The filter accepts the following options:
5397 This value is used to indicate if audio should be trimmed at beginning of
5398 the audio. A value of zero indicates no silence should be trimmed from the
5399 beginning. When specifying a non-zero value, it trims audio up until it
5400 finds non-silence. Normally, when trimming silence from beginning of audio
5401 the @var{start_periods} will be @code{1} but it can be increased to higher
5402 values to trim all audio up to specific count of non-silence periods.
5403 Default value is @code{0}.
5405 @item start_duration
5406 Specify the amount of time that non-silence must be detected before it stops
5407 trimming audio. By increasing the duration, bursts of noises can be treated
5408 as silence and trimmed off. Default value is @code{0}.
5410 @item start_threshold
5411 This indicates what sample value should be treated as silence. For digital
5412 audio, a value of @code{0} may be fine but for audio recorded from analog,
5413 you may wish to increase the value to account for background noise.
5414 Can be specified in dB (in case "dB" is appended to the specified value)
5415 or amplitude ratio. Default value is @code{0}.
5418 Specify max duration of silence at beginning that will be kept after
5419 trimming. Default is 0, which is equal to trimming all samples detected
5423 Specify mode of detection of silence end in start of multi-channel audio.
5424 Can be @var{any} or @var{all}. Default is @var{any}.
5425 With @var{any}, any sample that is detected as non-silence will cause
5426 stopped trimming of silence.
5427 With @var{all}, only if all channels are detected as non-silence will cause
5428 stopped trimming of silence.
5431 Set the count for trimming silence from the end of audio.
5432 To remove silence from the middle of a file, specify a @var{stop_periods}
5433 that is negative. This value is then treated as a positive value and is
5434 used to indicate the effect should restart processing as specified by
5435 @var{start_periods}, making it suitable for removing periods of silence
5436 in the middle of the audio.
5437 Default value is @code{0}.
5440 Specify a duration of silence that must exist before audio is not copied any
5441 more. By specifying a higher duration, silence that is wanted can be left in
5443 Default value is @code{0}.
5445 @item stop_threshold
5446 This is the same as @option{start_threshold} but for trimming silence from
5448 Can be specified in dB (in case "dB" is appended to the specified value)
5449 or amplitude ratio. Default value is @code{0}.
5452 Specify max duration of silence at end that will be kept after
5453 trimming. Default is 0, which is equal to trimming all samples detected
5457 Specify mode of detection of silence start in end of multi-channel audio.
5458 Can be @var{any} or @var{all}. Default is @var{any}.
5459 With @var{any}, any sample that is detected as non-silence will cause
5460 stopped trimming of silence.
5461 With @var{all}, only if all channels are detected as non-silence will cause
5462 stopped trimming of silence.
5465 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
5466 and works better with digital silence which is exactly 0.
5467 Default value is @code{rms}.
5470 Set duration in number of seconds used to calculate size of window in number
5471 of samples for detecting silence.
5472 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
5475 @subsection Examples
5479 The following example shows how this filter can be used to start a recording
5480 that does not contain the delay at the start which usually occurs between
5481 pressing the record button and the start of the performance:
5483 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
5487 Trim all silence encountered from beginning to end where there is more than 1
5488 second of silence in audio:
5490 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
5494 Trim all digital silence samples, using peak detection, from beginning to end
5495 where there is more than 0 samples of digital silence in audio and digital
5496 silence is detected in all channels at same positions in stream:
5498 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
5504 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
5505 loudspeakers around the user for binaural listening via headphones (audio
5506 formats up to 9 channels supported).
5507 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
5508 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
5509 Austrian Academy of Sciences.
5511 To enable compilation of this filter you need to configure FFmpeg with
5512 @code{--enable-libmysofa}.
5514 The filter accepts the following options:
5518 Set the SOFA file used for rendering.
5521 Set gain applied to audio. Value is in dB. Default is 0.
5524 Set rotation of virtual loudspeakers in deg. Default is 0.
5527 Set elevation of virtual speakers in deg. Default is 0.
5530 Set distance in meters between loudspeakers and the listener with near-field
5531 HRTFs. Default is 1.
5534 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
5535 processing audio in time domain which is slow.
5536 @var{freq} is processing audio in frequency domain which is fast.
5537 Default is @var{freq}.
5540 Set custom positions of virtual loudspeakers. Syntax for this option is:
5541 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
5542 Each virtual loudspeaker is described with short channel name following with
5543 azimuth and elevation in degrees.
5544 Each virtual loudspeaker description is separated by '|'.
5545 For example to override front left and front right channel positions use:
5546 'speakers=FL 45 15|FR 345 15'.
5547 Descriptions with unrecognised channel names are ignored.
5550 Set custom gain for LFE channels. Value is in dB. Default is 0.
5553 Set custom frame size in number of samples. Default is 1024.
5554 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5555 is set to @var{freq}.
5558 Should all IRs be normalized upon importing SOFA file.
5559 By default is enabled.
5562 Should nearest IRs be interpolated with neighbor IRs if exact position
5563 does not match. By default is disabled.
5566 Minphase all IRs upon loading of SOFA file. By default is disabled.
5569 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5572 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5575 @subsection Examples
5579 Using ClubFritz6 sofa file:
5581 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5585 Using ClubFritz12 sofa file and bigger radius with small rotation:
5587 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5591 Similar as above but with custom speaker positions for front left, front right, back left and back right
5592 and also with custom gain:
5594 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5601 This filter expands or compresses each half-cycle of audio samples
5602 (local set of samples all above or all below zero and between two nearest zero crossings) depending
5603 on threshold value, so audio reaches target peak value under conditions controlled by below options.
5605 The filter accepts the following options:
5609 Set the expansion target peak value. This specifies the highest allowed absolute amplitude
5610 level for the normalized audio input. Default value is 0.95. Allowed range is from 0.0 to 1.0.
5613 Set the maximum expansion factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5614 This option controls maximum local half-cycle of samples expansion. The maximum expansion
5615 would be such that local peak value reaches target peak value but never to surpass it and that
5616 ratio between new and previous peak value does not surpass this option value.
5618 @item compression, c
5619 Set the maximum compression factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5620 This option controls maximum local half-cycle of samples compression. This option is used
5621 only if @option{threshold} option is set to value greater than 0.0, then in such cases
5622 when local peak is lower or same as value set by @option{threshold} all samples belonging to
5623 that peak's half-cycle will be compressed by current compression factor.
5626 Set the threshold value. Default value is 0.0. Allowed range is from 0.0 to 1.0.
5627 This option specifies which half-cycles of samples will be compressed and which will be expanded.
5628 Any half-cycle samples with their local peak value below or same as this option value will be
5629 compressed by current compression factor, otherwise, if greater than threshold value they will be
5630 expanded with expansion factor so that it could reach peak target value but never surpass it.
5633 Set the expansion raising amount per each half-cycle of samples. Default value is 0.001.
5634 Allowed range is from 0.0 to 1.0. This controls how fast expansion factor is raised per
5635 each new half-cycle until it reaches @option{expansion} value.
5636 Setting this options too high may lead to distortions.
5639 Set the compression raising amount per each half-cycle of samples. Default value is 0.001.
5640 Allowed range is from 0.0 to 1.0. This controls how fast compression factor is raised per
5641 each new half-cycle until it reaches @option{compression} value.
5644 Specify which channels to filter, by default all available channels are filtered.
5647 Enable inverted filtering, by default is disabled. This inverts interpretation of @option{threshold}
5648 option. When enabled any half-cycle of samples with their local peak value below or same as
5649 @option{threshold} option will be expanded otherwise it will be compressed.
5652 Link channels when calculating gain applied to each filtered channel sample, by default is disabled.
5653 When disabled each filtered channel gain calculation is independent, otherwise when this option
5654 is enabled the minimum of all possible gains for each filtered channel is used.
5657 @subsection Commands
5659 This filter supports the all above options as @ref{commands}.
5661 @section stereotools
5663 This filter has some handy utilities to manage stereo signals, for converting
5664 M/S stereo recordings to L/R signal while having control over the parameters
5665 or spreading the stereo image of master track.
5667 The filter accepts the following options:
5671 Set input level before filtering for both channels. Defaults is 1.
5672 Allowed range is from 0.015625 to 64.
5675 Set output level after filtering for both channels. Defaults is 1.
5676 Allowed range is from 0.015625 to 64.
5679 Set input balance between both channels. Default is 0.
5680 Allowed range is from -1 to 1.
5683 Set output balance between both channels. Default is 0.
5684 Allowed range is from -1 to 1.
5687 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5688 clipping. Disabled by default.
5691 Mute the left channel. Disabled by default.
5694 Mute the right channel. Disabled by default.
5697 Change the phase of the left channel. Disabled by default.
5700 Change the phase of the right channel. Disabled by default.
5703 Set stereo mode. Available values are:
5707 Left/Right to Left/Right, this is default.
5710 Left/Right to Mid/Side.
5713 Mid/Side to Left/Right.
5716 Left/Right to Left/Left.
5719 Left/Right to Right/Right.
5722 Left/Right to Left + Right.
5725 Left/Right to Right/Left.
5728 Mid/Side to Left/Left.
5731 Mid/Side to Right/Right.
5734 Mid/Side to Right/Left.
5737 Left/Right to Left - Right.
5741 Set level of side signal. Default is 1.
5742 Allowed range is from 0.015625 to 64.
5745 Set balance of side signal. Default is 0.
5746 Allowed range is from -1 to 1.
5749 Set level of the middle signal. Default is 1.
5750 Allowed range is from 0.015625 to 64.
5753 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5756 Set stereo base between mono and inversed channels. Default is 0.
5757 Allowed range is from -1 to 1.
5760 Set delay in milliseconds how much to delay left from right channel and
5761 vice versa. Default is 0. Allowed range is from -20 to 20.
5764 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5767 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5769 @item bmode_in, bmode_out
5770 Set balance mode for balance_in/balance_out option.
5772 Can be one of the following:
5776 Classic balance mode. Attenuate one channel at time.
5777 Gain is raised up to 1.
5780 Similar as classic mode above but gain is raised up to 2.
5783 Equal power distribution, from -6dB to +6dB range.
5787 @subsection Commands
5789 This filter supports the all above options as @ref{commands}.
5791 @subsection Examples
5795 Apply karaoke like effect:
5797 stereotools=mlev=0.015625
5801 Convert M/S signal to L/R:
5803 "stereotools=mode=ms>lr"
5807 @section stereowiden
5809 This filter enhance the stereo effect by suppressing signal common to both
5810 channels and by delaying the signal of left into right and vice versa,
5811 thereby widening the stereo effect.
5813 The filter accepts the following options:
5817 Time in milliseconds of the delay of left signal into right and vice versa.
5818 Default is 20 milliseconds.
5821 Amount of gain in delayed signal into right and vice versa. Gives a delay
5822 effect of left signal in right output and vice versa which gives widening
5823 effect. Default is 0.3.
5826 Cross feed of left into right with inverted phase. This helps in suppressing
5827 the mono. If the value is 1 it will cancel all the signal common to both
5828 channels. Default is 0.3.
5831 Set level of input signal of original channel. Default is 0.8.
5834 @subsection Commands
5836 This filter supports the all above options except @code{delay} as @ref{commands}.
5838 @section superequalizer
5839 Apply 18 band equalizer.
5841 The filter accepts the following options:
5848 Set 131Hz band gain.
5850 Set 185Hz band gain.
5852 Set 262Hz band gain.
5854 Set 370Hz band gain.
5856 Set 523Hz band gain.
5858 Set 740Hz band gain.
5860 Set 1047Hz band gain.
5862 Set 1480Hz band gain.
5864 Set 2093Hz band gain.
5866 Set 2960Hz band gain.
5868 Set 4186Hz band gain.
5870 Set 5920Hz band gain.
5872 Set 8372Hz band gain.
5874 Set 11840Hz band gain.
5876 Set 16744Hz band gain.
5878 Set 20000Hz band gain.
5882 Apply audio surround upmix filter.
5884 This filter allows to produce multichannel output from audio stream.
5886 The filter accepts the following options:
5890 Set output channel layout. By default, this is @var{5.1}.
5892 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5893 for the required syntax.
5896 Set input channel layout. By default, this is @var{stereo}.
5898 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5899 for the required syntax.
5902 Set input volume level. By default, this is @var{1}.
5905 Set output volume level. By default, this is @var{1}.
5908 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5911 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5914 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5917 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5918 In @var{add} mode, LFE channel is created from input audio and added to output.
5919 In @var{sub} mode, LFE channel is created from input audio and added to output but
5920 also all non-LFE output channels are subtracted with output LFE channel.
5923 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5924 Default is @var{90}.
5927 Set front center input volume. By default, this is @var{1}.
5930 Set front center output volume. By default, this is @var{1}.
5933 Set front left input volume. By default, this is @var{1}.
5936 Set front left output volume. By default, this is @var{1}.
5939 Set front right input volume. By default, this is @var{1}.
5942 Set front right output volume. By default, this is @var{1}.
5945 Set side left input volume. By default, this is @var{1}.
5948 Set side left output volume. By default, this is @var{1}.
5951 Set side right input volume. By default, this is @var{1}.
5954 Set side right output volume. By default, this is @var{1}.
5957 Set back left input volume. By default, this is @var{1}.
5960 Set back left output volume. By default, this is @var{1}.
5963 Set back right input volume. By default, this is @var{1}.
5966 Set back right output volume. By default, this is @var{1}.
5969 Set back center input volume. By default, this is @var{1}.
5972 Set back center output volume. By default, this is @var{1}.
5975 Set LFE input volume. By default, this is @var{1}.
5978 Set LFE output volume. By default, this is @var{1}.
5981 Set spread usage of stereo image across X axis for all channels.
5984 Set spread usage of stereo image across Y axis for all channels.
5986 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5987 Set spread usage of stereo image across X axis for each channel.
5989 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5990 Set spread usage of stereo image across Y axis for each channel.
5993 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5996 Set window function.
5998 It accepts the following values:
6021 Default is @code{hann}.
6024 Set window overlap. If set to 1, the recommended overlap for selected
6025 window function will be picked. Default is @code{0.5}.
6028 @section treble, highshelf
6030 Boost or cut treble (upper) frequencies of the audio using a two-pole
6031 shelving filter with a response similar to that of a standard
6032 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
6034 The filter accepts the following options:
6038 Give the gain at whichever is the lower of ~22 kHz and the
6039 Nyquist frequency. Its useful range is about -20 (for a large cut)
6040 to +20 (for a large boost). Beware of clipping when using a positive gain.
6043 Set the filter's central frequency and so can be used
6044 to extend or reduce the frequency range to be boosted or cut.
6045 The default value is @code{3000} Hz.
6048 Set method to specify band-width of filter.
6063 Determine how steep is the filter's shelf transition.
6066 Set number of poles. Default is 2.
6069 How much to use filtered signal in output. Default is 1.
6070 Range is between 0 and 1.
6073 Specify which channels to filter, by default all available are filtered.
6076 Normalize biquad coefficients, by default is disabled.
6077 Enabling it will normalize magnitude response at DC to 0dB.
6080 Set transform type of IIR filter.
6089 Set precison of filtering.
6092 Pick automatic sample format depending on surround filters.
6094 Always use signed 16-bit.
6096 Always use signed 32-bit.
6098 Always use float 32-bit.
6100 Always use float 64-bit.
6104 @subsection Commands
6106 This filter supports the following commands:
6109 Change treble frequency.
6110 Syntax for the command is : "@var{frequency}"
6113 Change treble width_type.
6114 Syntax for the command is : "@var{width_type}"
6117 Change treble width.
6118 Syntax for the command is : "@var{width}"
6122 Syntax for the command is : "@var{gain}"
6126 Syntax for the command is : "@var{mix}"
6131 Sinusoidal amplitude modulation.
6133 The filter accepts the following options:
6137 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
6138 (20 Hz or lower) will result in a tremolo effect.
6139 This filter may also be used as a ring modulator by specifying
6140 a modulation frequency higher than 20 Hz.
6141 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6144 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6145 Default value is 0.5.
6150 Sinusoidal phase modulation.
6152 The filter accepts the following options:
6156 Modulation frequency in Hertz.
6157 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6160 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6161 Default value is 0.5.
6166 Adjust the input audio volume.
6168 It accepts the following parameters:
6172 Set audio volume expression.
6174 Output values are clipped to the maximum value.
6176 The output audio volume is given by the relation:
6178 @var{output_volume} = @var{volume} * @var{input_volume}
6181 The default value for @var{volume} is "1.0".
6184 This parameter represents the mathematical precision.
6186 It determines which input sample formats will be allowed, which affects the
6187 precision of the volume scaling.
6191 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
6193 32-bit floating-point; this limits input sample format to FLT. (default)
6195 64-bit floating-point; this limits input sample format to DBL.
6199 Choose the behaviour on encountering ReplayGain side data in input frames.
6203 Remove ReplayGain side data, ignoring its contents (the default).
6206 Ignore ReplayGain side data, but leave it in the frame.
6209 Prefer the track gain, if present.
6212 Prefer the album gain, if present.
6215 @item replaygain_preamp
6216 Pre-amplification gain in dB to apply to the selected replaygain gain.
6218 Default value for @var{replaygain_preamp} is 0.0.
6220 @item replaygain_noclip
6221 Prevent clipping by limiting the gain applied.
6223 Default value for @var{replaygain_noclip} is 1.
6226 Set when the volume expression is evaluated.
6228 It accepts the following values:
6231 only evaluate expression once during the filter initialization, or
6232 when the @samp{volume} command is sent
6235 evaluate expression for each incoming frame
6238 Default value is @samp{once}.
6241 The volume expression can contain the following parameters.
6245 frame number (starting at zero)
6248 @item nb_consumed_samples
6249 number of samples consumed by the filter
6251 number of samples in the current frame
6253 original frame position in the file
6259 PTS at start of stream
6261 time at start of stream
6267 last set volume value
6270 Note that when @option{eval} is set to @samp{once} only the
6271 @var{sample_rate} and @var{tb} variables are available, all other
6272 variables will evaluate to NAN.
6274 @subsection Commands
6276 This filter supports the following commands:
6279 Modify the volume expression.
6280 The command accepts the same syntax of the corresponding option.
6282 If the specified expression is not valid, it is kept at its current
6286 @subsection Examples
6290 Halve the input audio volume:
6294 volume=volume=-6.0206dB
6297 In all the above example the named key for @option{volume} can be
6298 omitted, for example like in:
6304 Increase input audio power by 6 decibels using fixed-point precision:
6306 volume=volume=6dB:precision=fixed
6310 Fade volume after time 10 with an annihilation period of 5 seconds:
6312 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
6316 @section volumedetect
6318 Detect the volume of the input video.
6320 The filter has no parameters. The input is not modified. Statistics about
6321 the volume will be printed in the log when the input stream end is reached.
6323 In particular it will show the mean volume (root mean square), maximum
6324 volume (on a per-sample basis), and the beginning of a histogram of the
6325 registered volume values (from the maximum value to a cumulated 1/1000 of
6328 All volumes are in decibels relative to the maximum PCM value.
6330 @subsection Examples
6332 Here is an excerpt of the output:
6334 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
6335 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
6336 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
6337 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
6338 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
6339 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
6340 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
6341 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
6342 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
6348 The mean square energy is approximately -27 dB, or 10^-2.7.
6350 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
6352 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
6355 In other words, raising the volume by +4 dB does not cause any clipping,
6356 raising it by +5 dB causes clipping for 6 samples, etc.
6358 @c man end AUDIO FILTERS
6360 @chapter Audio Sources
6361 @c man begin AUDIO SOURCES
6363 Below is a description of the currently available audio sources.
6367 Buffer audio frames, and make them available to the filter chain.
6369 This source is mainly intended for a programmatic use, in particular
6370 through the interface defined in @file{libavfilter/buffersrc.h}.
6372 It accepts the following parameters:
6376 The timebase which will be used for timestamps of submitted frames. It must be
6377 either a floating-point number or in @var{numerator}/@var{denominator} form.
6380 The sample rate of the incoming audio buffers.
6383 The sample format of the incoming audio buffers.
6384 Either a sample format name or its corresponding integer representation from
6385 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
6387 @item channel_layout
6388 The channel layout of the incoming audio buffers.
6389 Either a channel layout name from channel_layout_map in
6390 @file{libavutil/channel_layout.c} or its corresponding integer representation
6391 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
6394 The number of channels of the incoming audio buffers.
6395 If both @var{channels} and @var{channel_layout} are specified, then they
6400 @subsection Examples
6403 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
6406 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
6407 Since the sample format with name "s16p" corresponds to the number
6408 6 and the "stereo" channel layout corresponds to the value 0x3, this is
6411 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
6416 Generate an audio signal specified by an expression.
6418 This source accepts in input one or more expressions (one for each
6419 channel), which are evaluated and used to generate a corresponding
6422 This source accepts the following options:
6426 Set the '|'-separated expressions list for each separate channel. In case the
6427 @option{channel_layout} option is not specified, the selected channel layout
6428 depends on the number of provided expressions. Otherwise the last
6429 specified expression is applied to the remaining output channels.
6431 @item channel_layout, c
6432 Set the channel layout. The number of channels in the specified layout
6433 must be equal to the number of specified expressions.
6436 Set the minimum duration of the sourced audio. See
6437 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6438 for the accepted syntax.
6439 Note that the resulting duration may be greater than the specified
6440 duration, as the generated audio is always cut at the end of a
6443 If not specified, or the expressed duration is negative, the audio is
6444 supposed to be generated forever.
6447 Set the number of samples per channel per each output frame,
6450 @item sample_rate, s
6451 Specify the sample rate, default to 44100.
6454 Each expression in @var{exprs} can contain the following constants:
6458 number of the evaluated sample, starting from 0
6461 time of the evaluated sample expressed in seconds, starting from 0
6468 @subsection Examples
6478 Generate a sin signal with frequency of 440 Hz, set sample rate to
6481 aevalsrc="sin(440*2*PI*t):s=8000"
6485 Generate a two channels signal, specify the channel layout (Front
6486 Center + Back Center) explicitly:
6488 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
6492 Generate white noise:
6494 aevalsrc="-2+random(0)"
6498 Generate an amplitude modulated signal:
6500 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
6504 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
6506 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
6513 Generate a FIR coefficients using frequency sampling method.
6515 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6517 The filter accepts the following options:
6521 Set number of filter coefficents in output audio stream.
6522 Default value is 1025.
6525 Set frequency points from where magnitude and phase are set.
6526 This must be in non decreasing order, and first element must be 0, while last element
6527 must be 1. Elements are separated by white spaces.
6530 Set magnitude value for every frequency point set by @option{frequency}.
6531 Number of values must be same as number of frequency points.
6532 Values are separated by white spaces.
6535 Set phase value for every frequency point set by @option{frequency}.
6536 Number of values must be same as number of frequency points.
6537 Values are separated by white spaces.
6539 @item sample_rate, r
6540 Set sample rate, default is 44100.
6543 Set number of samples per each frame. Default is 1024.
6546 Set window function. Default is blackman.
6551 The null audio source, return unprocessed audio frames. It is mainly useful
6552 as a template and to be employed in analysis / debugging tools, or as
6553 the source for filters which ignore the input data (for example the sox
6556 This source accepts the following options:
6560 @item channel_layout, cl
6562 Specifies the channel layout, and can be either an integer or a string
6563 representing a channel layout. The default value of @var{channel_layout}
6566 Check the channel_layout_map definition in
6567 @file{libavutil/channel_layout.c} for the mapping between strings and
6568 channel layout values.
6570 @item sample_rate, r
6571 Specifies the sample rate, and defaults to 44100.
6574 Set the number of samples per requested frames.
6577 Set the duration of the sourced audio. See
6578 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6579 for the accepted syntax.
6581 If not specified, or the expressed duration is negative, the audio is
6582 supposed to be generated forever.
6585 @subsection Examples
6589 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
6591 anullsrc=r=48000:cl=4
6595 Do the same operation with a more obvious syntax:
6597 anullsrc=r=48000:cl=mono
6601 All the parameters need to be explicitly defined.
6605 Synthesize a voice utterance using the libflite library.
6607 To enable compilation of this filter you need to configure FFmpeg with
6608 @code{--enable-libflite}.
6610 Note that versions of the flite library prior to 2.0 are not thread-safe.
6612 The filter accepts the following options:
6617 If set to 1, list the names of the available voices and exit
6618 immediately. Default value is 0.
6621 Set the maximum number of samples per frame. Default value is 512.
6624 Set the filename containing the text to speak.
6627 Set the text to speak.
6630 Set the voice to use for the speech synthesis. Default value is
6631 @code{kal}. See also the @var{list_voices} option.
6634 @subsection Examples
6638 Read from file @file{speech.txt}, and synthesize the text using the
6639 standard flite voice:
6641 flite=textfile=speech.txt
6645 Read the specified text selecting the @code{slt} voice:
6647 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6651 Input text to ffmpeg:
6653 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6657 Make @file{ffplay} speak the specified text, using @code{flite} and
6658 the @code{lavfi} device:
6660 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
6664 For more information about libflite, check:
6665 @url{http://www.festvox.org/flite/}
6669 Generate a noise audio signal.
6671 The filter accepts the following options:
6674 @item sample_rate, r
6675 Specify the sample rate. Default value is 48000 Hz.
6678 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
6682 Specify the duration of the generated audio stream. Not specifying this option
6683 results in noise with an infinite length.
6685 @item color, colour, c
6686 Specify the color of noise. Available noise colors are white, pink, brown,
6687 blue, violet and velvet. Default color is white.
6690 Specify a value used to seed the PRNG.
6693 Set the number of samples per each output frame, default is 1024.
6696 @subsection Examples
6701 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6703 anoisesrc=d=60:c=pink:r=44100:a=0.5
6709 Generate odd-tap Hilbert transform FIR coefficients.
6711 The resulting stream can be used with @ref{afir} filter for phase-shifting
6712 the signal by 90 degrees.
6714 This is used in many matrix coding schemes and for analytic signal generation.
6715 The process is often written as a multiplication by i (or j), the imaginary unit.
6717 The filter accepts the following options:
6721 @item sample_rate, s
6722 Set sample rate, default is 44100.
6725 Set length of FIR filter, default is 22051.
6728 Set number of samples per each frame.
6731 Set window function to be used when generating FIR coefficients.
6736 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6738 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6740 The filter accepts the following options:
6743 @item sample_rate, r
6744 Set sample rate, default is 44100.
6747 Set number of samples per each frame. Default is 1024.
6750 Set high-pass frequency. Default is 0.
6753 Set low-pass frequency. Default is 0.
6754 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6755 is higher than 0 then filter will create band-pass filter coefficients,
6756 otherwise band-reject filter coefficients.
6759 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6762 Set Kaiser window beta.
6765 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6768 Enable rounding, by default is disabled.
6771 Set number of taps for high-pass filter.
6774 Set number of taps for low-pass filter.
6779 Generate an audio signal made of a sine wave with amplitude 1/8.
6781 The audio signal is bit-exact.
6783 The filter accepts the following options:
6788 Set the carrier frequency. Default is 440 Hz.
6790 @item beep_factor, b
6791 Enable a periodic beep every second with frequency @var{beep_factor} times
6792 the carrier frequency. Default is 0, meaning the beep is disabled.
6794 @item sample_rate, r
6795 Specify the sample rate, default is 44100.
6798 Specify the duration of the generated audio stream.
6800 @item samples_per_frame
6801 Set the number of samples per output frame.
6803 The expression can contain the following constants:
6807 The (sequential) number of the output audio frame, starting from 0.
6810 The PTS (Presentation TimeStamp) of the output audio frame,
6811 expressed in @var{TB} units.
6814 The PTS of the output audio frame, expressed in seconds.
6817 The timebase of the output audio frames.
6820 Default is @code{1024}.
6823 @subsection Examples
6828 Generate a simple 440 Hz sine wave:
6834 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6838 sine=frequency=220:beep_factor=4:duration=5
6842 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6845 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6849 @c man end AUDIO SOURCES
6851 @chapter Audio Sinks
6852 @c man begin AUDIO SINKS
6854 Below is a description of the currently available audio sinks.
6856 @section abuffersink
6858 Buffer audio frames, and make them available to the end of filter chain.
6860 This sink is mainly intended for programmatic use, in particular
6861 through the interface defined in @file{libavfilter/buffersink.h}
6862 or the options system.
6864 It accepts a pointer to an AVABufferSinkContext structure, which
6865 defines the incoming buffers' formats, to be passed as the opaque
6866 parameter to @code{avfilter_init_filter} for initialization.
6869 Null audio sink; do absolutely nothing with the input audio. It is
6870 mainly useful as a template and for use in analysis / debugging
6873 @c man end AUDIO SINKS
6875 @chapter Video Filters
6876 @c man begin VIDEO FILTERS
6878 When you configure your FFmpeg build, you can disable any of the
6879 existing filters using @code{--disable-filters}.
6880 The configure output will show the video filters included in your
6883 Below is a description of the currently available video filters.
6887 Mark a region of interest in a video frame.
6889 The frame data is passed through unchanged, but metadata is attached
6890 to the frame indicating regions of interest which can affect the
6891 behaviour of later encoding. Multiple regions can be marked by
6892 applying the filter multiple times.
6896 Region distance in pixels from the left edge of the frame.
6898 Region distance in pixels from the top edge of the frame.
6900 Region width in pixels.
6902 Region height in pixels.
6904 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6905 and may contain the following variables:
6908 Width of the input frame.
6910 Height of the input frame.
6914 Quantisation offset to apply within the region.
6916 This must be a real value in the range -1 to +1. A value of zero
6917 indicates no quality change. A negative value asks for better quality
6918 (less quantisation), while a positive value asks for worse quality
6919 (greater quantisation).
6921 The range is calibrated so that the extreme values indicate the
6922 largest possible offset - if the rest of the frame is encoded with the
6923 worst possible quality, an offset of -1 indicates that this region
6924 should be encoded with the best possible quality anyway. Intermediate
6925 values are then interpolated in some codec-dependent way.
6927 For example, in 10-bit H.264 the quantisation parameter varies between
6928 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6929 this region should be encoded with a QP around one-tenth of the full
6930 range better than the rest of the frame. So, if most of the frame
6931 were to be encoded with a QP of around 30, this region would get a QP
6932 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6933 An extreme value of -1 would indicate that this region should be
6934 encoded with the best possible quality regardless of the treatment of
6935 the rest of the frame - that is, should be encoded at a QP of -12.
6937 If set to true, remove any existing regions of interest marked on the
6938 frame before adding the new one.
6941 @subsection Examples
6945 Mark the centre quarter of the frame as interesting.
6947 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6950 Mark the 100-pixel-wide region on the left edge of the frame as very
6951 uninteresting (to be encoded at much lower quality than the rest of
6954 addroi=0:0:100:ih:+1/5
6958 @section alphaextract
6960 Extract the alpha component from the input as a grayscale video. This
6961 is especially useful with the @var{alphamerge} filter.
6965 Add or replace the alpha component of the primary input with the
6966 grayscale value of a second input. This is intended for use with
6967 @var{alphaextract} to allow the transmission or storage of frame
6968 sequences that have alpha in a format that doesn't support an alpha
6971 For example, to reconstruct full frames from a normal YUV-encoded video
6972 and a separate video created with @var{alphaextract}, you might use:
6974 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6979 Amplify differences between current pixel and pixels of adjacent frames in
6980 same pixel location.
6982 This filter accepts the following options:
6986 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6987 For example radius of 3 will instruct filter to calculate average of 7 frames.
6990 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6993 Set threshold for difference amplification. Any difference greater or equal to
6994 this value will not alter source pixel. Default is 10.
6995 Allowed range is from 0 to 65535.
6998 Set tolerance for difference amplification. Any difference lower to
6999 this value will not alter source pixel. Default is 0.
7000 Allowed range is from 0 to 65535.
7003 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
7004 This option controls maximum possible value that will decrease source pixel value.
7007 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
7008 This option controls maximum possible value that will increase source pixel value.
7011 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
7014 @subsection Commands
7016 This filter supports the following @ref{commands} that corresponds to option of same name:
7028 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
7029 and libavformat to work. On the other hand, it is limited to ASS (Advanced
7030 Substation Alpha) subtitles files.
7032 This filter accepts the following option in addition to the common options from
7033 the @ref{subtitles} filter:
7037 Set the shaping engine
7039 Available values are:
7042 The default libass shaping engine, which is the best available.
7044 Fast, font-agnostic shaper that can do only substitutions
7046 Slower shaper using OpenType for substitutions and positioning
7049 The default is @code{auto}.
7053 Apply an Adaptive Temporal Averaging Denoiser to the video input.
7055 The filter accepts the following options:
7059 Set threshold A for 1st plane. Default is 0.02.
7060 Valid range is 0 to 0.3.
7063 Set threshold B for 1st plane. Default is 0.04.
7064 Valid range is 0 to 5.
7067 Set threshold A for 2nd plane. Default is 0.02.
7068 Valid range is 0 to 0.3.
7071 Set threshold B for 2nd plane. Default is 0.04.
7072 Valid range is 0 to 5.
7075 Set threshold A for 3rd plane. Default is 0.02.
7076 Valid range is 0 to 0.3.
7079 Set threshold B for 3rd plane. Default is 0.04.
7080 Valid range is 0 to 5.
7082 Threshold A is designed to react on abrupt changes in the input signal and
7083 threshold B is designed to react on continuous changes in the input signal.
7086 Set number of frames filter will use for averaging. Default is 9. Must be odd
7087 number in range [5, 129].
7090 Set what planes of frame filter will use for averaging. Default is all.
7093 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
7094 Alternatively can be set to @code{s} serial.
7096 Parallel can be faster then serial, while other way around is never true.
7097 Parallel will abort early on first change being greater then thresholds, while serial
7098 will continue processing other side of frames if they are equal or below thresholds.
7103 Set sigma for 1st plane, 2nd plane or 3rd plane. Default is 32767.
7104 Valid range is from 0 to 32767.
7105 This options controls weight for each pixel in radius defined by size.
7106 Default value means every pixel have same weight.
7107 Setting this option to 0 effectively disables filtering.
7110 @subsection Commands
7111 This filter supports same @ref{commands} as options except option @code{s}.
7112 The command accepts the same syntax of the corresponding option.
7116 Apply average blur filter.
7118 The filter accepts the following options:
7122 Set horizontal radius size.
7125 Set which planes to filter. By default all planes are filtered.
7128 Set vertical radius size, if zero it will be same as @code{sizeX}.
7129 Default is @code{0}.
7132 @subsection Commands
7133 This filter supports same commands as options.
7134 The command accepts the same syntax of the corresponding option.
7136 If the specified expression is not valid, it is kept at its current
7141 Compute the bounding box for the non-black pixels in the input frame
7144 This filter computes the bounding box containing all the pixels with a
7145 luminance value greater than the minimum allowed value.
7146 The parameters describing the bounding box are printed on the filter
7149 The filter accepts the following option:
7153 Set the minimal luminance value. Default is @code{16}.
7156 @subsection Commands
7158 This filter supports the all above options as @ref{commands}.
7161 Apply bilateral filter, spatial smoothing while preserving edges.
7163 The filter accepts the following options:
7166 Set sigma of gaussian function to calculate spatial weight.
7167 Allowed range is 0 to 512. Default is 0.1.
7170 Set sigma of gaussian function to calculate range weight.
7171 Allowed range is 0 to 1. Default is 0.1.
7174 Set planes to filter. Default is first only.
7177 @subsection Commands
7179 This filter supports the all above options as @ref{commands}.
7181 @section bitplanenoise
7183 Show and measure bit plane noise.
7185 The filter accepts the following options:
7189 Set which plane to analyze. Default is @code{1}.
7192 Filter out noisy pixels from @code{bitplane} set above.
7193 Default is disabled.
7196 @section blackdetect
7198 Detect video intervals that are (almost) completely black. Can be
7199 useful to detect chapter transitions, commercials, or invalid
7202 The filter outputs its detection analysis to both the log as well as
7203 frame metadata. If a black segment of at least the specified minimum
7204 duration is found, a line with the start and end timestamps as well
7205 as duration is printed to the log with level @code{info}. In addition,
7206 a log line with level @code{debug} is printed per frame showing the
7207 black amount detected for that frame.
7209 The filter also attaches metadata to the first frame of a black
7210 segment with key @code{lavfi.black_start} and to the first frame
7211 after the black segment ends with key @code{lavfi.black_end}. The
7212 value is the frame's timestamp. This metadata is added regardless
7213 of the minimum duration specified.
7215 The filter accepts the following options:
7218 @item black_min_duration, d
7219 Set the minimum detected black duration expressed in seconds. It must
7220 be a non-negative floating point number.
7222 Default value is 2.0.
7224 @item picture_black_ratio_th, pic_th
7225 Set the threshold for considering a picture "black".
7226 Express the minimum value for the ratio:
7228 @var{nb_black_pixels} / @var{nb_pixels}
7231 for which a picture is considered black.
7232 Default value is 0.98.
7234 @item pixel_black_th, pix_th
7235 Set the threshold for considering a pixel "black".
7237 The threshold expresses the maximum pixel luminance value for which a
7238 pixel is considered "black". The provided value is scaled according to
7239 the following equation:
7241 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
7244 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
7245 the input video format, the range is [0-255] for YUV full-range
7246 formats and [16-235] for YUV non full-range formats.
7248 Default value is 0.10.
7251 The following example sets the maximum pixel threshold to the minimum
7252 value, and detects only black intervals of 2 or more seconds:
7254 blackdetect=d=2:pix_th=0.00
7259 Detect frames that are (almost) completely black. Can be useful to
7260 detect chapter transitions or commercials. Output lines consist of
7261 the frame number of the detected frame, the percentage of blackness,
7262 the position in the file if known or -1 and the timestamp in seconds.
7264 In order to display the output lines, you need to set the loglevel at
7265 least to the AV_LOG_INFO value.
7267 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
7268 The value represents the percentage of pixels in the picture that
7269 are below the threshold value.
7271 It accepts the following parameters:
7276 The percentage of the pixels that have to be below the threshold; it defaults to
7279 @item threshold, thresh
7280 The threshold below which a pixel value is considered black; it defaults to
7288 Blend two video frames into each other.
7290 The @code{blend} filter takes two input streams and outputs one
7291 stream, the first input is the "top" layer and second input is
7292 "bottom" layer. By default, the output terminates when the longest input terminates.
7294 The @code{tblend} (time blend) filter takes two consecutive frames
7295 from one single stream, and outputs the result obtained by blending
7296 the new frame on top of the old frame.
7298 A description of the accepted options follows.
7306 Set blend mode for specific pixel component or all pixel components in case
7307 of @var{all_mode}. Default value is @code{normal}.
7309 Available values for component modes are:
7351 Set blend opacity for specific pixel component or all pixel components in case
7352 of @var{all_opacity}. Only used in combination with pixel component blend modes.
7359 Set blend expression for specific pixel component or all pixel components in case
7360 of @var{all_expr}. Note that related mode options will be ignored if those are set.
7362 The expressions can use the following variables:
7366 The sequential number of the filtered frame, starting from @code{0}.
7370 the coordinates of the current sample
7374 the width and height of currently filtered plane
7378 Width and height scale for the plane being filtered. It is the
7379 ratio between the dimensions of the current plane to the luma plane,
7380 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
7381 the luma plane and @code{0.5,0.5} for the chroma planes.
7384 Time of the current frame, expressed in seconds.
7387 Value of pixel component at current location for first video frame (top layer).
7390 Value of pixel component at current location for second video frame (bottom layer).
7394 The @code{blend} filter also supports the @ref{framesync} options.
7396 @subsection Examples
7400 Apply transition from bottom layer to top layer in first 10 seconds:
7402 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
7406 Apply linear horizontal transition from top layer to bottom layer:
7408 blend=all_expr='A*(X/W)+B*(1-X/W)'
7412 Apply 1x1 checkerboard effect:
7414 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
7418 Apply uncover left effect:
7420 blend=all_expr='if(gte(N*SW+X,W),A,B)'
7424 Apply uncover down effect:
7426 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
7430 Apply uncover up-left effect:
7432 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
7436 Split diagonally video and shows top and bottom layer on each side:
7438 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
7442 Display differences between the current and the previous frame:
7444 tblend=all_mode=grainextract
7450 Denoise frames using Block-Matching 3D algorithm.
7452 The filter accepts the following options.
7456 Set denoising strength. Default value is 1.
7457 Allowed range is from 0 to 999.9.
7458 The denoising algorithm is very sensitive to sigma, so adjust it
7459 according to the source.
7462 Set local patch size. This sets dimensions in 2D.
7465 Set sliding step for processing blocks. Default value is 4.
7466 Allowed range is from 1 to 64.
7467 Smaller values allows processing more reference blocks and is slower.
7470 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
7471 When set to 1, no block matching is done. Larger values allows more blocks
7473 Allowed range is from 1 to 256.
7476 Set radius for search block matching. Default is 9.
7477 Allowed range is from 1 to INT32_MAX.
7480 Set step between two search locations for block matching. Default is 1.
7481 Allowed range is from 1 to 64. Smaller is slower.
7484 Set threshold of mean square error for block matching. Valid range is 0 to
7488 Set thresholding parameter for hard thresholding in 3D transformed domain.
7489 Larger values results in stronger hard-thresholding filtering in frequency
7493 Set filtering estimation mode. Can be @code{basic} or @code{final}.
7494 Default is @code{basic}.
7497 If enabled, filter will use 2nd stream for block matching.
7498 Default is disabled for @code{basic} value of @var{estim} option,
7499 and always enabled if value of @var{estim} is @code{final}.
7502 Set planes to filter. Default is all available except alpha.
7505 @subsection Examples
7509 Basic filtering with bm3d:
7511 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
7515 Same as above, but filtering only luma:
7517 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
7521 Same as above, but with both estimation modes:
7523 split[a][b],[a]bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic[a],[b][a]bm3d=sigma=3:block=4:bstep=2:group=16:estim=final:ref=1
7527 Same as above, but prefilter with @ref{nlmeans} filter instead:
7529 split[a][b],[a]nlmeans=s=3:r=7:p=3[a],[b][a]bm3d=sigma=3:block=4:bstep=2:group=16:estim=final:ref=1
7535 Apply a boxblur algorithm to the input video.
7537 It accepts the following parameters:
7541 @item luma_radius, lr
7542 @item luma_power, lp
7543 @item chroma_radius, cr
7544 @item chroma_power, cp
7545 @item alpha_radius, ar
7546 @item alpha_power, ap
7550 A description of the accepted options follows.
7553 @item luma_radius, lr
7554 @item chroma_radius, cr
7555 @item alpha_radius, ar
7556 Set an expression for the box radius in pixels used for blurring the
7557 corresponding input plane.
7559 The radius value must be a non-negative number, and must not be
7560 greater than the value of the expression @code{min(w,h)/2} for the
7561 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
7564 Default value for @option{luma_radius} is "2". If not specified,
7565 @option{chroma_radius} and @option{alpha_radius} default to the
7566 corresponding value set for @option{luma_radius}.
7568 The expressions can contain the following constants:
7572 The input width and height in pixels.
7576 The input chroma image width and height in pixels.
7580 The horizontal and vertical chroma subsample values. For example, for the
7581 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
7584 @item luma_power, lp
7585 @item chroma_power, cp
7586 @item alpha_power, ap
7587 Specify how many times the boxblur filter is applied to the
7588 corresponding plane.
7590 Default value for @option{luma_power} is 2. If not specified,
7591 @option{chroma_power} and @option{alpha_power} default to the
7592 corresponding value set for @option{luma_power}.
7594 A value of 0 will disable the effect.
7597 @subsection Examples
7601 Apply a boxblur filter with the luma, chroma, and alpha radii
7604 boxblur=luma_radius=2:luma_power=1
7609 Set the luma radius to 2, and alpha and chroma radius to 0:
7611 boxblur=2:1:cr=0:ar=0
7615 Set the luma and chroma radii to a fraction of the video dimension:
7617 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
7623 Deinterlace the input video ("bwdif" stands for "Bob Weaver
7624 Deinterlacing Filter").
7626 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
7627 interpolation algorithms.
7628 It accepts the following parameters:
7632 The interlacing mode to adopt. It accepts one of the following values:
7636 Output one frame for each frame.
7638 Output one frame for each field.
7641 The default value is @code{send_field}.
7644 The picture field parity assumed for the input interlaced video. It accepts one
7645 of the following values:
7649 Assume the top field is first.
7651 Assume the bottom field is first.
7653 Enable automatic detection of field parity.
7656 The default value is @code{auto}.
7657 If the interlacing is unknown or the decoder does not export this information,
7658 top field first will be assumed.
7661 Specify which frames to deinterlace. Accepts one of the following
7666 Deinterlace all frames.
7668 Only deinterlace frames marked as interlaced.
7671 The default value is @code{all}.
7676 Apply Contrast Adaptive Sharpen filter to video stream.
7678 The filter accepts the following options:
7682 Set the sharpening strength. Default value is 0.
7685 Set planes to filter. Default value is to filter all
7686 planes except alpha plane.
7689 @subsection Commands
7690 This filter supports same @ref{commands} as options.
7693 Remove all color information for all colors except for certain one.
7695 The filter accepts the following options:
7699 The color which will not be replaced with neutral chroma.
7702 Similarity percentage with the above color.
7703 0.01 matches only the exact key color, while 1.0 matches everything.
7707 0.0 makes pixels either fully gray, or not gray at all.
7708 Higher values result in more preserved color.
7711 Signals that the color passed is already in YUV instead of RGB.
7713 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7714 This can be used to pass exact YUV values as hexadecimal numbers.
7717 @subsection Commands
7718 This filter supports same @ref{commands} as options.
7719 The command accepts the same syntax of the corresponding option.
7721 If the specified expression is not valid, it is kept at its current
7725 YUV colorspace color/chroma keying.
7727 The filter accepts the following options:
7731 The color which will be replaced with transparency.
7734 Similarity percentage with the key color.
7736 0.01 matches only the exact key color, while 1.0 matches everything.
7741 0.0 makes pixels either fully transparent, or not transparent at all.
7743 Higher values result in semi-transparent pixels, with a higher transparency
7744 the more similar the pixels color is to the key color.
7747 Signals that the color passed is already in YUV instead of RGB.
7749 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7750 This can be used to pass exact YUV values as hexadecimal numbers.
7753 @subsection Commands
7754 This filter supports same @ref{commands} as options.
7755 The command accepts the same syntax of the corresponding option.
7757 If the specified expression is not valid, it is kept at its current
7760 @subsection Examples
7764 Make every green pixel in the input image transparent:
7766 ffmpeg -i input.png -vf chromakey=green out.png
7770 Overlay a greenscreen-video on top of a static black background.
7772 ffmpeg -f lavfi -i color=c=black:s=1280x720 -i video.mp4 -shortest -filter_complex "[1:v]chromakey=0x70de77:0.1:0.2[ckout];[0:v][ckout]overlay[out]" -map "[out]" output.mkv
7777 Reduce chrominance noise.
7779 The filter accepts the following options:
7783 Set threshold for averaging chrominance values.
7784 Sum of absolute difference of Y, U and V pixel components of current
7785 pixel and neighbour pixels lower than this threshold will be used in
7786 averaging. Luma component is left unchanged and is copied to output.
7787 Default value is 30. Allowed range is from 1 to 200.
7790 Set horizontal radius of rectangle used for averaging.
7791 Allowed range is from 1 to 100. Default value is 5.
7794 Set vertical radius of rectangle used for averaging.
7795 Allowed range is from 1 to 100. Default value is 5.
7798 Set horizontal step when averaging. Default value is 1.
7799 Allowed range is from 1 to 50.
7800 Mostly useful to speed-up filtering.
7803 Set vertical step when averaging. Default value is 1.
7804 Allowed range is from 1 to 50.
7805 Mostly useful to speed-up filtering.
7808 Set Y threshold for averaging chrominance values.
7809 Set finer control for max allowed difference between Y components
7810 of current pixel and neigbour pixels.
7811 Default value is 200. Allowed range is from 1 to 200.
7814 Set U threshold for averaging chrominance values.
7815 Set finer control for max allowed difference between U components
7816 of current pixel and neigbour pixels.
7817 Default value is 200. Allowed range is from 1 to 200.
7820 Set V threshold for averaging chrominance values.
7821 Set finer control for max allowed difference between V components
7822 of current pixel and neigbour pixels.
