1 @chapter Filtering Introduction
2 @c man begin FILTERING INTRODUCTION
4 Filtering in FFmpeg is enabled through the libavfilter library.
6 In libavfilter, a filter can have multiple inputs and multiple
8 To illustrate the sorts of things that are possible, we consider the
13 input --> split ---------------------> overlay --> output
16 +-----> crop --> vflip -------+
19 This filtergraph splits the input stream in two streams, then sends one
20 stream through the crop filter and the vflip filter, before merging it
21 back with the other stream by overlaying it on top. You can use the
22 following command to achieve this:
25 ffmpeg -i INPUT -vf "split [main][tmp]; [tmp] crop=iw:ih/2:0:0, vflip [flip]; [main][flip] overlay=0:H/2" OUTPUT
28 The result will be that the top half of the video is mirrored
29 onto the bottom half of the output video.
31 Filters in the same linear chain are separated by commas, and distinct
32 linear chains of filters are separated by semicolons. In our example,
33 @var{crop,vflip} are in one linear chain, @var{split} and
34 @var{overlay} are separately in another. The points where the linear
35 chains join are labelled by names enclosed in square brackets. In the
36 example, the split filter generates two outputs that are associated to
37 the labels @var{[main]} and @var{[tmp]}.
39 The stream sent to the second output of @var{split}, labelled as
40 @var{[tmp]}, is processed through the @var{crop} filter, which crops
41 away the lower half part of the video, and then vertically flipped. The
42 @var{overlay} filter takes in input the first unchanged output of the
43 split filter (which was labelled as @var{[main]}), and overlay on its
44 lower half the output generated by the @var{crop,vflip} filterchain.
46 Some filters take in input a list of parameters: they are specified
47 after the filter name and an equal sign, and are separated from each other
50 There exist so-called @var{source filters} that do not have an
51 audio/video input, and @var{sink filters} that will not have audio/video
54 @c man end FILTERING INTRODUCTION
57 @c man begin GRAPH2DOT
59 The @file{graph2dot} program included in the FFmpeg @file{tools}
60 directory can be used to parse a filtergraph description and issue a
61 corresponding textual representation in the dot language.
68 to see how to use @file{graph2dot}.
70 You can then pass the dot description to the @file{dot} program (from
71 the graphviz suite of programs) and obtain a graphical representation
74 For example the sequence of commands:
76 echo @var{GRAPH_DESCRIPTION} | \
77 tools/graph2dot -o graph.tmp && \
78 dot -Tpng graph.tmp -o graph.png && \
82 can be used to create and display an image representing the graph
83 described by the @var{GRAPH_DESCRIPTION} string. Note that this string must be
84 a complete self-contained graph, with its inputs and outputs explicitly defined.
85 For example if your command line is of the form:
87 ffmpeg -i infile -vf scale=640:360 outfile
89 your @var{GRAPH_DESCRIPTION} string will need to be of the form:
91 nullsrc,scale=640:360,nullsink
93 you may also need to set the @var{nullsrc} parameters and add a @var{format}
94 filter in order to simulate a specific input file.
98 @chapter Filtergraph description
99 @c man begin FILTERGRAPH DESCRIPTION
101 A filtergraph is a directed graph of connected filters. It can contain
102 cycles, and there can be multiple links between a pair of
103 filters. Each link has one input pad on one side connecting it to one
104 filter from which it takes its input, and one output pad on the other
105 side connecting it to one filter accepting its output.
107 Each filter in a filtergraph is an instance of a filter class
108 registered in the application, which defines the features and the
109 number of input and output pads of the filter.
111 A filter with no input pads is called a "source", and a filter with no
112 output pads is called a "sink".
114 @anchor{Filtergraph syntax}
115 @section Filtergraph syntax
117 A filtergraph has a textual representation, which is recognized by the
118 @option{-filter}/@option{-vf}/@option{-af} and
119 @option{-filter_complex} options in @command{ffmpeg} and
120 @option{-vf}/@option{-af} in @command{ffplay}, and by the
121 @code{avfilter_graph_parse_ptr()} function defined in
122 @file{libavfilter/avfilter.h}.
124 A filterchain consists of a sequence of connected filters, each one
125 connected to the previous one in the sequence. A filterchain is
126 represented by a list of ","-separated filter descriptions.