7823 Default value is 200. Allowed range is from 1 to 200.
7826 @subsection Commands
7827 This filter supports same @ref{commands} as options.
7828 The command accepts the same syntax of the corresponding option.
7830 @section chromashift
7831 Shift chroma pixels horizontally and/or vertically.
7833 The filter accepts the following options:
7836 Set amount to shift chroma-blue horizontally.
7838 Set amount to shift chroma-blue vertically.
7840 Set amount to shift chroma-red horizontally.
7842 Set amount to shift chroma-red vertically.
7844 Set edge mode, can be @var{smear}, default, or @var{warp}.
7847 @subsection Commands
7849 This filter supports the all above options as @ref{commands}.
7853 Display CIE color diagram with pixels overlaid onto it.
7855 The filter accepts the following options:
7870 @item uhdtv, rec2020
7884 Set what gamuts to draw.
7886 See @code{system} option for available values.
7889 Set ciescope size, by default set to 512.
7892 Set intensity used to map input pixel values to CIE diagram.
7895 Set contrast used to draw tongue colors that are out of active color system gamut.
7898 Correct gamma displayed on scope, by default enabled.
7901 Show white point on CIE diagram, by default disabled.
7904 Set input gamma. Used only with XYZ input color space.
7909 Visualize information exported by some codecs.
7911 Some codecs can export information through frames using side-data or other
7912 means. For example, some MPEG based codecs export motion vectors through the
7913 @var{export_mvs} flag in the codec @option{flags2} option.
7915 The filter accepts the following option:
7919 Set motion vectors to visualize.
7921 Available flags for @var{mv} are:
7925 forward predicted MVs of P-frames
7927 forward predicted MVs of B-frames
7929 backward predicted MVs of B-frames
7933 Display quantization parameters using the chroma planes.
7936 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7938 Available flags for @var{mv_type} are:
7942 forward predicted MVs
7944 backward predicted MVs
7947 @item frame_type, ft
7948 Set frame type to visualize motion vectors of.
7950 Available flags for @var{frame_type} are:
7954 intra-coded frames (I-frames)
7956 predicted frames (P-frames)
7958 bi-directionally predicted frames (B-frames)
7962 @subsection Examples
7966 Visualize forward predicted MVs of all frames using @command{ffplay}:
7968 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7972 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7974 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7978 @section colorbalance
7979 Modify intensity of primary colors (red, green and blue) of input frames.
7981 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7982 regions for the red-cyan, green-magenta or blue-yellow balance.
7984 A positive adjustment value shifts the balance towards the primary color, a negative
7985 value towards the complementary color.
7987 The filter accepts the following options:
7993 Adjust red, green and blue shadows (darkest pixels).
7998 Adjust red, green and blue midtones (medium pixels).
8003 Adjust red, green and blue highlights (brightest pixels).
8005 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
8008 Preserve lightness when changing color balance. Default is disabled.
8011 @subsection Examples
8015 Add red color cast to shadows:
8021 @subsection Commands
8023 This filter supports the all above options as @ref{commands}.
8025 @section colorchannelmixer
8027 Adjust video input frames by re-mixing color channels.
8029 This filter modifies a color channel by adding the values associated to
8030 the other channels of the same pixels. For example if the value to
8031 modify is red, the output value will be:
8033 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
8036 The filter accepts the following options:
8043 Adjust contribution of input red, green, blue and alpha channels for output red channel.
8044 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
8050 Adjust contribution of input red, green, blue and alpha channels for output green channel.
8051 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
8057 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
8058 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
8064 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
8065 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
8067 Allowed ranges for options are @code{[-2.0, 2.0]}.
8070 @subsection Examples
8074 Convert source to grayscale:
8076 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
8079 Simulate sepia tones:
8081 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
8085 @subsection Commands
8087 This filter supports the all above options as @ref{commands}.
8090 RGB colorspace color keying.
8092 The filter accepts the following options:
8096 The color which will be replaced with transparency.
8099 Similarity percentage with the key color.
8101 0.01 matches only the exact key color, while 1.0 matches everything.
8106 0.0 makes pixels either fully transparent, or not transparent at all.
8108 Higher values result in semi-transparent pixels, with a higher transparency
8109 the more similar the pixels color is to the key color.
8112 @subsection Examples
8116 Make every green pixel in the input image transparent:
8118 ffmpeg -i input.png -vf colorkey=green out.png
8122 Overlay a greenscreen-video on top of a static background image.
8124 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
8128 @subsection Commands
8129 This filter supports same @ref{commands} as options.
8130 The command accepts the same syntax of the corresponding option.
8132 If the specified expression is not valid, it is kept at its current
8136 Remove all color information for all RGB colors except for certain one.
8138 The filter accepts the following options:
8142 The color which will not be replaced with neutral gray.
8145 Similarity percentage with the above color.
8146 0.01 matches only the exact key color, while 1.0 matches everything.
8149 Blend percentage. 0.0 makes pixels fully gray.
8150 Higher values result in more preserved color.
8153 @subsection Commands
8154 This filter supports same @ref{commands} as options.
8155 The command accepts the same syntax of the corresponding option.
8157 If the specified expression is not valid, it is kept at its current
8160 @section colorlevels
8162 Adjust video input frames using levels.
8164 The filter accepts the following options:
8171 Adjust red, green, blue and alpha input black point.
8172 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
8178 Adjust red, green, blue and alpha input white point.
8179 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
8181 Input levels are used to lighten highlights (bright tones), darken shadows
8182 (dark tones), change the balance of bright and dark tones.
8188 Adjust red, green, blue and alpha output black point.
8189 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
8195 Adjust red, green, blue and alpha output white point.
8196 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
8198 Output levels allows manual selection of a constrained output level range.
8201 @subsection Examples
8205 Make video output darker:
8207 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
8213 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
8217 Make video output lighter:
8219 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
8223 Increase brightness:
8225 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
8229 @subsection Commands
8231 This filter supports the all above options as @ref{commands}.
8233 @section colormatrix
8235 Convert color matrix.
8237 The filter accepts the following options:
8242 Specify the source and destination color matrix. Both values must be
8245 The accepted values are:
8273 For example to convert from BT.601 to SMPTE-240M, use the command:
8275 colormatrix=bt601:smpte240m
8280 Convert colorspace, transfer characteristics or color primaries.
8281 Input video needs to have an even size.
8283 The filter accepts the following options:
8288 Specify all color properties at once.
8290 The accepted values are:
8320 Specify output colorspace.
8322 The accepted values are:
8331 BT.470BG or BT.601-6 625
8334 SMPTE-170M or BT.601-6 525
8343 BT.2020 with non-constant luminance
8349 Specify output transfer characteristics.
8351 The accepted values are:
8363 Constant gamma of 2.2
8366 Constant gamma of 2.8
8369 SMPTE-170M, BT.601-6 625 or BT.601-6 525
8387 BT.2020 for 10-bits content
8390 BT.2020 for 12-bits content
8396 Specify output color primaries.
8398 The accepted values are:
8407 BT.470BG or BT.601-6 625
8410 SMPTE-170M or BT.601-6 525
8434 Specify output color range.
8436 The accepted values are:
8439 TV (restricted) range
8442 MPEG (restricted) range
8453 Specify output color format.
8455 The accepted values are:
8458 YUV 4:2:0 planar 8-bits
8461 YUV 4:2:0 planar 10-bits
8464 YUV 4:2:0 planar 12-bits
8467 YUV 4:2:2 planar 8-bits
8470 YUV 4:2:2 planar 10-bits
8473 YUV 4:2:2 planar 12-bits
8476 YUV 4:4:4 planar 8-bits
8479 YUV 4:4:4 planar 10-bits
8482 YUV 4:4:4 planar 12-bits
8487 Do a fast conversion, which skips gamma/primary correction. This will take
8488 significantly less CPU, but will be mathematically incorrect. To get output
8489 compatible with that produced by the colormatrix filter, use fast=1.
8492 Specify dithering mode.
8494 The accepted values are:
8500 Floyd-Steinberg dithering
8504 Whitepoint adaptation mode.
8506 The accepted values are:
8509 Bradford whitepoint adaptation
8512 von Kries whitepoint adaptation
8515 identity whitepoint adaptation (i.e. no whitepoint adaptation)
8519 Override all input properties at once. Same accepted values as @ref{all}.
8522 Override input colorspace. Same accepted values as @ref{space}.
8525 Override input color primaries. Same accepted values as @ref{primaries}.
8528 Override input transfer characteristics. Same accepted values as @ref{trc}.
8531 Override input color range. Same accepted values as @ref{range}.
8535 The filter converts the transfer characteristics, color space and color
8536 primaries to the specified user values. The output value, if not specified,
8537 is set to a default value based on the "all" property. If that property is
8538 also not specified, the filter will log an error. The output color range and
8539 format default to the same value as the input color range and format. The
8540 input transfer characteristics, color space, color primaries and color range
8541 should be set on the input data. If any of these are missing, the filter will
8542 log an error and no conversion will take place.
8544 For example to convert the input to SMPTE-240M, use the command:
8546 colorspace=smpte240m
8549 @section convolution
8551 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
8553 The filter accepts the following options:
8560 Set matrix for each plane.
8561 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
8562 and from 1 to 49 odd number of signed integers in @var{row} mode.
8568 Set multiplier for calculated value for each plane.
8569 If unset or 0, it will be sum of all matrix elements.
8575 Set bias for each plane. This value is added to the result of the multiplication.
8576 Useful for making the overall image brighter or darker. Default is 0.0.
8582 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
8583 Default is @var{square}.
8586 @subsection Commands
8588 This filter supports the all above options as @ref{commands}.
8590 @subsection Examples
8596 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"
8602 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"
8608 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"
8614 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"
8618 Apply laplacian edge detector which includes diagonals:
8620 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"
8626 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"
8632 Apply 2D convolution of video stream in frequency domain using second stream
8635 The filter accepts the following options:
8639 Set which planes to process.
8642 Set which impulse video frames will be processed, can be @var{first}
8643 or @var{all}. Default is @var{all}.
8646 The @code{convolve} filter also supports the @ref{framesync} options.
8650 Copy the input video source unchanged to the output. This is mainly useful for
8655 Video filtering on GPU using Apple's CoreImage API on OSX.
8657 Hardware acceleration is based on an OpenGL context. Usually, this means it is
8658 processed by video hardware. However, software-based OpenGL implementations
8659 exist which means there is no guarantee for hardware processing. It depends on
8662 There are many filters and image generators provided by Apple that come with a
8663 large variety of options. The filter has to be referenced by its name along
8666 The coreimage filter accepts the following options:
8669 List all available filters and generators along with all their respective
8670 options as well as possible minimum and maximum values along with the default
8677 Specify all filters by their respective name and options.
8678 Use @var{list_filters} to determine all valid filter names and options.
8679 Numerical options are specified by a float value and are automatically clamped
8680 to their respective value range. Vector and color options have to be specified
8681 by a list of space separated float values. Character escaping has to be done.
8682 A special option name @code{default} is available to use default options for a
8685 It is required to specify either @code{default} or at least one of the filter options.
8686 All omitted options are used with their default values.
8687 The syntax of the filter string is as follows:
8689 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
8693 Specify a rectangle where the output of the filter chain is copied into the
8694 input image. It is given by a list of space separated float values:
8696 output_rect=x\ y\ width\ height
8698 If not given, the output rectangle equals the dimensions of the input image.
8699 The output rectangle is automatically cropped at the borders of the input
8700 image. Negative values are valid for each component.
8702 output_rect=25\ 25\ 100\ 100
8706 Several filters can be chained for successive processing without GPU-HOST
8707 transfers allowing for fast processing of complex filter chains.
8708 Currently, only filters with zero (generators) or exactly one (filters) input
8709 image and one output image are supported. Also, transition filters are not yet
8712 Some filters generate output images with additional padding depending on the
8713 respective filter kernel. The padding is automatically removed to ensure the
8714 filter output has the same size as the input image.
8716 For image generators, the size of the output image is determined by the
8717 previous output image of the filter chain or the input image of the whole
8718 filterchain, respectively. The generators do not use the pixel information of
8719 this image to generate their output. However, the generated output is
8720 blended onto this image, resulting in partial or complete coverage of the
8723 The @ref{coreimagesrc} video source can be used for generating input images
8724 which are directly fed into the filter chain. By using it, providing input
8725 images by another video source or an input video is not required.
8727 @subsection Examples
8732 List all filters available:
8734 coreimage=list_filters=true
8738 Use the CIBoxBlur filter with default options to blur an image:
8740 coreimage=filter=CIBoxBlur@@default
8744 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8745 its center at 100x100 and a radius of 50 pixels:
8747 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8751 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8752 given as complete and escaped command-line for Apple's standard bash shell:
8754 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8760 Cover a rectangular object
8762 It accepts the following options:
8766 Filepath of the optional cover image, needs to be in yuv420.
8771 It accepts the following values:
8774 cover it by the supplied image
8776 cover it by interpolating the surrounding pixels
8779 Default value is @var{blur}.
8782 @subsection Examples
8786 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8788 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8794 Crop the input video to given dimensions.
8796 It accepts the following parameters:
8800 The width of the output video. It defaults to @code{iw}.
8801 This expression is evaluated only once during the filter
8802 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8805 The height of the output video. It defaults to @code{ih}.
8806 This expression is evaluated only once during the filter
8807 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8810 The horizontal position, in the input video, of the left edge of the output
8811 video. It defaults to @code{(in_w-out_w)/2}.
8812 This expression is evaluated per-frame.
8815 The vertical position, in the input video, of the top edge of the output video.
8816 It defaults to @code{(in_h-out_h)/2}.
8817 This expression is evaluated per-frame.
8820 If set to 1 will force the output display aspect ratio
8821 to be the same of the input, by changing the output sample aspect
8822 ratio. It defaults to 0.
8825 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8826 width/height/x/y as specified and will not be rounded to nearest smaller value.
8830 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8831 expressions containing the following constants:
8836 The computed values for @var{x} and @var{y}. They are evaluated for
8841 The input width and height.
8845 These are the same as @var{in_w} and @var{in_h}.
8849 The output (cropped) width and height.
8853 These are the same as @var{out_w} and @var{out_h}.
8856 same as @var{iw} / @var{ih}
8859 input sample aspect ratio
8862 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8866 horizontal and vertical chroma subsample values. For example for the
8867 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8870 The number of the input frame, starting from 0.
8873 the position in the file of the input frame, NAN if unknown
8876 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8880 The expression for @var{out_w} may depend on the value of @var{out_h},
8881 and the expression for @var{out_h} may depend on @var{out_w}, but they
8882 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8883 evaluated after @var{out_w} and @var{out_h}.
8885 The @var{x} and @var{y} parameters specify the expressions for the
8886 position of the top-left corner of the output (non-cropped) area. They
8887 are evaluated for each frame. If the evaluated value is not valid, it
8888 is approximated to the nearest valid value.
8890 The expression for @var{x} may depend on @var{y}, and the expression
8891 for @var{y} may depend on @var{x}.
8893 @subsection Examples
8897 Crop area with size 100x100 at position (12,34).
8902 Using named options, the example above becomes:
8904 crop=w=100:h=100:x=12:y=34
8908 Crop the central input area with size 100x100:
8914 Crop the central input area with size 2/3 of the input video:
8916 crop=2/3*in_w:2/3*in_h
8920 Crop the input video central square:
8927 Delimit the rectangle with the top-left corner placed at position
8928 100:100 and the right-bottom corner corresponding to the right-bottom
8929 corner of the input image.
8931 crop=in_w-100:in_h-100:100:100
8935 Crop 10 pixels from the left and right borders, and 20 pixels from
8936 the top and bottom borders
8938 crop=in_w-2*10:in_h-2*20
8942 Keep only the bottom right quarter of the input image:
8944 crop=in_w/2:in_h/2:in_w/2:in_h/2
8948 Crop height for getting Greek harmony:
8950 crop=in_w:1/PHI*in_w
8954 Apply trembling effect:
8956 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)
8960 Apply erratic camera effect depending on timestamp:
8962 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)"
8966 Set x depending on the value of y:
8968 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8972 @subsection Commands
8974 This filter supports the following commands:
8980 Set width/height of the output video and the horizontal/vertical position
8982 The command accepts the same syntax of the corresponding option.
8984 If the specified expression is not valid, it is kept at its current
8990 Auto-detect the crop size.
8992 It calculates the necessary cropping parameters and prints the
8993 recommended parameters via the logging system. The detected dimensions
8994 correspond to the non-black area of the input video.
8996 It accepts the following parameters:
9001 Set higher black value threshold, which can be optionally specified
9002 from nothing (0) to everything (255 for 8-bit based formats). An intensity
9003 value greater to the set value is considered non-black. It defaults to 24.
9004 You can also specify a value between 0.0 and 1.0 which will be scaled depending
9005 on the bitdepth of the pixel format.
9008 The value which the width/height should be divisible by. It defaults to
9009 16. The offset is automatically adjusted to center the video. Use 2 to
9010 get only even dimensions (needed for 4:2:2 video). 16 is best when
9011 encoding to most video codecs.
9014 Set the number of initial frames for which evaluation is skipped.
9015 Default is 2. Range is 0 to INT_MAX.
9017 @item reset_count, reset
9018 Set the counter that determines after how many frames cropdetect will
9019 reset the previously detected largest video area and start over to
9020 detect the current optimal crop area. Default value is 0.
9022 This can be useful when channel logos distort the video area. 0
9023 indicates 'never reset', and returns the largest area encountered during
9030 Delay video filtering until a given wallclock timestamp. The filter first
9031 passes on @option{preroll} amount of frames, then it buffers at most
9032 @option{buffer} amount of frames and waits for the cue. After reaching the cue
9033 it forwards the buffered frames and also any subsequent frames coming in its
9036 The filter can be used synchronize the output of multiple ffmpeg processes for
9037 realtime output devices like decklink. By putting the delay in the filtering
9038 chain and pre-buffering frames the process can pass on data to output almost
9039 immediately after the target wallclock timestamp is reached.
9041 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
9047 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
9050 The duration of content to pass on as preroll expressed in seconds. Default is 0.
9053 The maximum duration of content to buffer before waiting for the cue expressed
9054 in seconds. Default is 0.
9061 Apply color adjustments using curves.
9063 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
9064 component (red, green and blue) has its values defined by @var{N} key points
9065 tied from each other using a smooth curve. The x-axis represents the pixel
9066 values from the input frame, and the y-axis the new pixel values to be set for
9069 By default, a component curve is defined by the two points @var{(0;0)} and
9070 @var{(1;1)}. This creates a straight line where each original pixel value is
9071 "adjusted" to its own value, which means no change to the image.
9073 The filter allows you to redefine these two points and add some more. A new
9074 curve (using a natural cubic spline interpolation) will be define to pass
9075 smoothly through all these new coordinates. The new defined points needs to be
9076 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
9077 be in the @var{[0;1]} interval. If the computed curves happened to go outside
9078 the vector spaces, the values will be clipped accordingly.
9080 The filter accepts the following options:
9084 Select one of the available color presets. This option can be used in addition
9085 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
9086 options takes priority on the preset values.
9087 Available presets are:
9090 @item color_negative
9093 @item increase_contrast
9095 @item linear_contrast
9096 @item medium_contrast
9098 @item strong_contrast
9101 Default is @code{none}.
9103 Set the master key points. These points will define a second pass mapping. It
9104 is sometimes called a "luminance" or "value" mapping. It can be used with
9105 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
9106 post-processing LUT.
9108 Set the key points for the red component.
9110 Set the key points for the green component.
9112 Set the key points for the blue component.
9114 Set the key points for all components (not including master).
9115 Can be used in addition to the other key points component
9116 options. In this case, the unset component(s) will fallback on this
9117 @option{all} setting.
9119 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
9121 Save Gnuplot script of the curves in specified file.
9124 To avoid some filtergraph syntax conflicts, each key points list need to be
9125 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
9127 @subsection Examples
9131 Increase slightly the middle level of blue:
9133 curves=blue='0/0 0.5/0.58 1/1'
9139 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'
9141 Here we obtain the following coordinates for each components:
9144 @code{(0;0.11) (0.42;0.51) (1;0.95)}
9146 @code{(0;0) (0.50;0.48) (1;1)}
9148 @code{(0;0.22) (0.49;0.44) (1;0.80)}
9152 The previous example can also be achieved with the associated built-in preset:
9154 curves=preset=vintage
9164 Use a Photoshop preset and redefine the points of the green component:
9166 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
9170 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
9171 and @command{gnuplot}:
9173 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
9174 gnuplot -p /tmp/curves.plt
9180 Video data analysis filter.
9182 This filter shows hexadecimal pixel values of part of video.
9184 The filter accepts the following options:
9188 Set output video size.
9191 Set x offset from where to pick pixels.
9194 Set y offset from where to pick pixels.
9197 Set scope mode, can be one of the following:
9200 Draw hexadecimal pixel values with white color on black background.
9203 Draw hexadecimal pixel values with input video pixel color on black
9207 Draw hexadecimal pixel values on color background picked from input video,
9208 the text color is picked in such way so its always visible.
9212 Draw rows and columns numbers on left and top of video.
9215 Set background opacity.
9218 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
9221 Set pixel components to display. By default all pixel components are displayed.
9225 Apply Directional blur filter.
9227 The filter accepts the following options:
9231 Set angle of directional blur. Default is @code{45}.
9234 Set radius of directional blur. Default is @code{5}.
9237 Set which planes to filter. By default all planes are filtered.
9240 @subsection Commands
9241 This filter supports same @ref{commands} as options.
9242 The command accepts the same syntax of the corresponding option.
9244 If the specified expression is not valid, it is kept at its current
9249 Denoise frames using 2D DCT (frequency domain filtering).
9251 This filter is not designed for real time.
9253 The filter accepts the following options:
9257 Set the noise sigma constant.
9259 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
9260 coefficient (absolute value) below this threshold with be dropped.
9262 If you need a more advanced filtering, see @option{expr}.
9264 Default is @code{0}.
9267 Set number overlapping pixels for each block. Since the filter can be slow, you
9268 may want to reduce this value, at the cost of a less effective filter and the
9269 risk of various artefacts.
9271 If the overlapping value doesn't permit processing the whole input width or
9272 height, a warning will be displayed and according borders won't be denoised.
9274 Default value is @var{blocksize}-1, which is the best possible setting.
9277 Set the coefficient factor expression.
9279 For each coefficient of a DCT block, this expression will be evaluated as a
9280 multiplier value for the coefficient.
9282 If this is option is set, the @option{sigma} option will be ignored.
9284 The absolute value of the coefficient can be accessed through the @var{c}
9288 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
9289 @var{blocksize}, which is the width and height of the processed blocks.
9291 The default value is @var{3} (8x8) and can be raised to @var{4} for a
9292 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
9293 on the speed processing. Also, a larger block size does not necessarily means a
9297 @subsection Examples
9299 Apply a denoise with a @option{sigma} of @code{4.5}:
9304 The same operation can be achieved using the expression system:
9306 dctdnoiz=e='gte(c, 4.5*3)'
9309 Violent denoise using a block size of @code{16x16}:
9316 Remove banding artifacts from input video.
9317 It works by replacing banded pixels with average value of referenced pixels.
9319 The filter accepts the following options:
9326 Set banding detection threshold for each plane. Default is 0.02.
9327 Valid range is 0.00003 to 0.5.
9328 If difference between current pixel and reference pixel is less than threshold,
9329 it will be considered as banded.
9332 Banding detection range in pixels. Default is 16. If positive, random number
9333 in range 0 to set value will be used. If negative, exact absolute value
9335 The range defines square of four pixels around current pixel.
9338 Set direction in radians from which four pixel will be compared. If positive,
9339 random direction from 0 to set direction will be picked. If negative, exact of
9340 absolute value will be picked. For example direction 0, -PI or -2*PI radians
9341 will pick only pixels on same row and -PI/2 will pick only pixels on same
9345 If enabled, current pixel is compared with average value of all four
9346 surrounding pixels. The default is enabled. If disabled current pixel is
9347 compared with all four surrounding pixels. The pixel is considered banded
9348 if only all four differences with surrounding pixels are less than threshold.
9351 If enabled, current pixel is changed if and only if all pixel components are banded,
9352 e.g. banding detection threshold is triggered for all color components.
9353 The default is disabled.
9358 Remove blocking artifacts from input video.
9360 The filter accepts the following options:
9364 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
9365 This controls what kind of deblocking is applied.
9368 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
9374 Set blocking detection thresholds. Allowed range is 0 to 1.
9375 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
9376 Using higher threshold gives more deblocking strength.
9377 Setting @var{alpha} controls threshold detection at exact edge of block.
9378 Remaining options controls threshold detection near the edge. Each one for
9379 below/above or left/right. Setting any of those to @var{0} disables
9383 Set planes to filter. Default is to filter all available planes.
9386 @subsection Examples
9390 Deblock using weak filter and block size of 4 pixels.
9392 deblock=filter=weak:block=4
9396 Deblock using strong filter, block size of 4 pixels and custom thresholds for
9397 deblocking more edges.
9399 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
9403 Similar as above, but filter only first plane.
9405 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
9409 Similar as above, but filter only second and third plane.
9411 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
9418 Drop duplicated frames at regular intervals.
9420 The filter accepts the following options:
9424 Set the number of frames from which one will be dropped. Setting this to
9425 @var{N} means one frame in every batch of @var{N} frames will be dropped.
9426 Default is @code{5}.
9429 Set the threshold for duplicate detection. If the difference metric for a frame
9430 is less than or equal to this value, then it is declared as duplicate. Default
9434 Set scene change threshold. Default is @code{15}.
9438 Set the size of the x and y-axis blocks used during metric calculations.
9439 Larger blocks give better noise suppression, but also give worse detection of
9440 small movements. Must be a power of two. Default is @code{32}.
9443 Mark main input as a pre-processed input and activate clean source input
9444 stream. This allows the input to be pre-processed with various filters to help
9445 the metrics calculation while keeping the frame selection lossless. When set to
9446 @code{1}, the first stream is for the pre-processed input, and the second
9447 stream is the clean source from where the kept frames are chosen. Default is
9451 Set whether or not chroma is considered in the metric calculations. Default is
9457 Apply 2D deconvolution of video stream in frequency domain using second stream
9460 The filter accepts the following options:
9464 Set which planes to process.
9467 Set which impulse video frames will be processed, can be @var{first}
9468 or @var{all}. Default is @var{all}.
9471 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
9472 and height are not same and not power of 2 or if stream prior to convolving
9476 The @code{deconvolve} filter also supports the @ref{framesync} options.
9480 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
9482 It accepts the following options:
9486 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
9487 @var{rainbows} for cross-color reduction.
9490 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
9493 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
9496 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
9499 Set temporal chroma threshold. Lower values increases reduction of cross-color.
9504 Apply deflate effect to the video.
9506 This filter replaces the pixel by the local(3x3) average by taking into account
9507 only values lower than the pixel.
9509 It accepts the following options:
9516 Limit the maximum change for each plane, default is 65535.
9517 If 0, plane will remain unchanged.
9520 @subsection Commands
9522 This filter supports the all above options as @ref{commands}.
9526 Remove temporal frame luminance variations.
9528 It accepts the following options:
9532 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
9535 Set averaging mode to smooth temporal luminance variations.
9537 Available values are:
9562 Do not actually modify frame. Useful when one only wants metadata.
9567 Remove judder produced by partially interlaced telecined content.
9569 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
9570 source was partially telecined content then the output of @code{pullup,dejudder}
9571 will have a variable frame rate. May change the recorded frame rate of the
9572 container. Aside from that change, this filter will not affect constant frame
9575 The option available in this filter is:
9579 Specify the length of the window over which the judder repeats.
9581 Accepts any integer greater than 1. Useful values are:
9585 If the original was telecined from 24 to 30 fps (Film to NTSC).
9588 If the original was telecined from 25 to 30 fps (PAL to NTSC).
9591 If a mixture of the two.
9594 The default is @samp{4}.
9599 Suppress a TV station logo by a simple interpolation of the surrounding
9600 pixels. Just set a rectangle covering the logo and watch it disappear
9601 (and sometimes something even uglier appear - your mileage may vary).
9603 It accepts the following parameters:
9608 Specify the top left corner coordinates of the logo. They must be
9613 Specify the width and height of the logo to clear. They must be
9617 Specify the thickness of the fuzzy edge of the rectangle (added to
9618 @var{w} and @var{h}). The default value is 1. This option is
9619 deprecated, setting higher values should no longer be necessary and
9623 When set to 1, a green rectangle is drawn on the screen to simplify
9624 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
9625 The default value is 0.
9627 The rectangle is drawn on the outermost pixels which will be (partly)
9628 replaced with interpolated values. The values of the next pixels
9629 immediately outside this rectangle in each direction will be used to
9630 compute the interpolated pixel values inside the rectangle.
9634 @subsection Examples
9638 Set a rectangle covering the area with top left corner coordinates 0,0
9639 and size 100x77, and a band of size 10:
9641 delogo=x=0:y=0:w=100:h=77:band=10
9649 Remove the rain in the input image/video by applying the derain methods based on
9650 convolutional neural networks. Supported models:
9654 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
9655 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
9658 Training as well as model generation scripts are provided in
9659 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
9661 Native model files (.model) can be generated from TensorFlow model
9662 files (.pb) by using tools/python/convert.py
9664 The filter accepts the following options:
9668 Specify which filter to use. This option accepts the following values:
9672 Derain filter. To conduct derain filter, you need to use a derain model.
9675 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
9677 Default value is @samp{derain}.
9680 Specify which DNN backend to use for model loading and execution. This option accepts
9681 the following values:
9685 Native implementation of DNN loading and execution.
9688 TensorFlow backend. To enable this backend you
9689 need to install the TensorFlow for C library (see
9690 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9691 @code{--enable-libtensorflow}
9693 Default value is @samp{native}.
9696 Set path to model file specifying network architecture and its parameters.
9697 Note that different backends use different file formats. TensorFlow and native
9698 backend can load files for only its format.
9701 It can also be finished with @ref{dnn_processing} filter.
9705 Attempt to fix small changes in horizontal and/or vertical shift. This
9706 filter helps remove camera shake from hand-holding a camera, bumping a
9707 tripod, moving on a vehicle, etc.
9709 The filter accepts the following options:
9717 Specify a rectangular area where to limit the search for motion
9719 If desired the search for motion vectors can be limited to a
9720 rectangular area of the frame defined by its top left corner, width
9721 and height. These parameters have the same meaning as the drawbox
9722 filter which can be used to visualise the position of the bounding
9725 This is useful when simultaneous movement of subjects within the frame
9726 might be confused for camera motion by the motion vector search.
9728 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
9729 then the full frame is used. This allows later options to be set
9730 without specifying the bounding box for the motion vector search.
9732 Default - search the whole frame.
9736 Specify the maximum extent of movement in x and y directions in the
9737 range 0-64 pixels. Default 16.
9740 Specify how to generate pixels to fill blanks at the edge of the
9741 frame. Available values are:
9744 Fill zeroes at blank locations
9746 Original image at blank locations
9748 Extruded edge value at blank locations
9750 Mirrored edge at blank locations
9752 Default value is @samp{mirror}.
9755 Specify the blocksize to use for motion search. Range 4-128 pixels,
9759 Specify the contrast threshold for blocks. Only blocks with more than
9760 the specified contrast (difference between darkest and lightest
9761 pixels) will be considered. Range 1-255, default 125.
9764 Specify the search strategy. Available values are:
9767 Set exhaustive search
9769 Set less exhaustive search.
9771 Default value is @samp{exhaustive}.
9774 If set then a detailed log of the motion search is written to the
9781 Remove unwanted contamination of foreground colors, caused by reflected color of
9782 greenscreen or bluescreen.
9784 This filter accepts the following options:
9788 Set what type of despill to use.
9791 Set how spillmap will be generated.
9794 Set how much to get rid of still remaining spill.
9797 Controls amount of red in spill area.
9800 Controls amount of green in spill area.
9801 Should be -1 for greenscreen.
9804 Controls amount of blue in spill area.
9805 Should be -1 for bluescreen.
9808 Controls brightness of spill area, preserving colors.
9811 Modify alpha from generated spillmap.
9814 @subsection Commands
9816 This filter supports the all above options as @ref{commands}.
9820 Apply an exact inverse of the telecine operation. It requires a predefined
9821 pattern specified using the pattern option which must be the same as that passed
9822 to the telecine filter.
9824 This filter accepts the following options:
9833 The default value is @code{top}.
9837 A string of numbers representing the pulldown pattern you wish to apply.
9838 The default value is @code{23}.
9841 A number representing position of the first frame with respect to the telecine
9842 pattern. This is to be used if the stream is cut. The default value is @code{0}.
9847 Apply dilation effect to the video.
9849 This filter replaces the pixel by the local(3x3) maximum.
9851 It accepts the following options:
9858 Limit the maximum change for each plane, default is 65535.
9859 If 0, plane will remain unchanged.
9862 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9865 Flags to local 3x3 coordinates maps like this:
9872 @subsection Commands
9874 This filter supports the all above options as @ref{commands}.
9878 Displace pixels as indicated by second and third input stream.
9880 It takes three input streams and outputs one stream, the first input is the
9881 source, and second and third input are displacement maps.
9883 The second input specifies how much to displace pixels along the
9884 x-axis, while the third input specifies how much to displace pixels
9886 If one of displacement map streams terminates, last frame from that
9887 displacement map will be used.
9889 Note that once generated, displacements maps can be reused over and over again.
9891 A description of the accepted options follows.
9895 Set displace behavior for pixels that are out of range.
9897 Available values are:
9900 Missing pixels are replaced by black pixels.
9903 Adjacent pixels will spread out to replace missing pixels.
9906 Out of range pixels are wrapped so they point to pixels of other side.
9909 Out of range pixels will be replaced with mirrored pixels.
9911 Default is @samp{smear}.
9915 @subsection Examples
9919 Add ripple effect to rgb input of video size hd720:
9921 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
9925 Add wave effect to rgb input of video size hd720:
9927 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
9931 @anchor{dnn_processing}
9932 @section dnn_processing
9934 Do image processing with deep neural networks. It works together with another filter
9935 which converts the pixel format of the Frame to what the dnn network requires.
9937 The filter accepts the following options:
9941 Specify which DNN backend to use for model loading and execution. This option accepts
9942 the following values:
9946 Native implementation of DNN loading and execution.
9949 TensorFlow backend. To enable this backend you
9950 need to install the TensorFlow for C library (see
9951 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9952 @code{--enable-libtensorflow}
9955 OpenVINO backend. To enable this backend you
9956 need to build and install the OpenVINO for C library (see
9957 @url{https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md}) and configure FFmpeg with
9958 @code{--enable-libopenvino} (--extra-cflags=-I... --extra-ldflags=-L... might
9959 be needed if the header files and libraries are not installed into system path)
9963 Default value is @samp{native}.
9966 Set path to model file specifying network architecture and its parameters.
9967 Note that different backends use different file formats. TensorFlow, OpenVINO and native
9968 backend can load files for only its format.
9970 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9973 Set the input name of the dnn network.
9976 Set the output name of the dnn network.
9979 use DNN async execution if set (default: set),
9980 roll back to sync execution if the backend does not support async.
9984 @subsection Examples
9988 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
9990 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
9994 Halve the pixel value of the frame with format gray32f:
9996 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
10000 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
10002 ./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
10006 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
10008 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
10015 Draw a colored box on the input image.
10017 It accepts the following parameters:
10022 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
10026 The expressions which specify the width and height of the box; if 0 they are interpreted as
10027 the input width and height. It defaults to 0.
10030 Specify the color of the box to write. For the general syntax of this option,
10031 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10032 value @code{invert} is used, the box edge color is the same as the
10033 video with inverted luma.
10036 The expression which sets the thickness of the box edge.
10037 A value of @code{fill} will create a filled box. Default value is @code{3}.
10039 See below for the list of accepted constants.
10042 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
10043 will overwrite the video's color and alpha pixels.
10044 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
10047 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10048 following constants:
10052 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10056 horizontal and vertical chroma subsample values. For example for the
10057 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10061 The input width and height.
10064 The input sample aspect ratio.
10068 The x and y offset coordinates where the box is drawn.
10072 The width and height of the drawn box.
10075 The thickness of the drawn box.
10077 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10078 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10082 @subsection Examples
10086 Draw a black box around the edge of the input image:
10092 Draw a box with color red and an opacity of 50%:
10094 drawbox=10:20:200:60:red@@0.5
10097 The previous example can be specified as:
10099 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
10103 Fill the box with pink color:
10105 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
10109 Draw a 2-pixel red 2.40:1 mask:
10111 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
10115 @subsection Commands
10116 This filter supports same commands as options.
10117 The command accepts the same syntax of the corresponding option.
10119 If the specified expression is not valid, it is kept at its current
10124 Draw a graph using input video metadata.
10126 It accepts the following parameters:
10130 Set 1st frame metadata key from which metadata values will be used to draw a graph.
10133 Set 1st foreground color expression.
10136 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
10139 Set 2nd foreground color expression.
10142 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
10145 Set 3rd foreground color expression.
10148 Set 4th frame metadata key from which metadata values will be used to draw a graph.
10151 Set 4th foreground color expression.
10154 Set minimal value of metadata value.
10157 Set maximal value of metadata value.
10160 Set graph background color. Default is white.
10165 Available values for mode is:
10172 Default is @code{line}.
10177 Available values for slide is:
10180 Draw new frame when right border is reached.
10183 Replace old columns with new ones.
10186 Scroll from right to left.
10189 Scroll from left to right.
10192 Draw single picture.
10195 Default is @code{frame}.
10198 Set size of graph video. For the syntax of this option, check the
10199 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
10200 The default value is @code{900x256}.
10203 Set the output frame rate. Default value is @code{25}.
10205 The foreground color expressions can use the following variables:
10208 Minimal value of metadata value.
10211 Maximal value of metadata value.
10214 Current metadata key value.
10217 The color is defined as 0xAABBGGRR.
10220 Example using metadata from @ref{signalstats} filter:
10222 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
10225 Example using metadata from @ref{ebur128} filter:
10227 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
10232 Draw a grid on the input image.
10234 It accepts the following parameters:
10239 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
10243 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
10244 input width and height, respectively, minus @code{thickness}, so image gets
10245 framed. Default to 0.
10248 Specify the color of the grid. For the general syntax of this option,
10249 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10250 value @code{invert} is used, the grid color is the same as the
10251 video with inverted luma.
10254 The expression which sets the thickness of the grid line. Default value is @code{1}.
10256 See below for the list of accepted constants.
10259 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
10260 will overwrite the video's color and alpha pixels.
10261 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
10264 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10265 following constants:
10269 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10273 horizontal and vertical chroma subsample values. For example for the
10274 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10278 The input grid cell width and height.
10281 The input sample aspect ratio.
10285 The x and y coordinates of some point of grid intersection (meant to configure offset).
10289 The width and height of the drawn cell.
10292 The thickness of the drawn cell.
10294 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10295 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10299 @subsection Examples
10303 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
10305 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
10309 Draw a white 3x3 grid with an opacity of 50%:
10311 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
10315 @subsection Commands
10316 This filter supports same commands as options.
10317 The command accepts the same syntax of the corresponding option.
10319 If the specified expression is not valid, it is kept at its current
10325 Draw a text string or text from a specified file on top of a video, using the
10326 libfreetype library.
10328 To enable compilation of this filter, you need to configure FFmpeg with
10329 @code{--enable-libfreetype}.
10330 To enable default font fallback and the @var{font} option you need to
10331 configure FFmpeg with @code{--enable-libfontconfig}.
10332 To enable the @var{text_shaping} option, you need to configure FFmpeg with
10333 @code{--enable-libfribidi}.
10337 It accepts the following parameters:
10342 Used to draw a box around text using the background color.
10343 The value must be either 1 (enable) or 0 (disable).
10344 The default value of @var{box} is 0.
10347 Set the width of the border to be drawn around the box using @var{boxcolor}.
10348 The default value of @var{boxborderw} is 0.
10351 The color to be used for drawing box around text. For the syntax of this
10352 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10354 The default value of @var{boxcolor} is "white".
10357 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
10358 The default value of @var{line_spacing} is 0.
10361 Set the width of the border to be drawn around the text using @var{bordercolor}.
10362 The default value of @var{borderw} is 0.
10365 Set the color to be used for drawing border around text. For the syntax of this
10366 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10368 The default value of @var{bordercolor} is "black".
10371 Select how the @var{text} is expanded. Can be either @code{none},
10372 @code{strftime} (deprecated) or
10373 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
10377 Set a start time for the count. Value is in microseconds. Only applied
10378 in the deprecated strftime expansion mode. To emulate in normal expansion
10379 mode use the @code{pts} function, supplying the start time (in seconds)
10380 as the second argument.
10383 If true, check and fix text coords to avoid clipping.
10386 The color to be used for drawing fonts. For the syntax of this option, check
10387 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10389 The default value of @var{fontcolor} is "black".
10391 @item fontcolor_expr
10392 String which is expanded the same way as @var{text} to obtain dynamic
10393 @var{fontcolor} value. By default this option has empty value and is not
10394 processed. When this option is set, it overrides @var{fontcolor} option.
10397 The font family to be used for drawing text. By default Sans.
10400 The font file to be used for drawing text. The path must be included.
10401 This parameter is mandatory if the fontconfig support is disabled.
10404 Draw the text applying alpha blending. The value can
10405 be a number between 0.0 and 1.0.
10406 The expression accepts the same variables @var{x, y} as well.
10407 The default value is 1.
10408 Please see @var{fontcolor_expr}.
10411 The font size to be used for drawing text.
10412 The default value of @var{fontsize} is 16.
10415 If set to 1, attempt to shape the text (for example, reverse the order of
10416 right-to-left text and join Arabic characters) before drawing it.
10417 Otherwise, just draw the text exactly as given.
10418 By default 1 (if supported).
10420 @item ft_load_flags
10421 The flags to be used for loading the fonts.
10423 The flags map the corresponding flags supported by libfreetype, and are
10424 a combination of the following values:
10431 @item vertical_layout
10432 @item force_autohint
10435 @item ignore_global_advance_width
10437 @item ignore_transform
10439 @item linear_design
10443 Default value is "default".
10445 For more information consult the documentation for the FT_LOAD_*
10449 The color to be used for drawing a shadow behind the drawn text. For the
10450 syntax of this option, check the @ref{color syntax,,"Color" section in the
10451 ffmpeg-utils manual,ffmpeg-utils}.
10453 The default value of @var{shadowcolor} is "black".
10457 The x and y offsets for the text shadow position with respect to the
10458 position of the text. They can be either positive or negative
10459 values. The default value for both is "0".
10462 The starting frame number for the n/frame_num variable. The default value
10466 The size in number of spaces to use for rendering the tab.
10467 Default value is 4.
10470 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
10471 format. It can be used with or without text parameter. @var{timecode_rate}
10472 option must be specified.
10474 @item timecode_rate, rate, r
10475 Set the timecode frame rate (timecode only). Value will be rounded to nearest
10476 integer. Minimum value is "1".
10477 Drop-frame timecode is supported for frame rates 30 & 60.
10480 If set to 1, the output of the timecode option will wrap around at 24 hours.
10481 Default is 0 (disabled).
10484 The text string to be drawn. The text must be a sequence of UTF-8
10485 encoded characters.
10486 This parameter is mandatory if no file is specified with the parameter
10490 A text file containing text to be drawn. The text must be a sequence
10491 of UTF-8 encoded characters.
10493 This parameter is mandatory if no text string is specified with the
10494 parameter @var{text}.
10496 If both @var{text} and @var{textfile} are specified, an error is thrown.
10499 If set to 1, the @var{textfile} will be reloaded before each frame.
10500 Be sure to update it atomically, or it may be read partially, or even fail.
10504 The expressions which specify the offsets where text will be drawn
10505 within the video frame. They are relative to the top/left border of the
10508 The default value of @var{x} and @var{y} is "0".
10510 See below for the list of accepted constants and functions.
10513 The parameters for @var{x} and @var{y} are expressions containing the
10514 following constants and functions:
10518 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
10522 horizontal and vertical chroma subsample values. For example for the
10523 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10526 the height of each text line
10534 @item max_glyph_a, ascent
10535 the maximum distance from the baseline to the highest/upper grid
10536 coordinate used to place a glyph outline point, for all the rendered
10538 It is a positive value, due to the grid's orientation with the Y axis
10541 @item max_glyph_d, descent
10542 the maximum distance from the baseline to the lowest grid coordinate
10543 used to place a glyph outline point, for all the rendered glyphs.
10544 This is a negative value, due to the grid's orientation, with the Y axis
10548 maximum glyph height, that is the maximum height for all the glyphs
10549 contained in the rendered text, it is equivalent to @var{ascent} -
10553 maximum glyph width, that is the maximum width for all the glyphs
10554 contained in the rendered text
10557 the number of input frame, starting from 0
10559 @item rand(min, max)
10560 return a random number included between @var{min} and @var{max}
10563 The input sample aspect ratio.
10566 timestamp expressed in seconds, NAN if the input timestamp is unknown
10569 the height of the rendered text
10572 the width of the rendered text
10576 the x and y offset coordinates where the text is drawn.
10578 These parameters allow the @var{x} and @var{y} expressions to refer
10579 to each other, so you can for example specify @code{y=x/dar}.
10582 A one character description of the current frame's picture type.
10585 The current packet's position in the input file or stream
10586 (in bytes, from the start of the input). A value of -1 indicates
10587 this info is not available.
10590 The current packet's duration, in seconds.
10593 The current packet's size (in bytes).
10596 @anchor{drawtext_expansion}
10597 @subsection Text expansion
10599 If @option{expansion} is set to @code{strftime},
10600 the filter recognizes strftime() sequences in the provided text and
10601 expands them accordingly. Check the documentation of strftime(). This
10602 feature is deprecated.
10604 If @option{expansion} is set to @code{none}, the text is printed verbatim.
10606 If @option{expansion} is set to @code{normal} (which is the default),
10607 the following expansion mechanism is used.
10609 The backslash character @samp{\}, followed by any character, always expands to
10610 the second character.
10612 Sequences of the form @code{%@{...@}} are expanded. The text between the
10613 braces is a function name, possibly followed by arguments separated by ':'.
10614 If the arguments contain special characters or delimiters (':' or '@}'),
10615 they should be escaped.
10617 Note that they probably must also be escaped as the value for the
10618 @option{text} option in the filter argument string and as the filter
10619 argument in the filtergraph description, and possibly also for the shell,
10620 that makes up to four levels of escaping; using a text file avoids these
10623 The following functions are available:
10628 The expression evaluation result.
10630 It must take one argument specifying the expression to be evaluated,
10631 which accepts the same constants and functions as the @var{x} and
10632 @var{y} values. Note that not all constants should be used, for
10633 example the text size is not known when evaluating the expression, so
10634 the constants @var{text_w} and @var{text_h} will have an undefined
10637 @item expr_int_format, eif
10638 Evaluate the expression's value and output as formatted integer.
10640 The first argument is the expression to be evaluated, just as for the @var{expr} function.
10641 The second argument specifies the output format. Allowed values are @samp{x},
10642 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
10643 @code{printf} function.
10644 The third parameter is optional and sets the number of positions taken by the output.
10645 It can be used to add padding with zeros from the left.
10648 The time at which the filter is running, expressed in UTC.
10649 It can accept an argument: a strftime() format string.
10652 The time at which the filter is running, expressed in the local time zone.
10653 It can accept an argument: a strftime() format string.
10656 Frame metadata. Takes one or two arguments.
10658 The first argument is mandatory and specifies the metadata key.
10660 The second argument is optional and specifies a default value, used when the
10661 metadata key is not found or empty.
10663 Available metadata can be identified by inspecting entries
10664 starting with TAG included within each frame section
10665 printed by running @code{ffprobe -show_frames}.
10667 String metadata generated in filters leading to
10668 the drawtext filter are also available.
10671 The frame number, starting from 0.
10674 A one character description of the current picture type.
10677 The timestamp of the current frame.
10678 It can take up to three arguments.
10680 The first argument is the format of the timestamp; it defaults to @code{flt}
10681 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
10682 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
10683 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
10684 @code{localtime} stands for the timestamp of the frame formatted as
10685 local time zone time.
10687 The second argument is an offset added to the timestamp.
10689 If the format is set to @code{hms}, a third argument @code{24HH} may be
10690 supplied to present the hour part of the formatted timestamp in 24h format
10693 If the format is set to @code{localtime} or @code{gmtime},
10694 a third argument may be supplied: a strftime() format string.
10695 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
10698 @subsection Commands
10700 This filter supports altering parameters via commands:
10703 Alter existing filter parameters.
10705 Syntax for the argument is the same as for filter invocation, e.g.
10708 fontsize=56:fontcolor=green:text='Hello World'
10711 Full filter invocation with sendcmd would look like this:
10714 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
10718 If the entire argument can't be parsed or applied as valid values then the filter will
10719 continue with its existing parameters.
10721 @subsection Examples
10725 Draw "Test Text" with font FreeSerif, using the default values for the
10726 optional parameters.
10729 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
10733 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
10734 and y=50 (counting from the top-left corner of the screen), text is
10735 yellow with a red box around it. Both the text and the box have an
10739 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
10740 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
10743 Note that the double quotes are not necessary if spaces are not used
10744 within the parameter list.