128 A filtergraph consists of a sequence of filterchains. A sequence of
129 filterchains is represented by a list of ";"-separated filterchain
132 A filter is represented by a string of the form:
133 [@var{in_link_1}]...[@var{in_link_N}]@var{filter_name}@@@var{id}=@var{arguments}[@var{out_link_1}]...[@var{out_link_M}]
135 @var{filter_name} is the name of the filter class of which the
136 described filter is an instance of, and has to be the name of one of
137 the filter classes registered in the program optionally followed by "@@@var{id}".
138 The name of the filter class is optionally followed by a string
141 @var{arguments} is a string which contains the parameters used to
142 initialize the filter instance. It may have one of two forms:
146 A ':'-separated list of @var{key=value} pairs.
149 A ':'-separated list of @var{value}. In this case, the keys are assumed to be
150 the option names in the order they are declared. E.g. the @code{fade} filter
151 declares three options in this order -- @option{type}, @option{start_frame} and
152 @option{nb_frames}. Then the parameter list @var{in:0:30} means that the value
153 @var{in} is assigned to the option @option{type}, @var{0} to
154 @option{start_frame} and @var{30} to @option{nb_frames}.
157 A ':'-separated list of mixed direct @var{value} and long @var{key=value}
158 pairs. The direct @var{value} must precede the @var{key=value} pairs, and
159 follow the same constraints order of the previous point. The following
160 @var{key=value} pairs can be set in any preferred order.
164 If the option value itself is a list of items (e.g. the @code{format} filter
165 takes a list of pixel formats), the items in the list are usually separated by
168 The list of arguments can be quoted using the character @samp{'} as initial
169 and ending mark, and the character @samp{\} for escaping the characters
170 within the quoted text; otherwise the argument string is considered
171 terminated when the next special character (belonging to the set
172 @samp{[]=;,}) is encountered.
174 The name and arguments of the filter are optionally preceded and
175 followed by a list of link labels.
176 A link label allows one to name a link and associate it to a filter output
177 or input pad. The preceding labels @var{in_link_1}
178 ... @var{in_link_N}, are associated to the filter input pads,
179 the following labels @var{out_link_1} ... @var{out_link_M}, are
180 associated to the output pads.
182 When two link labels with the same name are found in the
183 filtergraph, a link between the corresponding input and output pad is
186 If an output pad is not labelled, it is linked by default to the first
187 unlabelled input pad of the next filter in the filterchain.
188 For example in the filterchain
190 nullsrc, split[L1], [L2]overlay, nullsink
192 the split filter instance has two output pads, and the overlay filter
193 instance two input pads. The first output pad of split is labelled
194 "L1", the first input pad of overlay is labelled "L2", and the second
195 output pad of split is linked to the second input pad of overlay,
196 which are both unlabelled.
198 In a filter description, if the input label of the first filter is not
199 specified, "in" is assumed; if the output label of the last filter is not
200 specified, "out" is assumed.
202 In a complete filterchain all the unlabelled filter input and output
203 pads must be connected. A filtergraph is considered valid if all the
204 filter input and output pads of all the filterchains are connected.
206 Libavfilter will automatically insert @ref{scale} filters where format
207 conversion is required. It is possible to specify swscale flags
208 for those automatically inserted scalers by prepending
209 @code{sws_flags=@var{flags};}
210 to the filtergraph description.
212 Here is a BNF description of the filtergraph syntax:
214 @var{NAME} ::= sequence of alphanumeric characters and '_'
215 @var{FILTER_NAME} ::= @var{NAME}["@@"@var{NAME}]
216 @var{LINKLABEL} ::= "[" @var{NAME} "]"
217 @var{LINKLABELS} ::= @var{LINKLABEL} [@var{LINKLABELS}]
218 @var{FILTER_ARGUMENTS} ::= sequence of chars (possibly quoted)
219 @var{FILTER} ::= [@var{LINKLABELS}] @var{FILTER_NAME} ["=" @var{FILTER_ARGUMENTS}] [@var{LINKLABELS}]
220 @var{FILTERCHAIN} ::= @var{FILTER} [,@var{FILTERCHAIN}]
221 @var{FILTERGRAPH} ::= [sws_flags=@var{flags};] @var{FILTERCHAIN} [;@var{FILTERGRAPH}]
224 @anchor{filtergraph escaping}
225 @section Notes on filtergraph escaping
227 Filtergraph description composition entails several levels of
228 escaping. See @ref{quoting_and_escaping,,the "Quoting and escaping"
229 section in the ffmpeg-utils(1) manual,ffmpeg-utils} for more
230 information about the employed escaping procedure.
232 A first level escaping affects the content of each filter option
233 value, which may contain the special character @code{:} used to
234 separate values, or one of the escaping characters @code{\'}.
236 A second level escaping affects the whole filter description, which
237 may contain the escaping characters @code{\'} or the special
238 characters @code{[],;} used by the filtergraph description.
240 Finally, when you specify a filtergraph on a shell commandline, you
241 need to perform a third level escaping for the shell special
242 characters contained within it.
244 For example, consider the following string to be embedded in
245 the @ref{drawtext} filter description @option{text} value:
247 this is a 'string': may contain one, or more, special characters
250 This string contains the @code{'} special escaping character, and the
251 @code{:} special character, so it needs to be escaped in this way:
253 text=this is a \'string\'\: may contain one, or more, special characters
256 A second level of escaping is required when embedding the filter
257 description in a filtergraph description, in order to escape all the
258 filtergraph special characters. Thus the example above becomes:
260 drawtext=text=this is a \\\'string\\\'\\: may contain one\, or more\, special characters
262 (note that in addition to the @code{\'} escaping special characters,
263 also @code{,} needs to be escaped).
265 Finally an additional level of escaping is needed when writing the
266 filtergraph description in a shell command, which depends on the
267 escaping rules of the adopted shell. For example, assuming that
268 @code{\} is special and needs to be escaped with another @code{\}, the
269 previous string will finally result in:
271 -vf "drawtext=text=this is a \\\\\\'string\\\\\\'\\\\: may contain one\\, or more\\, special characters"
274 @chapter Timeline editing
276 Some filters support a generic @option{enable} option. For the filters
277 supporting timeline editing, this option can be set to an expression which is
278 evaluated before sending a frame to the filter. If the evaluation is non-zero,
279 the filter will be enabled, otherwise the frame will be sent unchanged to the
280 next filter in the filtergraph.
282 The expression accepts the following values:
285 timestamp expressed in seconds, NAN if the input timestamp is unknown
288 sequential number of the input frame, starting from 0
291 the position in the file of the input frame, NAN if unknown
295 width and height of the input frame if video
298 Additionally, these filters support an @option{enable} command that can be used
299 to re-define the expression.
301 Like any other filtering option, the @option{enable} option follows the same
304 For example, to enable a blur filter (@ref{smartblur}) from 10 seconds to 3
305 minutes, and a @ref{curves} filter starting at 3 seconds:
307 smartblur = enable='between(t,10,3*60)',
308 curves = enable='gte(t,3)' : preset=cross_process
311 See @code{ffmpeg -filters} to view which filters have timeline support.
313 @c man end FILTERGRAPH DESCRIPTION
316 @chapter Changing options at runtime with a command
318 Some options can be changed during the operation of the filter using
319 a command. These options are marked 'T' on the output of
320 @command{ffmpeg} @option{-h filter=<name of filter>}.
321 The name of the command is the name of the option and the argument is
325 @chapter Options for filters with several inputs (framesync)
326 @c man begin OPTIONS FOR FILTERS WITH SEVERAL INPUTS
328 Some filters with several inputs support a common set of options.
329 These options can only be set by name, not with the short notation.
333 The action to take when EOF is encountered on the secondary input; it accepts
334 one of the following values:
338 Repeat the last frame (the default).
342 Pass the main input through.
346 If set to 1, force the output to terminate when the shortest input
347 terminates. Default value is 0.
350 If set to 1, force the filter to extend the last frame of secondary streams
351 until the end of the primary stream. A value of 0 disables this behavior.
355 @c man end OPTIONS FOR FILTERS WITH SEVERAL INPUTS
357 @chapter Audio Filters
358 @c man begin AUDIO FILTERS
360 When you configure your FFmpeg build, you can disable any of the
361 existing filters using @code{--disable-filters}.
362 The configure output will show the audio filters included in your
365 Below is a description of the currently available audio filters.
369 A compressor is mainly used to reduce the dynamic range of a signal.
370 Especially modern music is mostly compressed at a high ratio to
371 improve the overall loudness. It's done to get the highest attention
372 of a listener, "fatten" the sound and bring more "power" to the track.
373 If a signal is compressed too much it may sound dull or "dead"
374 afterwards or it may start to "pump" (which could be a powerful effect
375 but can also destroy a track completely).
376 The right compression is the key to reach a professional sound and is
377 the high art of mixing and mastering. Because of its complex settings
378 it may take a long time to get the right feeling for this kind of effect.
380 Compression is done by detecting the volume above a chosen level
381 @code{threshold} and dividing it by the factor set with @code{ratio}.