10747 Show the text at the center of the video frame:
10749 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
10753 Show the text at a random position, switching to a new position every 30 seconds:
10755 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)"
10759 Show a text line sliding from right to left in the last row of the video
10760 frame. The file @file{LONG_LINE} is assumed to contain a single line
10763 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
10767 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
10769 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
10773 Draw a single green letter "g", at the center of the input video.
10774 The glyph baseline is placed at half screen height.
10776 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
10780 Show text for 1 second every 3 seconds:
10782 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
10786 Use fontconfig to set the font. Note that the colons need to be escaped.
10788 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
10792 Draw "Test Text" with font size dependent on height of the video.
10794 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
10798 Print the date of a real-time encoding (see strftime(3)):
10800 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
10804 Show text fading in and out (appearing/disappearing):
10807 DS=1.0 # display start
10808 DE=10.0 # display end
10809 FID=1.5 # fade in duration
10810 FOD=5 # fade out duration
10811 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 @}"
10815 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
10816 and the @option{fontsize} value are included in the @option{y} offset.
10818 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
10819 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
10823 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
10824 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
10825 must have option @option{-export_path_metadata 1} for the special metadata fields
10826 to be available for filters.
10828 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
10833 For more information about libfreetype, check:
10834 @url{http://www.freetype.org/}.
10836 For more information about fontconfig, check:
10837 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
10839 For more information about libfribidi, check:
10840 @url{http://fribidi.org/}.
10842 @section edgedetect
10844 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
10846 The filter accepts the following options:
10851 Set low and high threshold values used by the Canny thresholding
10854 The high threshold selects the "strong" edge pixels, which are then
10855 connected through 8-connectivity with the "weak" edge pixels selected
10856 by the low threshold.
10858 @var{low} and @var{high} threshold values must be chosen in the range
10859 [0,1], and @var{low} should be lesser or equal to @var{high}.
10861 Default value for @var{low} is @code{20/255}, and default value for @var{high}
10865 Define the drawing mode.
10869 Draw white/gray wires on black background.
10872 Mix the colors to create a paint/cartoon effect.
10875 Apply Canny edge detector on all selected planes.
10877 Default value is @var{wires}.
10880 Select planes for filtering. By default all available planes are filtered.
10883 @subsection Examples
10887 Standard edge detection with custom values for the hysteresis thresholding:
10889 edgedetect=low=0.1:high=0.4
10893 Painting effect without thresholding:
10895 edgedetect=mode=colormix:high=0
10901 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10903 For each input image, the filter will compute the optimal mapping from
10904 the input to the output given the codebook length, that is the number
10905 of distinct output colors.
10907 This filter accepts the following options.
10910 @item codebook_length, l
10911 Set codebook length. The value must be a positive integer, and
10912 represents the number of distinct output colors. Default value is 256.
10915 Set the maximum number of iterations to apply for computing the optimal
10916 mapping. The higher the value the better the result and the higher the
10917 computation time. Default value is 1.
10920 Set a random seed, must be an integer included between 0 and
10921 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10922 will try to use a good random seed on a best effort basis.
10925 Set pal8 output pixel format. This option does not work with codebook
10926 length greater than 256.
10931 Measure graylevel entropy in histogram of color channels of video frames.
10933 It accepts the following parameters:
10937 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10939 @var{diff} mode measures entropy of histogram delta values, absolute differences
10940 between neighbour histogram values.
10944 Set brightness, contrast, saturation and approximate gamma adjustment.
10946 The filter accepts the following options:
10950 Set the contrast expression. The value must be a float value in range
10951 @code{-1000.0} to @code{1000.0}. The default value is "1".
10954 Set the brightness expression. The value must be a float value in
10955 range @code{-1.0} to @code{1.0}. The default value is "0".
10958 Set the saturation expression. The value must be a float in
10959 range @code{0.0} to @code{3.0}. The default value is "1".
10962 Set the gamma expression. The value must be a float in range
10963 @code{0.1} to @code{10.0}. The default value is "1".
10966 Set the gamma expression for red. The value must be a float in
10967 range @code{0.1} to @code{10.0}. The default value is "1".
10970 Set the gamma expression for green. The value must be a float in range
10971 @code{0.1} to @code{10.0}. The default value is "1".
10974 Set the gamma expression for blue. The value must be a float in range
10975 @code{0.1} to @code{10.0}. The default value is "1".
10978 Set the gamma weight expression. It can be used to reduce the effect
10979 of a high gamma value on bright image areas, e.g. keep them from
10980 getting overamplified and just plain white. The value must be a float
10981 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
10982 gamma correction all the way down while @code{1.0} leaves it at its
10983 full strength. Default is "1".
10986 Set when the expressions for brightness, contrast, saturation and
10987 gamma expressions are evaluated.
10989 It accepts the following values:
10992 only evaluate expressions once during the filter initialization or
10993 when a command is processed
10996 evaluate expressions for each incoming frame
10999 Default value is @samp{init}.
11002 The expressions accept the following parameters:
11005 frame count of the input frame starting from 0
11008 byte position of the corresponding packet in the input file, NAN if
11012 frame rate of the input video, NAN if the input frame rate is unknown
11015 timestamp expressed in seconds, NAN if the input timestamp is unknown
11018 @subsection Commands
11019 The filter supports the following commands:
11023 Set the contrast expression.
11026 Set the brightness expression.
11029 Set the saturation expression.
11032 Set the gamma expression.
11035 Set the gamma_r expression.
11038 Set gamma_g expression.
11041 Set gamma_b expression.
11044 Set gamma_weight expression.
11046 The command accepts the same syntax of the corresponding option.
11048 If the specified expression is not valid, it is kept at its current
11055 Apply erosion effect to the video.
11057 This filter replaces the pixel by the local(3x3) minimum.
11059 It accepts the following options:
11066 Limit the maximum change for each plane, default is 65535.
11067 If 0, plane will remain unchanged.
11070 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
11073 Flags to local 3x3 coordinates maps like this:
11080 @subsection Commands
11082 This filter supports the all above options as @ref{commands}.
11086 Deinterlace the input video ("estdif" stands for "Edge Slope
11087 Tracing Deinterlacing Filter").
11089 Spatial only filter that uses edge slope tracing algorithm
11090 to interpolate missing lines.
11091 It accepts the following parameters:
11095 The interlacing mode to adopt. It accepts one of the following values:
11099 Output one frame for each frame.
11101 Output one frame for each field.
11104 The default value is @code{field}.
11107 The picture field parity assumed for the input interlaced video. It accepts one
11108 of the following values:
11112 Assume the top field is first.
11114 Assume the bottom field is first.
11116 Enable automatic detection of field parity.
11119 The default value is @code{auto}.
11120 If the interlacing is unknown or the decoder does not export this information,
11121 top field first will be assumed.
11124 Specify which frames to deinterlace. Accepts one of the following
11129 Deinterlace all frames.
11131 Only deinterlace frames marked as interlaced.
11134 The default value is @code{all}.
11137 Specify the search radius for edge slope tracing. Default value is 1.
11138 Allowed range is from 1 to 15.
11141 Specify the search radius for best edge matching. Default value is 2.
11142 Allowed range is from 0 to 15.
11145 Specify the interpolation used. Default is 4-point interpolation. It accepts one
11146 of the following values:
11150 Two-point interpolation.
11152 Four-point interpolation.
11154 Six-point interpolation.
11158 @subsection Commands
11159 This filter supports same @ref{commands} as options.
11161 @section extractplanes
11163 Extract color channel components from input video stream into
11164 separate grayscale video streams.
11166 The filter accepts the following option:
11170 Set plane(s) to extract.
11172 Available values for planes are:
11183 Choosing planes not available in the input will result in an error.
11184 That means you cannot select @code{r}, @code{g}, @code{b} planes
11185 with @code{y}, @code{u}, @code{v} planes at same time.
11188 @subsection Examples
11192 Extract luma, u and v color channel component from input video frame
11193 into 3 grayscale outputs:
11195 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
11201 Apply a fade-in/out effect to the input video.
11203 It accepts the following parameters:
11207 The effect type can be either "in" for a fade-in, or "out" for a fade-out
11209 Default is @code{in}.
11211 @item start_frame, s
11212 Specify the number of the frame to start applying the fade
11213 effect at. Default is 0.
11216 The number of frames that the fade effect lasts. At the end of the
11217 fade-in effect, the output video will have the same intensity as the input video.
11218 At the end of the fade-out transition, the output video will be filled with the
11219 selected @option{color}.
11223 If set to 1, fade only alpha channel, if one exists on the input.
11224 Default value is 0.
11226 @item start_time, st
11227 Specify the timestamp (in seconds) of the frame to start to apply the fade
11228 effect. If both start_frame and start_time are specified, the fade will start at
11229 whichever comes last. Default is 0.
11232 The number of seconds for which the fade effect has to last. At the end of the
11233 fade-in effect the output video will have the same intensity as the input video,
11234 at the end of the fade-out transition the output video will be filled with the
11235 selected @option{color}.
11236 If both duration and nb_frames are specified, duration is used. Default is 0
11237 (nb_frames is used by default).
11240 Specify the color of the fade. Default is "black".
11243 @subsection Examples
11247 Fade in the first 30 frames of video:
11252 The command above is equivalent to:
11258 Fade out the last 45 frames of a 200-frame video:
11261 fade=type=out:start_frame=155:nb_frames=45
11265 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
11267 fade=in:0:25, fade=out:975:25
11271 Make the first 5 frames yellow, then fade in from frame 5-24:
11273 fade=in:5:20:color=yellow
11277 Fade in alpha over first 25 frames of video:
11279 fade=in:0:25:alpha=1
11283 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
11285 fade=t=in:st=5.5:d=0.5
11291 Denoise frames using 3D FFT (frequency domain filtering).
11293 The filter accepts the following options:
11297 Set the noise sigma constant. This sets denoising strength.
11298 Default value is 1. Allowed range is from 0 to 30.
11299 Using very high sigma with low overlap may give blocking artifacts.
11302 Set amount of denoising. By default all detected noise is reduced.
11303 Default value is 1. Allowed range is from 0 to 1.
11306 Set size of block, Default is 4, can be 3, 4, 5 or 6.
11307 Actual size of block in pixels is 2 to power of @var{block}, so by default
11308 block size in pixels is 2^4 which is 16.
11311 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
11314 Set number of previous frames to use for denoising. By default is set to 0.
11317 Set number of next frames to to use for denoising. By default is set to 0.
11320 Set planes which will be filtered, by default are all available filtered
11325 Apply arbitrary expressions to samples in frequency domain
11329 Adjust the dc value (gain) of the luma plane of the image. The filter
11330 accepts an integer value in range @code{0} to @code{1000}. The default
11331 value is set to @code{0}.
11334 Adjust the dc value (gain) of the 1st chroma plane of the image. The
11335 filter accepts an integer value in range @code{0} to @code{1000}. The
11336 default value is set to @code{0}.
11339 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
11340 filter accepts an integer value in range @code{0} to @code{1000}. The
11341 default value is set to @code{0}.
11344 Set the frequency domain weight expression for the luma plane.
11347 Set the frequency domain weight expression for the 1st chroma plane.
11350 Set the frequency domain weight expression for the 2nd chroma plane.
11353 Set when the expressions are evaluated.
11355 It accepts the following values:
11358 Only evaluate expressions once during the filter initialization.
11361 Evaluate expressions for each incoming frame.
11364 Default value is @samp{init}.
11366 The filter accepts the following variables:
11369 The coordinates of the current sample.
11373 The width and height of the image.
11376 The number of input frame, starting from 0.
11379 @subsection Examples
11385 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
11391 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
11397 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
11403 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
11410 Extract a single field from an interlaced image using stride
11411 arithmetic to avoid wasting CPU time. The output frames are marked as
11414 The filter accepts the following options:
11418 Specify whether to extract the top (if the value is @code{0} or
11419 @code{top}) or the bottom field (if the value is @code{1} or
11425 Create new frames by copying the top and bottom fields from surrounding frames
11426 supplied as numbers by the hint file.
11430 Set file containing hints: absolute/relative frame numbers.
11432 There must be one line for each frame in a clip. Each line must contain two
11433 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
11434 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
11435 is current frame number for @code{absolute} mode or out of [-1, 1] range
11436 for @code{relative} mode. First number tells from which frame to pick up top
11437 field and second number tells from which frame to pick up bottom field.
11439 If optionally followed by @code{+} output frame will be marked as interlaced,
11440 else if followed by @code{-} output frame will be marked as progressive, else
11441 it will be marked same as input frame.
11442 If optionally followed by @code{t} output frame will use only top field, or in
11443 case of @code{b} it will use only bottom field.
11444 If line starts with @code{#} or @code{;} that line is skipped.
11447 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
11450 Example of first several lines of @code{hint} file for @code{relative} mode:
11452 0,0 - # first frame
11453 1,0 - # second frame, use third's frame top field and second's frame bottom field
11454 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
11469 @section fieldmatch
11471 Field matching filter for inverse telecine. It is meant to reconstruct the
11472 progressive frames from a telecined stream. The filter does not drop duplicated
11473 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
11474 followed by a decimation filter such as @ref{decimate} in the filtergraph.
11476 The separation of the field matching and the decimation is notably motivated by
11477 the possibility of inserting a de-interlacing filter fallback between the two.
11478 If the source has mixed telecined and real interlaced content,
11479 @code{fieldmatch} will not be able to match fields for the interlaced parts.
11480 But these remaining combed frames will be marked as interlaced, and thus can be
11481 de-interlaced by a later filter such as @ref{yadif} before decimation.
11483 In addition to the various configuration options, @code{fieldmatch} can take an
11484 optional second stream, activated through the @option{ppsrc} option. If
11485 enabled, the frames reconstruction will be based on the fields and frames from
11486 this second stream. This allows the first input to be pre-processed in order to
11487 help the various algorithms of the filter, while keeping the output lossless
11488 (assuming the fields are matched properly). Typically, a field-aware denoiser,
11489 or brightness/contrast adjustments can help.
11491 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
11492 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
11493 which @code{fieldmatch} is based on. While the semantic and usage are very
11494 close, some behaviour and options names can differ.
11496 The @ref{decimate} filter currently only works for constant frame rate input.
11497 If your input has mixed telecined (30fps) and progressive content with a lower
11498 framerate like 24fps use the following filterchain to produce the necessary cfr
11499 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
11501 The filter accepts the following options:
11505 Specify the assumed field order of the input stream. Available values are:
11509 Auto detect parity (use FFmpeg's internal parity value).
11511 Assume bottom field first.
11513 Assume top field first.
11516 Note that it is sometimes recommended not to trust the parity announced by the
11519 Default value is @var{auto}.
11522 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
11523 sense that it won't risk creating jerkiness due to duplicate frames when
11524 possible, but if there are bad edits or blended fields it will end up
11525 outputting combed frames when a good match might actually exist. On the other
11526 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
11527 but will almost always find a good frame if there is one. The other values are
11528 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
11529 jerkiness and creating duplicate frames versus finding good matches in sections
11530 with bad edits, orphaned fields, blended fields, etc.
11532 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
11534 Available values are:
11538 2-way matching (p/c)
11540 2-way matching, and trying 3rd match if still combed (p/c + n)
11542 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
11544 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
11545 still combed (p/c + n + u/b)
11547 3-way matching (p/c/n)
11549 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
11550 detected as combed (p/c/n + u/b)
11553 The parenthesis at the end indicate the matches that would be used for that
11554 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
11557 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
11560 Default value is @var{pc_n}.
11563 Mark the main input stream as a pre-processed input, and enable the secondary
11564 input stream as the clean source to pick the fields from. See the filter
11565 introduction for more details. It is similar to the @option{clip2} feature from
11568 Default value is @code{0} (disabled).
11571 Set the field to match from. It is recommended to set this to the same value as
11572 @option{order} unless you experience matching failures with that setting. In
11573 certain circumstances changing the field that is used to match from can have a
11574 large impact on matching performance. Available values are:
11578 Automatic (same value as @option{order}).
11580 Match from the bottom field.
11582 Match from the top field.
11585 Default value is @var{auto}.
11588 Set whether or not chroma is included during the match comparisons. In most
11589 cases it is recommended to leave this enabled. You should set this to @code{0}
11590 only if your clip has bad chroma problems such as heavy rainbowing or other
11591 artifacts. Setting this to @code{0} could also be used to speed things up at
11592 the cost of some accuracy.
11594 Default value is @code{1}.
11598 These define an exclusion band which excludes the lines between @option{y0} and
11599 @option{y1} from being included in the field matching decision. An exclusion
11600 band can be used to ignore subtitles, a logo, or other things that may
11601 interfere with the matching. @option{y0} sets the starting scan line and
11602 @option{y1} sets the ending line; all lines in between @option{y0} and
11603 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
11604 @option{y0} and @option{y1} to the same value will disable the feature.
11605 @option{y0} and @option{y1} defaults to @code{0}.
11608 Set the scene change detection threshold as a percentage of maximum change on
11609 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
11610 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
11611 @option{scthresh} is @code{[0.0, 100.0]}.
11613 Default value is @code{12.0}.
11616 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
11617 account the combed scores of matches when deciding what match to use as the
11618 final match. Available values are:
11622 No final matching based on combed scores.
11624 Combed scores are only used when a scene change is detected.
11626 Use combed scores all the time.
11629 Default is @var{sc}.
11632 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
11633 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
11634 Available values are:
11638 No forced calculation.
11640 Force p/c/n calculations.
11642 Force p/c/n/u/b calculations.
11645 Default value is @var{none}.
11648 This is the area combing threshold used for combed frame detection. This
11649 essentially controls how "strong" or "visible" combing must be to be detected.
11650 Larger values mean combing must be more visible and smaller values mean combing
11651 can be less visible or strong and still be detected. Valid settings are from
11652 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
11653 be detected as combed). This is basically a pixel difference value. A good
11654 range is @code{[8, 12]}.
11656 Default value is @code{9}.
11659 Sets whether or not chroma is considered in the combed frame decision. Only
11660 disable this if your source has chroma problems (rainbowing, etc.) that are
11661 causing problems for the combed frame detection with chroma enabled. Actually,
11662 using @option{chroma}=@var{0} is usually more reliable, except for the case
11663 where there is chroma only combing in the source.
11665 Default value is @code{0}.
11669 Respectively set the x-axis and y-axis size of the window used during combed
11670 frame detection. This has to do with the size of the area in which
11671 @option{combpel} pixels are required to be detected as combed for a frame to be
11672 declared combed. See the @option{combpel} parameter description for more info.
11673 Possible values are any number that is a power of 2 starting at 4 and going up
11676 Default value is @code{16}.
11679 The number of combed pixels inside any of the @option{blocky} by
11680 @option{blockx} size blocks on the frame for the frame to be detected as
11681 combed. While @option{cthresh} controls how "visible" the combing must be, this
11682 setting controls "how much" combing there must be in any localized area (a
11683 window defined by the @option{blockx} and @option{blocky} settings) on the
11684 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
11685 which point no frames will ever be detected as combed). This setting is known
11686 as @option{MI} in TFM/VFM vocabulary.
11688 Default value is @code{80}.
11691 @anchor{p/c/n/u/b meaning}
11692 @subsection p/c/n/u/b meaning
11694 @subsubsection p/c/n
11696 We assume the following telecined stream:
11699 Top fields: 1 2 2 3 4
11700 Bottom fields: 1 2 3 4 4
11703 The numbers correspond to the progressive frame the fields relate to. Here, the
11704 first two frames are progressive, the 3rd and 4th are combed, and so on.
11706 When @code{fieldmatch} is configured to run a matching from bottom
11707 (@option{field}=@var{bottom}) this is how this input stream get transformed:
11712 B 1 2 3 4 4 <-- matching reference
11721 As a result of the field matching, we can see that some frames get duplicated.
11722 To perform a complete inverse telecine, you need to rely on a decimation filter
11723 after this operation. See for instance the @ref{decimate} filter.
11725 The same operation now matching from top fields (@option{field}=@var{top})
11730 T 1 2 2 3 4 <-- matching reference
11740 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
11741 basically, they refer to the frame and field of the opposite parity:
11744 @item @var{p} matches the field of the opposite parity in the previous frame
11745 @item @var{c} matches the field of the opposite parity in the current frame
11746 @item @var{n} matches the field of the opposite parity in the next frame
11751 The @var{u} and @var{b} matching are a bit special in the sense that they match
11752 from the opposite parity flag. In the following examples, we assume that we are
11753 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
11754 'x' is placed above and below each matched fields.
11756 With bottom matching (@option{field}=@var{bottom}):
11761 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11762 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11770 With top matching (@option{field}=@var{top}):
11775 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11776 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11784 @subsection Examples
11786 Simple IVTC of a top field first telecined stream:
11788 fieldmatch=order=tff:combmatch=none, decimate
11791 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
11793 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
11796 @section fieldorder
11798 Transform the field order of the input video.
11800 It accepts the following parameters:
11805 The output field order. Valid values are @var{tff} for top field first or @var{bff}
11806 for bottom field first.
11809 The default value is @samp{tff}.
11811 The transformation is done by shifting the picture content up or down
11812 by one line, and filling the remaining line with appropriate picture content.
11813 This method is consistent with most broadcast field order converters.
11815 If the input video is not flagged as being interlaced, or it is already
11816 flagged as being of the required output field order, then this filter does
11817 not alter the incoming video.
11819 It is very useful when converting to or from PAL DV material,
11820 which is bottom field first.
11824 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
11827 @section fifo, afifo
11829 Buffer input images and send them when they are requested.
11831 It is mainly useful when auto-inserted by the libavfilter
11834 It does not take parameters.
11836 @section fillborders
11838 Fill borders of the input video, without changing video stream dimensions.
11839 Sometimes video can have garbage at the four edges and you may not want to
11840 crop video input to keep size multiple of some number.
11842 This filter accepts the following options:
11846 Number of pixels to fill from left border.
11849 Number of pixels to fill from right border.
11852 Number of pixels to fill from top border.
11855 Number of pixels to fill from bottom border.
11860 It accepts the following values:
11863 fill pixels using outermost pixels
11866 fill pixels using mirroring (half sample symmetric)
11869 fill pixels with constant value
11872 fill pixels using reflecting (whole sample symmetric)
11875 fill pixels using wrapping
11878 fade pixels to constant value
11881 Default is @var{smear}.
11884 Set color for pixels in fixed or fade mode. Default is @var{black}.
11887 @subsection Commands
11888 This filter supports same @ref{commands} as options.
11889 The command accepts the same syntax of the corresponding option.
11891 If the specified expression is not valid, it is kept at its current
11896 Find a rectangular object
11898 It accepts the following options:
11902 Filepath of the object image, needs to be in gray8.
11905 Detection threshold, default is 0.5.
11908 Number of mipmaps, default is 3.
11910 @item xmin, ymin, xmax, ymax
11911 Specifies the rectangle in which to search.
11914 @subsection Examples
11918 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
11920 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
11926 Flood area with values of same pixel components with another values.
11928 It accepts the following options:
11931 Set pixel x coordinate.
11934 Set pixel y coordinate.
11937 Set source #0 component value.
11940 Set source #1 component value.
11943 Set source #2 component value.
11946 Set source #3 component value.
11949 Set destination #0 component value.
11952 Set destination #1 component value.
11955 Set destination #2 component value.
11958 Set destination #3 component value.
11964 Convert the input video to one of the specified pixel formats.
11965 Libavfilter will try to pick one that is suitable as input to
11968 It accepts the following parameters:
11972 A '|'-separated list of pixel format names, such as
11973 "pix_fmts=yuv420p|monow|rgb24".
11977 @subsection Examples
11981 Convert the input video to the @var{yuv420p} format
11983 format=pix_fmts=yuv420p
11986 Convert the input video to any of the formats in the list
11988 format=pix_fmts=yuv420p|yuv444p|yuv410p
11995 Convert the video to specified constant frame rate by duplicating or dropping
11996 frames as necessary.
11998 It accepts the following parameters:
12002 The desired output frame rate. The default is @code{25}.
12005 Assume the first PTS should be the given value, in seconds. This allows for
12006 padding/trimming at the start of stream. By default, no assumption is made
12007 about the first frame's expected PTS, so no padding or trimming is done.
12008 For example, this could be set to 0 to pad the beginning with duplicates of
12009 the first frame if a video stream starts after the audio stream or to trim any
12010 frames with a negative PTS.
12013 Timestamp (PTS) rounding method.
12015 Possible values are:
12022 round towards -infinity
12024 round towards +infinity
12028 The default is @code{near}.
12031 Action performed when reading the last frame.
12033 Possible values are:
12036 Use same timestamp rounding method as used for other frames.
12038 Pass through last frame if input duration has not been reached yet.
12040 The default is @code{round}.
12044 Alternatively, the options can be specified as a flat string:
12045 @var{fps}[:@var{start_time}[:@var{round}]].
12047 See also the @ref{setpts} filter.
12049 @subsection Examples
12053 A typical usage in order to set the fps to 25:
12059 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
12061 fps=fps=film:round=near
12067 Pack two different video streams into a stereoscopic video, setting proper
12068 metadata on supported codecs. The two views should have the same size and
12069 framerate and processing will stop when the shorter video ends. Please note
12070 that you may conveniently adjust view properties with the @ref{scale} and
12073 It accepts the following parameters:
12077 The desired packing format. Supported values are:
12082 The views are next to each other (default).
12085 The views are on top of each other.
12088 The views are packed by line.
12091 The views are packed by column.
12094 The views are temporally interleaved.
12103 # Convert left and right views into a frame-sequential video
12104 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
12106 # Convert views into a side-by-side video with the same output resolution as the input
12107 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
12112 Change the frame rate by interpolating new video output frames from the source
12115 This filter is not designed to function correctly with interlaced media. If
12116 you wish to change the frame rate of interlaced media then you are required
12117 to deinterlace before this filter and re-interlace after this filter.
12119 A description of the accepted options follows.
12123 Specify the output frames per second. This option can also be specified
12124 as a value alone. The default is @code{50}.
12127 Specify the start of a range where the output frame will be created as a
12128 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12129 the default is @code{15}.
12132 Specify the end of a range where the output frame will be created as a
12133 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12134 the default is @code{240}.
12137 Specify the level at which a scene change is detected as a value between
12138 0 and 100 to indicate a new scene; a low value reflects a low
12139 probability for the current frame to introduce a new scene, while a higher
12140 value means the current frame is more likely to be one.
12141 The default is @code{8.2}.
12144 Specify flags influencing the filter process.
12146 Available value for @var{flags} is:
12149 @item scene_change_detect, scd
12150 Enable scene change detection using the value of the option @var{scene}.
12151 This flag is enabled by default.
12157 Select one frame every N-th frame.
12159 This filter accepts the following option:
12162 Select frame after every @code{step} frames.
12163 Allowed values are positive integers higher than 0. Default value is @code{1}.
12166 @section freezedetect
12168 Detect frozen video.
12170 This filter logs a message and sets frame metadata when it detects that the
12171 input video has no significant change in content during a specified duration.
12172 Video freeze detection calculates the mean average absolute difference of all
12173 the components of video frames and compares it to a noise floor.
12175 The printed times and duration are expressed in seconds. The
12176 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
12177 whose timestamp equals or exceeds the detection duration and it contains the
12178 timestamp of the first frame of the freeze. The
12179 @code{lavfi.freezedetect.freeze_duration} and
12180 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
12183 The filter accepts the following options:
12187 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
12188 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
12192 Set freeze duration until notification (default is 2 seconds).
12195 @section freezeframes
12197 Freeze video frames.
12199 This filter freezes video frames using frame from 2nd input.
12201 The filter accepts the following options:
12205 Set number of first frame from which to start freeze.
12208 Set number of last frame from which to end freeze.
12211 Set number of frame from 2nd input which will be used instead of replaced frames.
12217 Apply a frei0r effect to the input video.
12219 To enable the compilation of this filter, you need to install the frei0r
12220 header and configure FFmpeg with @code{--enable-frei0r}.
12222 It accepts the following parameters:
12227 The name of the frei0r effect to load. If the environment variable
12228 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
12229 directories specified by the colon-separated list in @env{FREI0R_PATH}.
12230 Otherwise, the standard frei0r paths are searched, in this order:
12231 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
12232 @file{/usr/lib/frei0r-1/}.
12234 @item filter_params
12235 A '|'-separated list of parameters to pass to the frei0r effect.
12239 A frei0r effect parameter can be a boolean (its value is either
12240 "y" or "n"), a double, a color (specified as
12241 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
12242 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
12243 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
12244 a position (specified as @var{X}/@var{Y}, where
12245 @var{X} and @var{Y} are floating point numbers) and/or a string.
12247 The number and types of parameters depend on the loaded effect. If an
12248 effect parameter is not specified, the default value is set.
12250 @subsection Examples
12254 Apply the distort0r effect, setting the first two double parameters:
12256 frei0r=filter_name=distort0r:filter_params=0.5|0.01
12260 Apply the colordistance effect, taking a color as the first parameter:
12262 frei0r=colordistance:0.2/0.3/0.4
12263 frei0r=colordistance:violet
12264 frei0r=colordistance:0x112233
12268 Apply the perspective effect, specifying the top left and top right image
12271 frei0r=perspective:0.2/0.2|0.8/0.2
12275 For more information, see
12276 @url{http://frei0r.dyne.org}
12278 @subsection Commands
12280 This filter supports the @option{filter_params} option as @ref{commands}.
12284 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
12286 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
12287 processing filter, one of them is performed once per block, not per pixel.
12288 This allows for much higher speed.
12290 The filter accepts the following options:
12294 Set quality. This option defines the number of levels for averaging. It accepts
12295 an integer in the range 4-5. Default value is @code{4}.
12298 Force a constant quantization parameter. It accepts an integer in range 0-63.
12299 If not set, the filter will use the QP from the video stream (if available).
12302 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
12303 more details but also more artifacts, while higher values make the image smoother
12304 but also blurrier. Default value is @code{0} − PSNR optimal.
12306 @item use_bframe_qp
12307 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
12308 option may cause flicker since the B-Frames have often larger QP. Default is
12309 @code{0} (not enabled).
12315 Apply Gaussian blur filter.
12317 The filter accepts the following options:
12321 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
12324 Set number of steps for Gaussian approximation. Default is @code{1}.
12327 Set which planes to filter. By default all planes are filtered.
12330 Set vertical sigma, if negative it will be same as @code{sigma}.
12331 Default is @code{-1}.
12334 @subsection Commands
12335 This filter supports same commands as options.
12336 The command accepts the same syntax of the corresponding option.
12338 If the specified expression is not valid, it is kept at its current
12343 Apply generic equation to each pixel.
12345 The filter accepts the following options:
12348 @item lum_expr, lum
12349 Set the luminance expression.
12351 Set the chrominance blue expression.
12353 Set the chrominance red expression.
12354 @item alpha_expr, a
12355 Set the alpha expression.
12357 Set the red expression.
12358 @item green_expr, g
12359 Set the green expression.
12361 Set the blue expression.
12364 The colorspace is selected according to the specified options. If one
12365 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
12366 options is specified, the filter will automatically select a YCbCr
12367 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
12368 @option{blue_expr} options is specified, it will select an RGB
12371 If one of the chrominance expression is not defined, it falls back on the other
12372 one. If no alpha expression is specified it will evaluate to opaque value.
12373 If none of chrominance expressions are specified, they will evaluate
12374 to the luminance expression.
12376 The expressions can use the following variables and functions:
12380 The sequential number of the filtered frame, starting from @code{0}.
12384 The coordinates of the current sample.
12388 The width and height of the image.
12392 Width and height scale depending on the currently filtered plane. It is the
12393 ratio between the corresponding luma plane number of pixels and the current
12394 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
12395 @code{0.5,0.5} for chroma planes.
12398 Time of the current frame, expressed in seconds.
12401 Return the value of the pixel at location (@var{x},@var{y}) of the current
12405 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
12409 Return the value of the pixel at location (@var{x},@var{y}) of the
12410 blue-difference chroma plane. Return 0 if there is no such plane.
12413 Return the value of the pixel at location (@var{x},@var{y}) of the
12414 red-difference chroma plane. Return 0 if there is no such plane.
12419 Return the value of the pixel at location (@var{x},@var{y}) of the
12420 red/green/blue component. Return 0 if there is no such component.
12423 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
12424 plane. Return 0 if there is no such plane.
12426 @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)
12427 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
12428 sums of samples within a rectangle. See the functions without the sum postfix.
12430 @item interpolation
12431 Set one of interpolation methods:
12436 Default is bilinear.
12439 For functions, if @var{x} and @var{y} are outside the area, the value will be
12440 automatically clipped to the closer edge.
12442 Please note that this filter can use multiple threads in which case each slice
12443 will have its own expression state. If you want to use only a single expression
12444 state because your expressions depend on previous state then you should limit
12445 the number of filter threads to 1.
12447 @subsection Examples
12451 Flip the image horizontally:
12457 Generate a bidimensional sine wave, with angle @code{PI/3} and a
12458 wavelength of 100 pixels:
12460 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
12464 Generate a fancy enigmatic moving light:
12466 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
12470 Generate a quick emboss effect:
12472 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
12476 Modify RGB components depending on pixel position:
12478 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
12482 Create a radial gradient that is the same size as the input (also see
12483 the @ref{vignette} filter):
12485 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
12491 Fix the banding artifacts that are sometimes introduced into nearly flat
12492 regions by truncation to 8-bit color depth.
12493 Interpolate the gradients that should go where the bands are, and
12496 It is designed for playback only. Do not use it prior to
12497 lossy compression, because compression tends to lose the dither and
12498 bring back the bands.
12500 It accepts the following parameters:
12505 The maximum amount by which the filter will change any one pixel. This is also
12506 the threshold for detecting nearly flat regions. Acceptable values range from
12507 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
12511 The neighborhood to fit the gradient to. A larger radius makes for smoother
12512 gradients, but also prevents the filter from modifying the pixels near detailed
12513 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
12514 values will be clipped to the valid range.
12518 Alternatively, the options can be specified as a flat string:
12519 @var{strength}[:@var{radius}]
12521 @subsection Examples
12525 Apply the filter with a @code{3.5} strength and radius of @code{8}:
12531 Specify radius, omitting the strength (which will fall-back to the default
12539 @anchor{graphmonitor}
12540 @section graphmonitor
12541 Show various filtergraph stats.
12543 With this filter one can debug complete filtergraph.
12544 Especially issues with links filling with queued frames.
12546 The filter accepts the following options:
12550 Set video output size. Default is @var{hd720}.
12553 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
12556 Set output mode, can be @var{fulll} or @var{compact}.
12557 In @var{compact} mode only filters with some queued frames have displayed stats.
12560 Set flags which enable which stats are shown in video.
12562 Available values for flags are:
12565 Display number of queued frames in each link.
12567 @item frame_count_in
12568 Display number of frames taken from filter.
12570 @item frame_count_out
12571 Display number of frames given out from filter.
12574 Display current filtered frame pts.
12577 Display current filtered frame time.
12580 Display time base for filter link.
12583 Display used format for filter link.
12586 Display video size or number of audio channels in case of audio used by filter link.
12589 Display video frame rate or sample rate in case of audio used by filter link.
12592 Display link output status.
12596 Set upper limit for video rate of output stream, Default value is @var{25}.
12597 This guarantee that output video frame rate will not be higher than this value.
12601 A color constancy variation filter which estimates scene illumination via grey edge algorithm
12602 and corrects the scene colors accordingly.
12604 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
12606 The filter accepts the following options:
12610 The order of differentiation to be applied on the scene. Must be chosen in the range
12611 [0,2] and default value is 1.
12614 The Minkowski parameter to be used for calculating the Minkowski distance. Must
12615 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
12616 max value instead of calculating Minkowski distance.
12619 The standard deviation of Gaussian blur to be applied on the scene. Must be
12620 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
12621 can't be equal to 0 if @var{difford} is greater than 0.
12624 @subsection Examples
12630 greyedge=difford=1:minknorm=5:sigma=2
12636 greyedge=difford=1:minknorm=0:sigma=2
12644 Apply a Hald CLUT to a video stream.
12646 First input is the video stream to process, and second one is the Hald CLUT.
12647 The Hald CLUT input can be a simple picture or a complete video stream.
12649 The filter accepts the following options:
12653 Force termination when the shortest input terminates. Default is @code{0}.
12655 Continue applying the last CLUT after the end of the stream. A value of
12656 @code{0} disable the filter after the last frame of the CLUT is reached.
12657 Default is @code{1}.
12660 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
12661 filters share the same internals).
12663 This filter also supports the @ref{framesync} options.
12665 More information about the Hald CLUT can be found on Eskil Steenberg's website
12666 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
12668 @subsection Workflow examples
12670 @subsubsection Hald CLUT video stream
12672 Generate an identity Hald CLUT stream altered with various effects:
12674 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
12677 Note: make sure you use a lossless codec.
12679 Then use it with @code{haldclut} to apply it on some random stream:
12681 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
12684 The Hald CLUT will be applied to the 10 first seconds (duration of
12685 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
12686 to the remaining frames of the @code{mandelbrot} stream.
12688 @subsubsection Hald CLUT with preview
12690 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
12691 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
12692 biggest possible square starting at the top left of the picture. The remaining
12693 padding pixels (bottom or right) will be ignored. This area can be used to add
12694 a preview of the Hald CLUT.
12696 Typically, the following generated Hald CLUT will be supported by the
12697 @code{haldclut} filter:
12700 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
12701 pad=iw+320 [padded_clut];
12702 smptebars=s=320x256, split [a][b];
12703 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
12704 [main][b] overlay=W-320" -frames:v 1 clut.png
12707 It contains the original and a preview of the effect of the CLUT: SMPTE color
12708 bars are displayed on the right-top, and below the same color bars processed by
12711 Then, the effect of this Hald CLUT can be visualized with:
12713 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
12718 Flip the input video horizontally.
12720 For example, to horizontally flip the input video with @command{ffmpeg}:
12722 ffmpeg -i in.avi -vf "hflip" out.avi
12726 This filter applies a global color histogram equalization on a
12729 It can be used to correct video that has a compressed range of pixel
12730 intensities. The filter redistributes the pixel intensities to
12731 equalize their distribution across the intensity range. It may be
12732 viewed as an "automatically adjusting contrast filter". This filter is
12733 useful only for correcting degraded or poorly captured source
12736 The filter accepts the following options:
12740 Determine the amount of equalization to be applied. As the strength
12741 is reduced, the distribution of pixel intensities more-and-more
12742 approaches that of the input frame. The value must be a float number
12743 in the range [0,1] and defaults to 0.200.
12746 Set the maximum intensity that can generated and scale the output
12747 values appropriately. The strength should be set as desired and then
12748 the intensity can be limited if needed to avoid washing-out. The value
12749 must be a float number in the range [0,1] and defaults to 0.210.
12752 Set the antibanding level. If enabled the filter will randomly vary
12753 the luminance of output pixels by a small amount to avoid banding of
12754 the histogram. Possible values are @code{none}, @code{weak} or
12755 @code{strong}. It defaults to @code{none}.
12761 Compute and draw a color distribution histogram for the input video.
12763 The computed histogram is a representation of the color component
12764 distribution in an image.
12766 Standard histogram displays the color components distribution in an image.
12767 Displays color graph for each color component. Shows distribution of
12768 the Y, U, V, A or R, G, B components, depending on input format, in the
12769 current frame. Below each graph a color component scale meter is shown.
12771 The filter accepts the following options:
12775 Set height of level. Default value is @code{200}.
12776 Allowed range is [50, 2048].
12779 Set height of color scale. Default value is @code{12}.
12780 Allowed range is [0, 40].
12784 It accepts the following values:
12787 Per color component graphs are placed below each other.
12790 Per color component graphs are placed side by side.
12793 Presents information identical to that in the @code{parade}, except
12794 that the graphs representing color components are superimposed directly
12797 Default is @code{stack}.
12800 Set mode. Can be either @code{linear}, or @code{logarithmic}.
12801 Default is @code{linear}.
12804 Set what color components to display.
12805 Default is @code{7}.
12808 Set foreground opacity. Default is @code{0.7}.
12811 Set background opacity. Default is @code{0.5}.
12814 @subsection Examples
12819 Calculate and draw histogram:
12821 ffplay -i input -vf histogram
12829 This is a high precision/quality 3d denoise filter. It aims to reduce
12830 image noise, producing smooth images and making still images really
12831 still. It should enhance compressibility.
12833 It accepts the following optional parameters:
12837 A non-negative floating point number which specifies spatial luma strength.
12838 It defaults to 4.0.
12840 @item chroma_spatial
12841 A non-negative floating point number which specifies spatial chroma strength.
12842 It defaults to 3.0*@var{luma_spatial}/4.0.
12845 A floating point number which specifies luma temporal strength. It defaults to
12846 6.0*@var{luma_spatial}/4.0.
12849 A floating point number which specifies chroma temporal strength. It defaults to
12850 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
12853 @subsection Commands
12854 This filter supports same @ref{commands} as options.
12855 The command accepts the same syntax of the corresponding option.
12857 If the specified expression is not valid, it is kept at its current
12860 @anchor{hwdownload}
12861 @section hwdownload
12863 Download hardware frames to system memory.
12865 The input must be in hardware frames, and the output a non-hardware format.
12866 Not all formats will be supported on the output - it may be necessary to insert
12867 an additional @option{format} filter immediately following in the graph to get
12868 the output in a supported format.
12872 Map hardware frames to system memory or to another device.
12874 This filter has several different modes of operation; which one is used depends
12875 on the input and output formats:
12878 Hardware frame input, normal frame output
12880 Map the input frames to system memory and pass them to the output. If the
12881 original hardware frame is later required (for example, after overlaying
12882 something else on part of it), the @option{hwmap} filter can be used again
12883 in the next mode to retrieve it.
12885 Normal frame input, hardware frame output
12887 If the input is actually a software-mapped hardware frame, then unmap it -
12888 that is, return the original hardware frame.
12890 Otherwise, a device must be provided. Create new hardware surfaces on that
12891 device for the output, then map them back to the software format at the input
12892 and give those frames to the preceding filter. This will then act like the
12893 @option{hwupload} filter, but may be able to avoid an additional copy when
12894 the input is already in a compatible format.
12896 Hardware frame input and output
12898 A device must be supplied for the output, either directly or with the
12899 @option{derive_device} option. The input and output devices must be of
12900 different types and compatible - the exact meaning of this is
12901 system-dependent, but typically it means that they must refer to the same
12902 underlying hardware context (for example, refer to the same graphics card).
12904 If the input frames were originally created on the output device, then unmap
12905 to retrieve the original frames.
12907 Otherwise, map the frames to the output device - create new hardware frames
12908 on the output corresponding to the frames on the input.
12911 The following additional parameters are accepted:
12915 Set the frame mapping mode. Some combination of:
12918 The mapped frame should be readable.
12920 The mapped frame should be writeable.
12922 The mapping will always overwrite the entire frame.
12924 This may improve performance in some cases, as the original contents of the
12925 frame need not be loaded.
12927 The mapping must not involve any copying.
12929 Indirect mappings to copies of frames are created in some cases where either
12930 direct mapping is not possible or it would have unexpected properties.
12931 Setting this flag ensures that the mapping is direct and will fail if that is
12934 Defaults to @var{read+write} if not specified.
12936 @item derive_device @var{type}
12937 Rather than using the device supplied at initialisation, instead derive a new
12938 device of type @var{type} from the device the input frames exist on.
12941 In a hardware to hardware mapping, map in reverse - create frames in the sink
12942 and map them back to the source. This may be necessary in some cases where
12943 a mapping in one direction is required but only the opposite direction is
12944 supported by the devices being used.
12946 This option is dangerous - it may break the preceding filter in undefined
12947 ways if there are any additional constraints on that filter's output.
12948 Do not use it without fully understanding the implications of its use.
12954 Upload system memory frames to hardware surfaces.
12956 The device to upload to must be supplied when the filter is initialised. If
12957 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
12958 option or with the @option{derive_device} option. The input and output devices
12959 must be of different types and compatible - the exact meaning of this is
12960 system-dependent, but typically it means that they must refer to the same
12961 underlying hardware context (for example, refer to the same graphics card).
12963 The following additional parameters are accepted:
12966 @item derive_device @var{type}
12967 Rather than using the device supplied at initialisation, instead derive a new
12968 device of type @var{type} from the device the input frames exist on.
12971 @anchor{hwupload_cuda}
12972 @section hwupload_cuda
12974 Upload system memory frames to a CUDA device.
12976 It accepts the following optional parameters:
12980 The number of the CUDA device to use
12985 Apply a high-quality magnification filter designed for pixel art. This filter
12986 was originally created by Maxim Stepin.
12988 It accepts the following option:
12992 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
12993 @code{hq3x} and @code{4} for @code{hq4x}.
12994 Default is @code{3}.
12998 Stack input videos horizontally.
13000 All streams must be of same pixel format and of same height.
13002 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
13003 to create same output.
13005 The filter accepts the following option:
13009 Set number of input streams. Default is 2.
13012 If set to 1, force the output to terminate when the shortest input
13013 terminates. Default value is 0.
13018 Modify the hue and/or the saturation of the input.
13020 It accepts the following parameters:
13024 Specify the hue angle as a number of degrees. It accepts an expression,
13025 and defaults to "0".
13028 Specify the saturation in the [-10,10] range. It accepts an expression and
13032 Specify the hue angle as a number of radians. It accepts an
13033 expression, and defaults to "0".
13036 Specify the brightness in the [-10,10] range. It accepts an expression and
13040 @option{h} and @option{H} are mutually exclusive, and can't be
13041 specified at the same time.
13043 The @option{b}, @option{h}, @option{H} and @option{s} option values are
13044 expressions containing the following constants:
13048 frame count of the input frame starting from 0
13051 presentation timestamp of the input frame expressed in time base units
13054 frame rate of the input video, NAN if the input frame rate is unknown
13057 timestamp expressed in seconds, NAN if the input timestamp is unknown
13060 time base of the input video
13063 @subsection Examples
13067 Set the hue to 90 degrees and the saturation to 1.0:
13073 Same command but expressing the hue in radians:
13079 Rotate hue and make the saturation swing between 0
13080 and 2 over a period of 1 second:
13082 hue="H=2*PI*t: s=sin(2*PI*t)+1"
13086 Apply a 3 seconds saturation fade-in effect starting at 0:
13088 hue="s=min(t/3\,1)"
13091 The general fade-in expression can be written as:
13093 hue="s=min(0\, max((t-START)/DURATION\, 1))"
13097 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
13099 hue="s=max(0\, min(1\, (8-t)/3))"
13102 The general fade-out expression can be written as:
13104 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
13109 @subsection Commands
13111 This filter supports the following commands:
13117 Modify the hue and/or the saturation and/or brightness of the input video.
13118 The command accepts the same syntax of the corresponding option.
13120 If the specified expression is not valid, it is kept at its current
13124 @section hysteresis
13126 Grow first stream into second stream by connecting components.
13127 This makes it possible to build more robust edge masks.
13129 This filter accepts the following options:
13133 Set which planes will be processed as bitmap, unprocessed planes will be
13134 copied from first stream.
13135 By default value 0xf, all planes will be processed.
13138 Set threshold which is used in filtering. If pixel component value is higher than
13139 this value filter algorithm for connecting components is activated.
13140 By default value is 0.
13143 The @code{hysteresis} filter also supports the @ref{framesync} options.
13147 Detect video interlacing type.
13149 This filter tries to detect if the input frames are interlaced, progressive,
13150 top or bottom field first. It will also try to detect fields that are
13151 repeated between adjacent frames (a sign of telecine).
13153 Single frame detection considers only immediately adjacent frames when classifying each frame.
13154 Multiple frame detection incorporates the classification history of previous frames.
13156 The filter will log these metadata values:
13159 @item single.current_frame
13160 Detected type of current frame using single-frame detection. One of:
13161 ``tff'' (top field first), ``bff'' (bottom field first),
13162 ``progressive'', or ``undetermined''
13165 Cumulative number of frames detected as top field first using single-frame detection.
13168 Cumulative number of frames detected as top field first using multiple-frame detection.
13171 Cumulative number of frames detected as bottom field first using single-frame detection.
13173 @item multiple.current_frame
13174 Detected type of current frame using multiple-frame detection. One of:
13175 ``tff'' (top field first), ``bff'' (bottom field first),
13176 ``progressive'', or ``undetermined''
13179 Cumulative number of frames detected as bottom field first using multiple-frame detection.
13181 @item single.progressive
13182 Cumulative number of frames detected as progressive using single-frame detection.
13184 @item multiple.progressive
13185 Cumulative number of frames detected as progressive using multiple-frame detection.
13187 @item single.undetermined
13188 Cumulative number of frames that could not be classified using single-frame detection.
13190 @item multiple.undetermined
13191 Cumulative number of frames that could not be classified using multiple-frame detection.
13193 @item repeated.current_frame
13194 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
13196 @item repeated.neither
13197 Cumulative number of frames with no repeated field.
13200 Cumulative number of frames with the top field repeated from the previous frame's top field.
13202 @item repeated.bottom
13203 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
13206 The filter accepts the following options:
13210 Set interlacing threshold.
13212 Set progressive threshold.
13214 Threshold for repeated field detection.
13216 Number of frames after which a given frame's contribution to the
13217 statistics is halved (i.e., it contributes only 0.5 to its
13218 classification). The default of 0 means that all frames seen are given
13219 full weight of 1.0 forever.
13220 @item analyze_interlaced_flag
13221 When this is not 0 then idet will use the specified number of frames to determine
13222 if the interlaced flag is accurate, it will not count undetermined frames.
13223 If the flag is found to be accurate it will be used without any further
13224 computations, if it is found to be inaccurate it will be cleared without any
13225 further computations. This allows inserting the idet filter as a low computational
13226 method to clean up the interlaced flag
13231 Deinterleave or interleave fields.
13233 This filter allows one to process interlaced images fields without
13234 deinterlacing them. Deinterleaving splits the input frame into 2
13235 fields (so called half pictures). Odd lines are moved to the top
13236 half of the output image, even lines to the bottom half.
13237 You can process (filter) them independently and then re-interleave them.
13239 The filter accepts the following options:
13243 @item chroma_mode, c
13244 @item alpha_mode, a
13245 Available values for @var{luma_mode}, @var{chroma_mode} and
13246 @var{alpha_mode} are:
13252 @item deinterleave, d
13253 Deinterleave fields, placing one above the other.
13255 @item interleave, i
13256 Interleave fields. Reverse the effect of deinterleaving.
13258 Default value is @code{none}.
13260 @item luma_swap, ls
13261 @item chroma_swap, cs
13262 @item alpha_swap, as
13263 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
13266 @subsection Commands
13268 This filter supports the all above options as @ref{commands}.
13272 Apply inflate effect to the video.
13274 This filter replaces the pixel by the local(3x3) average by taking into account
13275 only values higher than the pixel.
13277 It accepts the following options:
13284 Limit the maximum change for each plane, default is 65535.
13285 If 0, plane will remain unchanged.
13288 @subsection Commands
13290 This filter supports the all above options as @ref{commands}.
13294 Simple interlacing filter from progressive contents. This interleaves upper (or
13295 lower) lines from odd frames with lower (or upper) lines from even frames,
13296 halving the frame rate and preserving image height.
13299 Original Original New Frame
13300 Frame 'j' Frame 'j+1' (tff)
13301 ========== =========== ==================
13302 Line 0 --------------------> Frame 'j' Line 0
13303 Line 1 Line 1 ----> Frame 'j+1' Line 1
13304 Line 2 ---------------------> Frame 'j' Line 2
13305 Line 3 Line 3 ----> Frame 'j+1' Line 3
13307 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
13310 It accepts the following optional parameters:
13314 This determines whether the interlaced frame is taken from the even
13315 (tff - default) or odd (bff) lines of the progressive frame.
13318 Vertical lowpass filter to avoid twitter interlacing and
13319 reduce moire patterns.
13323 Disable vertical lowpass filter
13326 Enable linear filter (default)
13329 Enable complex filter. This will slightly less reduce twitter and moire
13330 but better retain detail and subjective sharpness impression.
13337 Deinterlace input video by applying Donald Graft's adaptive kernel
13338 deinterling. Work on interlaced parts of a video to produce
13339 progressive frames.
13341 The description of the accepted parameters follows.
13345 Set the threshold which affects the filter's tolerance when
13346 determining if a pixel line must be processed. It must be an integer
13347 in the range [0,255] and defaults to 10. A value of 0 will result in
13348 applying the process on every pixels.
13351 Paint pixels exceeding the threshold value to white if set to 1.
13355 Set the fields order. Swap fields if set to 1, leave fields alone if
13359 Enable additional sharpening if set to 1. Default is 0.
13362 Enable twoway sharpening if set to 1. Default is 0.
13365 @subsection Examples
13369 Apply default values:
13371 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
13375 Enable additional sharpening:
13381 Paint processed pixels in white:
13389 Slowly update darker pixels.
13391 This filter makes short flashes of light appear longer.
13392 This filter accepts the following options:
13396 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
13399 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
13402 @section lenscorrection
13404 Correct radial lens distortion
13406 This filter can be used to correct for radial distortion as can result from the use
13407 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
13408 one can use tools available for example as part of opencv or simply trial-and-error.
13409 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
13410 and extract the k1 and k2 coefficients from the resulting matrix.
13412 Note that effectively the same filter is available in the open-source tools Krita and
13413 Digikam from the KDE project.
13415 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
13416 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
13417 brightness distribution, so you may want to use both filters together in certain
13418 cases, though you will have to take care of ordering, i.e. whether vignetting should
13419 be applied before or after lens correction.
13421 @subsection Options
13423 The filter accepts the following options:
13427 Relative x-coordinate of the focal point of the image, and thereby the center of the
13428 distortion. This value has a range [0,1] and is expressed as fractions of the image
13429 width. Default is 0.5.
13431 Relative y-coordinate of the focal point of the image, and thereby the center of the
13432 distortion. This value has a range [0,1] and is expressed as fractions of the image
13433 height. Default is 0.5.
13435 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
13436 no correction. Default is 0.
13438 Coefficient of the double quadratic correction term. This value has a range [-1,1].
13439 0 means no correction. Default is 0.
13442 The formula that generates the correction is:
13444 @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)
13446 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
13447 distances from the focal point in the source and target images, respectively.
13451 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
13453 The @code{lensfun} filter requires the camera make, camera model, and lens model
13454 to apply the lens correction. The filter will load the lensfun database and
13455 query it to find the corresponding camera and lens entries in the database. As
13456 long as these entries can be found with the given options, the filter can
13457 perform corrections on frames. Note that incomplete strings will result in the
13458 filter choosing the best match with the given options, and the filter will
13459 output the chosen camera and lens models (logged with level "info"). You must
13460 provide the make, camera model, and lens model as they are required.
13462 The filter accepts the following options:
13466 The make of the camera (for example, "Canon"). This option is required.
13469 The model of the camera (for example, "Canon EOS 100D"). This option is
13473 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
13474 option is required.
13477 The type of correction to apply. The following values are valid options:
13481 Enables fixing lens vignetting.
13484 Enables fixing lens geometry. This is the default.
13487 Enables fixing chromatic aberrations.
13490 Enables fixing lens vignetting and lens geometry.
13493 Enables fixing lens vignetting and chromatic aberrations.
13496 Enables fixing both lens geometry and chromatic aberrations.
13499 Enables all possible corrections.
13503 The focal length of the image/video (zoom; expected constant for video). For
13504 example, a 18--55mm lens has focal length range of [18--55], so a value in that
13505 range should be chosen when using that lens. Default 18.
13508 The aperture of the image/video (expected constant for video). Note that
13509 aperture is only used for vignetting correction. Default 3.5.
13511 @item focus_distance
13512 The focus distance of the image/video (expected constant for video). Note that
13513 focus distance is only used for vignetting and only slightly affects the
13514 vignetting correction process. If unknown, leave it at the default value (which
13518 The scale factor which is applied after transformation. After correction the
13519 video is no longer necessarily rectangular. This parameter controls how much of
13520 the resulting image is visible. The value 0 means that a value will be chosen
13521 automatically such that there is little or no unmapped area in the output
13522 image. 1.0 means that no additional scaling is done. Lower values may result
13523 in more of the corrected image being visible, while higher values may avoid
13524 unmapped areas in the output.
13526 @item target_geometry
13527 The target geometry of the output image/video. The following values are valid
13531 @item rectilinear (default)
13534 @item equirectangular
13535 @item fisheye_orthographic
13536 @item fisheye_stereographic
13537 @item fisheye_equisolid
13538 @item fisheye_thoby
13541 Apply the reverse of image correction (instead of correcting distortion, apply
13544 @item interpolation
13545 The type of interpolation used when correcting distortion. The following values
13550 @item linear (default)
13555 @subsection Examples
13559 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
13560 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
13564 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
13568 Apply the same as before, but only for the first 5 seconds of video.
13571 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
13578 Obtain the VMAF (Video Multi-Method Assessment Fusion)
13579 score between two input videos.
13581 The obtained VMAF score is printed through the logging system.
13583 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
13584 After installing the library it can be enabled using:
13585 @code{./configure --enable-libvmaf}.
13586 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
13588 The filter has following options:
13592 Set the model path which is to be used for SVM.
13593 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
13596 Set the file path to be used to store logs.
13599 Set the format of the log file (csv, json or xml).
13601 @item enable_transform
13602 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
13603 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
13604 Default value: @code{false}
13607 Invokes the phone model which will generate VMAF scores higher than in the
13608 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
13609 Default value: @code{false}
13612 Enables computing psnr along with vmaf.
13613 Default value: @code{false}
13616 Enables computing ssim along with vmaf.
13617 Default value: @code{false}
13620 Enables computing ms_ssim along with vmaf.
13621 Default value: @code{false}
13624 Set the pool method to be used for computing vmaf.
13625 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
13628 Set number of threads to be used when computing vmaf.
13629 Default value: @code{0}, which makes use of all available logical processors.
13632 Set interval for frame subsampling used when computing vmaf.
13633 Default value: @code{1}
13635 @item enable_conf_interval
13636 Enables confidence interval.
13637 Default value: @code{false}
13640 This filter also supports the @ref{framesync} options.
13642 @subsection Examples
13645 On the below examples the input file @file{main.mpg} being processed is
13646 compared with the reference file @file{ref.mpg}.
13649 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
13653 Example with options:
13655 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
13659 Example with options and different containers:
13661 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 -
13667 Limits the pixel components values to the specified range [min, max].
13669 The filter accepts the following options:
13673 Lower bound. Defaults to the lowest allowed value for the input.
13676 Upper bound. Defaults to the highest allowed value for the input.
13679 Specify which planes will be processed. Defaults to all available.
13682 @subsection Commands
13684 This filter supports the all above options as @ref{commands}.
13690 The filter accepts the following options:
13694 Set the number of loops. Setting this value to -1 will result in infinite loops.
13698 Set maximal size in number of frames. Default is 0.
13701 Set first frame of loop. Default is 0.
13704 @subsection Examples
13708 Loop single first frame infinitely:
13710 loop=loop=-1:size=1:start=0
13714 Loop single first frame 10 times:
13716 loop=loop=10:size=1:start=0
13720 Loop 10 first frames 5 times:
13722 loop=loop=5:size=10:start=0
13728 Apply a 1D LUT to an input video.
13730 The filter accepts the following options:
13734 Set the 1D LUT file name.
13736 Currently supported formats:
13745 Select interpolation mode.
13747 Available values are:
13751 Use values from the nearest defined point.
13753 Interpolate values using the linear interpolation.
13755 Interpolate values using the cosine interpolation.
13757 Interpolate values using the cubic interpolation.
13759 Interpolate values using the spline interpolation.
13766 Apply a 3D LUT to an input video.
13768 The filter accepts the following options:
13772 Set the 3D LUT file name.
13774 Currently supported formats:
13788 Select interpolation mode.
13790 Available values are:
13794 Use values from the nearest defined point.
13796 Interpolate values using the 8 points defining a cube.
13798 Interpolate values using a tetrahedron.
13804 Turn certain luma values into transparency.
13806 The filter accepts the following options:
13810 Set the luma which will be used as base for transparency.
13811 Default value is @code{0}.
13814 Set the range of luma values to be keyed out.
13815 Default value is @code{0.01}.
13818 Set the range of softness. Default value is @code{0}.
13819 Use this to control gradual transition from zero to full transparency.
13822 @subsection Commands
13823 This filter supports same @ref{commands} as options.
13824 The command accepts the same syntax of the corresponding option.
13826 If the specified expression is not valid, it is kept at its current
13829 @section lut, lutrgb, lutyuv
13831 Compute a look-up table for binding each pixel component input value
13832 to an output value, and apply it to the input video.
13834 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
13835 to an RGB input video.
13837 These filters accept the following parameters:
13840 set first pixel component expression
13842 set second pixel component expression
13844 set third pixel component expression
13846 set fourth pixel component expression, corresponds to the alpha component
13849 set red component expression
13851 set green component expression
13853 set blue component expression
13855 alpha component expression
13858 set Y/luminance component expression
13860 set U/Cb component expression
13862 set V/Cr component expression
13865 Each of them specifies the expression to use for computing the lookup table for
13866 the corresponding pixel component values.
13868 The exact component associated to each of the @var{c*} options depends on the
13871 The @var{lut} filter requires either YUV or RGB pixel formats in input,
13872 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
13874 The expressions can contain the following constants and functions:
13879 The input width and height.
13882 The input value for the pixel component.
13885 The input value, clipped to the @var{minval}-@var{maxval} range.
13888 The maximum value for the pixel component.
13891 The minimum value for the pixel component.
13894 The negated value for the pixel component value, clipped to the
13895 @var{minval}-@var{maxval} range; it corresponds to the expression
13896 "maxval-clipval+minval".
13899 The computed value in @var{val}, clipped to the
13900 @var{minval}-@var{maxval} range.
13902 @item gammaval(gamma)
13903 The computed gamma correction value of the pixel component value,
13904 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
13906 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
13910 All expressions default to "val".
13912 @subsection Examples
13916 Negate input video:
13918 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
13919 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
13922 The above is the same as:
13924 lutrgb="r=negval:g=negval:b=negval"
13925 lutyuv="y=negval:u=negval:v=negval"
13935 Remove chroma components, turning the video into a graytone image:
13937 lutyuv="u=128:v=128"
13941 Apply a luma burning effect:
13947 Remove green and blue components:
13953 Set a constant alpha channel value on input:
13955 format=rgba,lutrgb=a="maxval-minval/2"
13959 Correct luminance gamma by a factor of 0.5:
13961 lutyuv=y=gammaval(0.5)
13965 Discard least significant bits of luma:
13967 lutyuv=y='bitand(val, 128+64+32)'
13971 Technicolor like effect:
13973 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
13977 @section lut2, tlut2
13979 The @code{lut2} filter takes two input streams and outputs one
13982 The @code{tlut2} (time lut2) filter takes two consecutive frames
13983 from one single stream.
13985 This filter accepts the following parameters:
13988 set first pixel component expression
13990 set second pixel component expression
13992 set third pixel component expression
13994 set fourth pixel component expression, corresponds to the alpha component
13997 set output bit depth, only available for @code{lut2} filter. By default is 0,
13998 which means bit depth is automatically picked from first input format.
14001 The @code{lut2} filter also supports the @ref{framesync} options.
14003 Each of them specifies the expression to use for computing the lookup table for
14004 the corresponding pixel component values.
14006 The exact component associated to each of the @var{c*} options depends on the
14009 The expressions can contain the following constants:
14014 The input width and height.
14017 The first input value for the pixel component.
14020 The second input value for the pixel component.
14023 The first input video bit depth.
14026 The second input video bit depth.
14029 All expressions default to "x".
14031 @subsection Examples
14035 Highlight differences between two RGB video streams:
14037 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)'
14041 Highlight differences between two YUV video streams:
14043 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)'
14047 Show max difference between two video streams:
14049 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)))'
14053 @section maskedclamp
14055 Clamp the first input stream with the second input and third input stream.
14057 Returns the value of first stream to be between second input
14058 stream - @code{undershoot} and third input stream + @code{overshoot}.
14060 This filter accepts the following options:
14063 Default value is @code{0}.
14066 Default value is @code{0}.
14069 Set which planes will be processed as bitmap, unprocessed planes will be
14070 copied from first stream.
14071 By default value 0xf, all planes will be processed.
14074 @subsection Commands
14076 This filter supports the all above options as @ref{commands}.
14080 Merge the second and third input stream into output stream using absolute differences
14081 between second input stream and first input stream and absolute difference between
14082 third input stream and first input stream. The picked value will be from second input
14083 stream if second absolute difference is greater than first one or from third input stream
14086 This filter accepts the following options:
14089 Set which planes will be processed as bitmap, unprocessed planes will be
14090 copied from first stream.
14091 By default value 0xf, all planes will be processed.
14094 @subsection Commands
14096 This filter supports the all above options as @ref{commands}.
14098 @section maskedmerge
14100 Merge the first input stream with the second input stream using per pixel
14101 weights in the third input stream.
14103 A value of 0 in the third stream pixel component means that pixel component
14104 from first stream is returned unchanged, while maximum value (eg. 255 for
14105 8-bit videos) means that pixel component from second stream is returned
14106 unchanged. Intermediate values define the amount of merging between both
14107 input stream's pixel components.
14109 This filter accepts the following options:
14112 Set which planes will be processed as bitmap, unprocessed planes will be
14113 copied from first stream.
14114 By default value 0xf, all planes will be processed.
14117 @subsection Commands
14119 This filter supports the all above options as @ref{commands}.
14123 Merge the second and third input stream into output stream using absolute differences
14124 between second input stream and first input stream and absolute difference between
14125 third input stream and first input stream. The picked value will be from second input
14126 stream if second absolute difference is less than first one or from third input stream
14129 This filter accepts the following options:
14132 Set which planes will be processed as bitmap, unprocessed planes will be
14133 copied from first stream.
14134 By default value 0xf, all planes will be processed.
14137 @subsection Commands
14139 This filter supports the all above options as @ref{commands}.
14141 @section maskedthreshold
14142 Pick pixels comparing absolute difference of two video streams with fixed
14145 If absolute difference between pixel component of first and second video
14146 stream is equal or lower than user supplied threshold than pixel component
14147 from first video stream is picked, otherwise pixel component from second
14148 video stream is picked.
14150 This filter accepts the following options:
14153 Set threshold used when picking pixels from absolute difference from two input
14157 Set which planes will be processed as bitmap, unprocessed planes will be
14158 copied from second stream.
14159 By default value 0xf, all planes will be processed.
14162 @subsection Commands
14164 This filter supports the all above options as @ref{commands}.
14167 Create mask from input video.
14169 For example it is useful to create motion masks after @code{tblend} filter.
14171 This filter accepts the following options:
14175 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
14178 Set high threshold. Any pixel component higher than this value will be set to max value
14179 allowed for current pixel format.
14182 Set planes to filter, by default all available planes are filtered.
14185 Fill all frame pixels with this value.
14188 Set max average pixel value for frame. If sum of all pixel components is higher that this
14189 average, output frame will be completely filled with value set by @var{fill} option.
14190 Typically useful for scene changes when used in combination with @code{tblend} filter.
14195 Apply motion-compensation deinterlacing.
14197 It needs one field per frame as input and must thus be used together
14198 with yadif=1/3 or equivalent.
14200 This filter accepts the following options:
14203 Set the deinterlacing mode.
14205 It accepts one of the following values:
14210 use iterative motion estimation
14212 like @samp{slow}, but use multiple reference frames.
14214 Default value is @samp{fast}.
14217 Set the picture field parity assumed for the input video. It must be
14218 one of the following values:
14222 assume top field first
14224 assume bottom field first
14227 Default value is @samp{bff}.
14230 Set per-block quantization parameter (QP) used by the internal
14233 Higher values should result in a smoother motion vector field but less
14234 optimal individual vectors. Default value is 1.
14239 Pick median pixel from certain rectangle defined by radius.
14241 This filter accepts the following options:
14245 Set horizontal radius size. Default value is @code{1}.
14246 Allowed range is integer from 1 to 127.
14249 Set which planes to process. Default is @code{15}, which is all available planes.
14252 Set vertical radius size. Default value is @code{0}.
14253 Allowed range is integer from 0 to 127.
14254 If it is 0, value will be picked from horizontal @code{radius} option.
14257 Set median percentile. Default value is @code{0.5}.
14258 Default value of @code{0.5} will pick always median values, while @code{0} will pick
14259 minimum values, and @code{1} maximum values.
14262 @subsection Commands
14263 This filter supports same @ref{commands} as options.
14264 The command accepts the same syntax of the corresponding option.
14266 If the specified expression is not valid, it is kept at its current
14269 @section mergeplanes
14271 Merge color channel components from several video streams.
14273 The filter accepts up to 4 input streams, and merge selected input
14274 planes to the output video.
14276 This filter accepts the following options:
14279 Set input to output plane mapping. Default is @code{0}.
14281 The mappings is specified as a bitmap. It should be specified as a
14282 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
14283 mapping for the first plane of the output stream. 'A' sets the number of
14284 the input stream to use (from 0 to 3), and 'a' the plane number of the
14285 corresponding input to use (from 0 to 3). The rest of the mappings is
14286 similar, 'Bb' describes the mapping for the output stream second
14287 plane, 'Cc' describes the mapping for the output stream third plane and
14288 'Dd' describes the mapping for the output stream fourth plane.
14291 Set output pixel format. Default is @code{yuva444p}.
14294 @subsection Examples
14298 Merge three gray video streams of same width and height into single video stream:
14300 [a0][a1][a2]mergeplanes=0x001020:yuv444p
14304 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
14306 [a0][a1]mergeplanes=0x00010210:yuva444p
14310 Swap Y and A plane in yuva444p stream:
14312 format=yuva444p,mergeplanes=0x03010200:yuva444p
14316 Swap U and V plane in yuv420p stream:
14318 format=yuv420p,mergeplanes=0x000201:yuv420p
14322 Cast a rgb24 clip to yuv444p:
14324 format=rgb24,mergeplanes=0x000102:yuv444p
14330 Estimate and export motion vectors using block matching algorithms.
14331 Motion vectors are stored in frame side data to be used by other filters.
14333 This filter accepts the following options:
14336 Specify the motion estimation method. Accepts one of the following values:
14340 Exhaustive search algorithm.
14342 Three step search algorithm.
14344 Two dimensional logarithmic search algorithm.
14346 New three step search algorithm.
14348 Four step search algorithm.
14350 Diamond search algorithm.
14352 Hexagon-based search algorithm.
14354 Enhanced predictive zonal search algorithm.
14356 Uneven multi-hexagon search algorithm.
14358 Default value is @samp{esa}.
14361 Macroblock size. Default @code{16}.
14364 Search parameter. Default @code{7}.
14367 @section midequalizer
14369 Apply Midway Image Equalization effect using two video streams.
14371 Midway Image Equalization adjusts a pair of images to have the same
14372 histogram, while maintaining their dynamics as much as possible. It's
14373 useful for e.g. matching exposures from a pair of stereo cameras.
14375 This filter has two inputs and one output, which must be of same pixel format, but
14376 may be of different sizes. The output of filter is first input adjusted with
14377 midway histogram of both inputs.
14379 This filter accepts the following option:
14383 Set which planes to process. Default is @code{15}, which is all available planes.
14386 @section minterpolate
14388 Convert the video to specified frame rate using motion interpolation.
14390 This filter accepts the following options:
14393 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}.
14396 Motion interpolation mode. Following values are accepted:
14399 Duplicate previous or next frame for interpolating new ones.
14401 Blend source frames. Interpolated frame is mean of previous and next frames.
14403 Motion compensated interpolation. Following options are effective when this mode is selected:
14407 Motion compensation mode. Following values are accepted:
14410 Overlapped block motion compensation.
14412 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
14414 Default mode is @samp{obmc}.
14417 Motion estimation mode. Following values are accepted:
14420 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
14422 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
14424 Default mode is @samp{bilat}.
14427 The algorithm to be used for motion estimation. Following values are accepted:
14430 Exhaustive search algorithm.
14432 Three step search algorithm.
14434 Two dimensional logarithmic search algorithm.
14436 New three step search algorithm.
14438 Four step search algorithm.
14440 Diamond search algorithm.
14442 Hexagon-based search algorithm.
14444 Enhanced predictive zonal search algorithm.
14446 Uneven multi-hexagon search algorithm.
14448 Default algorithm is @samp{epzs}.
14451 Macroblock size. Default @code{16}.
14454 Motion estimation search parameter. Default @code{32}.
14457 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).
14462 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:
14465 Disable scene change detection.
14467 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
14469 Default method is @samp{fdiff}.
14471 @item scd_threshold
14472 Scene change detection threshold. Default is @code{10.}.
14477 Mix several video input streams into one video stream.
14479 A description of the accepted options follows.
14483 The number of inputs. If unspecified, it defaults to 2.
14486 Specify weight of each input video stream as sequence.
14487 Each weight is separated by space. If number of weights
14488 is smaller than number of @var{frames} last specified
14489 weight will be used for all remaining unset weights.
14492 Specify scale, if it is set it will be multiplied with sum
14493 of each weight multiplied with pixel values to give final destination
14494 pixel value. By default @var{scale} is auto scaled to sum of weights.
14497 Specify how end of stream is determined.
14500 The duration of the longest input. (default)
14503 The duration of the shortest input.
14506 The duration of the first input.
14510 @section mpdecimate
14512 Drop frames that do not differ greatly from the previous frame in
14513 order to reduce frame rate.
14515 The main use of this filter is for very-low-bitrate encoding
14516 (e.g. streaming over dialup modem), but it could in theory be used for
14517 fixing movies that were inverse-telecined incorrectly.
14519 A description of the accepted options follows.
14523 Set the maximum number of consecutive frames which can be dropped (if
14524 positive), or the minimum interval between dropped frames (if
14525 negative). If the value is 0, the frame is dropped disregarding the
14526 number of previous sequentially dropped frames.
14528 Default value is 0.
14533 Set the dropping threshold values.
14535 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
14536 represent actual pixel value differences, so a threshold of 64
14537 corresponds to 1 unit of difference for each pixel, or the same spread
14538 out differently over the block.
14540 A frame is a candidate for dropping if no 8x8 blocks differ by more
14541 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
14542 meaning the whole image) differ by more than a threshold of @option{lo}.
14544 Default value for @option{hi} is 64*12, default value for @option{lo} is
14545 64*5, and default value for @option{frac} is 0.33.
14551 Negate (invert) the input video.
14553 It accepts the following option:
14558 With value 1, it negates the alpha component, if present. Default value is 0.
14564 Denoise frames using Non-Local Means algorithm.
14566 Each pixel is adjusted by looking for other pixels with similar contexts. This
14567 context similarity is defined by comparing their surrounding patches of size
14568 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
14571 Note that the research area defines centers for patches, which means some
14572 patches will be made of pixels outside that research area.
14574 The filter accepts the following options.
14578 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
14581 Set patch size. Default is 7. Must be odd number in range [0, 99].
14584 Same as @option{p} but for chroma planes.
14586 The default value is @var{0} and means automatic.
14589 Set research size. Default is 15. Must be odd number in range [0, 99].
14592 Same as @option{r} but for chroma planes.
14594 The default value is @var{0} and means automatic.
14599 Deinterlace video using neural network edge directed interpolation.
14601 This filter accepts the following options:
14605 Mandatory option, without binary file filter can not work.
14606 Currently file can be found here:
14607 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
14610 Set which frames to deinterlace, by default it is @code{all}.
14611 Can be @code{all} or @code{interlaced}.
14614 Set mode of operation.
14616 Can be one of the following:
14620 Use frame flags, both fields.
14622 Use frame flags, single field.
14624 Use top field only.
14626 Use bottom field only.
14628 Use both fields, top first.
14630 Use both fields, bottom first.
14634 Set which planes to process, by default filter process all frames.
14637 Set size of local neighborhood around each pixel, used by the predictor neural
14640 Can be one of the following:
14653 Set the number of neurons in predictor neural network.
14654 Can be one of the following:
14665 Controls the number of different neural network predictions that are blended
14666 together to compute the final output value. Can be @code{fast}, default or
14670 Set which set of weights to use in the predictor.
14671 Can be one of the following:
14675 weights trained to minimize absolute error
14677 weights trained to minimize squared error
14681 Controls whether or not the prescreener neural network is used to decide
14682 which pixels should be processed by the predictor neural network and which
14683 can be handled by simple cubic interpolation.
14684 The prescreener is trained to know whether cubic interpolation will be
14685 sufficient for a pixel or whether it should be predicted by the predictor nn.
14686 The computational complexity of the prescreener nn is much less than that of
14687 the predictor nn. Since most pixels can be handled by cubic interpolation,
14688 using the prescreener generally results in much faster processing.
14689 The prescreener is pretty accurate, so the difference between using it and not
14690 using it is almost always unnoticeable.
14692 Can be one of the following:
14702 Default is @code{new}.
14705 @subsection Commands
14706 This filter supports same @ref{commands} as options, excluding @var{weights} option.
14710 Force libavfilter not to use any of the specified pixel formats for the
14711 input to the next filter.
14713 It accepts the following parameters:
14717 A '|'-separated list of pixel format names, such as
14718 pix_fmts=yuv420p|monow|rgb24".
14722 @subsection Examples
14726 Force libavfilter to use a format different from @var{yuv420p} for the
14727 input to the vflip filter:
14729 noformat=pix_fmts=yuv420p,vflip
14733 Convert the input video to any of the formats not contained in the list:
14735 noformat=yuv420p|yuv444p|yuv410p
14741 Add noise on video input frame.
14743 The filter accepts the following options:
14751 Set noise seed for specific pixel component or all pixel components in case
14752 of @var{all_seed}. Default value is @code{123457}.
14754 @item all_strength, alls
14755 @item c0_strength, c0s
14756 @item c1_strength, c1s
14757 @item c2_strength, c2s
14758 @item c3_strength, c3s
14759 Set noise strength for specific pixel component or all pixel components in case
14760 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
14762 @item all_flags, allf
14763 @item c0_flags, c0f
14764 @item c1_flags, c1f
14765 @item c2_flags, c2f
14766 @item c3_flags, c3f
14767 Set pixel component flags or set flags for all components if @var{all_flags}.
14768 Available values for component flags are:
14771 averaged temporal noise (smoother)
14773 mix random noise with a (semi)regular pattern
14775 temporal noise (noise pattern changes between frames)
14777 uniform noise (gaussian otherwise)
14781 @subsection Examples
14783 Add temporal and uniform noise to input video:
14785 noise=alls=20:allf=t+u
14790 Normalize RGB video (aka histogram stretching, contrast stretching).
14791 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
14793 For each channel of each frame, the filter computes the input range and maps
14794 it linearly to the user-specified output range. The output range defaults
14795 to the full dynamic range from pure black to pure white.
14797 Temporal smoothing can be used on the input range to reduce flickering (rapid
14798 changes in brightness) caused when small dark or bright objects enter or leave
14799 the scene. This is similar to the auto-exposure (automatic gain control) on a
14800 video camera, and, like a video camera, it may cause a period of over- or
14801 under-exposure of the video.
14803 The R,G,B channels can be normalized independently, which may cause some
14804 color shifting, or linked together as a single channel, which prevents
14805 color shifting. Linked normalization preserves hue. Independent normalization
14806 does not, so it can be used to remove some color casts. Independent and linked
14807 normalization can be combined in any ratio.
14809 The normalize filter accepts the following options:
14814 Colors which define the output range. The minimum input value is mapped to
14815 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
14816 The defaults are black and white respectively. Specifying white for
14817 @var{blackpt} and black for @var{whitept} will give color-inverted,
14818 normalized video. Shades of grey can be used to reduce the dynamic range
14819 (contrast). Specifying saturated colors here can create some interesting
14823 The number of previous frames to use for temporal smoothing. The input range
14824 of each channel is smoothed using a rolling average over the current frame
14825 and the @var{smoothing} previous frames. The default is 0 (no temporal
14829 Controls the ratio of independent (color shifting) channel normalization to
14830 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
14831 independent. Defaults to 1.0 (fully independent).
14834 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
14835 expensive no-op. Defaults to 1.0 (full strength).
14839 @subsection Commands
14840 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
14841 The command accepts the same syntax of the corresponding option.
14843 If the specified expression is not valid, it is kept at its current
14846 @subsection Examples
14848 Stretch video contrast to use the full dynamic range, with no temporal
14849 smoothing; may flicker depending on the source content:
14851 normalize=blackpt=black:whitept=white:smoothing=0
14854 As above, but with 50 frames of temporal smoothing; flicker should be
14855 reduced, depending on the source content:
14857 normalize=blackpt=black:whitept=white:smoothing=50
14860 As above, but with hue-preserving linked channel normalization:
14862 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
14865 As above, but with half strength:
14867 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
14870 Map the darkest input color to red, the brightest input color to cyan:
14872 normalize=blackpt=red:whitept=cyan
14877 Pass the video source unchanged to the output.
14880 Optical Character Recognition
14882 This filter uses Tesseract for optical character recognition. To enable
14883 compilation of this filter, you need to configure FFmpeg with
14884 @code{--enable-libtesseract}.
14886 It accepts the following options:
14890 Set datapath to tesseract data. Default is to use whatever was
14891 set at installation.
14894 Set language, default is "eng".
14897 Set character whitelist.
14900 Set character blacklist.
14903 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
14904 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
14908 Apply a video transform using libopencv.
14910 To enable this filter, install the libopencv library and headers and
14911 configure FFmpeg with @code{--enable-libopencv}.
14913 It accepts the following parameters:
14918 The name of the libopencv filter to apply.
14920 @item filter_params
14921 The parameters to pass to the libopencv filter. If not specified, the default
14922 values are assumed.
14926 Refer to the official libopencv documentation for more precise
14928 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
14930 Several libopencv filters are supported; see the following subsections.
14935 Dilate an image by using a specific structuring element.
14936 It corresponds to the libopencv function @code{cvDilate}.
14938 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
14940 @var{struct_el} represents a structuring element, and has the syntax:
14941 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
14943 @var{cols} and @var{rows} represent the number of columns and rows of
14944 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
14945 point, and @var{shape} the shape for the structuring element. @var{shape}
14946 must be "rect", "cross", "ellipse", or "custom".
14948 If the value for @var{shape} is "custom", it must be followed by a
14949 string of the form "=@var{filename}". The file with name
14950 @var{filename} is assumed to represent a binary image, with each
14951 printable character corresponding to a bright pixel. When a custom
14952 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
14953 or columns and rows of the read file are assumed instead.
14955 The default value for @var{struct_el} is "3x3+0x0/rect".
14957 @var{nb_iterations} specifies the number of times the transform is
14958 applied to the image, and defaults to 1.
14962 # Use the default values
14965 # Dilate using a structuring element with a 5x5 cross, iterating two times
14966 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
14968 # Read the shape from the file diamond.shape, iterating two times.
14969 # The file diamond.shape may contain a pattern of characters like this
14975 # The specified columns and rows are ignored
14976 # but the anchor point coordinates are not
14977 ocv=dilate:0x0+2x2/custom=diamond.shape|2
14982 Erode an image by using a specific structuring element.
14983 It corresponds to the libopencv function @code{cvErode}.
14985 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
14986 with the same syntax and semantics as the @ref{dilate} filter.
14990 Smooth the input video.
14992 The filter takes the following parameters:
14993 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
14995 @var{type} is the type of smooth filter to apply, and must be one of
14996 the following values: "blur", "blur_no_scale", "median", "gaussian",
14997 or "bilateral". The default value is "gaussian".
14999 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
15000 depends on the smooth type. @var{param1} and
15001 @var{param2} accept integer positive values or 0. @var{param3} and
15002 @var{param4} accept floating point values.
15004 The default value for @var{param1} is 3. The default value for the
15005 other parameters is 0.
15007 These parameters correspond to the parameters assigned to the
15008 libopencv function @code{cvSmooth}.
15010 @section oscilloscope
15012 2D Video Oscilloscope.
15014 Useful to measure spatial impulse, step responses, chroma delays, etc.
15016 It accepts the following parameters:
15020 Set scope center x position.
15023 Set scope center y position.
15026 Set scope size, relative to frame diagonal.
15029 Set scope tilt/rotation.
15035 Set trace center x position.
15038 Set trace center y position.
15041 Set trace width, relative to width of frame.
15044 Set trace height, relative to height of frame.
15047 Set which components to trace. By default it traces first three components.
15050 Draw trace grid. By default is enabled.
15053 Draw some statistics. By default is enabled.
15056 Draw scope. By default is enabled.
15059 @subsection Commands
15060 This filter supports same @ref{commands} as options.
15061 The command accepts the same syntax of the corresponding option.
15063 If the specified expression is not valid, it is kept at its current
15066 @subsection Examples
15070 Inspect full first row of video frame.
15072 oscilloscope=x=0.5:y=0:s=1
15076 Inspect full last row of video frame.
15078 oscilloscope=x=0.5:y=1:s=1
15082 Inspect full 5th line of video frame of height 1080.
15084 oscilloscope=x=0.5:y=5/1080:s=1
15088 Inspect full last column of video frame.
15090 oscilloscope=x=1:y=0.5:s=1:t=1
15098 Overlay one video on top of another.
15100 It takes two inputs and has one output. The first input is the "main"
15101 video on which the second input is overlaid.
15103 It accepts the following parameters:
15105 A description of the accepted options follows.
15110 Set the expression for the x and y coordinates of the overlaid video
15111 on the main video. Default value is "0" for both expressions. In case
15112 the expression is invalid, it is set to a huge value (meaning that the
15113 overlay will not be displayed within the output visible area).
15116 See @ref{framesync}.
15119 Set when the expressions for @option{x}, and @option{y} are evaluated.
15121 It accepts the following values:
15124 only evaluate expressions once during the filter initialization or
15125 when a command is processed
15128 evaluate expressions for each incoming frame
15131 Default value is @samp{frame}.
15134 See @ref{framesync}.
15137 Set the format for the output video.
15139 It accepts the following values:
15142 force YUV420 output
15145 force YUV420p10 output
15148 force YUV422 output
15151 force YUV422p10 output
15154 force YUV444 output
15157 force packed RGB output
15160 force planar RGB output
15163 automatically pick format
15166 Default value is @samp{yuv420}.
15169 See @ref{framesync}.
15172 Set format of alpha of the overlaid video, it can be @var{straight} or
15173 @var{premultiplied}. Default is @var{straight}.
15176 The @option{x}, and @option{y} expressions can contain the following
15182 The main input width and height.
15186 The overlay input width and height.
15190 The computed values for @var{x} and @var{y}. They are evaluated for
15195 horizontal and vertical chroma subsample values of the output
15196 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
15200 the number of input frame, starting from 0
15203 the position in the file of the input frame, NAN if unknown
15206 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
15210 This filter also supports the @ref{framesync} options.
15212 Note that the @var{n}, @var{pos}, @var{t} variables are available only
15213 when evaluation is done @emph{per frame}, and will evaluate to NAN
15214 when @option{eval} is set to @samp{init}.
15216 Be aware that frames are taken from each input video in timestamp
15217 order, hence, if their initial timestamps differ, it is a good idea
15218 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
15219 have them begin in the same zero timestamp, as the example for
15220 the @var{movie} filter does.
15222 You can chain together more overlays but you should test the
15223 efficiency of such approach.
15225 @subsection Commands
15227 This filter supports the following commands:
15231 Modify the x and y of the overlay input.
15232 The command accepts the same syntax of the corresponding option.
15234 If the specified expression is not valid, it is kept at its current
15238 @subsection Examples
15242 Draw the overlay at 10 pixels from the bottom right corner of the main
15245 overlay=main_w-overlay_w-10:main_h-overlay_h-10
15248 Using named options the example above becomes:
15250 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
15254 Insert a transparent PNG logo in the bottom left corner of the input,
15255 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
15257 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
15261 Insert 2 different transparent PNG logos (second logo on bottom
15262 right corner) using the @command{ffmpeg} tool:
15264 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
15268 Add a transparent color layer on top of the main video; @code{WxH}
15269 must specify the size of the main input to the overlay filter:
15271 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
15275 Play an original video and a filtered version (here with the deshake
15276 filter) side by side using the @command{ffplay} tool:
15278 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
15281 The above command is the same as:
15283 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
15287 Make a sliding overlay appearing from the left to the right top part of the
15288 screen starting since time 2:
15290 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
15294 Compose output by putting two input videos side to side:
15296 ffmpeg -i left.avi -i right.avi -filter_complex "
15297 nullsrc=size=200x100 [background];
15298 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
15299 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
15300 [background][left] overlay=shortest=1 [background+left];
15301 [background+left][right] overlay=shortest=1:x=100 [left+right]
15306 Mask 10-20 seconds of a video by applying the delogo filter to a section
15308 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
15309 -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]'
15314 Chain several overlays in cascade:
15316 nullsrc=s=200x200 [bg];
15317 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
15318 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
15319 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
15320 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
15321 [in3] null, [mid2] overlay=100:100 [out0]
15326 @anchor{overlay_cuda}
15327 @section overlay_cuda
15329 Overlay one video on top of another.
15331 This is the CUDA variant of the @ref{overlay} filter.
15332 It only accepts CUDA frames. The underlying input pixel formats have to match.
15334 It takes two inputs and has one output. The first input is the "main"
15335 video on which the second input is overlaid.
15337 It accepts the following parameters:
15342 Set the x and y coordinates of the overlaid video on the main video.
15343 Default value is "0" for both expressions.
15346 See @ref{framesync}.
15349 See @ref{framesync}.
15352 See @ref{framesync}.
15356 This filter also supports the @ref{framesync} options.
15360 Apply Overcomplete Wavelet denoiser.
15362 The filter accepts the following options:
15368 Larger depth values will denoise lower frequency components more, but
15369 slow down filtering.
15371 Must be an int in the range 8-16, default is @code{8}.
15373 @item luma_strength, ls
15376 Must be a double value in the range 0-1000, default is @code{1.0}.
15378 @item chroma_strength, cs
15379 Set chroma strength.
15381 Must be a double value in the range 0-1000, default is @code{1.0}.
15387 Add paddings to the input image, and place the original input at the
15388 provided @var{x}, @var{y} coordinates.
15390 It accepts the following parameters:
15395 Specify an expression for the size of the output image with the
15396 paddings added. If the value for @var{width} or @var{height} is 0, the
15397 corresponding input size is used for the output.
15399 The @var{width} expression can reference the value set by the
15400 @var{height} expression, and vice versa.
15402 The default value of @var{width} and @var{height} is 0.
15406 Specify the offsets to place the input image at within the padded area,
15407 with respect to the top/left border of the output image.
15409 The @var{x} expression can reference the value set by the @var{y}
15410 expression, and vice versa.
15412 The default value of @var{x} and @var{y} is 0.
15414 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
15415 so the input image is centered on the padded area.
15418 Specify the color of the padded area. For the syntax of this option,
15419 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
15420 manual,ffmpeg-utils}.
15422 The default value of @var{color} is "black".
15425 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
15427 It accepts the following values:
15431 Only evaluate expressions once during the filter initialization or when
15432 a command is processed.
15435 Evaluate expressions for each incoming frame.
15439 Default value is @samp{init}.
15442 Pad to aspect instead to a resolution.
15446 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
15447 options are expressions containing the following constants:
15452 The input video width and height.
15456 These are the same as @var{in_w} and @var{in_h}.
15460 The output width and height (the size of the padded area), as
15461 specified by the @var{width} and @var{height} expressions.
15465 These are the same as @var{out_w} and @var{out_h}.
15469 The x and y offsets as specified by the @var{x} and @var{y}
15470 expressions, or NAN if not yet specified.
15473 same as @var{iw} / @var{ih}
15476 input sample aspect ratio
15479 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
15483 The horizontal and vertical chroma subsample values. For example for the
15484 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15487 @subsection Examples
15491 Add paddings with the color "violet" to the input video. The output video
15492 size is 640x480, and the top-left corner of the input video is placed at
15495 pad=640:480:0:40:violet
15498 The example above is equivalent to the following command:
15500 pad=width=640:height=480:x=0:y=40:color=violet
15504 Pad the input to get an output with dimensions increased by 3/2,
15505 and put the input video at the center of the padded area:
15507 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
15511 Pad the input to get a squared output with size equal to the maximum
15512 value between the input width and height, and put the input video at
15513 the center of the padded area:
15515 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
15519 Pad the input to get a final w/h ratio of 16:9:
15521 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
15525 In case of anamorphic video, in order to set the output display aspect
15526 correctly, it is necessary to use @var{sar} in the expression,
15527 according to the relation:
15529 (ih * X / ih) * sar = output_dar
15530 X = output_dar / sar
15533 Thus the previous example needs to be modified to:
15535 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
15539 Double the output size and put the input video in the bottom-right
15540 corner of the output padded area:
15542 pad="2*iw:2*ih:ow-iw:oh-ih"
15546 @anchor{palettegen}
15547 @section palettegen
15549 Generate one palette for a whole video stream.
15551 It accepts the following options:
15555 Set the maximum number of colors to quantize in the palette.
15556 Note: the palette will still contain 256 colors; the unused palette entries
15559 @item reserve_transparent
15560 Create a palette of 255 colors maximum and reserve the last one for
15561 transparency. Reserving the transparency color is useful for GIF optimization.
15562 If not set, the maximum of colors in the palette will be 256. You probably want
15563 to disable this option for a standalone image.
15566 @item transparency_color
15567 Set the color that will be used as background for transparency.
15570 Set statistics mode.
15572 It accepts the following values:
15575 Compute full frame histograms.
15577 Compute histograms only for the part that differs from previous frame. This
15578 might be relevant to give more importance to the moving part of your input if
15579 the background is static.
15581 Compute new histogram for each frame.
15584 Default value is @var{full}.
15587 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
15588 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
15589 color quantization of the palette. This information is also visible at
15590 @var{info} logging level.
15592 @subsection Examples
15596 Generate a representative palette of a given video using @command{ffmpeg}:
15598 ffmpeg -i input.mkv -vf palettegen palette.png
15602 @section paletteuse
15604 Use a palette to downsample an input video stream.
15606 The filter takes two inputs: one video stream and a palette. The palette must
15607 be a 256 pixels image.
15609 It accepts the following options:
15613 Select dithering mode. Available algorithms are:
15616 Ordered 8x8 bayer dithering (deterministic)
15618 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
15619 Note: this dithering is sometimes considered "wrong" and is included as a
15621 @item floyd_steinberg
15622 Floyd and Steingberg dithering (error diffusion)
15624 Frankie Sierra dithering v2 (error diffusion)
15626 Frankie Sierra dithering v2 "Lite" (error diffusion)
15629 Default is @var{sierra2_4a}.
15632 When @var{bayer} dithering is selected, this option defines the scale of the
15633 pattern (how much the crosshatch pattern is visible). A low value means more
15634 visible pattern for less banding, and higher value means less visible pattern
15635 at the cost of more banding.
15637 The option must be an integer value in the range [0,5]. Default is @var{2}.
15640 If set, define the zone to process
15644 Only the changing rectangle will be reprocessed. This is similar to GIF
15645 cropping/offsetting compression mechanism. This option can be useful for speed
15646 if only a part of the image is changing, and has use cases such as limiting the
15647 scope of the error diffusal @option{dither} to the rectangle that bounds the
15648 moving scene (it leads to more deterministic output if the scene doesn't change
15649 much, and as a result less moving noise and better GIF compression).
15652 Default is @var{none}.
15655 Take new palette for each output frame.
15657 @item alpha_threshold
15658 Sets the alpha threshold for transparency. Alpha values above this threshold
15659 will be treated as completely opaque, and values below this threshold will be
15660 treated as completely transparent.
15662 The option must be an integer value in the range [0,255]. Default is @var{128}.
15665 @subsection Examples
15669 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
15670 using @command{ffmpeg}:
15672 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
15676 @section perspective
15678 Correct perspective of video not recorded perpendicular to the screen.
15680 A description of the accepted parameters follows.
15691 Set coordinates expression for top left, top right, bottom left and bottom right corners.
15692 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
15693 If the @code{sense} option is set to @code{source}, then the specified points will be sent
15694 to the corners of the destination. If the @code{sense} option is set to @code{destination},
15695 then the corners of the source will be sent to the specified coordinates.
15697 The expressions can use the following variables:
15702 the width and height of video frame.
15706 Output frame count.
15709 @item interpolation
15710 Set interpolation for perspective correction.
15712 It accepts the following values:
15718 Default value is @samp{linear}.
15721 Set interpretation of coordinate options.
15723 It accepts the following values:
15727 Send point in the source specified by the given coordinates to
15728 the corners of the destination.
15730 @item 1, destination
15732 Send the corners of the source to the point in the destination specified
15733 by the given coordinates.
15735 Default value is @samp{source}.
15739 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
15741 It accepts the following values:
15744 only evaluate expressions once during the filter initialization or
15745 when a command is processed
15748 evaluate expressions for each incoming frame
15751 Default value is @samp{init}.
15756 Delay interlaced video by one field time so that the field order changes.
15758 The intended use is to fix PAL movies that have been captured with the
15759 opposite field order to the film-to-video transfer.
15761 A description of the accepted parameters follows.
15767 It accepts the following values:
15770 Capture field order top-first, transfer bottom-first.
15771 Filter will delay the bottom field.
15774 Capture field order bottom-first, transfer top-first.
15775 Filter will delay the top field.
15778 Capture and transfer with the same field order. This mode only exists
15779 for the documentation of the other options to refer to, but if you
15780 actually select it, the filter will faithfully do nothing.
15783 Capture field order determined automatically by field flags, transfer
15785 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
15786 basis using field flags. If no field information is available,
15787 then this works just like @samp{u}.
15790 Capture unknown or varying, transfer opposite.
15791 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
15792 analyzing the images and selecting the alternative that produces best
15793 match between the fields.
15796 Capture top-first, transfer unknown or varying.
15797 Filter selects among @samp{t} and @samp{p} using image analysis.
15800 Capture bottom-first, transfer unknown or varying.
15801 Filter selects among @samp{b} and @samp{p} using image analysis.
15804 Capture determined by field flags, transfer unknown or varying.
15805 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
15806 image analysis. If no field information is available, then this works just
15807 like @samp{U}. This is the default mode.