382 So if you set the threshold to -12dB and your signal reaches -6dB a ratio
383 of 2:1 will result in a signal at -9dB. Because an exact manipulation of
384 the signal would cause distortion of the waveform the reduction can be
385 levelled over the time. This is done by setting "Attack" and "Release".
386 @code{attack} determines how long the signal has to rise above the threshold
387 before any reduction will occur and @code{release} sets the time the signal
388 has to fall below the threshold to reduce the reduction again. Shorter signals
389 than the chosen attack time will be left untouched.
390 The overall reduction of the signal can be made up afterwards with the
391 @code{makeup} setting. So compressing the peaks of a signal about 6dB and
392 raising the makeup to this level results in a signal twice as loud than the
393 source. To gain a softer entry in the compression the @code{knee} flattens the
394 hard edge at the threshold in the range of the chosen decibels.
396 The filter accepts the following options:
400 Set input gain. Default is 1. Range is between 0.015625 and 64.
403 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
404 Default is @code{downward}.
407 If a signal of stream rises above this level it will affect the gain
409 By default it is 0.125. Range is between 0.00097563 and 1.
412 Set a ratio by which the signal is reduced. 1:2 means that if the level
413 rose 4dB above the threshold, it will be only 2dB above after the reduction.
414 Default is 2. Range is between 1 and 20.
417 Amount of milliseconds the signal has to rise above the threshold before gain
418 reduction starts. Default is 20. Range is between 0.01 and 2000.
421 Amount of milliseconds the signal has to fall below the threshold before
422 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
425 Set the amount by how much signal will be amplified after processing.
426 Default is 1. Range is from 1 to 64.
429 Curve the sharp knee around the threshold to enter gain reduction more softly.
430 Default is 2.82843. Range is between 1 and 8.
433 Choose if the @code{average} level between all channels of input stream
434 or the louder(@code{maximum}) channel of input stream affects the
435 reduction. Default is @code{average}.
438 Should the exact signal be taken in case of @code{peak} or an RMS one in case
439 of @code{rms}. Default is @code{rms} which is mostly smoother.
442 How much to use compressed signal in output. Default is 1.
443 Range is between 0 and 1.
448 This filter supports the all above options as @ref{commands}.
451 Simple audio dynamic range compression/expansion filter.
453 The filter accepts the following options:
457 Set contrast. Default is 33. Allowed range is between 0 and 100.
462 Copy the input audio source unchanged to the output. This is mainly useful for
467 Apply cross fade from one input audio stream to another input audio stream.
468 The cross fade is applied for specified duration near the end of first stream.
470 The filter accepts the following options:
474 Specify the number of samples for which the cross fade effect has to last.
475 At the end of the cross fade effect the first input audio will be completely
476 silent. Default is 44100.
479 Specify the duration of the cross fade effect. See
480 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
481 for the accepted syntax.
482 By default the duration is determined by @var{nb_samples}.
483 If set this option is used instead of @var{nb_samples}.
486 Should first stream end overlap with second stream start. Default is enabled.
489 Set curve for cross fade transition for first stream.
492 Set curve for cross fade transition for second stream.
494 For description of available curve types see @ref{afade} filter description.
501 Cross fade from one input to another:
503 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
507 Cross fade from one input to another but without overlapping:
509 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
514 Split audio stream into several bands.
516 This filter splits audio stream into two or more frequency ranges.
517 Summing all streams back will give flat output.
519 The filter accepts the following options:
523 Set split frequencies. Those must be positive and increasing.
526 Set filter order for each band split. This controls filter roll-off or steepness
527 of filter transfer function.
528 Available values are:
553 Default is @var{4th}.
556 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
559 Set output gain for each band. Default value is 1 for all bands.
566 Split input audio stream into two bands (low and high) with split frequency of 1500 Hz,
567 each band will be in separate stream:
569 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
573 Same as above, but with higher filter order:
575 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500:order=8th[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
579 Same as above, but also with additional middle band (frequencies between 1500 and 8000):
581 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500 8000:order=8th[LOW][MID][HIGH]' -map '[LOW]' low.wav -map '[MID]' mid.wav -map '[HIGH]' high.wav
587 Reduce audio bit resolution.
589 This filter is bit crusher with enhanced functionality. A bit crusher
590 is used to audibly reduce number of bits an audio signal is sampled
591 with. This doesn't change the bit depth at all, it just produces the
592 effect. Material reduced in bit depth sounds more harsh and "digital".
593 This filter is able to even round to continuous values instead of discrete
595 Additionally it has a D/C offset which results in different crushing of
596 the lower and the upper half of the signal.