15810 Both capture and transfer unknown or varying.
15811 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
15815 @subsection Commands
15817 This filter supports the all above options as @ref{commands}.
15819 @section photosensitivity
15820 Reduce various flashes in video, so to help users with epilepsy.
15822 It accepts the following options:
15825 Set how many frames to use when filtering. Default is 30.
15828 Set detection threshold factor. Default is 1.
15832 Set how many pixels to skip when sampling frames. Default is 1.
15833 Allowed range is from 1 to 1024.
15836 Leave frames unchanged. Default is disabled.
15839 @section pixdesctest
15841 Pixel format descriptor test filter, mainly useful for internal
15842 testing. The output video should be equal to the input video.
15846 format=monow, pixdesctest
15849 can be used to test the monowhite pixel format descriptor definition.
15853 Display sample values of color channels. Mainly useful for checking color
15854 and levels. Minimum supported resolution is 640x480.
15856 The filters accept the following options:
15860 Set scope X position, relative offset on X axis.
15863 Set scope Y position, relative offset on Y axis.
15872 Set window opacity. This window also holds statistics about pixel area.
15875 Set window X position, relative offset on X axis.
15878 Set window Y position, relative offset on Y axis.
15883 Enable the specified chain of postprocessing subfilters using libpostproc. This
15884 library should be automatically selected with a GPL build (@code{--enable-gpl}).
15885 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
15886 Each subfilter and some options have a short and a long name that can be used
15887 interchangeably, i.e. dr/dering are the same.
15889 The filters accept the following options:
15893 Set postprocessing subfilters string.
15896 All subfilters share common options to determine their scope:
15900 Honor the quality commands for this subfilter.
15903 Do chrominance filtering, too (default).
15906 Do luminance filtering only (no chrominance).
15909 Do chrominance filtering only (no luminance).
15912 These options can be appended after the subfilter name, separated by a '|'.
15914 Available subfilters are:
15917 @item hb/hdeblock[|difference[|flatness]]
15918 Horizontal deblocking filter
15921 Difference factor where higher values mean more deblocking (default: @code{32}).
15923 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15926 @item vb/vdeblock[|difference[|flatness]]
15927 Vertical deblocking filter
15930 Difference factor where higher values mean more deblocking (default: @code{32}).
15932 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15935 @item ha/hadeblock[|difference[|flatness]]
15936 Accurate horizontal deblocking filter
15939 Difference factor where higher values mean more deblocking (default: @code{32}).
15941 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15944 @item va/vadeblock[|difference[|flatness]]
15945 Accurate vertical deblocking filter
15948 Difference factor where higher values mean more deblocking (default: @code{32}).
15950 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15954 The horizontal and vertical deblocking filters share the difference and
15955 flatness values so you cannot set different horizontal and vertical
15959 @item h1/x1hdeblock
15960 Experimental horizontal deblocking filter
15962 @item v1/x1vdeblock
15963 Experimental vertical deblocking filter
15968 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
15971 larger -> stronger filtering
15973 larger -> stronger filtering
15975 larger -> stronger filtering
15978 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
15981 Stretch luminance to @code{0-255}.
15984 @item lb/linblenddeint
15985 Linear blend deinterlacing filter that deinterlaces the given block by
15986 filtering all lines with a @code{(1 2 1)} filter.
15988 @item li/linipoldeint
15989 Linear interpolating deinterlacing filter that deinterlaces the given block by
15990 linearly interpolating every second line.
15992 @item ci/cubicipoldeint
15993 Cubic interpolating deinterlacing filter deinterlaces the given block by
15994 cubically interpolating every second line.
15996 @item md/mediandeint
15997 Median deinterlacing filter that deinterlaces the given block by applying a
15998 median filter to every second line.
16000 @item fd/ffmpegdeint
16001 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
16002 second line with a @code{(-1 4 2 4 -1)} filter.
16005 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
16006 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
16008 @item fq/forceQuant[|quantizer]
16009 Overrides the quantizer table from the input with the constant quantizer you
16017 Default pp filter combination (@code{hb|a,vb|a,dr|a})
16020 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
16023 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
16026 @subsection Examples
16030 Apply horizontal and vertical deblocking, deringing and automatic
16031 brightness/contrast:
16037 Apply default filters without brightness/contrast correction:
16043 Apply default filters and temporal denoiser:
16045 pp=default/tmpnoise|1|2|3
16049 Apply deblocking on luminance only, and switch vertical deblocking on or off
16050 automatically depending on available CPU time:
16057 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
16058 similar to spp = 6 with 7 point DCT, where only the center sample is
16061 The filter accepts the following options:
16065 Force a constant quantization parameter. It accepts an integer in range
16066 0 to 63. If not set, the filter will use the QP from the video stream
16070 Set thresholding mode. Available modes are:
16074 Set hard thresholding.
16076 Set soft thresholding (better de-ringing effect, but likely blurrier).
16078 Set medium thresholding (good results, default).
16082 @section premultiply
16083 Apply alpha premultiply effect to input video stream using first plane
16084 of second stream as alpha.
16086 Both streams must have same dimensions and same pixel format.
16088 The filter accepts the following option:
16092 Set which planes will be processed, unprocessed planes will be copied.
16093 By default value 0xf, all planes will be processed.
16096 Do not require 2nd input for processing, instead use alpha plane from input stream.
16100 Apply prewitt operator to input video stream.
16102 The filter accepts the following option:
16106 Set which planes will be processed, unprocessed planes will be copied.
16107 By default value 0xf, all planes will be processed.
16110 Set value which will be multiplied with filtered result.
16113 Set value which will be added to filtered result.
16116 @subsection Commands
16118 This filter supports the all above options as @ref{commands}.
16120 @section pseudocolor
16122 Alter frame colors in video with pseudocolors.
16124 This filter accepts the following options:
16128 set pixel first component expression
16131 set pixel second component expression
16134 set pixel third component expression
16137 set pixel fourth component expression, corresponds to the alpha component
16140 set component to use as base for altering colors
16143 Each of them specifies the expression to use for computing the lookup table for
16144 the corresponding pixel component values.
16146 The expressions can contain the following constants and functions:
16151 The input width and height.
16154 The input value for the pixel component.
16156 @item ymin, umin, vmin, amin
16157 The minimum allowed component value.
16159 @item ymax, umax, vmax, amax
16160 The maximum allowed component value.
16163 All expressions default to "val".
16165 @subsection Examples
16169 Change too high luma values to gradient:
16171 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'"
16177 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
16178 Ratio) between two input videos.
16180 This filter takes in input two input videos, the first input is
16181 considered the "main" source and is passed unchanged to the
16182 output. The second input is used as a "reference" video for computing
16185 Both video inputs must have the same resolution and pixel format for
16186 this filter to work correctly. Also it assumes that both inputs
16187 have the same number of frames, which are compared one by one.
16189 The obtained average PSNR is printed through the logging system.
16191 The filter stores the accumulated MSE (mean squared error) of each
16192 frame, and at the end of the processing it is averaged across all frames
16193 equally, and the following formula is applied to obtain the PSNR:
16196 PSNR = 10*log10(MAX^2/MSE)
16199 Where MAX is the average of the maximum values of each component of the
16202 The description of the accepted parameters follows.
16205 @item stats_file, f
16206 If specified the filter will use the named file to save the PSNR of
16207 each individual frame. When filename equals "-" the data is sent to
16210 @item stats_version
16211 Specifies which version of the stats file format to use. Details of
16212 each format are written below.
16213 Default value is 1.
16215 @item stats_add_max
16216 Determines whether the max value is output to the stats log.
16217 Default value is 0.
16218 Requires stats_version >= 2. If this is set and stats_version < 2,
16219 the filter will return an error.
16222 This filter also supports the @ref{framesync} options.
16224 The file printed if @var{stats_file} is selected, contains a sequence of
16225 key/value pairs of the form @var{key}:@var{value} for each compared
16228 If a @var{stats_version} greater than 1 is specified, a header line precedes
16229 the list of per-frame-pair stats, with key value pairs following the frame
16230 format with the following parameters:
16233 @item psnr_log_version
16234 The version of the log file format. Will match @var{stats_version}.
16237 A comma separated list of the per-frame-pair parameters included in
16241 A description of each shown per-frame-pair parameter follows:
16245 sequential number of the input frame, starting from 1
16248 Mean Square Error pixel-by-pixel average difference of the compared
16249 frames, averaged over all the image components.
16251 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
16252 Mean Square Error pixel-by-pixel average difference of the compared
16253 frames for the component specified by the suffix.
16255 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
16256 Peak Signal to Noise ratio of the compared frames for the component
16257 specified by the suffix.
16259 @item max_avg, max_y, max_u, max_v
16260 Maximum allowed value for each channel, and average over all
16264 @subsection Examples
16269 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
16270 [main][ref] psnr="stats_file=stats.log" [out]
16273 On this example the input file being processed is compared with the
16274 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
16275 is stored in @file{stats.log}.
16278 Another example with different containers:
16280 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 -
16287 Pulldown reversal (inverse telecine) filter, capable of handling mixed
16288 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
16291 The pullup filter is designed to take advantage of future context in making
16292 its decisions. This filter is stateless in the sense that it does not lock
16293 onto a pattern to follow, but it instead looks forward to the following
16294 fields in order to identify matches and rebuild progressive frames.
16296 To produce content with an even framerate, insert the fps filter after
16297 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
16298 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
16300 The filter accepts the following options:
16307 These options set the amount of "junk" to ignore at the left, right, top, and
16308 bottom of the image, respectively. Left and right are in units of 8 pixels,
16309 while top and bottom are in units of 2 lines.
16310 The default is 8 pixels on each side.
16313 Set the strict breaks. Setting this option to 1 will reduce the chances of
16314 filter generating an occasional mismatched frame, but it may also cause an
16315 excessive number of frames to be dropped during high motion sequences.
16316 Conversely, setting it to -1 will make filter match fields more easily.
16317 This may help processing of video where there is slight blurring between
16318 the fields, but may also cause there to be interlaced frames in the output.
16319 Default value is @code{0}.
16322 Set the metric plane to use. It accepts the following values:
16328 Use chroma blue plane.
16331 Use chroma red plane.
16334 This option may be set to use chroma plane instead of the default luma plane
16335 for doing filter's computations. This may improve accuracy on very clean
16336 source material, but more likely will decrease accuracy, especially if there
16337 is chroma noise (rainbow effect) or any grayscale video.
16338 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
16339 load and make pullup usable in realtime on slow machines.
16342 For best results (without duplicated frames in the output file) it is
16343 necessary to change the output frame rate. For example, to inverse
16344 telecine NTSC input:
16346 ffmpeg -i input -vf pullup -r 24000/1001 ...
16351 Change video quantization parameters (QP).
16353 The filter accepts the following option:
16357 Set expression for quantization parameter.
16360 The expression is evaluated through the eval API and can contain, among others,
16361 the following constants:
16365 1 if index is not 129, 0 otherwise.
16368 Sequential index starting from -129 to 128.
16371 @subsection Examples
16375 Some equation like:
16383 Flush video frames from internal cache of frames into a random order.
16384 No frame is discarded.
16385 Inspired by @ref{frei0r} nervous filter.
16389 Set size in number of frames of internal cache, in range from @code{2} to
16390 @code{512}. Default is @code{30}.
16393 Set seed for random number generator, must be an integer included between
16394 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
16395 less than @code{0}, the filter will try to use a good random seed on a
16399 @section readeia608
16401 Read closed captioning (EIA-608) information from the top lines of a video frame.
16403 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
16404 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
16405 with EIA-608 data (starting from 0). A description of each metadata value follows:
16408 @item lavfi.readeia608.X.cc
16409 The two bytes stored as EIA-608 data (printed in hexadecimal).
16411 @item lavfi.readeia608.X.line
16412 The number of the line on which the EIA-608 data was identified and read.
16415 This filter accepts the following options:
16419 Set the line to start scanning for EIA-608 data. Default is @code{0}.
16422 Set the line to end scanning for EIA-608 data. Default is @code{29}.
16425 Set the ratio of width reserved for sync code detection.
16426 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
16429 Enable checking the parity bit. In the event of a parity error, the filter will output
16430 @code{0x00} for that character. Default is false.
16433 Lowpass lines prior to further processing. Default is enabled.
16436 @subsection Commands
16438 This filter supports the all above options as @ref{commands}.
16440 @subsection Examples
16444 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
16446 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
16452 Read vertical interval timecode (VITC) information from the top lines of a
16455 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
16456 timecode value, if a valid timecode has been detected. Further metadata key
16457 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
16458 timecode data has been found or not.
16460 This filter accepts the following options:
16464 Set the maximum number of lines to scan for VITC data. If the value is set to
16465 @code{-1} the full video frame is scanned. Default is @code{45}.
16468 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
16469 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
16472 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
16473 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
16476 @subsection Examples
16480 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
16481 draw @code{--:--:--:--} as a placeholder:
16483 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
16489 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
16491 Destination pixel at position (X, Y) will be picked from source (x, y) position
16492 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
16493 value for pixel will be used for destination pixel.
16495 Xmap and Ymap input video streams must be of same dimensions. Output video stream
16496 will have Xmap/Ymap video stream dimensions.
16497 Xmap and Ymap input video streams are 16bit depth, single channel.
16501 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
16502 Default is @code{color}.
16505 Specify the color of the unmapped pixels. For the syntax of this option,
16506 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
16507 manual,ffmpeg-utils}. Default color is @code{black}.
16510 @section removegrain
16512 The removegrain filter is a spatial denoiser for progressive video.
16516 Set mode for the first plane.
16519 Set mode for the second plane.
16522 Set mode for the third plane.
16525 Set mode for the fourth plane.
16528 Range of mode is from 0 to 24. Description of each mode follows:
16532 Leave input plane unchanged. Default.
16535 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
16538 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
16541 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
16544 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
16545 This is equivalent to a median filter.
16548 Line-sensitive clipping giving the minimal change.
16551 Line-sensitive clipping, intermediate.
16554 Line-sensitive clipping, intermediate.
16557 Line-sensitive clipping, intermediate.
16560 Line-sensitive clipping on a line where the neighbours pixels are the closest.
16563 Replaces the target pixel with the closest neighbour.
16566 [1 2 1] horizontal and vertical kernel blur.
16572 Bob mode, interpolates top field from the line where the neighbours
16573 pixels are the closest.
16576 Bob mode, interpolates bottom field from the line where the neighbours
16577 pixels are the closest.
16580 Bob mode, interpolates top field. Same as 13 but with a more complicated
16581 interpolation formula.
16584 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
16585 interpolation formula.
16588 Clips the pixel with the minimum and maximum of respectively the maximum and
16589 minimum of each pair of opposite neighbour pixels.
16592 Line-sensitive clipping using opposite neighbours whose greatest distance from
16593 the current pixel is minimal.
16596 Replaces the pixel with the average of its 8 neighbours.
16599 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
16602 Clips pixels using the averages of opposite neighbour.
16605 Same as mode 21 but simpler and faster.
16608 Small edge and halo removal, but reputed useless.
16614 @section removelogo
16616 Suppress a TV station logo, using an image file to determine which
16617 pixels comprise the logo. It works by filling in the pixels that
16618 comprise the logo with neighboring pixels.
16620 The filter accepts the following options:
16624 Set the filter bitmap file, which can be any image format supported by
16625 libavformat. The width and height of the image file must match those of the
16626 video stream being processed.
16629 Pixels in the provided bitmap image with a value of zero are not
16630 considered part of the logo, non-zero pixels are considered part of
16631 the logo. If you use white (255) for the logo and black (0) for the
16632 rest, you will be safe. For making the filter bitmap, it is
16633 recommended to take a screen capture of a black frame with the logo
16634 visible, and then using a threshold filter followed by the erode
16635 filter once or twice.
16637 If needed, little splotches can be fixed manually. Remember that if
16638 logo pixels are not covered, the filter quality will be much
16639 reduced. Marking too many pixels as part of the logo does not hurt as
16640 much, but it will increase the amount of blurring needed to cover over
16641 the image and will destroy more information than necessary, and extra
16642 pixels will slow things down on a large logo.
16644 @section repeatfields
16646 This filter uses the repeat_field flag from the Video ES headers and hard repeats
16647 fields based on its value.
16651 Reverse a video clip.
16653 Warning: This filter requires memory to buffer the entire clip, so trimming
16656 @subsection Examples
16660 Take the first 5 seconds of a clip, and reverse it.
16667 Shift R/G/B/A pixels horizontally and/or vertically.
16669 The filter accepts the following options:
16672 Set amount to shift red horizontally.
16674 Set amount to shift red vertically.
16676 Set amount to shift green horizontally.
16678 Set amount to shift green vertically.
16680 Set amount to shift blue horizontally.
16682 Set amount to shift blue vertically.
16684 Set amount to shift alpha horizontally.
16686 Set amount to shift alpha vertically.
16688 Set edge mode, can be @var{smear}, default, or @var{warp}.
16691 @subsection Commands
16693 This filter supports the all above options as @ref{commands}.
16696 Apply roberts cross operator to input video stream.
16698 The filter accepts the following option:
16702 Set which planes will be processed, unprocessed planes will be copied.
16703 By default value 0xf, all planes will be processed.
16706 Set value which will be multiplied with filtered result.
16709 Set value which will be added to filtered result.
16712 @subsection Commands
16714 This filter supports the all above options as @ref{commands}.
16718 Rotate video by an arbitrary angle expressed in radians.
16720 The filter accepts the following options:
16722 A description of the optional parameters follows.
16725 Set an expression for the angle by which to rotate the input video
16726 clockwise, expressed as a number of radians. A negative value will
16727 result in a counter-clockwise rotation. By default it is set to "0".
16729 This expression is evaluated for each frame.
16732 Set the output width expression, default value is "iw".
16733 This expression is evaluated just once during configuration.
16736 Set the output height expression, default value is "ih".
16737 This expression is evaluated just once during configuration.
16740 Enable bilinear interpolation if set to 1, a value of 0 disables
16741 it. Default value is 1.
16744 Set the color used to fill the output area not covered by the rotated
16745 image. For the general syntax of this option, check the
16746 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
16747 If the special value "none" is selected then no
16748 background is printed (useful for example if the background is never shown).
16750 Default value is "black".
16753 The expressions for the angle and the output size can contain the
16754 following constants and functions:
16758 sequential number of the input frame, starting from 0. It is always NAN
16759 before the first frame is filtered.
16762 time in seconds of the input frame, it is set to 0 when the filter is
16763 configured. It is always NAN before the first frame is filtered.
16767 horizontal and vertical chroma subsample values. For example for the
16768 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16772 the input video width and height
16776 the output width and height, that is the size of the padded area as
16777 specified by the @var{width} and @var{height} expressions
16781 the minimal width/height required for completely containing the input
16782 video rotated by @var{a} radians.
16784 These are only available when computing the @option{out_w} and
16785 @option{out_h} expressions.
16788 @subsection Examples
16792 Rotate the input by PI/6 radians clockwise:
16798 Rotate the input by PI/6 radians counter-clockwise:
16804 Rotate the input by 45 degrees clockwise:
16810 Apply a constant rotation with period T, starting from an angle of PI/3:
16812 rotate=PI/3+2*PI*t/T
16816 Make the input video rotation oscillating with a period of T
16817 seconds and an amplitude of A radians:
16819 rotate=A*sin(2*PI/T*t)
16823 Rotate the video, output size is chosen so that the whole rotating
16824 input video is always completely contained in the output:
16826 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
16830 Rotate the video, reduce the output size so that no background is ever
16833 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
16837 @subsection Commands
16839 The filter supports the following commands:
16843 Set the angle expression.
16844 The command accepts the same syntax of the corresponding option.
16846 If the specified expression is not valid, it is kept at its current
16852 Apply Shape Adaptive Blur.
16854 The filter accepts the following options:
16857 @item luma_radius, lr
16858 Set luma blur filter strength, must be a value in range 0.1-4.0, default
16859 value is 1.0. A greater value will result in a more blurred image, and
16860 in slower processing.
16862 @item luma_pre_filter_radius, lpfr
16863 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
16866 @item luma_strength, ls
16867 Set luma maximum difference between pixels to still be considered, must
16868 be a value in the 0.1-100.0 range, default value is 1.0.
16870 @item chroma_radius, cr
16871 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
16872 greater value will result in a more blurred image, and in slower
16875 @item chroma_pre_filter_radius, cpfr
16876 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
16878 @item chroma_strength, cs
16879 Set chroma maximum difference between pixels to still be considered,
16880 must be a value in the -0.9-100.0 range.
16883 Each chroma option value, if not explicitly specified, is set to the
16884 corresponding luma option value.
16889 Scale (resize) the input video, using the libswscale library.
16891 The scale filter forces the output display aspect ratio to be the same
16892 of the input, by changing the output sample aspect ratio.
16894 If the input image format is different from the format requested by
16895 the next filter, the scale filter will convert the input to the
16898 @subsection Options
16899 The filter accepts the following options, or any of the options
16900 supported by the libswscale scaler.
16902 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
16903 the complete list of scaler options.
16908 Set the output video dimension expression. Default value is the input
16911 If the @var{width} or @var{w} value is 0, the input width is used for
16912 the output. If the @var{height} or @var{h} value is 0, the input height
16913 is used for the output.
16915 If one and only one of the values is -n with n >= 1, the scale filter
16916 will use a value that maintains the aspect ratio of the input image,
16917 calculated from the other specified dimension. After that it will,
16918 however, make sure that the calculated dimension is divisible by n and
16919 adjust the value if necessary.
16921 If both values are -n with n >= 1, the behavior will be identical to
16922 both values being set to 0 as previously detailed.
16924 See below for the list of accepted constants for use in the dimension
16928 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
16932 Only evaluate expressions once during the filter initialization or when a command is processed.
16935 Evaluate expressions for each incoming frame.
16939 Default value is @samp{init}.
16943 Set the interlacing mode. It accepts the following values:
16947 Force interlaced aware scaling.
16950 Do not apply interlaced scaling.
16953 Select interlaced aware scaling depending on whether the source frames
16954 are flagged as interlaced or not.
16957 Default value is @samp{0}.
16960 Set libswscale scaling flags. See
16961 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16962 complete list of values. If not explicitly specified the filter applies
16966 @item param0, param1
16967 Set libswscale input parameters for scaling algorithms that need them. See
16968 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16969 complete documentation. If not explicitly specified the filter applies
16975 Set the video size. For the syntax of this option, check the
16976 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16978 @item in_color_matrix
16979 @item out_color_matrix
16980 Set in/output YCbCr color space type.
16982 This allows the autodetected value to be overridden as well as allows forcing
16983 a specific value used for the output and encoder.
16985 If not specified, the color space type depends on the pixel format.
16991 Choose automatically.
16994 Format conforming to International Telecommunication Union (ITU)
16995 Recommendation BT.709.
16998 Set color space conforming to the United States Federal Communications
16999 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
17004 Set color space conforming to:
17008 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
17011 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
17014 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
17019 Set color space conforming to SMPTE ST 240:1999.
17022 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
17027 Set in/output YCbCr sample range.
17029 This allows the autodetected value to be overridden as well as allows forcing
17030 a specific value used for the output and encoder. If not specified, the
17031 range depends on the pixel format. Possible values:
17035 Choose automatically.
17038 Set full range (0-255 in case of 8-bit luma).
17040 @item mpeg/limited/tv
17041 Set "MPEG" range (16-235 in case of 8-bit luma).
17044 @item force_original_aspect_ratio
17045 Enable decreasing or increasing output video width or height if necessary to
17046 keep the original aspect ratio. Possible values:
17050 Scale the video as specified and disable this feature.
17053 The output video dimensions will automatically be decreased if needed.
17056 The output video dimensions will automatically be increased if needed.
17060 One useful instance of this option is that when you know a specific device's
17061 maximum allowed resolution, you can use this to limit the output video to
17062 that, while retaining the aspect ratio. For example, device A allows
17063 1280x720 playback, and your video is 1920x800. Using this option (set it to
17064 decrease) and specifying 1280x720 to the command line makes the output
17067 Please note that this is a different thing than specifying -1 for @option{w}
17068 or @option{h}, you still need to specify the output resolution for this option
17071 @item force_divisible_by
17072 Ensures that both the output dimensions, width and height, are divisible by the
17073 given integer when used together with @option{force_original_aspect_ratio}. This
17074 works similar to using @code{-n} in the @option{w} and @option{h} options.
17076 This option respects the value set for @option{force_original_aspect_ratio},
17077 increasing or decreasing the resolution accordingly. The video's aspect ratio
17078 may be slightly modified.
17080 This option can be handy if you need to have a video fit within or exceed
17081 a defined resolution using @option{force_original_aspect_ratio} but also have
17082 encoder restrictions on width or height divisibility.
17086 The values of the @option{w} and @option{h} options are expressions
17087 containing the following constants:
17092 The input width and height
17096 These are the same as @var{in_w} and @var{in_h}.
17100 The output (scaled) width and height
17104 These are the same as @var{out_w} and @var{out_h}
17107 The same as @var{iw} / @var{ih}
17110 input sample aspect ratio
17113 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
17117 horizontal and vertical input chroma subsample values. For example for the
17118 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17122 horizontal and vertical output chroma subsample values. For example for the
17123 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17126 The (sequential) number of the input frame, starting from 0.
17127 Only available with @code{eval=frame}.
17130 The presentation timestamp of the input frame, expressed as a number of
17131 seconds. Only available with @code{eval=frame}.
17134 The position (byte offset) of the frame in the input stream, or NaN if
17135 this information is unavailable and/or meaningless (for example in case of synthetic video).
17136 Only available with @code{eval=frame}.
17139 @subsection Examples
17143 Scale the input video to a size of 200x100
17148 This is equivalent to:
17159 Specify a size abbreviation for the output size:
17164 which can also be written as:
17170 Scale the input to 2x:
17172 scale=w=2*iw:h=2*ih
17176 The above is the same as:
17178 scale=2*in_w:2*in_h
17182 Scale the input to 2x with forced interlaced scaling:
17184 scale=2*iw:2*ih:interl=1
17188 Scale the input to half size:
17190 scale=w=iw/2:h=ih/2
17194 Increase the width, and set the height to the same size:
17200 Seek Greek harmony:
17207 Increase the height, and set the width to 3/2 of the height:
17209 scale=w=3/2*oh:h=3/5*ih
17213 Increase the size, making the size a multiple of the chroma
17216 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
17220 Increase the width to a maximum of 500 pixels,
17221 keeping the same aspect ratio as the input:
17223 scale=w='min(500\, iw*3/2):h=-1'
17227 Make pixels square by combining scale and setsar:
17229 scale='trunc(ih*dar):ih',setsar=1/1
17233 Make pixels square by combining scale and setsar,
17234 making sure the resulting resolution is even (required by some codecs):
17236 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
17240 @subsection Commands
17242 This filter supports the following commands:
17246 Set the output video dimension expression.
17247 The command accepts the same syntax of the corresponding option.
17249 If the specified expression is not valid, it is kept at its current
17255 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
17256 format conversion on CUDA video frames. Setting the output width and height
17257 works in the same way as for the @var{scale} filter.
17259 The following additional options are accepted:
17262 The pixel format of the output CUDA frames. If set to the string "same" (the
17263 default), the input format will be kept. Note that automatic format negotiation
17264 and conversion is not yet supported for hardware frames
17267 The interpolation algorithm used for resizing. One of the following:
17274 @item cubic2p_bspline
17275 2-parameter cubic (B=1, C=0)
17277 @item cubic2p_catmullrom
17278 2-parameter cubic (B=0, C=1/2)
17280 @item cubic2p_b05c03
17281 2-parameter cubic (B=1/2, C=3/10)
17289 @item force_original_aspect_ratio
17290 Enable decreasing or increasing output video width or height if necessary to
17291 keep the original aspect ratio. Possible values:
17295 Scale the video as specified and disable this feature.
17298 The output video dimensions will automatically be decreased if needed.
17301 The output video dimensions will automatically be increased if needed.
17305 One useful instance of this option is that when you know a specific device's
17306 maximum allowed resolution, you can use this to limit the output video to
17307 that, while retaining the aspect ratio. For example, device A allows
17308 1280x720 playback, and your video is 1920x800. Using this option (set it to
17309 decrease) and specifying 1280x720 to the command line makes the output
17312 Please note that this is a different thing than specifying -1 for @option{w}
17313 or @option{h}, you still need to specify the output resolution for this option
17316 @item force_divisible_by
17317 Ensures that both the output dimensions, width and height, are divisible by the
17318 given integer when used together with @option{force_original_aspect_ratio}. This
17319 works similar to using @code{-n} in the @option{w} and @option{h} options.
17321 This option respects the value set for @option{force_original_aspect_ratio},
17322 increasing or decreasing the resolution accordingly. The video's aspect ratio
17323 may be slightly modified.
17325 This option can be handy if you need to have a video fit within or exceed
17326 a defined resolution using @option{force_original_aspect_ratio} but also have
17327 encoder restrictions on width or height divisibility.
17333 Scale (resize) the input video, based on a reference video.
17335 See the scale filter for available options, scale2ref supports the same but
17336 uses the reference video instead of the main input as basis. scale2ref also
17337 supports the following additional constants for the @option{w} and
17338 @option{h} options:
17343 The main input video's width and height
17346 The same as @var{main_w} / @var{main_h}
17349 The main input video's sample aspect ratio
17351 @item main_dar, mdar
17352 The main input video's display aspect ratio. Calculated from
17353 @code{(main_w / main_h) * main_sar}.
17357 The main input video's horizontal and vertical chroma subsample values.
17358 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
17362 The (sequential) number of the main input frame, starting from 0.
17363 Only available with @code{eval=frame}.
17366 The presentation timestamp of the main input frame, expressed as a number of
17367 seconds. Only available with @code{eval=frame}.
17370 The position (byte offset) of the frame in the main input stream, or NaN if
17371 this information is unavailable and/or meaningless (for example in case of synthetic video).
17372 Only available with @code{eval=frame}.
17375 @subsection Examples
17379 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
17381 'scale2ref[b][a];[a][b]overlay'
17385 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
17387 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
17391 @subsection Commands
17393 This filter supports the following commands:
17397 Set the output video dimension expression.
17398 The command accepts the same syntax of the corresponding option.
17400 If the specified expression is not valid, it is kept at its current
17405 Scroll input video horizontally and/or vertically by constant speed.
17407 The filter accepts the following options:
17409 @item horizontal, h
17410 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
17411 Negative values changes scrolling direction.
17414 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
17415 Negative values changes scrolling direction.
17418 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
17421 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
17424 @subsection Commands
17426 This filter supports the following @ref{commands}:
17428 @item horizontal, h
17429 Set the horizontal scrolling speed.
17431 Set the vertical scrolling speed.
17437 Detect video scene change.
17439 This filter sets frame metadata with mafd between frame, the scene score, and
17440 forward the frame to the next filter, so they can use these metadata to detect
17441 scene change or others.
17443 In addition, this filter logs a message and sets frame metadata when it detects
17444 a scene change by @option{threshold}.
17446 @code{lavfi.scd.mafd} metadata keys are set with mafd for every frame.
17448 @code{lavfi.scd.score} metadata keys are set with scene change score for every frame
17449 to detect scene change.
17451 @code{lavfi.scd.time} metadata keys are set with current filtered frame time which
17452 detect scene change with @option{threshold}.
17454 The filter accepts the following options:
17458 Set the scene change detection threshold as a percentage of maximum change. Good
17459 values are in the @code{[8.0, 14.0]} range. The range for @option{threshold} is
17462 Default value is @code{10.}.
17465 Set the flag to pass scene change frames to the next filter. Default value is @code{0}
17466 You can enable it if you want to get snapshot of scene change frames only.
17469 @anchor{selectivecolor}
17470 @section selectivecolor
17472 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
17473 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
17474 by the "purity" of the color (that is, how saturated it already is).
17476 This filter is similar to the Adobe Photoshop Selective Color tool.
17478 The filter accepts the following options:
17481 @item correction_method
17482 Select color correction method.
17484 Available values are:
17487 Specified adjustments are applied "as-is" (added/subtracted to original pixel
17490 Specified adjustments are relative to the original component value.
17492 Default is @code{absolute}.
17494 Adjustments for red pixels (pixels where the red component is the maximum)
17496 Adjustments for yellow pixels (pixels where the blue component is the minimum)
17498 Adjustments for green pixels (pixels where the green component is the maximum)
17500 Adjustments for cyan pixels (pixels where the red component is the minimum)
17502 Adjustments for blue pixels (pixels where the blue component is the maximum)
17504 Adjustments for magenta pixels (pixels where the green component is the minimum)
17506 Adjustments for white pixels (pixels where all components are greater than 128)
17508 Adjustments for all pixels except pure black and pure white
17510 Adjustments for black pixels (pixels where all components are lesser than 128)
17512 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
17515 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
17516 4 space separated floating point adjustment values in the [-1,1] range,
17517 respectively to adjust the amount of cyan, magenta, yellow and black for the
17518 pixels of its range.
17520 @subsection Examples
17524 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
17525 increase magenta by 27% in blue areas:
17527 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
17531 Use a Photoshop selective color preset:
17533 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
17537 @anchor{separatefields}
17538 @section separatefields
17540 The @code{separatefields} takes a frame-based video input and splits
17541 each frame into its components fields, producing a new half height clip
17542 with twice the frame rate and twice the frame count.
17544 This filter use field-dominance information in frame to decide which
17545 of each pair of fields to place first in the output.
17546 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
17548 @section setdar, setsar
17550 The @code{setdar} filter sets the Display Aspect Ratio for the filter
17553 This is done by changing the specified Sample (aka Pixel) Aspect
17554 Ratio, according to the following equation:
17556 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
17559 Keep in mind that the @code{setdar} filter does not modify the pixel
17560 dimensions of the video frame. Also, the display aspect ratio set by
17561 this filter may be changed by later filters in the filterchain,
17562 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
17565 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
17566 the filter output video.
17568 Note that as a consequence of the application of this filter, the
17569 output display aspect ratio will change according to the equation
17572 Keep in mind that the sample aspect ratio set by the @code{setsar}
17573 filter may be changed by later filters in the filterchain, e.g. if
17574 another "setsar" or a "setdar" filter is applied.
17576 It accepts the following parameters:
17579 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
17580 Set the aspect ratio used by the filter.
17582 The parameter can be a floating point number string, an expression, or
17583 a string of the form @var{num}:@var{den}, where @var{num} and
17584 @var{den} are the numerator and denominator of the aspect ratio. If
17585 the parameter is not specified, it is assumed the value "0".
17586 In case the form "@var{num}:@var{den}" is used, the @code{:} character
17590 Set the maximum integer value to use for expressing numerator and
17591 denominator when reducing the expressed aspect ratio to a rational.
17592 Default value is @code{100}.
17596 The parameter @var{sar} is an expression containing
17597 the following constants:
17601 These are approximated values for the mathematical constants e
17602 (Euler's number), pi (Greek pi), and phi (the golden ratio).
17605 The input width and height.
17608 These are the same as @var{w} / @var{h}.
17611 The input sample aspect ratio.
17614 The input display aspect ratio. It is the same as
17615 (@var{w} / @var{h}) * @var{sar}.
17618 Horizontal and vertical chroma subsample values. For example, for the
17619 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17622 @subsection Examples
17627 To change the display aspect ratio to 16:9, specify one of the following:
17634 To change the sample aspect ratio to 10:11, specify:
17640 To set a display aspect ratio of 16:9, and specify a maximum integer value of
17641 1000 in the aspect ratio reduction, use the command:
17643 setdar=ratio=16/9:max=1000
17651 Force field for the output video frame.
17653 The @code{setfield} filter marks the interlace type field for the
17654 output frames. It does not change the input frame, but only sets the
17655 corresponding property, which affects how the frame is treated by
17656 following filters (e.g. @code{fieldorder} or @code{yadif}).
17658 The filter accepts the following options:
17663 Available values are:
17667 Keep the same field property.
17670 Mark the frame as bottom-field-first.
17673 Mark the frame as top-field-first.
17676 Mark the frame as progressive.
17683 Force frame parameter for the output video frame.
17685 The @code{setparams} filter marks interlace and color range for the
17686 output frames. It does not change the input frame, but only sets the
17687 corresponding property, which affects how the frame is treated by
17692 Available values are:
17696 Keep the same field property (default).
17699 Mark the frame as bottom-field-first.
17702 Mark the frame as top-field-first.
17705 Mark the frame as progressive.
17709 Available values are:
17713 Keep the same color range property (default).
17715 @item unspecified, unknown
17716 Mark the frame as unspecified color range.
17718 @item limited, tv, mpeg
17719 Mark the frame as limited range.
17721 @item full, pc, jpeg
17722 Mark the frame as full range.
17725 @item color_primaries
17726 Set the color primaries.
17727 Available values are:
17731 Keep the same color primaries property (default).
17748 Set the color transfer.
17749 Available values are:
17753 Keep the same color trc property (default).
17775 Set the colorspace.
17776 Available values are:
17780 Keep the same colorspace property (default).
17793 @item chroma-derived-nc
17794 @item chroma-derived-c
17801 Show a line containing various information for each input video frame.
17802 The input video is not modified.
17804 This filter supports the following options:
17808 Calculate checksums of each plane. By default enabled.
17811 The shown line contains a sequence of key/value pairs of the form
17812 @var{key}:@var{value}.
17814 The following values are shown in the output:
17818 The (sequential) number of the input frame, starting from 0.
17821 The Presentation TimeStamp of the input frame, expressed as a number of
17822 time base units. The time base unit depends on the filter input pad.
17825 The Presentation TimeStamp of the input frame, expressed as a number of
17829 The position of the frame in the input stream, or -1 if this information is
17830 unavailable and/or meaningless (for example in case of synthetic video).
17833 The pixel format name.
17836 The sample aspect ratio of the input frame, expressed in the form
17837 @var{num}/@var{den}.
17840 The size of the input frame. For the syntax of this option, check the
17841 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17844 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
17845 for bottom field first).
17848 This is 1 if the frame is a key frame, 0 otherwise.
17851 The picture type of the input frame ("I" for an I-frame, "P" for a
17852 P-frame, "B" for a B-frame, or "?" for an unknown type).
17853 Also refer to the documentation of the @code{AVPictureType} enum and of
17854 the @code{av_get_picture_type_char} function defined in
17855 @file{libavutil/avutil.h}.
17858 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
17860 @item plane_checksum
17861 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
17862 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
17865 The mean value of pixels in each plane of the input frame, expressed in the form
17866 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
17869 The standard deviation of pixel values in each plane of the input frame, expressed
17870 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
17874 @section showpalette
17876 Displays the 256 colors palette of each frame. This filter is only relevant for
17877 @var{pal8} pixel format frames.
17879 It accepts the following option:
17883 Set the size of the box used to represent one palette color entry. Default is
17884 @code{30} (for a @code{30x30} pixel box).
17887 @section shuffleframes
17889 Reorder and/or duplicate and/or drop video frames.
17891 It accepts the following parameters:
17895 Set the destination indexes of input frames.
17896 This is space or '|' separated list of indexes that maps input frames to output
17897 frames. Number of indexes also sets maximal value that each index may have.
17898 '-1' index have special meaning and that is to drop frame.
17901 The first frame has the index 0. The default is to keep the input unchanged.
17903 @subsection Examples
17907 Swap second and third frame of every three frames of the input:
17909 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
17913 Swap 10th and 1st frame of every ten frames of the input:
17915 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
17919 @section shufflepixels
17921 Reorder pixels in video frames.
17923 This filter accepts the following options:
17927 Set shuffle direction. Can be forward or inverse direction.
17928 Default direction is forward.
17931 Set shuffle mode. Can be horizontal, vertical or block mode.
17935 Set shuffle block_size. In case of horizontal shuffle mode only width
17936 part of size is used, and in case of vertical shuffle mode only height
17937 part of size is used.
17940 Set random seed used with shuffling pixels. Mainly useful to set to be able
17941 to reverse filtering process to get original input.
17942 For example, to reverse forward shuffle you need to use same parameters
17943 and exact same seed and to set direction to inverse.
17946 @section shuffleplanes
17948 Reorder and/or duplicate video planes.
17950 It accepts the following parameters:
17955 The index of the input plane to be used as the first output plane.
17958 The index of the input plane to be used as the second output plane.
17961 The index of the input plane to be used as the third output plane.
17964 The index of the input plane to be used as the fourth output plane.
17968 The first plane has the index 0. The default is to keep the input unchanged.
17970 @subsection Examples
17974 Swap the second and third planes of the input:
17976 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
17980 @anchor{signalstats}
17981 @section signalstats
17982 Evaluate various visual metrics that assist in determining issues associated
17983 with the digitization of analog video media.
17985 By default the filter will log these metadata values:
17989 Display the minimal Y value contained within the input frame. Expressed in
17993 Display the Y value at the 10% percentile within the input frame. Expressed in
17997 Display the average Y value within the input frame. Expressed in range of
18001 Display the Y value at the 90% percentile within the input frame. Expressed in
18005 Display the maximum Y value contained within the input frame. Expressed in
18009 Display the minimal U value contained within the input frame. Expressed in
18013 Display the U value at the 10% percentile within the input frame. Expressed in
18017 Display the average U value within the input frame. Expressed in range of
18021 Display the U value at the 90% percentile within the input frame. Expressed in
18025 Display the maximum U value contained within the input frame. Expressed in
18029 Display the minimal V value contained within the input frame. Expressed in
18033 Display the V value at the 10% percentile within the input frame. Expressed in
18037 Display the average V value within the input frame. Expressed in range of
18041 Display the V value at the 90% percentile within the input frame. Expressed in
18045 Display the maximum V value contained within the input frame. Expressed in
18049 Display the minimal saturation value contained within the input frame.
18050 Expressed in range of [0-~181.02].
18053 Display the saturation value at the 10% percentile within the input frame.
18054 Expressed in range of [0-~181.02].
18057 Display the average saturation value within the input frame. Expressed in range
18061 Display the saturation value at the 90% percentile within the input frame.
18062 Expressed in range of [0-~181.02].
18065 Display the maximum saturation value contained within the input frame.
18066 Expressed in range of [0-~181.02].
18069 Display the median value for hue within the input frame. Expressed in range of
18073 Display the average value for hue within the input frame. Expressed in range of
18077 Display the average of sample value difference between all values of the Y
18078 plane in the current frame and corresponding values of the previous input frame.
18079 Expressed in range of [0-255].
18082 Display the average of sample value difference between all values of the U
18083 plane in the current frame and corresponding values of the previous input frame.
18084 Expressed in range of [0-255].
18087 Display the average of sample value difference between all values of the V
18088 plane in the current frame and corresponding values of the previous input frame.
18089 Expressed in range of [0-255].
18092 Display bit depth of Y plane in current frame.
18093 Expressed in range of [0-16].
18096 Display bit depth of U plane in current frame.
18097 Expressed in range of [0-16].
18100 Display bit depth of V plane in current frame.
18101 Expressed in range of [0-16].
18104 The filter accepts the following options:
18110 @option{stat} specify an additional form of image analysis.
18111 @option{out} output video with the specified type of pixel highlighted.
18113 Both options accept the following values:
18117 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
18118 unlike the neighboring pixels of the same field. Examples of temporal outliers
18119 include the results of video dropouts, head clogs, or tape tracking issues.
18122 Identify @var{vertical line repetition}. Vertical line repetition includes
18123 similar rows of pixels within a frame. In born-digital video vertical line
18124 repetition is common, but this pattern is uncommon in video digitized from an
18125 analog source. When it occurs in video that results from the digitization of an
18126 analog source it can indicate concealment from a dropout compensator.
18129 Identify pixels that fall outside of legal broadcast range.
18133 Set the highlight color for the @option{out} option. The default color is
18137 @subsection Examples
18141 Output data of various video metrics:
18143 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
18147 Output specific data about the minimum and maximum values of the Y plane per frame:
18149 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
18153 Playback video while highlighting pixels that are outside of broadcast range in red.
18155 ffplay example.mov -vf signalstats="out=brng:color=red"
18159 Playback video with signalstats metadata drawn over the frame.
18161 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
18164 The contents of signalstat_drawtext.txt used in the command are:
18167 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
18168 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
18169 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
18170 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
18178 Calculates the MPEG-7 Video Signature. The filter can handle more than one
18179 input. In this case the matching between the inputs can be calculated additionally.
18180 The filter always passes through the first input. The signature of each stream can
18181 be written into a file.
18183 It accepts the following options:
18187 Enable or disable the matching process.
18189 Available values are:
18193 Disable the calculation of a matching (default).
18195 Calculate the matching for the whole video and output whether the whole video
18196 matches or only parts.
18198 Calculate only until a matching is found or the video ends. Should be faster in
18203 Set the number of inputs. The option value must be a non negative integer.
18204 Default value is 1.
18207 Set the path to which the output is written. If there is more than one input,
18208 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
18209 integer), that will be replaced with the input number. If no filename is
18210 specified, no output will be written. This is the default.
18213 Choose the output format.
18215 Available values are:
18219 Use the specified binary representation (default).
18221 Use the specified xml representation.
18225 Set threshold to detect one word as similar. The option value must be an integer
18226 greater than zero. The default value is 9000.
18229 Set threshold to detect all words as similar. The option value must be an integer
18230 greater than zero. The default value is 60000.
18233 Set threshold to detect frames as similar. The option value must be an integer
18234 greater than zero. The default value is 116.
18237 Set the minimum length of a sequence in frames to recognize it as matching
18238 sequence. The option value must be a non negative integer value.
18239 The default value is 0.
18242 Set the minimum relation, that matching frames to all frames must have.
18243 The option value must be a double value between 0 and 1. The default value is 0.5.
18246 @subsection Examples
18250 To calculate the signature of an input video and store it in signature.bin:
18252 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
18256 To detect whether two videos match and store the signatures in XML format in
18257 signature0.xml and signature1.xml:
18259 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 -
18267 Blur the input video without impacting the outlines.
18269 It accepts the following options:
18272 @item luma_radius, lr
18273 Set the luma radius. The option value must be a float number in
18274 the range [0.1,5.0] that specifies the variance of the gaussian filter
18275 used to blur the image (slower if larger). Default value is 1.0.
18277 @item luma_strength, ls
18278 Set the luma strength. The option value must be a float number
18279 in the range [-1.0,1.0] that configures the blurring. A value included
18280 in [0.0,1.0] will blur the image whereas a value included in
18281 [-1.0,0.0] will sharpen the image. Default value is 1.0.
18283 @item luma_threshold, lt
18284 Set the luma threshold used as a coefficient to determine
18285 whether a pixel should be blurred or not. The option value must be an
18286 integer in the range [-30,30]. A value of 0 will filter all the image,
18287 a value included in [0,30] will filter flat areas and a value included
18288 in [-30,0] will filter edges. Default value is 0.
18290 @item chroma_radius, cr
18291 Set the chroma radius. The option value must be a float number in
18292 the range [0.1,5.0] that specifies the variance of the gaussian filter
18293 used to blur the image (slower if larger). Default value is @option{luma_radius}.
18295 @item chroma_strength, cs
18296 Set the chroma strength. The option value must be a float number
18297 in the range [-1.0,1.0] that configures the blurring. A value included
18298 in [0.0,1.0] will blur the image whereas a value included in
18299 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
18301 @item chroma_threshold, ct
18302 Set the chroma threshold used as a coefficient to determine
18303 whether a pixel should be blurred or not. The option value must be an
18304 integer in the range [-30,30]. A value of 0 will filter all the image,
18305 a value included in [0,30] will filter flat areas and a value included
18306 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
18309 If a chroma option is not explicitly set, the corresponding luma value
18313 Apply sobel operator to input video stream.
18315 The filter accepts the following option:
18319 Set which planes will be processed, unprocessed planes will be copied.
18320 By default value 0xf, all planes will be processed.
18323 Set value which will be multiplied with filtered result.
18326 Set value which will be added to filtered result.
18329 @subsection Commands
18331 This filter supports the all above options as @ref{commands}.
18336 Apply a simple postprocessing filter that compresses and decompresses the image
18337 at several (or - in the case of @option{quality} level @code{6} - all) shifts
18338 and average the results.
18340 The filter accepts the following options:
18344 Set quality. This option defines the number of levels for averaging. It accepts
18345 an integer in the range 0-6. If set to @code{0}, the filter will have no
18346 effect. A value of @code{6} means the higher quality. For each increment of
18347 that value the speed drops by a factor of approximately 2. Default value is
18351 Force a constant quantization parameter. If not set, the filter will use the QP
18352 from the video stream (if available).
18355 Set thresholding mode. Available modes are:
18359 Set hard thresholding (default).
18361 Set soft thresholding (better de-ringing effect, but likely blurrier).
18364 @item use_bframe_qp
18365 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
18366 option may cause flicker since the B-Frames have often larger QP. Default is
18367 @code{0} (not enabled).
18370 @subsection Commands
18372 This filter supports the following commands:
18374 @item quality, level
18375 Set quality level. The value @code{max} can be used to set the maximum level,
18376 currently @code{6}.
18382 Scale the input by applying one of the super-resolution methods based on
18383 convolutional neural networks. Supported models:
18387 Super-Resolution Convolutional Neural Network model (SRCNN).
18388 See @url{https://arxiv.org/abs/1501.00092}.
18391 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
18392 See @url{https://arxiv.org/abs/1609.05158}.
18395 Training scripts as well as scripts for model file (.pb) saving can be found at
18396 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
18397 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
18399 Native model files (.model) can be generated from TensorFlow model
18400 files (.pb) by using tools/python/convert.py
18402 The filter accepts the following options:
18406 Specify which DNN backend to use for model loading and execution. This option accepts
18407 the following values:
18411 Native implementation of DNN loading and execution.
18414 TensorFlow backend. To enable this backend you
18415 need to install the TensorFlow for C library (see
18416 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
18417 @code{--enable-libtensorflow}
18420 Default value is @samp{native}.
18423 Set path to model file specifying network architecture and its parameters.
18424 Note that different backends use different file formats. TensorFlow backend
18425 can load files for both formats, while native backend can load files for only
18429 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
18430 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
18431 input upscaled using bicubic upscaling with proper scale factor.
18434 This feature can also be finished with @ref{dnn_processing} filter.
18438 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
18440 This filter takes in input two input videos, the first input is
18441 considered the "main" source and is passed unchanged to the
18442 output. The second input is used as a "reference" video for computing
18445 Both video inputs must have the same resolution and pixel format for
18446 this filter to work correctly. Also it assumes that both inputs
18447 have the same number of frames, which are compared one by one.
18449 The filter stores the calculated SSIM of each frame.
18451 The description of the accepted parameters follows.
18454 @item stats_file, f
18455 If specified the filter will use the named file to save the SSIM of
18456 each individual frame. When filename equals "-" the data is sent to
18460 The file printed if @var{stats_file} is selected, contains a sequence of
18461 key/value pairs of the form @var{key}:@var{value} for each compared
18464 A description of each shown parameter follows:
18468 sequential number of the input frame, starting from 1
18470 @item Y, U, V, R, G, B
18471 SSIM of the compared frames for the component specified by the suffix.
18474 SSIM of the compared frames for the whole frame.
18477 Same as above but in dB representation.
18480 This filter also supports the @ref{framesync} options.
18482 @subsection Examples
18487 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
18488 [main][ref] ssim="stats_file=stats.log" [out]
18491 On this example the input file being processed is compared with the
18492 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
18493 is stored in @file{stats.log}.
18496 Another example with both psnr and ssim at same time:
18498 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
18502 Another example with different containers:
18504 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 -
18510 Convert between different stereoscopic image formats.
18512 The filters accept the following options:
18516 Set stereoscopic image format of input.
18518 Available values for input image formats are:
18521 side by side parallel (left eye left, right eye right)
18524 side by side crosseye (right eye left, left eye right)
18527 side by side parallel with half width resolution
18528 (left eye left, right eye right)
18531 side by side crosseye with half width resolution
18532 (right eye left, left eye right)
18536 above-below (left eye above, right eye below)
18540 above-below (right eye above, left eye below)
18544 above-below with half height resolution
18545 (left eye above, right eye below)
18549 above-below with half height resolution
18550 (right eye above, left eye below)
18553 alternating frames (left eye first, right eye second)
18556 alternating frames (right eye first, left eye second)
18559 interleaved rows (left eye has top row, right eye starts on next row)
18562 interleaved rows (right eye has top row, left eye starts on next row)
18565 interleaved columns, left eye first
18568 interleaved columns, right eye first
18570 Default value is @samp{sbsl}.
18574 Set stereoscopic image format of output.
18578 side by side parallel (left eye left, right eye right)
18581 side by side crosseye (right eye left, left eye right)
18584 side by side parallel with half width resolution
18585 (left eye left, right eye right)
18588 side by side crosseye with half width resolution
18589 (right eye left, left eye right)
18593 above-below (left eye above, right eye below)
18597 above-below (right eye above, left eye below)
18601 above-below with half height resolution
18602 (left eye above, right eye below)
18606 above-below with half height resolution
18607 (right eye above, left eye below)
18610 alternating frames (left eye first, right eye second)
18613 alternating frames (right eye first, left eye second)
18616 interleaved rows (left eye has top row, right eye starts on next row)
18619 interleaved rows (right eye has top row, left eye starts on next row)
18622 anaglyph red/blue gray
18623 (red filter on left eye, blue filter on right eye)
18626 anaglyph red/green gray
18627 (red filter on left eye, green filter on right eye)
18630 anaglyph red/cyan gray
18631 (red filter on left eye, cyan filter on right eye)
18634 anaglyph red/cyan half colored
18635 (red filter on left eye, cyan filter on right eye)
18638 anaglyph red/cyan color
18639 (red filter on left eye, cyan filter on right eye)
18642 anaglyph red/cyan color optimized with the least squares projection of dubois
18643 (red filter on left eye, cyan filter on right eye)
18646 anaglyph green/magenta gray
18647 (green filter on left eye, magenta filter on right eye)
18650 anaglyph green/magenta half colored
18651 (green filter on left eye, magenta filter on right eye)
18654 anaglyph green/magenta colored
18655 (green filter on left eye, magenta filter on right eye)
18658 anaglyph green/magenta color optimized with the least squares projection of dubois
18659 (green filter on left eye, magenta filter on right eye)
18662 anaglyph yellow/blue gray
18663 (yellow filter on left eye, blue filter on right eye)
18666 anaglyph yellow/blue half colored
18667 (yellow filter on left eye, blue filter on right eye)
18670 anaglyph yellow/blue colored
18671 (yellow filter on left eye, blue filter on right eye)
18674 anaglyph yellow/blue color optimized with the least squares projection of dubois
18675 (yellow filter on left eye, blue filter on right eye)
18678 mono output (left eye only)
18681 mono output (right eye only)
18684 checkerboard, left eye first
18687 checkerboard, right eye first
18690 interleaved columns, left eye first
18693 interleaved columns, right eye first
18699 Default value is @samp{arcd}.
18702 @subsection Examples
18706 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
18712 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
18718 @section streamselect, astreamselect
18719 Select video or audio streams.
18721 The filter accepts the following options:
18725 Set number of inputs. Default is 2.
18728 Set input indexes to remap to outputs.
18731 @subsection Commands
18733 The @code{streamselect} and @code{astreamselect} filter supports the following
18738 Set input indexes to remap to outputs.
18741 @subsection Examples
18745 Select first 5 seconds 1st stream and rest of time 2nd stream:
18747 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
18751 Same as above, but for audio:
18753 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
18760 Draw subtitles on top of input video using the libass library.
18762 To enable compilation of this filter you need to configure FFmpeg with
18763 @code{--enable-libass}. This filter also requires a build with libavcodec and
18764 libavformat to convert the passed subtitles file to ASS (Advanced Substation
18765 Alpha) subtitles format.
18767 The filter accepts the following options:
18771 Set the filename of the subtitle file to read. It must be specified.
18773 @item original_size
18774 Specify the size of the original video, the video for which the ASS file
18775 was composed. For the syntax of this option, check the
18776 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18777 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
18778 correctly scale the fonts if the aspect ratio has been changed.
18781 Set a directory path containing fonts that can be used by the filter.
18782 These fonts will be used in addition to whatever the font provider uses.
18785 Process alpha channel, by default alpha channel is untouched.
18788 Set subtitles input character encoding. @code{subtitles} filter only. Only
18789 useful if not UTF-8.
18791 @item stream_index, si
18792 Set subtitles stream index. @code{subtitles} filter only.
18795 Override default style or script info parameters of the subtitles. It accepts a
18796 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
18799 If the first key is not specified, it is assumed that the first value
18800 specifies the @option{filename}.
18802 For example, to render the file @file{sub.srt} on top of the input
18803 video, use the command:
18808 which is equivalent to:
18810 subtitles=filename=sub.srt
18813 To render the default subtitles stream from file @file{video.mkv}, use:
18815 subtitles=video.mkv
18818 To render the second subtitles stream from that file, use:
18820 subtitles=video.mkv:si=1
18823 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
18824 @code{DejaVu Serif}, use:
18826 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
18829 @section super2xsai
18831 Scale the input by 2x and smooth using the Super2xSaI (Scale and
18832 Interpolate) pixel art scaling algorithm.
18834 Useful for enlarging pixel art images without reducing sharpness.
18838 Swap two rectangular objects in video.
18840 This filter accepts the following options:
18850 Set 1st rect x coordinate.
18853 Set 1st rect y coordinate.
18856 Set 2nd rect x coordinate.
18859 Set 2nd rect y coordinate.
18861 All expressions are evaluated once for each frame.
18864 The all options are expressions containing the following constants:
18869 The input width and height.
18872 same as @var{w} / @var{h}
18875 input sample aspect ratio
18878 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
18881 The number of the input frame, starting from 0.
18884 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
18887 the position in the file of the input frame, NAN if unknown
18894 Blend successive video frames.
18900 Apply telecine process to the video.
18902 This filter accepts the following options:
18911 The default value is @code{top}.
18915 A string of numbers representing the pulldown pattern you wish to apply.
18916 The default value is @code{23}.
18920 Some typical patterns:
18925 24p: 2332 (preferred)
18932 24p: 222222222223 ("Euro pulldown")
18937 @section thistogram
18939 Compute and draw a color distribution histogram for the input video across time.
18941 Unlike @ref{histogram} video filter which only shows histogram of single input frame
18942 at certain time, this filter shows also past histograms of number of frames defined
18943 by @code{width} option.
18945 The computed histogram is a representation of the color component
18946 distribution in an image.
18948 The filter accepts the following options:
18952 Set width of single color component output. Default value is @code{0}.
18953 Value of @code{0} means width will be picked from input video.
18954 This also set number of passed histograms to keep.
18955 Allowed range is [0, 8192].
18957 @item display_mode, d
18959 It accepts the following values:
18962 Per color component graphs are placed below each other.
18965 Per color component graphs are placed side by side.
18968 Presents information identical to that in the @code{parade}, except
18969 that the graphs representing color components are superimposed directly
18972 Default is @code{stack}.
18974 @item levels_mode, m
18975 Set mode. Can be either @code{linear}, or @code{logarithmic}.
18976 Default is @code{linear}.
18978 @item components, c
18979 Set what color components to display.
18980 Default is @code{7}.
18983 Set background opacity. Default is @code{0.9}.
18986 Show envelope. Default is disabled.
18989 Set envelope color. Default is @code{gold}.
18994 Available values for slide is:
18997 Draw new frame when right border is reached.
19000 Replace old columns with new ones.
19003 Scroll from right to left.
19006 Scroll from left to right.
19009 Draw single picture.
19012 Default is @code{replace}.
19017 Apply threshold effect to video stream.
19019 This filter needs four video streams to perform thresholding.
19020 First stream is stream we are filtering.
19021 Second stream is holding threshold values, third stream is holding min values,
19022 and last, fourth stream is holding max values.
19024 The filter accepts the following option:
19028 Set which planes will be processed, unprocessed planes will be copied.
19029 By default value 0xf, all planes will be processed.
19032 For example if first stream pixel's component value is less then threshold value
19033 of pixel component from 2nd threshold stream, third stream value will picked,
19034 otherwise fourth stream pixel component value will be picked.
19036 Using color source filter one can perform various types of thresholding:
19038 @subsection Examples
19042 Binary threshold, using gray color as threshold:
19044 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
19048 Inverted binary threshold, using gray color as threshold:
19050 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
19054 Truncate binary threshold, using gray color as threshold:
19056 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
19060 Threshold to zero, using gray color as threshold:
19062 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
19066 Inverted threshold to zero, using gray color as threshold:
19068 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
19073 Select the most representative frame in a given sequence of consecutive frames.
19075 The filter accepts the following options:
19079 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
19080 will pick one of them, and then handle the next batch of @var{n} frames until
19081 the end. Default is @code{100}.
19084 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
19085 value will result in a higher memory usage, so a high value is not recommended.
19087 @subsection Examples
19091 Extract one picture each 50 frames:
19097 Complete example of a thumbnail creation with @command{ffmpeg}:
19099 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
19106 Tile several successive frames together.
19108 The @ref{untile} filter can do the reverse.
19110 The filter accepts the following options:
19115 Set the grid size (i.e. the number of lines and columns). For the syntax of
19116 this option, check the
19117 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19120 Set the maximum number of frames to render in the given area. It must be less
19121 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
19122 the area will be used.
19125 Set the outer border margin in pixels.
19128 Set the inner border thickness (i.e. the number of pixels between frames). For
19129 more advanced padding options (such as having different values for the edges),
19130 refer to the pad video filter.
19133 Specify the color of the unused area. For the syntax of this option, check the
19134 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
19135 The default value of @var{color} is "black".
19138 Set the number of frames to overlap when tiling several successive frames together.
19139 The value must be between @code{0} and @var{nb_frames - 1}.
19142 Set the number of frames to initially be empty before displaying first output frame.
19143 This controls how soon will one get first output frame.
19144 The value must be between @code{0} and @var{nb_frames - 1}.
19147 @subsection Examples
19151 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
19153 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
19155 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
19156 duplicating each output frame to accommodate the originally detected frame
19160 Display @code{5} pictures in an area of @code{3x2} frames,
19161 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
19162 mixed flat and named options:
19164 tile=3x2:nb_frames=5:padding=7:margin=2
19168 @section tinterlace
19170 Perform various types of temporal field interlacing.
19172 Frames are counted starting from 1, so the first input frame is
19175 The filter accepts the following options:
19180 Specify the mode of the interlacing. This option can also be specified
19181 as a value alone. See below for a list of values for this option.
19183 Available values are:
19187 Move odd frames into the upper field, even into the lower field,
19188 generating a double height frame at half frame rate.
19192 Frame 1 Frame 2 Frame 3 Frame 4
19194 11111 22222 33333 44444
19195 11111 22222 33333 44444
19196 11111 22222 33333 44444
19197 11111 22222 33333 44444
19211 Only output odd frames, even frames are dropped, generating a frame with
19212 unchanged height at half frame rate.
19217 Frame 1 Frame 2 Frame 3 Frame 4
19219 11111 22222 33333 44444
19220 11111 22222 33333 44444
19221 11111 22222 33333 44444
19222 11111 22222 33333 44444
19232 Only output even frames, odd frames are dropped, generating a frame with
19233 unchanged height at half frame rate.
19238 Frame 1 Frame 2 Frame 3 Frame 4
19240 11111 22222 33333 44444
19241 11111 22222 33333 44444
19242 11111 22222 33333 44444
19243 11111 22222 33333 44444
19253 Expand each frame to full height, but pad alternate lines with black,
19254 generating a frame with double height at the same input frame rate.
19259 Frame 1 Frame 2 Frame 3 Frame 4
19261 11111 22222 33333 44444
19262 11111 22222 33333 44444
19263 11111 22222 33333 44444
19264 11111 22222 33333 44444
19267 11111 ..... 33333 .....
19268 ..... 22222 ..... 44444
19269 11111 ..... 33333 .....
19270 ..... 22222 ..... 44444
19271 11111 ..... 33333 .....
19272 ..... 22222 ..... 44444
19273 11111 ..... 33333 .....
19274 ..... 22222 ..... 44444
19278 @item interleave_top, 4
19279 Interleave the upper field from odd frames with the lower field from
19280 even frames, generating a frame with unchanged height at half frame rate.
19285 Frame 1 Frame 2 Frame 3 Frame 4
19287 11111<- 22222 33333<- 44444
19288 11111 22222<- 33333 44444<-
19289 11111<- 22222 33333<- 44444
19290 11111 22222<- 33333 44444<-
19300 @item interleave_bottom, 5
19301 Interleave the lower field from odd frames with the upper field from
19302 even frames, generating a frame with unchanged height at half frame rate.
19307 Frame 1 Frame 2 Frame 3 Frame 4
19309 11111 22222<- 33333 44444<-
19310 11111<- 22222 33333<- 44444
19311 11111 22222<- 33333 44444<-
19312 11111<- 22222 33333<- 44444
19322 @item interlacex2, 6
19323 Double frame rate with unchanged height. Frames are inserted each
19324 containing the second temporal field from the previous input frame and
19325 the first temporal field from the next input frame. This mode relies on
19326 the top_field_first flag. Useful for interlaced video displays with no
19327 field synchronisation.
19332 Frame 1 Frame 2 Frame 3 Frame 4
19334 11111 22222 33333 44444
19335 11111 22222 33333 44444
19336 11111 22222 33333 44444
19337 11111 22222 33333 44444
19340 11111 22222 22222 33333 33333 44444 44444
19341 11111 11111 22222 22222 33333 33333 44444
19342 11111 22222 22222 33333 33333 44444 44444
19343 11111 11111 22222 22222 33333 33333 44444
19348 Move odd frames into the upper field, even into the lower field,
19349 generating a double height frame at same frame rate.
19354 Frame 1 Frame 2 Frame 3 Frame 4
19356 11111 22222 33333 44444
19357 11111 22222 33333 44444
19358 11111 22222 33333 44444
19359 11111 22222 33333 44444
19362 11111 33333 33333 55555
19363 22222 22222 44444 44444
19364 11111 33333 33333 55555
19365 22222 22222 44444 44444
19366 11111 33333 33333 55555
19367 22222 22222 44444 44444
19368 11111 33333 33333 55555
19369 22222 22222 44444 44444
19374 Numeric values are deprecated but are accepted for backward
19375 compatibility reasons.
19377 Default mode is @code{merge}.
19380 Specify flags influencing the filter process.
19382 Available value for @var{flags} is:
19385 @item low_pass_filter, vlpf
19386 Enable linear vertical low-pass filtering in the filter.
19387 Vertical low-pass filtering is required when creating an interlaced
19388 destination from a progressive source which contains high-frequency
19389 vertical detail. Filtering will reduce interlace 'twitter' and Moire
19392 @item complex_filter, cvlpf
19393 Enable complex vertical low-pass filtering.
19394 This will slightly less reduce interlace 'twitter' and Moire
19395 patterning but better retain detail and subjective sharpness impression.
19398 Bypass already interlaced frames, only adjust the frame rate.
19401 Vertical low-pass filtering and bypassing already interlaced frames can only be
19402 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
19407 Pick median pixels from several successive input video frames.
19409 The filter accepts the following options:
19413 Set radius of median filter.
19414 Default is 1. Allowed range is from 1 to 127.
19417 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
19420 Set median percentile. Default value is @code{0.5}.
19421 Default value of @code{0.5} will pick always median values, while @code{0} will pick
19422 minimum values, and @code{1} maximum values.
19425 @subsection Commands
19427 This filter supports all above options as @ref{commands}, excluding option @code{radius}.
19429 @section tmidequalizer
19431 Apply Temporal Midway Video Equalization effect.
19433 Midway Video Equalization adjusts a sequence of video frames to have the same
19434 histograms, while maintaining their dynamics as much as possible. It's
19435 useful for e.g. matching exposures from a video frames sequence.
19437 This filter accepts the following option:
19441 Set filtering radius. Default is @code{5}. Allowed range is from 1 to 127.
19444 Set filtering sigma. Default is @code{0.5}. This controls strength of filtering.
19445 Setting this option to 0 effectively does nothing.
19448 Set which planes to process. Default is @code{15}, which is all available planes.
19453 Mix successive video frames.
19455 A description of the accepted options follows.
19459 The number of successive frames to mix. If unspecified, it defaults to 3.
19462 Specify weight of each input video frame.
19463 Each weight is separated by space. If number of weights is smaller than
19464 number of @var{frames} last specified weight will be used for all remaining
19468 Specify scale, if it is set it will be multiplied with sum
19469 of each weight multiplied with pixel values to give final destination
19470 pixel value. By default @var{scale} is auto scaled to sum of weights.
19473 @subsection Examples
19477 Average 7 successive frames:
19479 tmix=frames=7:weights="1 1 1 1 1 1 1"
19483 Apply simple temporal convolution:
19485 tmix=frames=3:weights="-1 3 -1"
19489 Similar as above but only showing temporal differences:
19491 tmix=frames=3:weights="-1 2 -1":scale=1
19497 Tone map colors from different dynamic ranges.
19499 This filter expects data in single precision floating point, as it needs to
19500 operate on (and can output) out-of-range values. Another filter, such as
19501 @ref{zscale}, is needed to convert the resulting frame to a usable format.
19503 The tonemapping algorithms implemented only work on linear light, so input
19504 data should be linearized beforehand (and possibly correctly tagged).
19507 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
19510 @subsection Options
19511 The filter accepts the following options.
19515 Set the tone map algorithm to use.
19517 Possible values are:
19520 Do not apply any tone map, only desaturate overbright pixels.
19523 Hard-clip any out-of-range values. Use it for perfect color accuracy for
19524 in-range values, while distorting out-of-range values.
19527 Stretch the entire reference gamut to a linear multiple of the display.
19530 Fit a logarithmic transfer between the tone curves.
19533 Preserve overall image brightness with a simple curve, using nonlinear
19534 contrast, which results in flattening details and degrading color accuracy.
19537 Preserve both dark and bright details better than @var{reinhard}, at the cost
19538 of slightly darkening everything. Use it when detail preservation is more
19539 important than color and brightness accuracy.
19542 Smoothly map out-of-range values, while retaining contrast and colors for
19543 in-range material as much as possible. Use it when color accuracy is more
19544 important than detail preservation.
19550 Tune the tone mapping algorithm.
19552 This affects the following algorithms:
19558 Specifies the scale factor to use while stretching.
19562 Specifies the exponent of the function.
19566 Specify an extra linear coefficient to multiply into the signal before clipping.
19570 Specify the local contrast coefficient at the display peak.
19571 Default to 0.5, which means that in-gamut values will be about half as bright
19578 Specify the transition point from linear to mobius transform. Every value
19579 below this point is guaranteed to be mapped 1:1. The higher the value, the
19580 more accurate the result will be, at the cost of losing bright details.
19581 Default to 0.3, which due to the steep initial slope still preserves in-range
19582 colors fairly accurately.
19586 Apply desaturation for highlights that exceed this level of brightness. The
19587 higher the parameter, the more color information will be preserved. This
19588 setting helps prevent unnaturally blown-out colors for super-highlights, by
19589 (smoothly) turning into white instead. This makes images feel more natural,
19590 at the cost of reducing information about out-of-range colors.
19592 The default of 2.0 is somewhat conservative and will mostly just apply to
19593 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
19595 This option works only if the input frame has a supported color tag.
19598 Override signal/nominal/reference peak with this value. Useful when the
19599 embedded peak information in display metadata is not reliable or when tone
19600 mapping from a lower range to a higher range.
19605 Temporarily pad video frames.
19607 The filter accepts the following options:
19611 Specify number of delay frames before input video stream. Default is 0.
19614 Specify number of padding frames after input video stream.
19615 Set to -1 to pad indefinitely. Default is 0.
19618 Set kind of frames added to beginning of stream.
19619 Can be either @var{add} or @var{clone}.
19620 With @var{add} frames of solid-color are added.
19621 With @var{clone} frames are clones of first frame.
19622 Default is @var{add}.
19625 Set kind of frames added to end of stream.
19626 Can be either @var{add} or @var{clone}.
19627 With @var{add} frames of solid-color are added.
19628 With @var{clone} frames are clones of last frame.
19629 Default is @var{add}.
19631 @item start_duration, stop_duration
19632 Specify the duration of the start/stop delay. See
19633 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19634 for the accepted syntax.
19635 These options override @var{start} and @var{stop}. Default is 0.
19638 Specify the color of the padded area. For the syntax of this option,
19639 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
19640 manual,ffmpeg-utils}.
19642 The default value of @var{color} is "black".
19648 Transpose rows with columns in the input video and optionally flip it.
19650 It accepts the following parameters:
19655 Specify the transposition direction.
19657 Can assume the following values:
19659 @item 0, 4, cclock_flip
19660 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
19668 Rotate by 90 degrees clockwise, that is:
19676 Rotate by 90 degrees counterclockwise, that is:
19683 @item 3, 7, clock_flip
19684 Rotate by 90 degrees clockwise and vertically flip, that is:
19692 For values between 4-7, the transposition is only done if the input
19693 video geometry is portrait and not landscape. These values are
19694 deprecated, the @code{passthrough} option should be used instead.
19696 Numerical values are deprecated, and should be dropped in favor of
19697 symbolic constants.
19700 Do not apply the transposition if the input geometry matches the one
19701 specified by the specified value. It accepts the following values:
19704 Always apply transposition.
19706 Preserve portrait geometry (when @var{height} >= @var{width}).
19708 Preserve landscape geometry (when @var{width} >= @var{height}).
19711 Default value is @code{none}.
19714 For example to rotate by 90 degrees clockwise and preserve portrait
19717 transpose=dir=1:passthrough=portrait
19720 The command above can also be specified as:
19722 transpose=1:portrait
19725 @section transpose_npp
19727 Transpose rows with columns in the input video and optionally flip it.
19728 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
19730 It accepts the following parameters:
19735 Specify the transposition direction.
19737 Can assume the following values:
19740 Rotate by 90 degrees counterclockwise and vertically flip. (default)
19743 Rotate by 90 degrees clockwise.
19746 Rotate by 90 degrees counterclockwise.
19749 Rotate by 90 degrees clockwise and vertically flip.
19753 Do not apply the transposition if the input geometry matches the one
19754 specified by the specified value. It accepts the following values:
19757 Always apply transposition. (default)
19759 Preserve portrait geometry (when @var{height} >= @var{width}).
19761 Preserve landscape geometry (when @var{width} >= @var{height}).
19767 Trim the input so that the output contains one continuous subpart of the input.
19769 It accepts the following parameters:
19772 Specify the time of the start of the kept section, i.e. the frame with the
19773 timestamp @var{start} will be the first frame in the output.
19776 Specify the time of the first frame that will be dropped, i.e. the frame
19777 immediately preceding the one with the timestamp @var{end} will be the last
19778 frame in the output.
19781 This is the same as @var{start}, except this option sets the start timestamp
19782 in timebase units instead of seconds.
19785 This is the same as @var{end}, except this option sets the end timestamp
19786 in timebase units instead of seconds.
19789 The maximum duration of the output in seconds.
19792 The number of the first frame that should be passed to the output.
19795 The number of the first frame that should be dropped.
19798 @option{start}, @option{end}, and @option{duration} are expressed as time
19799 duration specifications; see
19800 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19801 for the accepted syntax.
19803 Note that the first two sets of the start/end options and the @option{duration}
19804 option look at the frame timestamp, while the _frame variants simply count the
19805 frames that pass through the filter. Also note that this filter does not modify
19806 the timestamps. If you wish for the output timestamps to start at zero, insert a
19807 setpts filter after the trim filter.
19809 If multiple start or end options are set, this filter tries to be greedy and
19810 keep all the frames that match at least one of the specified constraints. To keep
19811 only the part that matches all the constraints at once, chain multiple trim
19814 The defaults are such that all the input is kept. So it is possible to set e.g.
19815 just the end values to keep everything before the specified time.
19820 Drop everything except the second minute of input:
19822 ffmpeg -i INPUT -vf trim=60:120
19826 Keep only the first second:
19828 ffmpeg -i INPUT -vf trim=duration=1
19833 @section unpremultiply
19834 Apply alpha unpremultiply effect to input video stream using first plane
19835 of second stream as alpha.
19837 Both streams must have same dimensions and same pixel format.
19839 The filter accepts the following option:
19843 Set which planes will be processed, unprocessed planes will be copied.
19844 By default value 0xf, all planes will be processed.
19846 If the format has 1 or 2 components, then luma is bit 0.
19847 If the format has 3 or 4 components:
19848 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
19849 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
19850 If present, the alpha channel is always the last bit.
19853 Do not require 2nd input for processing, instead use alpha plane from input stream.
19859 Sharpen or blur the input video.
19861 It accepts the following parameters:
19864 @item luma_msize_x, lx
19865 Set the luma matrix horizontal size. It must be an odd integer between
19866 3 and 23. The default value is 5.
19868 @item luma_msize_y, ly
19869 Set the luma matrix vertical size. It must be an odd integer between 3
19870 and 23. The default value is 5.
19872 @item luma_amount, la
19873 Set the luma effect strength. It must be a floating point number, reasonable
19874 values lay between -1.5 and 1.5.
19876 Negative values will blur the input video, while positive values will
19877 sharpen it, a value of zero will disable the effect.
19879 Default value is 1.0.
19881 @item chroma_msize_x, cx
19882 Set the chroma matrix horizontal size. It must be an odd integer
19883 between 3 and 23. The default value is 5.
19885 @item chroma_msize_y, cy
19886 Set the chroma matrix vertical size. It must be an odd integer
19887 between 3 and 23. The default value is 5.
19889 @item chroma_amount, ca
19890 Set the chroma effect strength. It must be a floating point number, reasonable
19891 values lay between -1.5 and 1.5.
19893 Negative values will blur the input video, while positive values will
19894 sharpen it, a value of zero will disable the effect.
19896 Default value is 0.0.
19900 All parameters are optional and default to the equivalent of the
19901 string '5:5:1.0:5:5:0.0'.
19903 @subsection Examples
19907 Apply strong luma sharpen effect:
19909 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
19913 Apply a strong blur of both luma and chroma parameters:
19915 unsharp=7:7:-2:7:7:-2
19922 Decompose a video made of tiled images into the individual images.
19924 The frame rate of the output video is the frame rate of the input video
19925 multiplied by the number of tiles.
19927 This filter does the reverse of @ref{tile}.
19929 The filter accepts the following options:
19934 Set the grid size (i.e. the number of lines and columns). For the syntax of
19935 this option, check the
19936 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19939 @subsection Examples
19943 Produce a 1-second video from a still image file made of 25 frames stacked
19944 vertically, like an analogic film reel:
19946 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
19952 Apply ultra slow/simple postprocessing filter that compresses and decompresses
19953 the image at several (or - in the case of @option{quality} level @code{8} - all)
19954 shifts and average the results.
19956 The way this differs from the behavior of spp is that uspp actually encodes &
19957 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
19958 DCT similar to MJPEG.
19960 The filter accepts the following options:
19964 Set quality. This option defines the number of levels for averaging. It accepts
19965 an integer in the range 0-8. If set to @code{0}, the filter will have no
19966 effect. A value of @code{8} means the higher quality. For each increment of
19967 that value the speed drops by a factor of approximately 2. Default value is
19971 Force a constant quantization parameter. If not set, the filter will use the QP
19972 from the video stream (if available).
19977 Convert 360 videos between various formats.
19979 The filter accepts the following options:
19985 Set format of the input/output video.
19993 Equirectangular projection.
19998 Cubemap with 3x2/6x1/1x6 layout.
20000 Format specific options:
20005 Set padding proportion for the input/output cubemap. Values in decimals.
20012 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)
20015 Default value is @b{@samp{0}}.
20016 Maximum value is @b{@samp{0.1}}.
20020 Set fixed padding for the input/output cubemap. Values in pixels.
20022 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
20026 Set order of faces for the input/output cubemap. Choose one direction for each position.
20028 Designation of directions:
20044 Default value is @b{@samp{rludfb}}.
20048 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
20050 Designation of angles:
20053 0 degrees clockwise
20055 90 degrees clockwise
20057 180 degrees clockwise
20059 270 degrees clockwise
20062 Default value is @b{@samp{000000}}.
20066 Equi-Angular Cubemap.
20073 Format specific options:
20078 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20080 If diagonal field of view is set it overrides horizontal and vertical field of view.
20085 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20087 If diagonal field of view is set it overrides horizontal and vertical field of view.
20093 Format specific options:
20098 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20100 If diagonal field of view is set it overrides horizontal and vertical field of view.
20105 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20107 If diagonal field of view is set it overrides horizontal and vertical field of view.
20113 Facebook's 360 formats.
20116 Stereographic format.
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.
20139 Ball format, gives significant distortion toward the back.
20142 Hammer-Aitoff map projection format.
20145 Sinusoidal map projection format.
20148 Fisheye projection.
20150 Format specific options:
20155 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20157 If diagonal field of view is set it overrides horizontal and vertical field of view.
20162 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20164 If diagonal field of view is set it overrides horizontal and vertical field of view.
20168 Pannini projection.
20170 Format specific options:
20173 Set output pannini parameter.
20176 Set input pannini parameter.
20180 Cylindrical projection.
20182 Format specific options:
20187 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20189 If diagonal field of view is set it overrides horizontal and vertical field of view.
20194 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20196 If diagonal field of view is set it overrides horizontal and vertical field of view.
20200 Perspective projection. @i{(output only)}
20202 Format specific options:
20205 Set perspective parameter.
20209 Tetrahedron projection.
20212 Truncated square pyramid projection.
20216 Half equirectangular projection.
20221 Format specific options:
20226 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20228 If diagonal field of view is set it overrides horizontal and vertical field of view.
20233 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20235 If diagonal field of view is set it overrides horizontal and vertical field of view.
20239 Orthographic format.
20241 Format specific options:
20246 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20248 If diagonal field of view is set it overrides horizontal and vertical field of view.
20253 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20255 If diagonal field of view is set it overrides horizontal and vertical field of view.
20259 Octahedron projection.
20263 Set interpolation method.@*
20264 @i{Note: more complex interpolation methods require much more memory to run.}
20274 Bilinear interpolation.
20276 Lagrange9 interpolation.
20279 Bicubic interpolation.
20282 Lanczos interpolation.
20285 Spline16 interpolation.
20288 Gaussian interpolation.
20290 Mitchell interpolation.
20293 Default value is @b{@samp{line}}.
20297 Set the output video resolution.
20299 Default resolution depends on formats.
20303 Set the input/output stereo format.
20314 Default value is @b{@samp{2d}} for input and output format.
20319 Set rotation for the output video. Values in degrees.
20322 Set rotation order for the output video. Choose one item for each position.
20333 Default value is @b{@samp{ypr}}.
20338 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
20342 Set if input video is flipped horizontally/vertically. Boolean values.
20345 Set if input video is transposed. Boolean value, by default disabled.
20348 Set if output video needs to be transposed. Boolean value, by default disabled.
20351 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
20354 @subsection Examples
20358 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
20360 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
20363 Extract back view of Equi-Angular Cubemap:
20365 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
20368 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
20370 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
20374 @subsection Commands
20376 This filter supports subset of above options as @ref{commands}.
20378 @section vaguedenoiser
20380 Apply a wavelet based denoiser.
20382 It transforms each frame from the video input into the wavelet domain,
20383 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
20384 the obtained coefficients. It does an inverse wavelet transform after.
20385 Due to wavelet properties, it should give a nice smoothed result, and
20386 reduced noise, without blurring picture features.
20388 This filter accepts the following options:
20392 The filtering strength. The higher, the more filtered the video will be.
20393 Hard thresholding can use a higher threshold than soft thresholding
20394 before the video looks overfiltered. Default value is 2.
20397 The filtering method the filter will use.
20399 It accepts the following values:
20402 All values under the threshold will be zeroed.
20405 All values under the threshold will be zeroed. All values above will be
20406 reduced by the threshold.
20409 Scales or nullifies coefficients - intermediary between (more) soft and
20410 (less) hard thresholding.
20413 Default is garrote.
20416 Number of times, the wavelet will decompose the picture. Picture can't
20417 be decomposed beyond a particular point (typically, 8 for a 640x480
20418 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
20421 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
20424 A list of the planes to process. By default all planes are processed.
20427 The threshold type the filter will use.
20429 It accepts the following values:
20432 Threshold used is same for all decompositions.
20435 Threshold used depends also on each decomposition coefficients.
20438 Default is universal.
20441 @section vectorscope
20443 Display 2 color component values in the two dimensional graph (which is called
20446 This filter accepts the following options:
20450 Set vectorscope mode.
20452 It accepts the following values:
20456 Gray values are displayed on graph, higher brightness means more pixels have
20457 same component color value on location in graph. This is the default mode.
20460 Gray values are displayed on graph. Surrounding pixels values which are not
20461 present in video frame are drawn in gradient of 2 color components which are
20462 set by option @code{x} and @code{y}. The 3rd color component is static.
20465 Actual color components values present in video frame are displayed on graph.
20468 Similar as color2 but higher frequency of same values @code{x} and @code{y}
20469 on graph increases value of another color component, which is luminance by
20470 default values of @code{x} and @code{y}.
20473 Actual colors present in video frame are displayed on graph. If two different
20474 colors map to same position on graph then color with higher value of component
20475 not present in graph is picked.
20478 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
20479 component picked from radial gradient.
20483 Set which color component will be represented on X-axis. Default is @code{1}.
20486 Set which color component will be represented on Y-axis. Default is @code{2}.
20489 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
20490 of color component which represents frequency of (X, Y) location in graph.
20495 No envelope, this is default.
20498 Instant envelope, even darkest single pixel will be clearly highlighted.
20501 Hold maximum and minimum values presented in graph over time. This way you
20502 can still spot out of range values without constantly looking at vectorscope.
20505 Peak and instant envelope combined together.
20509 Set what kind of graticule to draw.
20518 Set graticule opacity.
20521 Set graticule flags.
20525 Draw graticule for white point.
20528 Draw graticule for black point.
20531 Draw color points short names.
20535 Set background opacity.
20537 @item lthreshold, l
20538 Set low threshold for color component not represented on X or Y axis.
20539 Values lower than this value will be ignored. Default is 0.
20540 Note this value is multiplied with actual max possible value one pixel component
20541 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
20544 @item hthreshold, h
20545 Set high threshold for color component not represented on X or Y axis.
20546 Values higher than this value will be ignored. Default is 1.
20547 Note this value is multiplied with actual max possible value one pixel component
20548 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
20549 is 0.9 * 255 = 230.
20551 @item colorspace, c
20552 Set what kind of colorspace to use when drawing graticule.
20562 Set color tint for gray/tint vectorscope mode. By default both options are zero.
20563 This means no tint, and output will remain gray.
20566 @anchor{vidstabdetect}
20567 @section vidstabdetect
20569 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
20570 @ref{vidstabtransform} for pass 2.
20572 This filter generates a file with relative translation and rotation
20573 transform information about subsequent frames, which is then used by
20574 the @ref{vidstabtransform} filter.
20576 To enable compilation of this filter you need to configure FFmpeg with
20577 @code{--enable-libvidstab}.
20579 This filter accepts the following options:
20583 Set the path to the file used to write the transforms information.
20584 Default value is @file{transforms.trf}.
20587 Set how shaky the video is and how quick the camera is. It accepts an
20588 integer in the range 1-10, a value of 1 means little shakiness, a
20589 value of 10 means strong shakiness. Default value is 5.
20592 Set the accuracy of the detection process. It must be a value in the
20593 range 1-15. A value of 1 means low accuracy, a value of 15 means high
20594 accuracy. Default value is 15.
20597 Set stepsize of the search process. The region around minimum is
20598 scanned with 1 pixel resolution. Default value is 6.
20601 Set minimum contrast. Below this value a local measurement field is
20602 discarded. Must be a floating point value in the range 0-1. Default
20606 Set reference frame number for tripod mode.
20608 If enabled, the motion of the frames is compared to a reference frame
20609 in the filtered stream, identified by the specified number. The idea
20610 is to compensate all movements in a more-or-less static scene and keep
20611 the camera view absolutely still.
20613 If set to 0, it is disabled. The frames are counted starting from 1.
20616 Show fields and transforms in the resulting frames. It accepts an
20617 integer in the range 0-2. Default value is 0, which disables any
20621 @subsection Examples
20625 Use default values:
20631 Analyze strongly shaky movie and put the results in file
20632 @file{mytransforms.trf}:
20634 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
20638 Visualize the result of internal transformations in the resulting
20641 vidstabdetect=show=1
20645 Analyze a video with medium shakiness using @command{ffmpeg}:
20647 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
20651 @anchor{vidstabtransform}
20652 @section vidstabtransform
20654 Video stabilization/deshaking: pass 2 of 2,
20655 see @ref{vidstabdetect} for pass 1.
20657 Read a file with transform information for each frame and
20658 apply/compensate them. Together with the @ref{vidstabdetect}
20659 filter this can be used to deshake videos. See also
20660 @url{http://public.hronopik.de/vid.stab}. It is important to also use
20661 the @ref{unsharp} filter, see below.
20663 To enable compilation of this filter you need to configure FFmpeg with
20664 @code{--enable-libvidstab}.
20666 @subsection Options
20670 Set path to the file used to read the transforms. Default value is
20671 @file{transforms.trf}.
20674 Set the number of frames (value*2 + 1) used for lowpass filtering the
20675 camera movements. Default value is 10.
20677 For example a number of 10 means that 21 frames are used (10 in the
20678 past and 10 in the future) to smoothen the motion in the video. A
20679 larger value leads to a smoother video, but limits the acceleration of
20680 the camera (pan/tilt movements). 0 is a special case where a static
20681 camera is simulated.
20684 Set the camera path optimization algorithm.
20686 Accepted values are:
20689 gaussian kernel low-pass filter on camera motion (default)
20691 averaging on transformations
20695 Set maximal number of pixels to translate frames. Default value is -1,
20699 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
20700 value is -1, meaning no limit.
20703 Specify how to deal with borders that may be visible due to movement
20706 Available values are:
20709 keep image information from previous frame (default)
20711 fill the border black
20715 Invert transforms if set to 1. Default value is 0.
20718 Consider transforms as relative to previous frame if set to 1,
20719 absolute if set to 0. Default value is 0.
20722 Set percentage to zoom. A positive value will result in a zoom-in
20723 effect, a negative value in a zoom-out effect. Default value is 0 (no
20727 Set optimal zooming to avoid borders.
20729 Accepted values are:
20734 optimal static zoom value is determined (only very strong movements
20735 will lead to visible borders) (default)
20737 optimal adaptive zoom value is determined (no borders will be
20738 visible), see @option{zoomspeed}
20741 Note that the value given at zoom is added to the one calculated here.
20744 Set percent to zoom maximally each frame (enabled when
20745 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
20749 Specify type of interpolation.
20751 Available values are:
20756 linear only horizontal
20758 linear in both directions (default)
20760 cubic in both directions (slow)
20764 Enable virtual tripod mode if set to 1, which is equivalent to
20765 @code{relative=0:smoothing=0}. Default value is 0.
20767 Use also @code{tripod} option of @ref{vidstabdetect}.
20770 Increase log verbosity if set to 1. Also the detected global motions
20771 are written to the temporary file @file{global_motions.trf}. Default
20775 @subsection Examples
20779 Use @command{ffmpeg} for a typical stabilization with default values:
20781 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
20784 Note the use of the @ref{unsharp} filter which is always recommended.
20787 Zoom in a bit more and load transform data from a given file:
20789 vidstabtransform=zoom=5:input="mytransforms.trf"
20793 Smoothen the video even more:
20795 vidstabtransform=smoothing=30
20801 Flip the input video vertically.
20803 For example, to vertically flip a video with @command{ffmpeg}:
20805 ffmpeg -i in.avi -vf "vflip" out.avi
20810 Detect variable frame rate video.
20812 This filter tries to detect if the input is variable or constant frame rate.
20814 At end it will output number of frames detected as having variable delta pts,
20815 and ones with constant delta pts.
20816 If there was frames with variable delta, than it will also show min, max and
20817 average delta encountered.
20821 Boost or alter saturation.
20823 The filter accepts the following options:
20826 Set strength of boost if positive value or strength of alter if negative value.
20827 Default is 0. Allowed range is from -2 to 2.
20830 Set the red balance. Default is 1. Allowed range is from -10 to 10.
20833 Set the green balance. Default is 1. Allowed range is from -10 to 10.
20836 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
20839 Set the red luma coefficient.
20842 Set the green luma coefficient.
20845 Set the blue luma coefficient.
20848 If @code{intensity} is negative and this is set to 1, colors will change,
20849 otherwise colors will be less saturated, more towards gray.
20852 @subsection Commands
20854 This filter supports the all above options as @ref{commands}.
20859 Make or reverse a natural vignetting effect.
20861 The filter accepts the following options:
20865 Set lens angle expression as a number of radians.
20867 The value is clipped in the @code{[0,PI/2]} range.
20869 Default value: @code{"PI/5"}
20873 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
20877 Set forward/backward mode.
20879 Available modes are:
20882 The larger the distance from the central point, the darker the image becomes.
20885 The larger the distance from the central point, the brighter the image becomes.
20886 This can be used to reverse a vignette effect, though there is no automatic
20887 detection to extract the lens @option{angle} and other settings (yet). It can
20888 also be used to create a burning effect.
20891 Default value is @samp{forward}.
20894 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
20896 It accepts the following values:
20899 Evaluate expressions only once during the filter initialization.
20902 Evaluate expressions for each incoming frame. This is way slower than the
20903 @samp{init} mode since it requires all the scalers to be re-computed, but it
20904 allows advanced dynamic expressions.
20907 Default value is @samp{init}.
20910 Set dithering to reduce the circular banding effects. Default is @code{1}
20914 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
20915 Setting this value to the SAR of the input will make a rectangular vignetting
20916 following the dimensions of the video.
20918 Default is @code{1/1}.
20921 @subsection Expressions
20923 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
20924 following parameters.
20929 input width and height
20932 the number of input frame, starting from 0
20935 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
20936 @var{TB} units, NAN if undefined
20939 frame rate of the input video, NAN if the input frame rate is unknown
20942 the PTS (Presentation TimeStamp) of the filtered video frame,
20943 expressed in seconds, NAN if undefined
20946 time base of the input video
20950 @subsection Examples
20954 Apply simple strong vignetting effect:
20960 Make a flickering vignetting:
20962 vignette='PI/4+random(1)*PI/50':eval=frame
20967 @section vmafmotion
20969 Obtain the average VMAF motion score of a video.
20970 It is one of the component metrics of VMAF.
20972 The obtained average motion score is printed through the logging system.
20974 The filter accepts the following options:
20978 If specified, the filter will use the named file to save the motion score of
20979 each frame with respect to the previous frame.
20980 When filename equals "-" the data is sent to standard output.
20985 ffmpeg -i ref.mpg -vf vmafmotion -f null -
20989 Stack input videos vertically.
20991 All streams must be of same pixel format and of same width.
20993 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
20994 to create same output.
20996 The filter accepts the following options:
21000 Set number of input streams. Default is 2.
21003 If set to 1, force the output to terminate when the shortest input
21004 terminates. Default value is 0.
21009 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
21010 Deinterlacing Filter").
21012 Based on the process described by Martin Weston for BBC R&D, and
21013 implemented based on the de-interlace algorithm written by Jim
21014 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
21015 uses filter coefficients calculated by BBC R&D.
21017 This filter uses field-dominance information in frame to decide which
21018 of each pair of fields to place first in the output.
21019 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
21021 There are two sets of filter coefficients, so called "simple"
21022 and "complex". Which set of filter coefficients is used can
21023 be set by passing an optional parameter:
21027 Set the interlacing filter coefficients. Accepts one of the following values:
21031 Simple filter coefficient set.
21033 More-complex filter coefficient set.
21035 Default value is @samp{complex}.
21038 The interlacing mode to adopt. It accepts one of the following values:
21042 Output one frame for each frame.
21044 Output one frame for each field.
21047 The default value is @code{field}.
21050 The picture field parity assumed for the input interlaced video. It accepts one
21051 of the following values:
21055 Assume the top field is first.
21057 Assume the bottom field is first.
21059 Enable automatic detection of field parity.
21062 The default value is @code{auto}.
21063 If the interlacing is unknown or the decoder does not export this information,
21064 top field first will be assumed.
21067 Specify which frames to deinterlace. Accepts one of the following values:
21071 Deinterlace all frames,
21073 Only deinterlace frames marked as interlaced.
21076 Default value is @samp{all}.
21079 @subsection Commands
21080 This filter supports same @ref{commands} as options.
21083 Video waveform monitor.
21085 The waveform monitor plots color component intensity. By default luminance
21086 only. Each column of the waveform corresponds to a column of pixels in the
21089 It accepts the following options:
21093 Can be either @code{row}, or @code{column}. Default is @code{column}.
21094 In row mode, the graph on the left side represents color component value 0 and
21095 the right side represents value = 255. In column mode, the top side represents
21096 color component value = 0 and bottom side represents value = 255.
21099 Set intensity. Smaller values are useful to find out how many values of the same
21100 luminance are distributed across input rows/columns.
21101 Default value is @code{0.04}. Allowed range is [0, 1].
21104 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
21105 In mirrored mode, higher values will be represented on the left
21106 side for @code{row} mode and at the top for @code{column} mode. Default is
21107 @code{1} (mirrored).
21111 It accepts the following values:
21114 Presents information identical to that in the @code{parade}, except
21115 that the graphs representing color components are superimposed directly
21118 This display mode makes it easier to spot relative differences or similarities
21119 in overlapping areas of the color components that are supposed to be identical,
21120 such as neutral whites, grays, or blacks.
21123 Display separate graph for the color components side by side in
21124 @code{row} mode or one below the other in @code{column} mode.
21127 Display separate graph for the color components side by side in
21128 @code{column} mode or one below the other in @code{row} mode.
21130 Using this display mode makes it easy to spot color casts in the highlights
21131 and shadows of an image, by comparing the contours of the top and the bottom
21132 graphs of each waveform. Since whites, grays, and blacks are characterized
21133 by exactly equal amounts of red, green, and blue, neutral areas of the picture
21134 should display three waveforms of roughly equal width/height. If not, the
21135 correction is easy to perform by making level adjustments the three waveforms.