597 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
599 Another feature of this filter is the logarithmic mode.
600 This setting switches from linear distances between bits to logarithmic ones.
601 The result is a much more "natural" sounding crusher which doesn't gate low
602 signals for example. The human ear has a logarithmic perception,
603 so this kind of crushing is much more pleasant.
604 Logarithmic crushing is also able to get anti-aliased.
606 The filter accepts the following options:
622 Can be linear: @code{lin} or logarithmic: @code{log}.
631 Set sample reduction.
634 Enable LFO. By default disabled.
645 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
649 Remove impulsive noise from input audio.
651 Samples detected as impulsive noise are replaced by interpolated samples using
652 autoregressive modelling.
656 Set window size, in milliseconds. Allowed range is from @code{10} to
657 @code{100}. Default value is @code{55} milliseconds.
658 This sets size of window which will be processed at once.
661 Set window overlap, in percentage of window size. Allowed range is from
662 @code{50} to @code{95}. Default value is @code{75} percent.
663 Setting this to a very high value increases impulsive noise removal but makes
664 whole process much slower.
667 Set autoregression order, in percentage of window size. Allowed range is from
668 @code{0} to @code{25}. Default value is @code{2} percent. This option also
669 controls quality of interpolated samples using neighbour good samples.
672 Set threshold value. Allowed range is from @code{1} to @code{100}.
673 Default value is @code{2}.
674 This controls the strength of impulsive noise which is going to be removed.
675 The lower value, the more samples will be detected as impulsive noise.
678 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
679 @code{10}. Default value is @code{2}.
680 If any two samples detected as noise are spaced less than this value then any
681 sample between those two samples will be also detected as noise.
686 It accepts the following values:
689 Select overlap-add method. Even not interpolated samples are slightly
690 changed with this method.
693 Select overlap-save method. Not interpolated samples remain unchanged.
696 Default value is @code{a}.
700 Remove clipped samples from input audio.
702 Samples detected as clipped are replaced by interpolated samples using
703 autoregressive modelling.
707 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
708 Default value is @code{55} milliseconds.
709 This sets size of window which will be processed at once.
712 Set window overlap, in percentage of window size. Allowed range is from @code{50}
713 to @code{95}. Default value is @code{75} percent.
716 Set autoregression order, in percentage of window size. Allowed range is from
717 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
718 quality of interpolated samples using neighbour good samples.
721 Set threshold value. Allowed range is from @code{1} to @code{100}.
722 Default value is @code{10}. Higher values make clip detection less aggressive.
725 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
726 Default value is @code{1000}. Higher values make clip detection less aggressive.
731 It accepts the following values:
734 Select overlap-add method. Even not interpolated samples are slightly changed
738 Select overlap-save method. Not interpolated samples remain unchanged.
741 Default value is @code{a}.
746 Delay one or more audio channels.
748 Samples in delayed channel are filled with silence.
750 The filter accepts the following option:
754 Set list of delays in milliseconds for each channel separated by '|'.
755 Unused delays will be silently ignored. If number of given delays is
756 smaller than number of channels all remaining channels will not be delayed.
757 If you want to delay exact number of samples, append 'S' to number.
758 If you want instead to delay in seconds, append 's' to number.
761 Use last set delay for all remaining channels. By default is disabled.
762 This option if enabled changes how option @code{delays} is interpreted.
769 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
770 the second channel (and any other channels that may be present) unchanged.
776 Delay second channel by 500 samples, the third channel by 700 samples and leave
777 the first channel (and any other channels that may be present) unchanged.
783 Delay all channels by same number of samples:
785 adelay=delays=64S:all=1
790 Remedy denormals in audio by adding extremely low-level noise.
792 This filter shall be placed before any filter that can produce denormals.
794 A description of the accepted parameters follows.
798 Set level of added noise in dB. Default is @code{-351}.
799 Allowed range is from -451 to -90.
802 Set type of added noise.
815 Default is @code{dc}.
820 This filter supports the all above options as @ref{commands}.
822 @section aderivative, aintegral
824 Compute derivative/integral of audio stream.
826 Applying both filters one after another produces original audio.
830 Apply echoing to the input audio.
832 Echoes are reflected sound and can occur naturally amongst mountains
833 (and sometimes large buildings) when talking or shouting; digital echo
834 effects emulate this behaviour and are often used to help fill out the
835 sound of a single instrument or vocal. The time difference between the
836 original signal and the reflection is the @code{delay}, and the
837 loudness of the reflected signal is the @code{decay}.