21137 Default is @code{stack}.
21139 @item components, c
21140 Set which color components to display. Default is 1, which means only luminance
21141 or red color component if input is in RGB colorspace. If is set for example to
21142 7 it will display all 3 (if) available color components.
21147 No envelope, this is default.
21150 Instant envelope, minimum and maximum values presented in graph will be easily
21151 visible even with small @code{step} value.
21154 Hold minimum and maximum values presented in graph across time. This way you
21155 can still spot out of range values without constantly looking at waveforms.
21158 Peak and instant envelope combined together.
21164 No filtering, this is default.
21167 Luma and chroma combined together.
21170 Similar as above, but shows difference between blue and red chroma.
21173 Similar as above, but use different colors.
21176 Similar as above, but again with different colors.
21179 Displays only chroma.
21182 Displays actual color value on waveform.
21185 Similar as above, but with luma showing frequency of chroma values.
21189 Set which graticule to display.
21193 Do not display graticule.
21196 Display green graticule showing legal broadcast ranges.
21199 Display orange graticule showing legal broadcast ranges.
21202 Display invert graticule showing legal broadcast ranges.
21206 Set graticule opacity.
21209 Set graticule flags.
21213 Draw numbers above lines. By default enabled.
21216 Draw dots instead of lines.
21220 Set scale used for displaying graticule.
21227 Default is digital.
21230 Set background opacity.
21234 Set tint for output.
21235 Only used with lowpass filter and when display is not overlay and input
21236 pixel formats are not RGB.
21239 @section weave, doubleweave
21241 The @code{weave} takes a field-based video input and join
21242 each two sequential fields into single frame, producing a new double
21243 height clip with half the frame rate and half the frame count.
21245 The @code{doubleweave} works same as @code{weave} but without
21246 halving frame rate and frame count.
21248 It accepts the following option:
21252 Set first field. Available values are:
21256 Set the frame as top-field-first.
21259 Set the frame as bottom-field-first.
21263 @subsection Examples
21267 Interlace video using @ref{select} and @ref{separatefields} filter:
21269 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
21274 Apply the xBR high-quality magnification filter which is designed for pixel
21275 art. It follows a set of edge-detection rules, see
21276 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
21278 It accepts the following option:
21282 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
21283 @code{3xBR} and @code{4} for @code{4xBR}.
21284 Default is @code{3}.
21289 Apply cross fade from one input video stream to another input video stream.
21290 The cross fade is applied for specified duration.
21292 The filter accepts the following options:
21296 Set one of available transition effects:
21344 Default transition effect is fade.
21347 Set cross fade duration in seconds.
21348 Default duration is 1 second.
21351 Set cross fade start relative to first input stream in seconds.
21352 Default offset is 0.
21355 Set expression for custom transition effect.
21357 The expressions can use the following variables and functions:
21362 The coordinates of the current sample.
21366 The width and height of the image.
21369 Progress of transition effect.
21372 Currently processed plane.
21375 Return value of first input at current location and plane.
21378 Return value of second input at current location and plane.
21384 Return the value of the pixel at location (@var{x},@var{y}) of the
21385 first/second/third/fourth component of first input.
21391 Return the value of the pixel at location (@var{x},@var{y}) of the
21392 first/second/third/fourth component of second input.
21396 @subsection Examples
21400 Cross fade from one input video to another input video, with fade transition and duration of transition
21401 of 2 seconds starting at offset of 5 seconds:
21403 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
21408 Pick median pixels from several input videos.
21410 The filter accepts the following options:
21414 Set number of inputs.
21415 Default is 3. Allowed range is from 3 to 255.
21416 If number of inputs is even number, than result will be mean value between two median values.
21419 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
21422 Set median percentile. Default value is @code{0.5}.
21423 Default value of @code{0.5} will pick always median values, while @code{0} will pick
21424 minimum values, and @code{1} maximum values.
21427 @subsection Commands
21429 This filter supports all above options as @ref{commands}, excluding option @code{inputs}.
21432 Stack video inputs into custom layout.
21434 All streams must be of same pixel format.
21436 The filter accepts the following options:
21440 Set number of input streams. Default is 2.
21443 Specify layout of inputs.
21444 This option requires the desired layout configuration to be explicitly set by the user.
21445 This sets position of each video input in output. Each input
21446 is separated by '|'.
21447 The first number represents the column, and the second number represents the row.
21448 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
21449 where X is video input from which to take width or height.
21450 Multiple values can be used when separated by '+'. In such
21451 case values are summed together.
21453 Note that if inputs are of different sizes gaps may appear, as not all of
21454 the output video frame will be filled. Similarly, videos can overlap each
21455 other if their position doesn't leave enough space for the full frame of
21458 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
21459 a layout must be set by the user.
21462 If set to 1, force the output to terminate when the shortest input
21463 terminates. Default value is 0.
21466 If set to valid color, all unused pixels will be filled with that color.
21467 By default fill is set to none, so it is disabled.
21470 @subsection Examples
21474 Display 4 inputs into 2x2 grid.
21478 input1(0, 0) | input3(w0, 0)
21479 input2(0, h0) | input4(w0, h0)
21483 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
21486 Note that if inputs are of different sizes, gaps or overlaps may occur.
21489 Display 4 inputs into 1x4 grid.
21496 input4(0, h0+h1+h2)
21500 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
21503 Note that if inputs are of different widths, unused space will appear.
21506 Display 9 inputs into 3x3 grid.
21510 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
21511 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
21512 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
21516 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
21519 Note that if inputs are of different sizes, gaps or overlaps may occur.
21522 Display 16 inputs into 4x4 grid.
21526 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
21527 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
21528 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
21529 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
21533 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|
21534 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
21537 Note that if inputs are of different sizes, gaps or overlaps may occur.
21544 Deinterlace the input video ("yadif" means "yet another deinterlacing
21547 It accepts the following parameters:
21553 The interlacing mode to adopt. It accepts one of the following values:
21556 @item 0, send_frame
21557 Output one frame for each frame.
21558 @item 1, send_field
21559 Output one frame for each field.
21560 @item 2, send_frame_nospatial
21561 Like @code{send_frame}, but it skips the spatial interlacing check.
21562 @item 3, send_field_nospatial
21563 Like @code{send_field}, but it skips the spatial interlacing check.
21566 The default value is @code{send_frame}.
21569 The picture field parity assumed for the input interlaced video. It accepts one
21570 of the following values:
21574 Assume the top field is first.
21576 Assume the bottom field is first.
21578 Enable automatic detection of field parity.
21581 The default value is @code{auto}.
21582 If the interlacing is unknown or the decoder does not export this information,
21583 top field first will be assumed.
21586 Specify which frames to deinterlace. Accepts one of the following
21591 Deinterlace all frames.
21592 @item 1, interlaced
21593 Only deinterlace frames marked as interlaced.
21596 The default value is @code{all}.
21599 @section yadif_cuda
21601 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
21602 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
21605 It accepts the following parameters:
21611 The interlacing mode to adopt. It accepts one of the following values:
21614 @item 0, send_frame
21615 Output one frame for each frame.
21616 @item 1, send_field
21617 Output one frame for each field.
21618 @item 2, send_frame_nospatial
21619 Like @code{send_frame}, but it skips the spatial interlacing check.
21620 @item 3, send_field_nospatial
21621 Like @code{send_field}, but it skips the spatial interlacing check.
21624 The default value is @code{send_frame}.
21627 The picture field parity assumed for the input interlaced video. It accepts one
21628 of the following values:
21632 Assume the top field is first.
21634 Assume the bottom field is first.
21636 Enable automatic detection of field parity.
21639 The default value is @code{auto}.
21640 If the interlacing is unknown or the decoder does not export this information,
21641 top field first will be assumed.
21644 Specify which frames to deinterlace. Accepts one of the following
21649 Deinterlace all frames.
21650 @item 1, interlaced
21651 Only deinterlace frames marked as interlaced.
21654 The default value is @code{all}.
21659 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
21660 The algorithm is described in
21661 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
21663 It accepts the following parameters:
21667 Set the window radius. Default value is 3.
21670 Set which planes to filter. Default is only the first plane.
21673 Set blur strength. Default value is 128.
21676 @subsection Commands
21677 This filter supports same @ref{commands} as options.
21681 Apply Zoom & Pan effect.
21683 This filter accepts the following options:
21687 Set the zoom expression. Range is 1-10. Default is 1.
21691 Set the x and y expression. Default is 0.
21694 Set the duration expression in number of frames.
21695 This sets for how many number of frames effect will last for
21696 single input image.
21699 Set the output image size, default is 'hd720'.
21702 Set the output frame rate, default is '25'.
21705 Each expression can contain the following constants:
21724 Output frame count.
21727 The input timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
21729 @item out_time, time, ot
21730 The output timestamp expressed in seconds.
21734 Last calculated 'x' and 'y' position from 'x' and 'y' expression
21735 for current input frame.
21739 'x' and 'y' of last output frame of previous input frame or 0 when there was
21740 not yet such frame (first input frame).
21743 Last calculated zoom from 'z' expression for current input frame.
21746 Last calculated zoom of last output frame of previous input frame.
21749 Number of output frames for current input frame. Calculated from 'd' expression
21750 for each input frame.
21753 number of output frames created for previous input frame
21756 Rational number: input width / input height
21759 sample aspect ratio
21762 display aspect ratio
21766 @subsection Examples
21770 Zoom in up to 1.5x and pan at same time to some spot near center of picture:
21772 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
21776 Zoom in up to 1.5x and pan always at center of picture:
21778 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21782 Same as above but without pausing:
21784 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21788 Zoom in 2x into center of picture only for the first second of the input video:
21790 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21797 Scale (resize) the input video, using the z.lib library:
21798 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
21799 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
21801 The zscale filter forces the output display aspect ratio to be the same
21802 as the input, by changing the output sample aspect ratio.
21804 If the input image format is different from the format requested by
21805 the next filter, the zscale filter will convert the input to the
21808 @subsection Options
21809 The filter accepts the following options.
21814 Set the output video dimension expression. Default value is the input
21817 If the @var{width} or @var{w} value is 0, the input width is used for
21818 the output. If the @var{height} or @var{h} value is 0, the input height
21819 is used for the output.
21821 If one and only one of the values is -n with n >= 1, the zscale filter
21822 will use a value that maintains the aspect ratio of the input image,
21823 calculated from the other specified dimension. After that it will,
21824 however, make sure that the calculated dimension is divisible by n and
21825 adjust the value if necessary.
21827 If both values are -n with n >= 1, the behavior will be identical to
21828 both values being set to 0 as previously detailed.
21830 See below for the list of accepted constants for use in the dimension
21834 Set the video size. For the syntax of this option, check the
21835 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21838 Set the dither type.
21840 Possible values are:
21845 @item error_diffusion
21851 Set the resize filter type.
21853 Possible values are:
21863 Default is bilinear.
21866 Set the color range.
21868 Possible values are:
21875 Default is same as input.
21878 Set the color primaries.
21880 Possible values are:
21890 Default is same as input.
21893 Set the transfer characteristics.
21895 Possible values are:
21909 Default is same as input.
21912 Set the colorspace matrix.
21914 Possible value are:
21925 Default is same as input.
21928 Set the input color range.
21930 Possible values are:
21937 Default is same as input.
21939 @item primariesin, pin
21940 Set the input color primaries.
21942 Possible values are:
21952 Default is same as input.
21954 @item transferin, tin
21955 Set the input transfer characteristics.
21957 Possible values are:
21968 Default is same as input.
21970 @item matrixin, min
21971 Set the input colorspace matrix.
21973 Possible value are:
21985 Set the output chroma location.
21987 Possible values are:
21998 @item chromalin, cin
21999 Set the input chroma location.
22001 Possible values are:
22013 Set the nominal peak luminance.
22016 The values of the @option{w} and @option{h} options are expressions
22017 containing the following constants:
22022 The input width and height
22026 These are the same as @var{in_w} and @var{in_h}.
22030 The output (scaled) width and height
22034 These are the same as @var{out_w} and @var{out_h}
22037 The same as @var{iw} / @var{ih}
22040 input sample aspect ratio
22043 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
22047 horizontal and vertical input chroma subsample values. For example for the
22048 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
22052 horizontal and vertical output chroma subsample values. For example for the
22053 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
22056 @subsection Commands
22058 This filter supports the following commands:
22062 Set the output video dimension expression.
22063 The command accepts the same syntax of the corresponding option.
22065 If the specified expression is not valid, it is kept at its current
22069 @c man end VIDEO FILTERS
22071 @chapter OpenCL Video Filters
22072 @c man begin OPENCL VIDEO FILTERS
22074 Below is a description of the currently available OpenCL video filters.
22076 To enable compilation of these filters you need to configure FFmpeg with
22077 @code{--enable-opencl}.
22079 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
22082 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
22083 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
22084 given device parameters.
22086 @item -filter_hw_device @var{name}
22087 Pass the hardware device called @var{name} to all filters in any filter graph.
22091 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
22095 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
22097 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
22101 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.
22103 @section avgblur_opencl
22105 Apply average blur filter.
22107 The filter accepts the following options:
22111 Set horizontal radius size.
22112 Range is @code{[1, 1024]} and default value is @code{1}.
22115 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22118 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
22121 @subsection Example
22125 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.
22127 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
22131 @section boxblur_opencl
22133 Apply a boxblur algorithm to the input video.
22135 It accepts the following parameters:
22139 @item luma_radius, lr
22140 @item luma_power, lp
22141 @item chroma_radius, cr
22142 @item chroma_power, cp
22143 @item alpha_radius, ar
22144 @item alpha_power, ap
22148 A description of the accepted options follows.
22151 @item luma_radius, lr
22152 @item chroma_radius, cr
22153 @item alpha_radius, ar
22154 Set an expression for the box radius in pixels used for blurring the
22155 corresponding input plane.
22157 The radius value must be a non-negative number, and must not be
22158 greater than the value of the expression @code{min(w,h)/2} for the
22159 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
22162 Default value for @option{luma_radius} is "2". If not specified,
22163 @option{chroma_radius} and @option{alpha_radius} default to the
22164 corresponding value set for @option{luma_radius}.
22166 The expressions can contain the following constants:
22170 The input width and height in pixels.
22174 The input chroma image width and height in pixels.
22178 The horizontal and vertical chroma subsample values. For example, for the
22179 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
22182 @item luma_power, lp
22183 @item chroma_power, cp
22184 @item alpha_power, ap
22185 Specify how many times the boxblur filter is applied to the
22186 corresponding plane.
22188 Default value for @option{luma_power} is 2. If not specified,
22189 @option{chroma_power} and @option{alpha_power} default to the
22190 corresponding value set for @option{luma_power}.
22192 A value of 0 will disable the effect.
22195 @subsection Examples
22197 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.
22201 Apply a boxblur filter with the luma, chroma, and alpha radius
22202 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.
22204 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
22205 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
22209 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.
22211 For the luma plane, a 2x2 box radius will be run once.
22213 For the chroma plane, a 4x4 box radius will be run 5 times.
22215 For the alpha plane, a 3x3 box radius will be run 7 times.
22217 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
22221 @section colorkey_opencl
22222 RGB colorspace color keying.
22224 The filter accepts the following options:
22228 The color which will be replaced with transparency.
22231 Similarity percentage with the key color.
22233 0.01 matches only the exact key color, while 1.0 matches everything.
22238 0.0 makes pixels either fully transparent, or not transparent at all.
22240 Higher values result in semi-transparent pixels, with a higher transparency
22241 the more similar the pixels color is to the key color.
22244 @subsection Examples
22248 Make every semi-green pixel in the input transparent with some slight blending:
22250 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
22254 @section convolution_opencl
22256 Apply convolution of 3x3, 5x5, 7x7 matrix.
22258 The filter accepts the following options:
22265 Set matrix for each plane.
22266 Matrix is sequence of 9, 25 or 49 signed numbers.
22267 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
22273 Set multiplier for calculated value for each plane.
22274 If unset or 0, it will be sum of all matrix elements.
22275 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
22281 Set bias for each plane. This value is added to the result of the multiplication.
22282 Useful for making the overall image brighter or darker.
22283 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
22287 @subsection Examples
22293 -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
22299 -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
22303 Apply edge enhance:
22305 -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
22311 -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
22315 Apply laplacian edge detector which includes diagonals:
22317 -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
22323 -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
22327 @section erosion_opencl
22329 Apply erosion effect to the video.
22331 This filter replaces the pixel by the local(3x3) minimum.
22333 It accepts the following options:
22340 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22341 If @code{0}, plane will remain unchanged.
22344 Flag which specifies the pixel to refer to.
22345 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22347 Flags to local 3x3 coordinates region centered on @code{x}:
22356 @subsection Example
22360 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.
22362 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22366 @section deshake_opencl
22367 Feature-point based video stabilization filter.
22369 The filter accepts the following options:
22373 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
22376 Whether or not additional debug info should be displayed, both in the processed output and in the console.
22378 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
22380 Viewing point matches in the output video is only supported for RGB input.
22382 Defaults to @code{0}.
22384 @item adaptive_crop
22385 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
22387 Defaults to @code{1}.
22389 @item refine_features
22390 Whether or not feature points should be refined at a sub-pixel level.
22392 This can be turned off for a slight performance gain at the cost of precision.
22394 Defaults to @code{1}.
22396 @item smooth_strength
22397 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
22399 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
22401 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
22403 Defaults to @code{0.0}.
22405 @item smooth_window_multiplier
22406 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
22408 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
22410 Acceptable values range from @code{0.1} to @code{10.0}.
22412 Larger values increase the amount of motion data available for determining how to smooth the camera path,
22413 potentially improving smoothness, but also increase latency and memory usage.
22415 Defaults to @code{2.0}.
22419 @subsection Examples
22423 Stabilize a video with a fixed, medium smoothing strength:
22425 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
22429 Stabilize a video with debugging (both in console and in rendered video):
22431 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
22435 @section dilation_opencl
22437 Apply dilation effect to the video.
22439 This filter replaces the pixel by the local(3x3) maximum.
22441 It accepts the following options:
22448 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22449 If @code{0}, plane will remain unchanged.
22452 Flag which specifies the pixel to refer to.
22453 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22455 Flags to local 3x3 coordinates region centered on @code{x}:
22464 @subsection Example
22468 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.
22470 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22474 @section nlmeans_opencl
22476 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
22478 @section overlay_opencl
22480 Overlay one video on top of another.
22482 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
22483 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
22485 The filter accepts the following options:
22490 Set the x coordinate of the overlaid video on the main video.
22491 Default value is @code{0}.
22494 Set the y coordinate of the overlaid video on the main video.
22495 Default value is @code{0}.
22499 @subsection Examples
22503 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
22505 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22508 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
22510 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22515 @section pad_opencl
22517 Add paddings to the input image, and place the original input at the
22518 provided @var{x}, @var{y} coordinates.
22520 It accepts the following options:
22525 Specify an expression for the size of the output image with the
22526 paddings added. If the value for @var{width} or @var{height} is 0, the
22527 corresponding input size is used for the output.
22529 The @var{width} expression can reference the value set by the
22530 @var{height} expression, and vice versa.
22532 The default value of @var{width} and @var{height} is 0.
22536 Specify the offsets to place the input image at within the padded area,
22537 with respect to the top/left border of the output image.
22539 The @var{x} expression can reference the value set by the @var{y}
22540 expression, and vice versa.
22542 The default value of @var{x} and @var{y} is 0.
22544 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
22545 so the input image is centered on the padded area.
22548 Specify the color of the padded area. For the syntax of this option,
22549 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
22550 manual,ffmpeg-utils}.
22553 Pad to an aspect instead to a resolution.
22556 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
22557 options are expressions containing the following constants:
22562 The input video width and height.
22566 These are the same as @var{in_w} and @var{in_h}.
22570 The output width and height (the size of the padded area), as
22571 specified by the @var{width} and @var{height} expressions.
22575 These are the same as @var{out_w} and @var{out_h}.
22579 The x and y offsets as specified by the @var{x} and @var{y}
22580 expressions, or NAN if not yet specified.
22583 same as @var{iw} / @var{ih}
22586 input sample aspect ratio
22589 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
22592 @section prewitt_opencl
22594 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
22596 The filter accepts the following option:
22600 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22603 Set value which will be multiplied with filtered result.
22604 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22607 Set value which will be added to filtered result.
22608 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22611 @subsection Example
22615 Apply the Prewitt operator with scale set to 2 and delta set to 10.
22617 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
22621 @anchor{program_opencl}
22622 @section program_opencl
22624 Filter video using an OpenCL program.
22629 OpenCL program source file.
22632 Kernel name in program.
22635 Number of inputs to the filter. Defaults to 1.
22638 Size of output frames. Defaults to the same as the first input.
22642 The @code{program_opencl} filter also supports the @ref{framesync} options.
22644 The program source file must contain a kernel function with the given name,
22645 which will be run once for each plane of the output. Each run on a plane
22646 gets enqueued as a separate 2D global NDRange with one work-item for each
22647 pixel to be generated. The global ID offset for each work-item is therefore
22648 the coordinates of a pixel in the destination image.
22650 The kernel function needs to take the following arguments:
22653 Destination image, @var{__write_only image2d_t}.
22655 This image will become the output; the kernel should write all of it.
22657 Frame index, @var{unsigned int}.
22659 This is a counter starting from zero and increasing by one for each frame.
22661 Source images, @var{__read_only image2d_t}.
22663 These are the most recent images on each input. The kernel may read from
22664 them to generate the output, but they can't be written to.
22671 Copy the input to the output (output must be the same size as the input).
22673 __kernel void copy(__write_only image2d_t destination,
22674 unsigned int index,
22675 __read_only image2d_t source)
22677 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
22679 int2 location = (int2)(get_global_id(0), get_global_id(1));
22681 float4 value = read_imagef(source, sampler, location);
22683 write_imagef(destination, location, value);
22688 Apply a simple transformation, rotating the input by an amount increasing
22689 with the index counter. Pixel values are linearly interpolated by the
22690 sampler, and the output need not have the same dimensions as the input.
22692 __kernel void rotate_image(__write_only image2d_t dst,
22693 unsigned int index,
22694 __read_only image2d_t src)
22696 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22697 CLK_FILTER_LINEAR);
22699 float angle = (float)index / 100.0f;
22701 float2 dst_dim = convert_float2(get_image_dim(dst));
22702 float2 src_dim = convert_float2(get_image_dim(src));
22704 float2 dst_cen = dst_dim / 2.0f;
22705 float2 src_cen = src_dim / 2.0f;
22707 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22709 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
22711 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
22712 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
22714 src_pos = src_pos * src_dim / dst_dim;
22716 float2 src_loc = src_pos + src_cen;
22718 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
22719 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
22720 write_imagef(dst, dst_loc, 0.5f);
22722 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
22727 Blend two inputs together, with the amount of each input used varying
22728 with the index counter.
22730 __kernel void blend_images(__write_only image2d_t dst,
22731 unsigned int index,
22732 __read_only image2d_t src1,
22733 __read_only image2d_t src2)
22735 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22736 CLK_FILTER_LINEAR);
22738 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
22740 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22741 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
22742 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
22744 float4 val1 = read_imagef(src1, sampler, src1_loc);
22745 float4 val2 = read_imagef(src2, sampler, src2_loc);
22747 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
22753 @section roberts_opencl
22754 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
22756 The filter accepts the following option:
22760 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22763 Set value which will be multiplied with filtered result.
22764 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22767 Set value which will be added to filtered result.
22768 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22771 @subsection Example
22775 Apply the Roberts cross operator with scale set to 2 and delta set to 10
22777 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
22781 @section sobel_opencl
22783 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
22785 The filter accepts the following option:
22789 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22792 Set value which will be multiplied with filtered result.
22793 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22796 Set value which will be added to filtered result.
22797 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22800 @subsection Example
22804 Apply sobel operator with scale set to 2 and delta set to 10
22806 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
22810 @section tonemap_opencl
22812 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
22814 It accepts the following parameters:
22818 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
22821 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
22824 Apply desaturation for highlights that exceed this level of brightness. The
22825 higher the parameter, the more color information will be preserved. This
22826 setting helps prevent unnaturally blown-out colors for super-highlights, by
22827 (smoothly) turning into white instead. This makes images feel more natural,
22828 at the cost of reducing information about out-of-range colors.
22830 The default value is 0.5, and the algorithm here is a little different from
22831 the cpu version tonemap currently. A setting of 0.0 disables this option.
22834 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
22835 is used to detect whether the scene has changed or not. If the distance between
22836 the current frame average brightness and the current running average exceeds
22837 a threshold value, we would re-calculate scene average and peak brightness.
22838 The default value is 0.2.
22841 Specify the output pixel format.
22843 Currently supported formats are:
22850 Set the output color range.
22852 Possible values are:
22858 Default is same as input.
22861 Set the output color primaries.
22863 Possible values are:
22869 Default is same as input.
22872 Set the output transfer characteristics.
22874 Possible values are:
22883 Set the output colorspace matrix.
22885 Possible value are:
22891 Default is same as input.
22895 @subsection Example
22899 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
22901 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
22905 @section unsharp_opencl
22907 Sharpen or blur the input video.
22909 It accepts the following parameters:
22912 @item luma_msize_x, lx
22913 Set the luma matrix horizontal size.
22914 Range is @code{[1, 23]} and default value is @code{5}.
22916 @item luma_msize_y, ly
22917 Set the luma matrix vertical size.
22918 Range is @code{[1, 23]} and default value is @code{5}.
22920 @item luma_amount, la
22921 Set the luma effect strength.
22922 Range is @code{[-10, 10]} and default value is @code{1.0}.
22924 Negative values will blur the input video, while positive values will
22925 sharpen it, a value of zero will disable the effect.
22927 @item chroma_msize_x, cx
22928 Set the chroma matrix horizontal size.
22929 Range is @code{[1, 23]} and default value is @code{5}.
22931 @item chroma_msize_y, cy
22932 Set the chroma matrix vertical size.
22933 Range is @code{[1, 23]} and default value is @code{5}.
22935 @item chroma_amount, ca
22936 Set the chroma effect strength.
22937 Range is @code{[-10, 10]} and default value is @code{0.0}.
22939 Negative values will blur the input video, while positive values will
22940 sharpen it, a value of zero will disable the effect.
22944 All parameters are optional and default to the equivalent of the
22945 string '5:5:1.0:5:5:0.0'.
22947 @subsection Examples
22951 Apply strong luma sharpen effect:
22953 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
22957 Apply a strong blur of both luma and chroma parameters:
22959 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
22963 @section xfade_opencl
22965 Cross fade two videos with custom transition effect by using OpenCL.
22967 It accepts the following options:
22971 Set one of possible transition effects.
22975 Select custom transition effect, the actual transition description
22976 will be picked from source and kernel options.
22988 Default transition is fade.
22992 OpenCL program source file for custom transition.
22995 Set name of kernel to use for custom transition from program source file.
22998 Set duration of video transition.
23001 Set time of start of transition relative to first video.
23004 The program source file must contain a kernel function with the given name,
23005 which will be run once for each plane of the output. Each run on a plane
23006 gets enqueued as a separate 2D global NDRange with one work-item for each
23007 pixel to be generated. The global ID offset for each work-item is therefore
23008 the coordinates of a pixel in the destination image.
23010 The kernel function needs to take the following arguments:
23013 Destination image, @var{__write_only image2d_t}.
23015 This image will become the output; the kernel should write all of it.
23018 First Source image, @var{__read_only image2d_t}.
23019 Second Source image, @var{__read_only image2d_t}.
23021 These are the most recent images on each input. The kernel may read from
23022 them to generate the output, but they can't be written to.
23025 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
23032 Apply dots curtain transition effect:
23034 __kernel void blend_images(__write_only image2d_t dst,
23035 __read_only image2d_t src1,
23036 __read_only image2d_t src2,
23039 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
23040 CLK_FILTER_LINEAR);
23041 int2 p = (int2)(get_global_id(0), get_global_id(1));
23042 float2 rp = (float2)(get_global_id(0), get_global_id(1));
23043 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
23046 float2 dots = (float2)(20.0, 20.0);
23047 float2 center = (float2)(0,0);
23050 float4 val1 = read_imagef(src1, sampler, p);
23051 float4 val2 = read_imagef(src2, sampler, p);
23052 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
23054 write_imagef(dst, p, next ? val1 : val2);
23060 @c man end OPENCL VIDEO FILTERS
23062 @chapter VAAPI Video Filters
23063 @c man begin VAAPI VIDEO FILTERS
23065 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
23067 To enable compilation of these filters you need to configure FFmpeg with
23068 @code{--enable-vaapi}.
23070 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}
23072 @section tonemap_vaapi
23074 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
23075 It maps the dynamic range of HDR10 content to the SDR content.
23076 It currently only accepts HDR10 as input.
23078 It accepts the following parameters:
23082 Specify the output pixel format.
23084 Currently supported formats are:
23093 Set the output color primaries.
23095 Default is same as input.
23098 Set the output transfer characteristics.
23103 Set the output colorspace matrix.
23105 Default is same as input.
23109 @subsection Example
23113 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
23115 tonemap_vaapi=format=p010:t=bt2020-10
23119 @c man end VAAPI VIDEO FILTERS
23121 @chapter Video Sources
23122 @c man begin VIDEO SOURCES
23124 Below is a description of the currently available video sources.
23128 Buffer video frames, and make them available to the filter chain.
23130 This source is mainly intended for a programmatic use, in particular
23131 through the interface defined in @file{libavfilter/buffersrc.h}.
23133 It accepts the following parameters:
23138 Specify the size (width and height) of the buffered video frames. For the
23139 syntax of this option, check the
23140 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23143 The input video width.
23146 The input video height.
23149 A string representing the pixel format of the buffered video frames.
23150 It may be a number corresponding to a pixel format, or a pixel format
23154 Specify the timebase assumed by the timestamps of the buffered frames.
23157 Specify the frame rate expected for the video stream.
23159 @item pixel_aspect, sar
23160 The sample (pixel) aspect ratio of the input video.
23163 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
23164 to the filtergraph description to specify swscale flags for automatically
23165 inserted scalers. See @ref{Filtergraph syntax}.
23167 @item hw_frames_ctx
23168 When using a hardware pixel format, this should be a reference to an
23169 AVHWFramesContext describing input frames.
23174 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
23177 will instruct the source to accept video frames with size 320x240 and
23178 with format "yuv410p", assuming 1/24 as the timestamps timebase and
23179 square pixels (1:1 sample aspect ratio).
23180 Since the pixel format with name "yuv410p" corresponds to the number 6
23181 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
23182 this example corresponds to:
23184 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
23187 Alternatively, the options can be specified as a flat string, but this
23188 syntax is deprecated:
23190 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
23194 Create a pattern generated by an elementary cellular automaton.
23196 The initial state of the cellular automaton can be defined through the
23197 @option{filename} and @option{pattern} options. If such options are
23198 not specified an initial state is created randomly.
23200 At each new frame a new row in the video is filled with the result of
23201 the cellular automaton next generation. The behavior when the whole
23202 frame is filled is defined by the @option{scroll} option.
23204 This source accepts the following options:
23208 Read the initial cellular automaton state, i.e. the starting row, from
23209 the specified file.
23210 In the file, each non-whitespace character is considered an alive
23211 cell, a newline will terminate the row, and further characters in the
23212 file will be ignored.
23215 Read the initial cellular automaton state, i.e. the starting row, from
23216 the specified string.
23218 Each non-whitespace character in the string is considered an alive
23219 cell, a newline will terminate the row, and further characters in the
23220 string will be ignored.
23223 Set the video rate, that is the number of frames generated per second.
23226 @item random_fill_ratio, ratio
23227 Set the random fill ratio for the initial cellular automaton row. It
23228 is a floating point number value ranging from 0 to 1, defaults to
23231 This option is ignored when a file or a pattern is specified.
23233 @item random_seed, seed
23234 Set the seed for filling randomly the initial row, must be an integer
23235 included between 0 and UINT32_MAX. If not specified, or if explicitly
23236 set to -1, the filter will try to use a good random seed on a best
23240 Set the cellular automaton rule, it is a number ranging from 0 to 255.
23241 Default value is 110.
23244 Set the size of the output video. For the syntax of this option, check the
23245 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23247 If @option{filename} or @option{pattern} is specified, the size is set
23248 by default to the width of the specified initial state row, and the
23249 height is set to @var{width} * PHI.
23251 If @option{size} is set, it must contain the width of the specified
23252 pattern string, and the specified pattern will be centered in the
23255 If a filename or a pattern string is not specified, the size value
23256 defaults to "320x518" (used for a randomly generated initial state).
23259 If set to 1, scroll the output upward when all the rows in the output
23260 have been already filled. If set to 0, the new generated row will be
23261 written over the top row just after the bottom row is filled.
23264 @item start_full, full
23265 If set to 1, completely fill the output with generated rows before
23266 outputting the first frame.
23267 This is the default behavior, for disabling set the value to 0.
23270 If set to 1, stitch the left and right row edges together.
23271 This is the default behavior, for disabling set the value to 0.
23274 @subsection Examples
23278 Read the initial state from @file{pattern}, and specify an output of
23281 cellauto=f=pattern:s=200x400
23285 Generate a random initial row with a width of 200 cells, with a fill
23288 cellauto=ratio=2/3:s=200x200
23292 Create a pattern generated by rule 18 starting by a single alive cell
23293 centered on an initial row with width 100:
23295 cellauto=p=@@:s=100x400:full=0:rule=18
23299 Specify a more elaborated initial pattern:
23301 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
23306 @anchor{coreimagesrc}
23307 @section coreimagesrc
23308 Video source generated on GPU using Apple's CoreImage API on OSX.
23310 This video source is a specialized version of the @ref{coreimage} video filter.
23311 Use a core image generator at the beginning of the applied filterchain to
23312 generate the content.
23314 The coreimagesrc video source accepts the following options:
23316 @item list_generators
23317 List all available generators along with all their respective options as well as
23318 possible minimum and maximum values along with the default values.
23320 list_generators=true
23324 Specify the size of the sourced video. For the syntax of this option, check the
23325 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23326 The default value is @code{320x240}.
23329 Specify the frame rate of the sourced video, as the number of frames
23330 generated per second. It has to be a string in the format
23331 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23332 number or a valid video frame rate abbreviation. The default value is
23336 Set the sample aspect ratio of the sourced video.
23339 Set the duration of the sourced video. See
23340 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23341 for the accepted syntax.
23343 If not specified, or the expressed duration is negative, the video is
23344 supposed to be generated forever.
23347 Additionally, all options of the @ref{coreimage} video filter are accepted.
23348 A complete filterchain can be used for further processing of the
23349 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
23350 and examples for details.
23352 @subsection Examples
23357 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
23358 given as complete and escaped command-line for Apple's standard bash shell:
23360 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
23362 This example is equivalent to the QRCode example of @ref{coreimage} without the
23363 need for a nullsrc video source.
23368 Generate several gradients.
23372 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23373 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23376 Set frame rate, expressed as number of frames per second. Default
23379 @item c0, c1, c2, c3, c4, c5, c6, c7
23380 Set 8 colors. Default values for colors is to pick random one.
23382 @item x0, y0, y0, y1
23383 Set gradient line source and destination points. If negative or out of range, random ones
23387 Set number of colors to use at once. Allowed range is from 2 to 8. Default value is 2.
23390 Set seed for picking gradient line points.
23393 Set the duration of the sourced video. See
23394 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23395 for the accepted syntax.
23397 If not specified, or the expressed duration is negative, the video is
23398 supposed to be generated forever.
23401 Set speed of gradients rotation.
23405 @section mandelbrot
23407 Generate a Mandelbrot set fractal, and progressively zoom towards the
23408 point specified with @var{start_x} and @var{start_y}.
23410 This source accepts the following options:
23415 Set the terminal pts value. Default value is 400.
23418 Set the terminal scale value.
23419 Must be a floating point value. Default value is 0.3.
23422 Set the inner coloring mode, that is the algorithm used to draw the
23423 Mandelbrot fractal internal region.
23425 It shall assume one of the following values:
23430 Show time until convergence.
23432 Set color based on point closest to the origin of the iterations.
23437 Default value is @var{mincol}.
23440 Set the bailout value. Default value is 10.0.
23443 Set the maximum of iterations performed by the rendering
23444 algorithm. Default value is 7189.
23447 Set outer coloring mode.
23448 It shall assume one of following values:
23450 @item iteration_count
23451 Set iteration count mode.
23452 @item normalized_iteration_count
23453 set normalized iteration count mode.
23455 Default value is @var{normalized_iteration_count}.
23458 Set frame rate, expressed as number of frames per second. Default
23462 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23463 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23466 Set the initial scale value. Default value is 3.0.
23469 Set the initial x position. Must be a floating point value between
23470 -100 and 100. Default value is -0.743643887037158704752191506114774.
23473 Set the initial y position. Must be a floating point value between
23474 -100 and 100. Default value is -0.131825904205311970493132056385139.
23479 Generate various test patterns, as generated by the MPlayer test filter.
23481 The size of the generated video is fixed, and is 256x256.
23482 This source is useful in particular for testing encoding features.
23484 This source accepts the following options:
23489 Specify the frame rate of the sourced video, as the number of frames
23490 generated per second. It has to be a string in the format
23491 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23492 number or a valid video frame rate abbreviation. The default value is
23496 Set the duration of the sourced video. See
23497 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23498 for the accepted syntax.
23500 If not specified, or the expressed duration is negative, the video is
23501 supposed to be generated forever.
23505 Set the number or the name of the test to perform. Supported tests are:
23519 @item max_frames, m
23520 Set the maximum number of frames generated for each test, default value is 30.
23524 Default value is "all", which will cycle through the list of all tests.
23529 mptestsrc=t=dc_luma
23532 will generate a "dc_luma" test pattern.
23534 @section frei0r_src
23536 Provide a frei0r source.
23538 To enable compilation of this filter you need to install the frei0r
23539 header and configure FFmpeg with @code{--enable-frei0r}.
23541 This source accepts the following parameters:
23546 The size of the video to generate. For the syntax of this option, check the
23547 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23550 The framerate of the generated video. It may be a string of the form
23551 @var{num}/@var{den} or a frame rate abbreviation.
23554 The name to the frei0r source to load. For more information regarding frei0r and
23555 how to set the parameters, read the @ref{frei0r} section in the video filters
23558 @item filter_params
23559 A '|'-separated list of parameters to pass to the frei0r source.
23563 For example, to generate a frei0r partik0l source with size 200x200
23564 and frame rate 10 which is overlaid on the overlay filter main input:
23566 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
23571 Generate a life pattern.
23573 This source is based on a generalization of John Conway's life game.
23575 The sourced input represents a life grid, each pixel represents a cell
23576 which can be in one of two possible states, alive or dead. Every cell
23577 interacts with its eight neighbours, which are the cells that are
23578 horizontally, vertically, or diagonally adjacent.
23580 At each interaction the grid evolves according to the adopted rule,
23581 which specifies the number of neighbor alive cells which will make a
23582 cell stay alive or born. The @option{rule} option allows one to specify
23585 This source accepts the following options:
23589 Set the file from which to read the initial grid state. In the file,
23590 each non-whitespace character is considered an alive cell, and newline
23591 is used to delimit the end of each row.
23593 If this option is not specified, the initial grid is generated
23597 Set the video rate, that is the number of frames generated per second.
23600 @item random_fill_ratio, ratio
23601 Set the random fill ratio for the initial random grid. It is a
23602 floating point number value ranging from 0 to 1, defaults to 1/PHI.
23603 It is ignored when a file is specified.
23605 @item random_seed, seed
23606 Set the seed for filling the initial random grid, must be an integer
23607 included between 0 and UINT32_MAX. If not specified, or if explicitly
23608 set to -1, the filter will try to use a good random seed on a best
23614 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
23615 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
23616 @var{NS} specifies the number of alive neighbor cells which make a
23617 live cell stay alive, and @var{NB} the number of alive neighbor cells
23618 which make a dead cell to become alive (i.e. to "born").
23619 "s" and "b" can be used in place of "S" and "B", respectively.
23621 Alternatively a rule can be specified by an 18-bits integer. The 9
23622 high order bits are used to encode the next cell state if it is alive
23623 for each number of neighbor alive cells, the low order bits specify
23624 the rule for "borning" new cells. Higher order bits encode for an
23625 higher number of neighbor cells.
23626 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
23627 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
23629 Default value is "S23/B3", which is the original Conway's game of life
23630 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
23631 cells, and will born a new cell if there are three alive cells around
23635 Set the size of the output video. For the syntax of this option, check the
23636 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23638 If @option{filename} is specified, the size is set by default to the
23639 same size of the input file. If @option{size} is set, it must contain
23640 the size specified in the input file, and the initial grid defined in
23641 that file is centered in the larger resulting area.
23643 If a filename is not specified, the size value defaults to "320x240"
23644 (used for a randomly generated initial grid).
23647 If set to 1, stitch the left and right grid edges together, and the
23648 top and bottom edges also. Defaults to 1.
23651 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
23652 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
23653 value from 0 to 255.
23656 Set the color of living (or new born) cells.
23659 Set the color of dead cells. If @option{mold} is set, this is the first color
23660 used to represent a dead cell.
23663 Set mold color, for definitely dead and moldy cells.
23665 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
23666 ffmpeg-utils manual,ffmpeg-utils}.
23669 @subsection Examples
23673 Read a grid from @file{pattern}, and center it on a grid of size
23676 life=f=pattern:s=300x300
23680 Generate a random grid of size 200x200, with a fill ratio of 2/3:
23682 life=ratio=2/3:s=200x200
23686 Specify a custom rule for evolving a randomly generated grid:
23692 Full example with slow death effect (mold) using @command{ffplay}:
23694 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
23701 @anchor{haldclutsrc}
23704 @anchor{pal100bars}
23705 @anchor{rgbtestsrc}
23707 @anchor{smptehdbars}
23710 @anchor{yuvtestsrc}
23711 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
23713 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
23715 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
23717 The @code{color} source provides an uniformly colored input.
23719 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
23720 @ref{haldclut} filter.
23722 The @code{nullsrc} source returns unprocessed video frames. It is
23723 mainly useful to be employed in analysis / debugging tools, or as the
23724 source for filters which ignore the input data.
23726 The @code{pal75bars} source generates a color bars pattern, based on
23727 EBU PAL recommendations with 75% color levels.
23729 The @code{pal100bars} source generates a color bars pattern, based on
23730 EBU PAL recommendations with 100% color levels.
23732 The @code{rgbtestsrc} source generates an RGB test pattern useful for
23733 detecting RGB vs BGR issues. You should see a red, green and blue
23734 stripe from top to bottom.
23736 The @code{smptebars} source generates a color bars pattern, based on
23737 the SMPTE Engineering Guideline EG 1-1990.
23739 The @code{smptehdbars} source generates a color bars pattern, based on
23740 the SMPTE RP 219-2002.
23742 The @code{testsrc} source generates a test video pattern, showing a
23743 color pattern, a scrolling gradient and a timestamp. This is mainly
23744 intended for testing purposes.
23746 The @code{testsrc2} source is similar to testsrc, but supports more
23747 pixel formats instead of just @code{rgb24}. This allows using it as an
23748 input for other tests without requiring a format conversion.
23750 The @code{yuvtestsrc} source generates an YUV test pattern. You should
23751 see a y, cb and cr stripe from top to bottom.
23753 The sources accept the following parameters:
23758 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
23759 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
23760 pixels to be used as identity matrix for 3D lookup tables. Each component is
23761 coded on a @code{1/(N*N)} scale.
23764 Specify the color of the source, only available in the @code{color}
23765 source. For the syntax of this option, check the
23766 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
23769 Specify the size of the sourced video. For the syntax of this option, check the
23770 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23771 The default value is @code{320x240}.
23773 This option is not available with the @code{allrgb}, @code{allyuv}, and
23774 @code{haldclutsrc} filters.
23777 Specify the frame rate of the sourced video, as the number of frames
23778 generated per second. It has to be a string in the format
23779 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23780 number or a valid video frame rate abbreviation. The default value is
23784 Set the duration of the sourced video. See
23785 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23786 for the accepted syntax.
23788 If not specified, or the expressed duration is negative, the video is
23789 supposed to be generated forever.
23791 Since the frame rate is used as time base, all frames including the last one
23792 will have their full duration. If the specified duration is not a multiple
23793 of the frame duration, it will be rounded up.
23796 Set the sample aspect ratio of the sourced video.
23799 Specify the alpha (opacity) of the background, only available in the
23800 @code{testsrc2} source. The value must be between 0 (fully transparent) and
23801 255 (fully opaque, the default).
23804 Set the number of decimals to show in the timestamp, only available in the
23805 @code{testsrc} source.
23807 The displayed timestamp value will correspond to the original
23808 timestamp value multiplied by the power of 10 of the specified
23809 value. Default value is 0.
23812 @subsection Examples
23816 Generate a video with a duration of 5.3 seconds, with size
23817 176x144 and a frame rate of 10 frames per second:
23819 testsrc=duration=5.3:size=qcif:rate=10
23823 The following graph description will generate a red source
23824 with an opacity of 0.2, with size "qcif" and a frame rate of 10
23827 color=c=red@@0.2:s=qcif:r=10
23831 If the input content is to be ignored, @code{nullsrc} can be used. The
23832 following command generates noise in the luminance plane by employing
23833 the @code{geq} filter:
23835 nullsrc=s=256x256, geq=random(1)*255:128:128
23839 @subsection Commands
23841 The @code{color} source supports the following commands:
23845 Set the color of the created image. Accepts the same syntax of the
23846 corresponding @option{color} option.
23851 Generate video using an OpenCL program.
23856 OpenCL program source file.
23859 Kernel name in program.
23862 Size of frames to generate. This must be set.
23865 Pixel format to use for the generated frames. This must be set.
23868 Number of frames generated every second. Default value is '25'.
23872 For details of how the program loading works, see the @ref{program_opencl}
23879 Generate a colour ramp by setting pixel values from the position of the pixel
23880 in the output image. (Note that this will work with all pixel formats, but
23881 the generated output will not be the same.)
23883 __kernel void ramp(__write_only image2d_t dst,
23884 unsigned int index)
23886 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23889 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
23891 write_imagef(dst, loc, val);
23896 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
23898 __kernel void sierpinski_carpet(__write_only image2d_t dst,
23899 unsigned int index)
23901 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23903 float4 value = 0.0f;
23904 int x = loc.x + index;
23905 int y = loc.y + index;
23906 while (x > 0 || y > 0) {
23907 if (x % 3 == 1 && y % 3 == 1) {
23915 write_imagef(dst, loc, value);
23921 @section sierpinski
23923 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
23925 This source accepts the following options:
23929 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23930 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23933 Set frame rate, expressed as number of frames per second. Default
23937 Set seed which is used for random panning.
23940 Set max jump for single pan destination. Allowed range is from 1 to 10000.
23943 Set fractal type, can be default @code{carpet} or @code{triangle}.
23946 @c man end VIDEO SOURCES
23948 @chapter Video Sinks
23949 @c man begin VIDEO SINKS
23951 Below is a description of the currently available video sinks.
23953 @section buffersink
23955 Buffer video frames, and make them available to the end of the filter
23958 This sink is mainly intended for programmatic use, in particular
23959 through the interface defined in @file{libavfilter/buffersink.h}
23960 or the options system.
23962 It accepts a pointer to an AVBufferSinkContext structure, which
23963 defines the incoming buffers' formats, to be passed as the opaque
23964 parameter to @code{avfilter_init_filter} for initialization.
23968 Null video sink: do absolutely nothing with the input video. It is
23969 mainly useful as a template and for use in analysis / debugging
23972 @c man end VIDEO SINKS
23974 @chapter Multimedia Filters
23975 @c man begin MULTIMEDIA FILTERS
23977 Below is a description of the currently available multimedia filters.
23981 Convert input audio to a video output, displaying the audio bit scope.
23983 The filter accepts the following options:
23987 Set frame rate, expressed as number of frames per second. Default
23991 Specify the video size for the output. For the syntax of this option, check the
23992 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23993 Default value is @code{1024x256}.
23996 Specify list of colors separated by space or by '|' which will be used to
23997 draw channels. Unrecognized or missing colors will be replaced
24001 @section adrawgraph
24002 Draw a graph using input audio metadata.
24004 See @ref{drawgraph}
24006 @section agraphmonitor
24008 See @ref{graphmonitor}.
24010 @section ahistogram
24012 Convert input audio to a video output, displaying the volume histogram.
24014 The filter accepts the following options:
24018 Specify how histogram is calculated.
24020 It accepts the following values:
24023 Use single histogram for all channels.
24025 Use separate histogram for each channel.
24027 Default is @code{single}.
24030 Set frame rate, expressed as number of frames per second. Default
24034 Specify the video size for the output. For the syntax of this option, check the
24035 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24036 Default value is @code{hd720}.