838 Multiple echoes can have different delays and decays.
840 A description of the accepted parameters follows.
844 Set input gain of reflected signal. Default is @code{0.6}.
847 Set output gain of reflected signal. Default is @code{0.3}.
850 Set list of time intervals in milliseconds between original signal and reflections
851 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
852 Default is @code{1000}.
855 Set list of loudness of reflected signals separated by '|'.
856 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
857 Default is @code{0.5}.
864 Make it sound as if there are twice as many instruments as are actually playing:
866 aecho=0.8:0.88:60:0.4
870 If delay is very short, then it sounds like a (metallic) robot playing music:
876 A longer delay will sound like an open air concert in the mountains:
878 aecho=0.8:0.9:1000:0.3
882 Same as above but with one more mountain:
884 aecho=0.8:0.9:1000|1800:0.3|0.25
889 Audio emphasis filter creates or restores material directly taken from LPs or
890 emphased CDs with different filter curves. E.g. to store music on vinyl the
891 signal has to be altered by a filter first to even out the disadvantages of
892 this recording medium.
893 Once the material is played back the inverse filter has to be applied to
894 restore the distortion of the frequency response.
896 The filter accepts the following options:
906 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
907 use @code{production} mode. Default is @code{reproduction} mode.
910 Set filter type. Selects medium. Can be one of the following:
922 select Compact Disc (CD).
928 select 50µs (FM-KF).
930 select 75µs (FM-KF).
936 This filter supports the all above options as @ref{commands}.
940 Modify an audio signal according to the specified expressions.
942 This filter accepts one or more expressions (one for each channel),
943 which are evaluated and used to modify a corresponding audio signal.
945 It accepts the following parameters:
949 Set the '|'-separated expressions list for each separate channel. If
950 the number of input channels is greater than the number of
951 expressions, the last specified expression is used for the remaining
954 @item channel_layout, c
955 Set output channel layout. If not specified, the channel layout is
956 specified by the number of expressions. If set to @samp{same}, it will
957 use by default the same input channel layout.
960 Each expression in @var{exprs} can contain the following constants and functions:
964 channel number of the current expression
967 number of the evaluated sample, starting from 0
973 time of the evaluated sample expressed in seconds
976 @item nb_out_channels
977 input and output number of channels
980 the value of input channel with number @var{CH}
983 Note: this filter is slow. For faster processing you should use a
992 aeval=val(ch)/2:c=same
996 Invert phase of the second channel:
1005 Apply fade-in/out effect to input audio.
1007 A description of the accepted parameters follows.
1011 Specify the effect type, can be either @code{in} for fade-in, or
1012 @code{out} for a fade-out effect. Default is @code{in}.
1014 @item start_sample, ss
1015 Specify the number of the start sample for starting to apply the fade
1016 effect. Default is 0.
1018 @item nb_samples, ns
1019 Specify the number of samples for which the fade effect has to last. At
1020 the end of the fade-in effect the output audio will have the same
1021 volume as the input audio, at the end of the fade-out transition
1022 the output audio will be silence. Default is 44100.
1024 @item start_time, st
1025 Specify the start time of the fade effect. Default is 0.
1026 The value must be specified as a time duration; see
1027 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1028 for the accepted syntax.
1029 If set this option is used instead of @var{start_sample}.
1032 Specify the duration of the fade effect. See
1033 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1034 for the accepted syntax.
1035 At the end of the fade-in effect the output audio will have the same
1036 volume as the input audio, at the end of the fade-out transition
1037 the output audio will be silence.
1038 By default the duration is determined by @var{nb_samples}.
1039 If set this option is used instead of @var{nb_samples}.
1042 Set curve for fade transition.
1044 It accepts the following values:
1047 select triangular, linear slope (default)
1049 select quarter of sine wave
1051 select half of sine wave
1053 select exponential sine wave
1057 select inverted parabola
1071 select inverted quarter of sine wave
1073 select inverted half of sine wave
1075 select double-exponential seat
1077 select double-exponential sigmoid
1079 select logistic sigmoid
1081 select sine cardinal function
1083 select inverted sine cardinal function
1089 @subsection Commands
1091 This filter supports the all above options as @ref{commands}.
1093 @subsection Examples
1097 Fade in first 15 seconds of audio:
1099 afade=t=in:ss=0:d=15
1103 Fade out last 25 seconds of a 900 seconds audio:
1105 afade=t=out:st=875:d=25
1110 Denoise audio samples with FFT.
1112 A description of the accepted parameters follows.
1116 Set the noise reduction in dB, allowed range is 0.01 to 97.
1117 Default value is 12 dB.
1120 Set the noise floor in dB, allowed range is -80 to -20.
1121 Default value is -50 dB.
1126 It accepts the following values:
1135 Select shellac noise.
1138 Select custom noise, defined in @code{bn} option.
1140 Default value is white noise.
1144 Set custom band noise for every one of 15 bands.
1145 Bands are separated by ' ' or '|'.
1148 Set the residual floor in dB, allowed range is -80 to -20.
1149 Default value is -38 dB.
1152 Enable noise tracking. By default is disabled.
1153 With this enabled, noise floor is automatically adjusted.
1156 Enable residual tracking. By default is disabled.
1159 Set the output mode.
1161 It accepts the following values:
1164 Pass input unchanged.
1167 Pass noise filtered out.
1172 Default value is @var{o}.
1176 @subsection Commands
1178 This filter supports the following commands:
1180 @item sample_noise, sn
1181 Start or stop measuring noise profile.
1182 Syntax for the command is : "start" or "stop" string.
1183 After measuring noise profile is stopped it will be
1184 automatically applied in filtering.
1186 @item noise_reduction, nr
1187 Change noise reduction. Argument is single float number.
1188 Syntax for the command is : "@var{noise_reduction}"
1190 @item noise_floor, nf
1191 Change noise floor. Argument is single float number.
1192 Syntax for the command is : "@var{noise_floor}"
1194 @item output_mode, om
1195 Change output mode operation.
1196 Syntax for the command is : "i", "o" or "n" string.
1200 Apply arbitrary expressions to samples in frequency domain.
1204 Set frequency domain real expression for each separate channel separated
1205 by '|'. Default is "re".
1206 If the number of input channels is greater than the number of
1207 expressions, the last specified expression is used for the remaining
1211 Set frequency domain imaginary expression for each separate channel
1212 separated by '|'. Default is "im".
1214 Each expression in @var{real} and @var{imag} can contain the following
1215 constants and functions:
1222 current frequency bin number
1225 number of available bins
1228 channel number of the current expression
1237 current real part of frequency bin of current channel
1240 current imaginary part of frequency bin of current channel
1243 Return the value of real part of frequency bin at location (@var{bin},@var{channel})
1246 Return the value of imaginary part of frequency bin at location (@var{bin},@var{channel})
1250 Set window size. Allowed range is from 16 to 131072.
1251 Default is @code{4096}
1254 Set window function. Default is @code{hann}.
1257 Set window overlap. If set to 1, the recommended overlap for selected
1258 window function will be picked. Default is @code{0.75}.
1261 @subsection Examples
1265 Leave almost only low frequencies in audio:
1267 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1271 Apply robotize effect:
1273 afftfilt="real='hypot(re,im)*sin(0)':imag='hypot(re,im)*cos(0)':win_size=512:overlap=0.75"
1277 Apply whisper effect:
1279 afftfilt="real='hypot(re,im)*cos((random(0)*2-1)*2*3.14)':imag='hypot(re,im)*sin((random(1)*2-1)*2*3.14)':win_size=128:overlap=0.8"
1286 Apply an arbitrary Finite Impulse Response filter.
1288 This filter is designed for applying long FIR filters,
1289 up to 60 seconds long.
1291 It can be used as component for digital crossover filters,
1292 room equalization, cross talk cancellation, wavefield synthesis,
1293 auralization, ambiophonics, ambisonics and spatialization.
1295 This filter uses the streams higher than first one as FIR coefficients.
1296 If the non-first stream holds a single channel, it will be used
1297 for all input channels in the first stream, otherwise
1298 the number of channels in the non-first stream must be same as
1299 the number of channels in the first stream.
1301 It accepts the following parameters:
1305 Set dry gain. This sets input gain.
1308 Set wet gain. This sets final output gain.
1311 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1314 Enable applying gain measured from power of IR.
1316 Set which approach to use for auto gain measurement.
1320 Do not apply any gain.
1323 select peak gain, very conservative approach. This is default value.
1326 select DC gain, limited application.
1329 select gain to noise approach, this is most popular one.
1333 Set gain to be applied to IR coefficients before filtering.
1334 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1337 Set format of IR stream. Can be @code{mono} or @code{input}.
1338 Default is @code{input}.
1341 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1342 Allowed range is 0.1 to 60 seconds.
1345 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1346 By default it is disabled.
1349 Set for which IR channel to display frequency response. By default is first channel
1350 displayed. This option is used only when @var{response} is enabled.