24041 It accepts the following values:
24052 reverse logarithmic
24054 Default is @code{log}.
24057 Set amplitude scale.
24059 It accepts the following values:
24066 Default is @code{log}.
24069 Set how much frames to accumulate in histogram.
24070 Default is 1. Setting this to -1 accumulates all frames.
24073 Set histogram ratio of window height.
24076 Set sonogram sliding.
24078 It accepts the following values:
24081 replace old rows with new ones.
24083 scroll from top to bottom.
24085 Default is @code{replace}.
24088 @section aphasemeter
24090 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
24091 representing mean phase of current audio frame. A video output can also be produced and is
24092 enabled by default. The audio is passed through as first output.
24094 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
24095 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
24096 and @code{1} means channels are in phase.
24098 The filter accepts the following options, all related to its video output:
24102 Set the output frame rate. Default value is @code{25}.
24105 Set the video size for the output. For the syntax of this option, check the
24106 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24107 Default value is @code{800x400}.
24112 Specify the red, green, blue contrast. Default values are @code{2},
24113 @code{7} and @code{1}.
24114 Allowed range is @code{[0, 255]}.
24117 Set color which will be used for drawing median phase. If color is
24118 @code{none} which is default, no median phase value will be drawn.
24121 Enable video output. Default is enabled.
24124 @subsection phasing detection
24126 The filter also detects out of phase and mono sequences in stereo streams.
24127 It logs the sequence start, end and duration when it lasts longer or as long as the minimum set.
24129 The filter accepts the following options for this detection:
24133 Enable mono and out of phase detection. Default is disabled.
24136 Set phase tolerance for mono detection, in amplitude ratio. Default is @code{0}.
24137 Allowed range is @code{[0, 1]}.
24140 Set angle threshold for out of phase detection, in degree. Default is @code{170}.
24141 Allowed range is @code{[90, 180]}.
24144 Set mono or out of phase duration until notification, expressed in seconds. Default is @code{2}.
24147 @subsection Examples
24151 Complete example with @command{ffmpeg} to detect 1 second of mono with 0.001 phase tolerance:
24153 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
24157 @section avectorscope
24159 Convert input audio to a video output, representing the audio vector
24162 The filter is used to measure the difference between channels of stereo
24163 audio stream. A monaural signal, consisting of identical left and right
24164 signal, results in straight vertical line. Any stereo separation is visible
24165 as a deviation from this line, creating a Lissajous figure.
24166 If the straight (or deviation from it) but horizontal line appears this
24167 indicates that the left and right channels are out of phase.
24169 The filter accepts the following options:
24173 Set the vectorscope mode.
24175 Available values are:
24178 Lissajous rotated by 45 degrees.
24181 Same as above but not rotated.
24184 Shape resembling half of circle.
24187 Default value is @samp{lissajous}.
24190 Set the video size for the output. For the syntax of this option, check the
24191 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24192 Default value is @code{400x400}.
24195 Set the output frame rate. Default value is @code{25}.
24201 Specify the red, green, blue and alpha contrast. Default values are @code{40},
24202 @code{160}, @code{80} and @code{255}.
24203 Allowed range is @code{[0, 255]}.
24209 Specify the red, green, blue and alpha fade. Default values are @code{15},
24210 @code{10}, @code{5} and @code{5}.
24211 Allowed range is @code{[0, 255]}.
24214 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
24215 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
24218 Set the vectorscope drawing mode.
24220 Available values are:
24223 Draw dot for each sample.
24226 Draw line between previous and current sample.
24229 Default value is @samp{dot}.
24232 Specify amplitude scale of audio samples.
24234 Available values are:
24250 Swap left channel axis with right channel axis.
24260 Mirror only x axis.
24263 Mirror only y axis.
24271 @subsection Examples
24275 Complete example using @command{ffplay}:
24277 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
24278 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
24282 @section bench, abench
24284 Benchmark part of a filtergraph.
24286 The filter accepts the following options:
24290 Start or stop a timer.
24292 Available values are:
24295 Get the current time, set it as frame metadata (using the key
24296 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
24299 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
24300 the input frame metadata to get the time difference. Time difference, average,
24301 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
24302 @code{min}) are then printed. The timestamps are expressed in seconds.
24306 @subsection Examples
24310 Benchmark @ref{selectivecolor} filter:
24312 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
24318 Concatenate audio and video streams, joining them together one after the
24321 The filter works on segments of synchronized video and audio streams. All
24322 segments must have the same number of streams of each type, and that will
24323 also be the number of streams at output.
24325 The filter accepts the following options:
24330 Set the number of segments. Default is 2.
24333 Set the number of output video streams, that is also the number of video
24334 streams in each segment. Default is 1.
24337 Set the number of output audio streams, that is also the number of audio
24338 streams in each segment. Default is 0.
24341 Activate unsafe mode: do not fail if segments have a different format.
24345 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
24346 @var{a} audio outputs.
24348 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
24349 segment, in the same order as the outputs, then the inputs for the second
24352 Related streams do not always have exactly the same duration, for various
24353 reasons including codec frame size or sloppy authoring. For that reason,
24354 related synchronized streams (e.g. a video and its audio track) should be
24355 concatenated at once. The concat filter will use the duration of the longest
24356 stream in each segment (except the last one), and if necessary pad shorter
24357 audio streams with silence.
24359 For this filter to work correctly, all segments must start at timestamp 0.
24361 All corresponding streams must have the same parameters in all segments; the
24362 filtering system will automatically select a common pixel format for video
24363 streams, and a common sample format, sample rate and channel layout for
24364 audio streams, but other settings, such as resolution, must be converted
24365 explicitly by the user.
24367 Different frame rates are acceptable but will result in variable frame rate
24368 at output; be sure to configure the output file to handle it.
24370 @subsection Examples
24374 Concatenate an opening, an episode and an ending, all in bilingual version
24375 (video in stream 0, audio in streams 1 and 2):
24377 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
24378 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
24379 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
24380 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
24384 Concatenate two parts, handling audio and video separately, using the
24385 (a)movie sources, and adjusting the resolution:
24387 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
24388 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
24389 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
24391 Note that a desync will happen at the stitch if the audio and video streams
24392 do not have exactly the same duration in the first file.
24396 @subsection Commands
24398 This filter supports the following commands:
24401 Close the current segment and step to the next one
24407 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
24408 level. By default, it logs a message at a frequency of 10Hz with the
24409 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
24410 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
24412 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
24413 sample format is double-precision floating point. The input stream will be converted to
24414 this specification, if needed. Users may need to insert aformat and/or aresample filters
24415 after this filter to obtain the original parameters.
24417 The filter also has a video output (see the @var{video} option) with a real
24418 time graph to observe the loudness evolution. The graphic contains the logged
24419 message mentioned above, so it is not printed anymore when this option is set,
24420 unless the verbose logging is set. The main graphing area contains the
24421 short-term loudness (3 seconds of analysis), and the gauge on the right is for
24422 the momentary loudness (400 milliseconds), but can optionally be configured
24423 to instead display short-term loudness (see @var{gauge}).
24425 The green area marks a +/- 1LU target range around the target loudness
24426 (-23LUFS by default, unless modified through @var{target}).
24428 More information about the Loudness Recommendation EBU R128 on
24429 @url{http://tech.ebu.ch/loudness}.
24431 The filter accepts the following options:
24436 Activate the video output. The audio stream is passed unchanged whether this
24437 option is set or no. The video stream will be the first output stream if
24438 activated. Default is @code{0}.
24441 Set the video size. This option is for video only. For the syntax of this
24443 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24444 Default and minimum resolution is @code{640x480}.
24447 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
24448 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
24449 other integer value between this range is allowed.
24452 Set metadata injection. If set to @code{1}, the audio input will be segmented
24453 into 100ms output frames, each of them containing various loudness information
24454 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
24456 Default is @code{0}.
24459 Force the frame logging level.
24461 Available values are:
24464 information logging level
24466 verbose logging level
24469 By default, the logging level is set to @var{info}. If the @option{video} or
24470 the @option{metadata} options are set, it switches to @var{verbose}.
24475 Available modes can be cumulated (the option is a @code{flag} type). Possible
24479 Disable any peak mode (default).
24481 Enable sample-peak mode.
24483 Simple peak mode looking for the higher sample value. It logs a message
24484 for sample-peak (identified by @code{SPK}).
24486 Enable true-peak mode.
24488 If enabled, the peak lookup is done on an over-sampled version of the input
24489 stream for better peak accuracy. It logs a message for true-peak.
24490 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
24491 This mode requires a build with @code{libswresample}.
24495 Treat mono input files as "dual mono". If a mono file is intended for playback
24496 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
24497 If set to @code{true}, this option will compensate for this effect.
24498 Multi-channel input files are not affected by this option.
24501 Set a specific pan law to be used for the measurement of dual mono files.
24502 This parameter is optional, and has a default value of -3.01dB.
24505 Set a specific target level (in LUFS) used as relative zero in the visualization.
24506 This parameter is optional and has a default value of -23LUFS as specified
24507 by EBU R128. However, material published online may prefer a level of -16LUFS
24508 (e.g. for use with podcasts or video platforms).
24511 Set the value displayed by the gauge. Valid values are @code{momentary} and s
24512 @code{shortterm}. By default the momentary value will be used, but in certain
24513 scenarios it may be more useful to observe the short term value instead (e.g.
24517 Sets the display scale for the loudness. Valid parameters are @code{absolute}
24518 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
24519 video output, not the summary or continuous log output.
24522 @subsection Examples
24526 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
24528 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
24532 Run an analysis with @command{ffmpeg}:
24534 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
24538 @section interleave, ainterleave
24540 Temporally interleave frames from several inputs.
24542 @code{interleave} works with video inputs, @code{ainterleave} with audio.
24544 These filters read frames from several inputs and send the oldest
24545 queued frame to the output.
24547 Input streams must have well defined, monotonically increasing frame
24550 In order to submit one frame to output, these filters need to enqueue
24551 at least one frame for each input, so they cannot work in case one
24552 input is not yet terminated and will not receive incoming frames.
24554 For example consider the case when one input is a @code{select} filter
24555 which always drops input frames. The @code{interleave} filter will keep
24556 reading from that input, but it will never be able to send new frames
24557 to output until the input sends an end-of-stream signal.
24559 Also, depending on inputs synchronization, the filters will drop
24560 frames in case one input receives more frames than the other ones, and
24561 the queue is already filled.
24563 These filters accept the following options:
24567 Set the number of different inputs, it is 2 by default.
24570 How to determine the end-of-stream.
24574 The duration of the longest input. (default)
24577 The duration of the shortest input.
24580 The duration of the first input.
24585 @subsection Examples
24589 Interleave frames belonging to different streams using @command{ffmpeg}:
24591 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
24595 Add flickering blur effect:
24597 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
24601 @section metadata, ametadata
24603 Manipulate frame metadata.
24605 This filter accepts the following options:
24609 Set mode of operation of the filter.
24611 Can be one of the following:
24615 If both @code{value} and @code{key} is set, select frames
24616 which have such metadata. If only @code{key} is set, select
24617 every frame that has such key in metadata.
24620 Add new metadata @code{key} and @code{value}. If key is already available
24624 Modify value of already present key.
24627 If @code{value} is set, delete only keys that have such value.
24628 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
24632 Print key and its value if metadata was found. If @code{key} is not set print all
24633 metadata values available in frame.
24637 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
24640 Set metadata value which will be used. This option is mandatory for
24641 @code{modify} and @code{add} mode.
24644 Which function to use when comparing metadata value and @code{value}.
24646 Can be one of following:
24650 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
24653 Values are interpreted as strings, returns true if metadata value starts with
24654 the @code{value} option string.
24657 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
24660 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
24663 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
24666 Values are interpreted as floats, returns true if expression from option @code{expr}
24670 Values are interpreted as strings, returns true if metadata value ends with
24671 the @code{value} option string.
24675 Set expression which is used when @code{function} is set to @code{expr}.
24676 The expression is evaluated through the eval API and can contain the following
24681 Float representation of @code{value} from metadata key.
24684 Float representation of @code{value} as supplied by user in @code{value} option.
24688 If specified in @code{print} mode, output is written to the named file. Instead of
24689 plain filename any writable url can be specified. Filename ``-'' is a shorthand
24690 for standard output. If @code{file} option is not set, output is written to the log
24691 with AV_LOG_INFO loglevel.
24694 Reduces buffering in print mode when output is written to a URL set using @var{file}.
24698 @subsection Examples
24702 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
24705 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
24708 Print silencedetect output to file @file{metadata.txt}.
24710 silencedetect,ametadata=mode=print:file=metadata.txt
24713 Direct all metadata to a pipe with file descriptor 4.
24715 metadata=mode=print:file='pipe\:4'
24719 @section perms, aperms
24721 Set read/write permissions for the output frames.
24723 These filters are mainly aimed at developers to test direct path in the
24724 following filter in the filtergraph.
24726 The filters accept the following options:
24730 Select the permissions mode.
24732 It accepts the following values:
24735 Do nothing. This is the default.
24737 Set all the output frames read-only.
24739 Set all the output frames directly writable.
24741 Make the frame read-only if writable, and writable if read-only.
24743 Set each output frame read-only or writable randomly.
24747 Set the seed for the @var{random} mode, must be an integer included between
24748 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
24749 @code{-1}, the filter will try to use a good random seed on a best effort
24753 Note: in case of auto-inserted filter between the permission filter and the
24754 following one, the permission might not be received as expected in that
24755 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
24756 perms/aperms filter can avoid this problem.
24758 @section realtime, arealtime
24760 Slow down filtering to match real time approximately.
24762 These filters will pause the filtering for a variable amount of time to
24763 match the output rate with the input timestamps.
24764 They are similar to the @option{re} option to @code{ffmpeg}.
24766 They accept the following options:
24770 Time limit for the pauses. Any pause longer than that will be considered
24771 a timestamp discontinuity and reset the timer. Default is 2 seconds.
24773 Speed factor for processing. The value must be a float larger than zero.
24774 Values larger than 1.0 will result in faster than realtime processing,
24775 smaller will slow processing down. The @var{limit} is automatically adapted
24776 accordingly. Default is 1.0.
24778 A processing speed faster than what is possible without these filters cannot
24783 @section select, aselect
24785 Select frames to pass in output.
24787 This filter accepts the following options:
24792 Set expression, which is evaluated for each input frame.
24794 If the expression is evaluated to zero, the frame is discarded.
24796 If the evaluation result is negative or NaN, the frame is sent to the
24797 first output; otherwise it is sent to the output with index
24798 @code{ceil(val)-1}, assuming that the input index starts from 0.
24800 For example a value of @code{1.2} corresponds to the output with index
24801 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
24804 Set the number of outputs. The output to which to send the selected
24805 frame is based on the result of the evaluation. Default value is 1.
24808 The expression can contain the following constants:
24812 The (sequential) number of the filtered frame, starting from 0.
24815 The (sequential) number of the selected frame, starting from 0.
24817 @item prev_selected_n
24818 The sequential number of the last selected frame. It's NAN if undefined.
24821 The timebase of the input timestamps.
24824 The PTS (Presentation TimeStamp) of the filtered video frame,
24825 expressed in @var{TB} units. It's NAN if undefined.
24828 The PTS of the filtered video frame,
24829 expressed in seconds. It's NAN if undefined.
24832 The PTS of the previously filtered video frame. It's NAN if undefined.
24834 @item prev_selected_pts
24835 The PTS of the last previously filtered video frame. It's NAN if undefined.
24837 @item prev_selected_t
24838 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
24841 The PTS of the first video frame in the video. It's NAN if undefined.
24844 The time of the first video frame in the video. It's NAN if undefined.
24846 @item pict_type @emph{(video only)}
24847 The type of the filtered frame. It can assume one of the following
24859 @item interlace_type @emph{(video only)}
24860 The frame interlace type. It can assume one of the following values:
24863 The frame is progressive (not interlaced).
24865 The frame is top-field-first.
24867 The frame is bottom-field-first.
24870 @item consumed_sample_n @emph{(audio only)}
24871 the number of selected samples before the current frame
24873 @item samples_n @emph{(audio only)}
24874 the number of samples in the current frame
24876 @item sample_rate @emph{(audio only)}
24877 the input sample rate
24880 This is 1 if the filtered frame is a key-frame, 0 otherwise.
24883 the position in the file of the filtered frame, -1 if the information
24884 is not available (e.g. for synthetic video)
24886 @item scene @emph{(video only)}
24887 value between 0 and 1 to indicate a new scene; a low value reflects a low
24888 probability for the current frame to introduce a new scene, while a higher
24889 value means the current frame is more likely to be one (see the example below)
24891 @item concatdec_select
24892 The concat demuxer can select only part of a concat input file by setting an
24893 inpoint and an outpoint, but the output packets may not be entirely contained
24894 in the selected interval. By using this variable, it is possible to skip frames
24895 generated by the concat demuxer which are not exactly contained in the selected
24898 This works by comparing the frame pts against the @var{lavf.concat.start_time}
24899 and the @var{lavf.concat.duration} packet metadata values which are also
24900 present in the decoded frames.
24902 The @var{concatdec_select} variable is -1 if the frame pts is at least
24903 start_time and either the duration metadata is missing or the frame pts is less
24904 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
24907 That basically means that an input frame is selected if its pts is within the
24908 interval set by the concat demuxer.
24912 The default value of the select expression is "1".
24914 @subsection Examples
24918 Select all frames in input:
24923 The example above is the same as:
24935 Select only I-frames:
24937 select='eq(pict_type\,I)'
24941 Select one frame every 100:
24943 select='not(mod(n\,100))'
24947 Select only frames contained in the 10-20 time interval:
24949 select=between(t\,10\,20)
24953 Select only I-frames contained in the 10-20 time interval:
24955 select=between(t\,10\,20)*eq(pict_type\,I)
24959 Select frames with a minimum distance of 10 seconds:
24961 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
24965 Use aselect to select only audio frames with samples number > 100:
24967 aselect='gt(samples_n\,100)'
24971 Create a mosaic of the first scenes:
24973 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
24976 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
24980 Send even and odd frames to separate outputs, and compose them:
24982 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
24986 Select useful frames from an ffconcat file which is using inpoints and
24987 outpoints but where the source files are not intra frame only.
24989 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
24993 @section sendcmd, asendcmd
24995 Send commands to filters in the filtergraph.
24997 These filters read commands to be sent to other filters in the
25000 @code{sendcmd} must be inserted between two video filters,
25001 @code{asendcmd} must be inserted between two audio filters, but apart
25002 from that they act the same way.
25004 The specification of commands can be provided in the filter arguments
25005 with the @var{commands} option, or in a file specified by the
25006 @var{filename} option.
25008 These filters accept the following options:
25011 Set the commands to be read and sent to the other filters.
25013 Set the filename of the commands to be read and sent to the other
25017 @subsection Commands syntax
25019 A commands description consists of a sequence of interval
25020 specifications, comprising a list of commands to be executed when a
25021 particular event related to that interval occurs. The occurring event
25022 is typically the current frame time entering or leaving a given time
25025 An interval is specified by the following syntax:
25027 @var{START}[-@var{END}] @var{COMMANDS};
25030 The time interval is specified by the @var{START} and @var{END} times.
25031 @var{END} is optional and defaults to the maximum time.
25033 The current frame time is considered within the specified interval if
25034 it is included in the interval [@var{START}, @var{END}), that is when
25035 the time is greater or equal to @var{START} and is lesser than
25038 @var{COMMANDS} consists of a sequence of one or more command
25039 specifications, separated by ",", relating to that interval. The
25040 syntax of a command specification is given by:
25042 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
25045 @var{FLAGS} is optional and specifies the type of events relating to
25046 the time interval which enable sending the specified command, and must
25047 be a non-null sequence of identifier flags separated by "+" or "|" and
25048 enclosed between "[" and "]".
25050 The following flags are recognized:
25053 The command is sent when the current frame timestamp enters the
25054 specified interval. In other words, the command is sent when the
25055 previous frame timestamp was not in the given interval, and the
25059 The command is sent when the current frame timestamp leaves the
25060 specified interval. In other words, the command is sent when the
25061 previous frame timestamp was in the given interval, and the
25065 The command @var{ARG} is interpreted as expression and result of
25066 expression is passed as @var{ARG}.
25068 The expression is evaluated through the eval API and can contain the following
25073 Original position in the file of the frame, or undefined if undefined
25074 for the current frame.
25077 The presentation timestamp in input.
25080 The count of the input frame for video or audio, starting from 0.
25083 The time in seconds of the current frame.
25086 The start time in seconds of the current command interval.
25089 The end time in seconds of the current command interval.
25092 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
25097 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
25100 @var{TARGET} specifies the target of the command, usually the name of
25101 the filter class or a specific filter instance name.
25103 @var{COMMAND} specifies the name of the command for the target filter.
25105 @var{ARG} is optional and specifies the optional list of argument for
25106 the given @var{COMMAND}.
25108 Between one interval specification and another, whitespaces, or
25109 sequences of characters starting with @code{#} until the end of line,
25110 are ignored and can be used to annotate comments.
25112 A simplified BNF description of the commands specification syntax
25115 @var{COMMAND_FLAG} ::= "enter" | "leave"
25116 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
25117 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
25118 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
25119 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
25120 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
25123 @subsection Examples
25127 Specify audio tempo change at second 4:
25129 asendcmd=c='4.0 atempo tempo 1.5',atempo
25133 Target a specific filter instance:
25135 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
25139 Specify a list of drawtext and hue commands in a file.
25141 # show text in the interval 5-10
25142 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
25143 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
25145 # desaturate the image in the interval 15-20
25146 15.0-20.0 [enter] hue s 0,
25147 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
25149 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
25151 # apply an exponential saturation fade-out effect, starting from time 25
25152 25 [enter] hue s exp(25-t)
25155 A filtergraph allowing to read and process the above command list
25156 stored in a file @file{test.cmd}, can be specified with:
25158 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
25163 @section setpts, asetpts
25165 Change the PTS (presentation timestamp) of the input frames.
25167 @code{setpts} works on video frames, @code{asetpts} on audio frames.
25169 This filter accepts the following options:
25174 The expression which is evaluated for each frame to construct its timestamp.
25178 The expression is evaluated through the eval API and can contain the following
25182 @item FRAME_RATE, FR
25183 frame rate, only defined for constant frame-rate video
25186 The presentation timestamp in input
25189 The count of the input frame for video or the number of consumed samples,
25190 not including the current frame for audio, starting from 0.
25192 @item NB_CONSUMED_SAMPLES
25193 The number of consumed samples, not including the current frame (only
25196 @item NB_SAMPLES, S
25197 The number of samples in the current frame (only audio)
25199 @item SAMPLE_RATE, SR
25200 The audio sample rate.
25203 The PTS of the first frame.
25206 the time in seconds of the first frame
25209 State whether the current frame is interlaced.
25212 the time in seconds of the current frame
25215 original position in the file of the frame, or undefined if undefined
25216 for the current frame
25219 The previous input PTS.
25222 previous input time in seconds
25225 The previous output PTS.
25228 previous output time in seconds
25231 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
25235 The wallclock (RTC) time at the start of the movie in microseconds.
25238 The timebase of the input timestamps.
25242 @subsection Examples
25246 Start counting PTS from zero
25248 setpts=PTS-STARTPTS
25252 Apply fast motion effect:
25258 Apply slow motion effect:
25264 Set fixed rate of 25 frames per second:
25270 Set fixed rate 25 fps with some jitter:
25272 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
25276 Apply an offset of 10 seconds to the input PTS:
25282 Generate timestamps from a "live source" and rebase onto the current timebase:
25284 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
25288 Generate timestamps by counting samples:
25297 Force color range for the output video frame.
25299 The @code{setrange} filter marks the color range property for the
25300 output frames. It does not change the input frame, but only sets the
25301 corresponding property, which affects how the frame is treated by
25304 The filter accepts the following options:
25309 Available values are:
25313 Keep the same color range property.
25315 @item unspecified, unknown
25316 Set the color range as unspecified.
25318 @item limited, tv, mpeg
25319 Set the color range as limited.
25321 @item full, pc, jpeg
25322 Set the color range as full.
25326 @section settb, asettb
25328 Set the timebase to use for the output frames timestamps.
25329 It is mainly useful for testing timebase configuration.
25331 It accepts the following parameters:
25336 The expression which is evaluated into the output timebase.
25340 The value for @option{tb} is an arithmetic expression representing a
25341 rational. The expression can contain the constants "AVTB" (the default
25342 timebase), "intb" (the input timebase) and "sr" (the sample rate,
25343 audio only). Default value is "intb".
25345 @subsection Examples
25349 Set the timebase to 1/25:
25355 Set the timebase to 1/10:
25361 Set the timebase to 1001/1000:
25367 Set the timebase to 2*intb:
25373 Set the default timebase value:
25380 Convert input audio to a video output representing frequency spectrum
25381 logarithmically using Brown-Puckette constant Q transform algorithm with
25382 direct frequency domain coefficient calculation (but the transform itself
25383 is not really constant Q, instead the Q factor is actually variable/clamped),
25384 with musical tone scale, from E0 to D#10.
25386 The filter accepts the following options:
25390 Specify the video size for the output. It must be even. For the syntax of this option,
25391 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25392 Default value is @code{1920x1080}.
25395 Set the output frame rate. Default value is @code{25}.
25398 Set the bargraph height. It must be even. Default value is @code{-1} which
25399 computes the bargraph height automatically.
25402 Set the axis height. It must be even. Default value is @code{-1} which computes
25403 the axis height automatically.
25406 Set the sonogram height. It must be even. Default value is @code{-1} which
25407 computes the sonogram height automatically.
25410 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
25411 instead. Default value is @code{1}.
25413 @item sono_v, volume
25414 Specify the sonogram volume expression. It can contain variables:
25417 the @var{bar_v} evaluated expression
25418 @item frequency, freq, f
25419 the frequency where it is evaluated
25420 @item timeclamp, tc
25421 the value of @var{timeclamp} option
25425 @item a_weighting(f)
25426 A-weighting of equal loudness
25427 @item b_weighting(f)
25428 B-weighting of equal loudness
25429 @item c_weighting(f)
25430 C-weighting of equal loudness.
25432 Default value is @code{16}.
25434 @item bar_v, volume2
25435 Specify the bargraph volume expression. It can contain variables:
25438 the @var{sono_v} evaluated expression
25439 @item frequency, freq, f
25440 the frequency where it is evaluated
25441 @item timeclamp, tc
25442 the value of @var{timeclamp} option
25446 @item a_weighting(f)
25447 A-weighting of equal loudness
25448 @item b_weighting(f)
25449 B-weighting of equal loudness
25450 @item c_weighting(f)
25451 C-weighting of equal loudness.
25453 Default value is @code{sono_v}.
25455 @item sono_g, gamma
25456 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
25457 higher gamma makes the spectrum having more range. Default value is @code{3}.
25458 Acceptable range is @code{[1, 7]}.
25460 @item bar_g, gamma2
25461 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
25465 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
25466 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
25468 @item timeclamp, tc
25469 Specify the transform timeclamp. At low frequency, there is trade-off between
25470 accuracy in time domain and frequency domain. If timeclamp is lower,
25471 event in time domain is represented more accurately (such as fast bass drum),
25472 otherwise event in frequency domain is represented more accurately
25473 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
25476 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
25477 limits future samples by applying asymmetric windowing in time domain, useful
25478 when low latency is required. Accepted range is @code{[0, 1]}.
25481 Specify the transform base frequency. Default value is @code{20.01523126408007475},
25482 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
25485 Specify the transform end frequency. Default value is @code{20495.59681441799654},
25486 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
25489 This option is deprecated and ignored.
25492 Specify the transform length in time domain. Use this option to control accuracy
25493 trade-off between time domain and frequency domain at every frequency sample.
25494 It can contain variables:
25496 @item frequency, freq, f
25497 the frequency where it is evaluated
25498 @item timeclamp, tc
25499 the value of @var{timeclamp} option.
25501 Default value is @code{384*tc/(384+tc*f)}.
25504 Specify the transform count for every video frame. Default value is @code{6}.
25505 Acceptable range is @code{[1, 30]}.
25508 Specify the transform count for every single pixel. Default value is @code{0},
25509 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
25512 Specify font file for use with freetype to draw the axis. If not specified,
25513 use embedded font. Note that drawing with font file or embedded font is not
25514 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
25518 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
25519 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
25523 Specify font color expression. This is arithmetic expression that should return
25524 integer value 0xRRGGBB. It can contain variables:
25526 @item frequency, freq, f
25527 the frequency where it is evaluated
25528 @item timeclamp, tc
25529 the value of @var{timeclamp} option
25534 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
25535 @item r(x), g(x), b(x)
25536 red, green, and blue value of intensity x.
25538 Default value is @code{st(0, (midi(f)-59.5)/12);
25539 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
25540 r(1-ld(1)) + b(ld(1))}.
25543 Specify image file to draw the axis. This option override @var{fontfile} and
25544 @var{fontcolor} option.
25547 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
25548 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
25549 Default value is @code{1}.
25552 Set colorspace. The accepted values are:
25555 Unspecified (default)
25564 BT.470BG or BT.601-6 625
25567 SMPTE-170M or BT.601-6 525
25573 BT.2020 with non-constant luminance
25578 Set spectrogram color scheme. This is list of floating point values with format
25579 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
25580 The default is @code{1|0.5|0|0|0.5|1}.
25584 @subsection Examples
25588 Playing audio while showing the spectrum:
25590 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
25594 Same as above, but with frame rate 30 fps:
25596 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
25600 Playing at 1280x720:
25602 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
25606 Disable sonogram display:
25612 A1 and its harmonics: A1, A2, (near)E3, A3:
25614 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),
25615 asplit[a][out1]; [a] showcqt [out0]'
25619 Same as above, but with more accuracy in frequency domain:
25621 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),
25622 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
25628 bar_v=10:sono_v=bar_v*a_weighting(f)
25632 Custom gamma, now spectrum is linear to the amplitude.
25638 Custom tlength equation:
25640 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)))'
25644 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
25646 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
25650 Custom font using fontconfig:
25652 font='Courier New,Monospace,mono|bold'
25656 Custom frequency range with custom axis using image file:
25658 axisfile=myaxis.png:basefreq=40:endfreq=10000
25664 Convert input audio to video output representing the audio power spectrum.
25665 Audio amplitude is on Y-axis while frequency is on X-axis.
25667 The filter accepts the following options:
25671 Specify size of video. For the syntax of this option, check the
25672 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25673 Default is @code{1024x512}.
25677 This set how each frequency bin will be represented.
25679 It accepts the following values:
25685 Default is @code{bar}.
25688 Set amplitude scale.
25690 It accepts the following values:
25704 Default is @code{log}.
25707 Set frequency scale.
25709 It accepts the following values:
25718 Reverse logarithmic scale.
25720 Default is @code{lin}.
25723 Set window size. Allowed range is from 16 to 65536.
25725 Default is @code{2048}
25728 Set windowing function.
25730 It accepts the following values:
25753 Default is @code{hanning}.
25756 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25757 which means optimal overlap for selected window function will be picked.
25760 Set time averaging. Setting this to 0 will display current maximal peaks.
25761 Default is @code{1}, which means time averaging is disabled.
25764 Specify list of colors separated by space or by '|' which will be used to
25765 draw channel frequencies. Unrecognized or missing colors will be replaced
25769 Set channel display mode.
25771 It accepts the following values:
25776 Default is @code{combined}.
25779 Set minimum amplitude used in @code{log} amplitude scaler.
25782 Set data display mode.
25784 It accepts the following values:
25790 Default is @code{magnitude}.
25793 @section showspatial
25795 Convert stereo input audio to a video output, representing the spatial relationship
25796 between two channels.
25798 The filter accepts the following options:
25802 Specify the video size for the output. For the syntax of this option, check the
25803 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25804 Default value is @code{512x512}.
25807 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
25810 Set window function.
25812 It accepts the following values:
25837 Default value is @code{hann}.
25840 Set ratio of overlap window. Default value is @code{0.5}.
25841 When value is @code{1} overlap is set to recommended size for specific
25842 window function currently used.
25845 @anchor{showspectrum}
25846 @section showspectrum
25848 Convert input audio to a video output, representing the audio frequency
25851 The filter accepts the following options:
25855 Specify the video size for the output. For the syntax of this option, check the
25856 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25857 Default value is @code{640x512}.
25860 Specify how the spectrum should slide along the window.
25862 It accepts the following values:
25865 the samples start again on the left when they reach the right
25867 the samples scroll from right to left
25869 frames are only produced when the samples reach the right
25871 the samples scroll from left to right
25874 Default value is @code{replace}.
25877 Specify display mode.
25879 It accepts the following values:
25882 all channels are displayed in the same row
25884 all channels are displayed in separate rows
25887 Default value is @samp{combined}.
25890 Specify display color mode.
25892 It accepts the following values:
25895 each channel is displayed in a separate color
25897 each channel is displayed using the same color scheme
25899 each channel is displayed using the rainbow color scheme
25901 each channel is displayed using the moreland color scheme
25903 each channel is displayed using the nebulae color scheme
25905 each channel is displayed using the fire color scheme
25907 each channel is displayed using the fiery color scheme
25909 each channel is displayed using the fruit color scheme
25911 each channel is displayed using the cool color scheme
25913 each channel is displayed using the magma color scheme
25915 each channel is displayed using the green color scheme
25917 each channel is displayed using the viridis color scheme
25919 each channel is displayed using the plasma color scheme
25921 each channel is displayed using the cividis color scheme
25923 each channel is displayed using the terrain color scheme
25926 Default value is @samp{channel}.
25929 Specify scale used for calculating intensity color values.
25931 It accepts the following values:
25936 square root, default
25947 Default value is @samp{sqrt}.
25950 Specify frequency scale.
25952 It accepts the following values:
25960 Default value is @samp{lin}.
25963 Set saturation modifier for displayed colors. Negative values provide
25964 alternative color scheme. @code{0} is no saturation at all.
25965 Saturation must be in [-10.0, 10.0] range.
25966 Default value is @code{1}.
25969 Set window function.
25971 It accepts the following values:
25996 Default value is @code{hann}.
25999 Set orientation of time vs frequency axis. Can be @code{vertical} or
26000 @code{horizontal}. Default is @code{vertical}.
26003 Set ratio of overlap window. Default value is @code{0}.
26004 When value is @code{1} overlap is set to recommended size for specific
26005 window function currently used.
26008 Set scale gain for calculating intensity color values.
26009 Default value is @code{1}.
26012 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
26015 Set color rotation, must be in [-1.0, 1.0] range.
26016 Default value is @code{0}.
26019 Set start frequency from which to display spectrogram. Default is @code{0}.
26022 Set stop frequency to which to display spectrogram. Default is @code{0}.
26025 Set upper frame rate limit. Default is @code{auto}, unlimited.
26028 Draw time and frequency axes and legends. Default is disabled.
26031 The usage is very similar to the showwaves filter; see the examples in that
26034 @subsection Examples
26038 Large window with logarithmic color scaling:
26040 showspectrum=s=1280x480:scale=log
26044 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
26046 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
26047 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
26051 @section showspectrumpic
26053 Convert input audio to a single video frame, representing the audio frequency
26056 The filter accepts the following options:
26060 Specify the video size for the output. For the syntax of this option, check the
26061 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26062 Default value is @code{4096x2048}.
26065 Specify display mode.
26067 It accepts the following values:
26070 all channels are displayed in the same row
26072 all channels are displayed in separate rows
26074 Default value is @samp{combined}.
26077 Specify display color mode.
26079 It accepts the following values:
26082 each channel is displayed in a separate color
26084 each channel is displayed using the same color scheme
26086 each channel is displayed using the rainbow color scheme
26088 each channel is displayed using the moreland color scheme
26090 each channel is displayed using the nebulae color scheme
26092 each channel is displayed using the fire color scheme
26094 each channel is displayed using the fiery color scheme
26096 each channel is displayed using the fruit color scheme
26098 each channel is displayed using the cool color scheme
26100 each channel is displayed using the magma color scheme
26102 each channel is displayed using the green color scheme
26104 each channel is displayed using the viridis color scheme
26106 each channel is displayed using the plasma color scheme
26108 each channel is displayed using the cividis color scheme
26110 each channel is displayed using the terrain color scheme
26112 Default value is @samp{intensity}.
26115 Specify scale used for calculating intensity color values.
26117 It accepts the following values:
26122 square root, default
26132 Default value is @samp{log}.
26135 Specify frequency scale.
26137 It accepts the following values:
26145 Default value is @samp{lin}.
26148 Set saturation modifier for displayed colors. Negative values provide
26149 alternative color scheme. @code{0} is no saturation at all.
26150 Saturation must be in [-10.0, 10.0] range.
26151 Default value is @code{1}.
26154 Set window function.
26156 It accepts the following values:
26180 Default value is @code{hann}.
26183 Set orientation of time vs frequency axis. Can be @code{vertical} or
26184 @code{horizontal}. Default is @code{vertical}.
26187 Set scale gain for calculating intensity color values.
26188 Default value is @code{1}.
26191 Draw time and frequency axes and legends. Default is enabled.
26194 Set color rotation, must be in [-1.0, 1.0] range.
26195 Default value is @code{0}.
26198 Set start frequency from which to display spectrogram. Default is @code{0}.
26201 Set stop frequency to which to display spectrogram. Default is @code{0}.
26204 @subsection Examples
26208 Extract an audio spectrogram of a whole audio track
26209 in a 1024x1024 picture using @command{ffmpeg}:
26211 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
26215 @section showvolume
26217 Convert input audio volume to a video output.
26219 The filter accepts the following options:
26226 Set border width, allowed range is [0, 5]. Default is 1.
26229 Set channel width, allowed range is [80, 8192]. Default is 400.
26232 Set channel height, allowed range is [1, 900]. Default is 20.
26235 Set fade, allowed range is [0, 1]. Default is 0.95.
26238 Set volume color expression.
26240 The expression can use the following variables:
26244 Current max volume of channel in dB.
26250 Current channel number, starting from 0.
26254 If set, displays channel names. Default is enabled.
26257 If set, displays volume values. Default is enabled.
26260 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
26261 default is @code{h}.
26264 Set step size, allowed range is [0, 5]. Default is 0, which means
26268 Set background opacity, allowed range is [0, 1]. Default is 0.
26271 Set metering mode, can be peak: @code{p} or rms: @code{r},
26272 default is @code{p}.
26275 Set display scale, can be linear: @code{lin} or log: @code{log},
26276 default is @code{lin}.
26280 If set to > 0., display a line for the max level
26281 in the previous seconds.
26282 default is disabled: @code{0.}
26285 The color of the max line. Use when @code{dm} option is set to > 0.
26286 default is: @code{orange}
26291 Convert input audio to a video output, representing the samples waves.
26293 The filter accepts the following options:
26297 Specify the video size for the output. For the syntax of this option, check the
26298 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26299 Default value is @code{600x240}.
26304 Available values are:
26307 Draw a point for each sample.
26310 Draw a vertical line for each sample.
26313 Draw a point for each sample and a line between them.
26316 Draw a centered vertical line for each sample.
26319 Default value is @code{point}.
26322 Set the number of samples which are printed on the same column. A
26323 larger value will decrease the frame rate. Must be a positive
26324 integer. This option can be set only if the value for @var{rate}
26325 is not explicitly specified.
26328 Set the (approximate) output frame rate. This is done by setting the
26329 option @var{n}. Default value is "25".
26331 @item split_channels
26332 Set if channels should be drawn separately or overlap. Default value is 0.
26335 Set colors separated by '|' which are going to be used for drawing of each channel.
26338 Set amplitude scale.
26340 Available values are:
26358 Set the draw mode. This is mostly useful to set for high @var{n}.
26360 Available values are:
26363 Scale pixel values for each drawn sample.
26366 Draw every sample directly.
26369 Default value is @code{scale}.
26372 @subsection Examples
26376 Output the input file audio and the corresponding video representation
26379 amovie=a.mp3,asplit[out0],showwaves[out1]
26383 Create a synthetic signal and show it with showwaves, forcing a
26384 frame rate of 30 frames per second:
26386 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
26390 @section showwavespic
26392 Convert input audio to a single video frame, representing the samples waves.
26394 The filter accepts the following options:
26398 Specify the video size for the output. For the syntax of this option, check the
26399 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26400 Default value is @code{600x240}.
26402 @item split_channels
26403 Set if channels should be drawn separately or overlap. Default value is 0.
26406 Set colors separated by '|' which are going to be used for drawing of each channel.
26409 Set amplitude scale.
26411 Available values are:
26431 Available values are:
26434 Scale pixel values for each drawn sample.
26437 Draw every sample directly.
26440 Default value is @code{scale}.
26443 Set the filter mode.
26445 Available values are:
26448 Use average samples values for each drawn sample.
26451 Use peak samples values for each drawn sample.
26454 Default value is @code{average}.
26457 @subsection Examples
26461 Extract a channel split representation of the wave form of a whole audio track
26462 in a 1024x800 picture using @command{ffmpeg}:
26464 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
26468 @section sidedata, asidedata
26470 Delete frame side data, or select frames based on it.
26472 This filter accepts the following options:
26476 Set mode of operation of the filter.
26478 Can be one of the following:
26482 Select every frame with side data of @code{type}.
26485 Delete side data of @code{type}. If @code{type} is not set, delete all side
26491 Set side data type used with all modes. Must be set for @code{select} mode. For
26492 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
26493 in @file{libavutil/frame.h}. For example, to choose
26494 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
26498 @section spectrumsynth
26500 Synthesize audio from 2 input video spectrums, first input stream represents
26501 magnitude across time and second represents phase across time.
26502 The filter will transform from frequency domain as displayed in videos back
26503 to time domain as presented in audio output.
26505 This filter is primarily created for reversing processed @ref{showspectrum}
26506 filter outputs, but can synthesize sound from other spectrograms too.
26507 But in such case results are going to be poor if the phase data is not
26508 available, because in such cases phase data need to be recreated, usually
26509 it's just recreated from random noise.
26510 For best results use gray only output (@code{channel} color mode in
26511 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
26512 @code{lin} scale for phase video. To produce phase, for 2nd video, use
26513 @code{data} option. Inputs videos should generally use @code{fullframe}
26514 slide mode as that saves resources needed for decoding video.
26516 The filter accepts the following options:
26520 Specify sample rate of output audio, the sample rate of audio from which
26521 spectrum was generated may differ.
26524 Set number of channels represented in input video spectrums.
26527 Set scale which was used when generating magnitude input spectrum.
26528 Can be @code{lin} or @code{log}. Default is @code{log}.
26531 Set slide which was used when generating inputs spectrums.
26532 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
26533 Default is @code{fullframe}.
26536 Set window function used for resynthesis.
26539 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
26540 which means optimal overlap for selected window function will be picked.
26543 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
26544 Default is @code{vertical}.
26547 @subsection Examples
26551 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
26552 then resynthesize videos back to audio with spectrumsynth:
26554 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
26555 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
26556 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
26560 @section split, asplit
26562 Split input into several identical outputs.
26564 @code{asplit} works with audio input, @code{split} with video.
26566 The filter accepts a single parameter which specifies the number of outputs. If
26567 unspecified, it defaults to 2.
26569 @subsection Examples
26573 Create two separate outputs from the same input:
26575 [in] split [out0][out1]
26579 To create 3 or more outputs, you need to specify the number of
26582 [in] asplit=3 [out0][out1][out2]
26586 Create two separate outputs from the same input, one cropped and
26589 [in] split [splitout1][splitout2];
26590 [splitout1] crop=100:100:0:0 [cropout];
26591 [splitout2] pad=200:200:100:100 [padout];
26595 Create 5 copies of the input audio with @command{ffmpeg}:
26597 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
26603 Receive commands sent through a libzmq client, and forward them to
26604 filters in the filtergraph.
26606 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
26607 must be inserted between two video filters, @code{azmq} between two
26608 audio filters. Both are capable to send messages to any filter type.
26610 To enable these filters you need to install the libzmq library and
26611 headers and configure FFmpeg with @code{--enable-libzmq}.
26613 For more information about libzmq see:
26614 @url{http://www.zeromq.org/}
26616 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
26617 receives messages sent through a network interface defined by the
26618 @option{bind_address} (or the abbreviation "@option{b}") option.
26619 Default value of this option is @file{tcp://localhost:5555}. You may
26620 want to alter this value to your needs, but do not forget to escape any
26621 ':' signs (see @ref{filtergraph escaping}).
26623 The received message must be in the form:
26625 @var{TARGET} @var{COMMAND} [@var{ARG}]
26628 @var{TARGET} specifies the target of the command, usually the name of
26629 the filter class or a specific filter instance name. The default
26630 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
26631 but you can override this by using the @samp{filter_name@@id} syntax
26632 (see @ref{Filtergraph syntax}).
26634 @var{COMMAND} specifies the name of the command for the target filter.
26636 @var{ARG} is optional and specifies the optional argument list for the
26637 given @var{COMMAND}.
26639 Upon reception, the message is processed and the corresponding command
26640 is injected into the filtergraph. Depending on the result, the filter
26641 will send a reply to the client, adopting the format:
26643 @var{ERROR_CODE} @var{ERROR_REASON}
26647 @var{MESSAGE} is optional.
26649 @subsection Examples
26651 Look at @file{tools/zmqsend} for an example of a zmq client which can
26652 be used to send commands processed by these filters.
26654 Consider the following filtergraph generated by @command{ffplay}.
26655 In this example the last overlay filter has an instance name. All other
26656 filters will have default instance names.
26659 ffplay -dumpgraph 1 -f lavfi "
26660 color=s=100x100:c=red [l];
26661 color=s=100x100:c=blue [r];
26662 nullsrc=s=200x100, zmq [bg];
26663 [bg][l] overlay [bg+l];
26664 [bg+l][r] overlay@@my=x=100 "
26667 To change the color of the left side of the video, the following
26668 command can be used:
26670 echo Parsed_color_0 c yellow | tools/zmqsend
26673 To change the right side:
26675 echo Parsed_color_1 c pink | tools/zmqsend
26678 To change the position of the right side:
26680 echo overlay@@my x 150 | tools/zmqsend
26684 @c man end MULTIMEDIA FILTERS
26686 @chapter Multimedia Sources
26687 @c man begin MULTIMEDIA SOURCES
26689 Below is a description of the currently available multimedia sources.
26693 This is the same as @ref{movie} source, except it selects an audio
26699 Read audio and/or video stream(s) from a movie container.
26701 It accepts the following parameters:
26705 The name of the resource to read (not necessarily a file; it can also be a
26706 device or a stream accessed through some protocol).
26708 @item format_name, f
26709 Specifies the format assumed for the movie to read, and can be either
26710 the name of a container or an input device. If not specified, the
26711 format is guessed from @var{movie_name} or by probing.
26713 @item seek_point, sp
26714 Specifies the seek point in seconds. The frames will be output
26715 starting from this seek point. The parameter is evaluated with
26716 @code{av_strtod}, so the numerical value may be suffixed by an IS
26717 postfix. The default value is "0".
26720 Specifies the streams to read. Several streams can be specified,
26721 separated by "+". The source will then have as many outputs, in the
26722 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
26723 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
26724 respectively the default (best suited) video and audio stream. Default
26725 is "dv", or "da" if the filter is called as "amovie".
26727 @item stream_index, si
26728 Specifies the index of the video stream to read. If the value is -1,
26729 the most suitable video stream will be automatically selected. The default
26730 value is "-1". Deprecated. If the filter is called "amovie", it will select
26731 audio instead of video.
26734 Specifies how many times to read the stream in sequence.
26735 If the value is 0, the stream will be looped infinitely.
26736 Default value is "1".
26738 Note that when the movie is looped the source timestamps are not
26739 changed, so it will generate non monotonically increasing timestamps.
26741 @item discontinuity
26742 Specifies the time difference between frames above which the point is
26743 considered a timestamp discontinuity which is removed by adjusting the later
26747 It allows overlaying a second video on top of the main input of
26748 a filtergraph, as shown in this graph:
26750 input -----------> deltapts0 --> overlay --> output
26753 movie --> scale--> deltapts1 -------+
26755 @subsection Examples
26759 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
26760 on top of the input labelled "in":
26762 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
26763 [in] setpts=PTS-STARTPTS [main];
26764 [main][over] overlay=16:16 [out]
26768 Read from a video4linux2 device, and overlay it on top of the input
26771 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
26772 [in] setpts=PTS-STARTPTS [main];
26773 [main][over] overlay=16:16 [out]
26777 Read the first video stream and the audio stream with id 0x81 from
26778 dvd.vob; the video is connected to the pad named "video" and the audio is
26779 connected to the pad named "audio":
26781 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
26785 @subsection Commands
26787 Both movie and amovie support the following commands:
26790 Perform seek using "av_seek_frame".
26791 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
26794 @var{stream_index}: If stream_index is -1, a default
26795 stream is selected, and @var{timestamp} is automatically converted
26796 from AV_TIME_BASE units to the stream specific time_base.
26798 @var{timestamp}: Timestamp in AVStream.time_base units
26799 or, if no stream is specified, in AV_TIME_BASE units.
26801 @var{flags}: Flags which select direction and seeking mode.
26805 Get movie duration in AV_TIME_BASE units.
26809 @c man end MULTIMEDIA SOURCES