1353 Set video stream size. This option is used only when @var{response} is enabled.
1356 Set video stream frame rate. This option is used only when @var{response} is enabled.
1359 Set minimal partition size used for convolution. Default is @var{8192}.
1360 Allowed range is from @var{1} to @var{32768}.
1361 Lower values decreases latency at cost of higher CPU usage.
1364 Set maximal partition size used for convolution. Default is @var{8192}.
1365 Allowed range is from @var{8} to @var{32768}.
1366 Lower values may increase CPU usage.
1369 Set number of input impulse responses streams which will be switchable at runtime.
1370 Allowed range is from @var{1} to @var{32}. Default is @var{1}.
1373 Set IR stream which will be used for convolution, starting from @var{0}, should always be
1374 lower than supplied value by @code{nbirs} option. Default is @var{0}.
1375 This option can be changed at runtime via @ref{commands}.
1378 @subsection Examples
1382 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1384 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1391 Set output format constraints for the input audio. The framework will
1392 negotiate the most appropriate format to minimize conversions.
1394 It accepts the following parameters:
1397 @item sample_fmts, f
1398 A '|'-separated list of requested sample formats.
1400 @item sample_rates, r
1401 A '|'-separated list of requested sample rates.
1403 @item channel_layouts, cl
1404 A '|'-separated list of requested channel layouts.
1406 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1407 for the required syntax.
1410 If a parameter is omitted, all values are allowed.
1412 Force the output to either unsigned 8-bit or signed 16-bit stereo
1414 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1418 Apply frequency shift to input audio samples.
1420 The filter accepts the following options:
1424 Specify frequency shift. Allowed range is -INT_MAX to INT_MAX.
1425 Default value is 0.0.
1428 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
1429 Default value is 1.0.
1432 @subsection Commands
1434 This filter supports the all above options as @ref{commands}.
1438 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1439 processing reduces disturbing noise between useful signals.
1441 Gating is done by detecting the volume below a chosen level @var{threshold}
1442 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1443 floor is set via @var{range}. Because an exact manipulation of the signal
1444 would cause distortion of the waveform the reduction can be levelled over
1445 time. This is done by setting @var{attack} and @var{release}.
1447 @var{attack} determines how long the signal has to fall below the threshold
1448 before any reduction will occur and @var{release} sets the time the signal
1449 has to rise above the threshold to reduce the reduction again.
1450 Shorter signals than the chosen attack time will be left untouched.
1454 Set input level before filtering.
1455 Default is 1. Allowed range is from 0.015625 to 64.
1458 Set the mode of operation. Can be @code{upward} or @code{downward}.
1459 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
1460 will be amplified, expanding dynamic range in upward direction.
1461 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
1464 Set the level of gain reduction when the signal is below the threshold.
1465 Default is 0.06125. Allowed range is from 0 to 1.
1466 Setting this to 0 disables reduction and then filter behaves like expander.
1469 If a signal rises above this level the gain reduction is released.
1470 Default is 0.125. Allowed range is from 0 to 1.
1473 Set a ratio by which the signal is reduced.
1474 Default is 2. Allowed range is from 1 to 9000.
1477 Amount of milliseconds the signal has to rise above the threshold before gain
1479 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1482 Amount of milliseconds the signal has to fall below the threshold before the
1483 reduction is increased again. Default is 250 milliseconds.
1484 Allowed range is from 0.01 to 9000.
1487 Set amount of amplification of signal after processing.
1488 Default is 1. Allowed range is from 1 to 64.
1491 Curve the sharp knee around the threshold to enter gain reduction more softly.
1492 Default is 2.828427125. Allowed range is from 1 to 8.
1495 Choose if exact signal should be taken for detection or an RMS like one.
1496 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1499 Choose if the average level between all channels or the louder channel affects
1501 Default is @code{average}. Can be @code{average} or @code{maximum}.
1504 @subsection Commands
1506 This filter supports the all above options as @ref{commands}.
1510 Apply an arbitrary Infinite Impulse Response filter.
1512 It accepts the following parameters:
1516 Set B/numerator/zeros/reflection coefficients.
1519 Set A/denominator/poles/ladder coefficients.
1531 Set coefficients format.
1535 lattice-ladder function
1537 analog transfer function
1539 digital transfer function
1541 Z-plane zeros/poles, cartesian (default)
1543 Z-plane zeros/poles, polar radians
1545 Z-plane zeros/poles, polar degrees
1551 Set type of processing.
1563 Set filtering precision.
1567 double-precision floating-point (default)
1569 single-precision floating-point
1577 Normalize filter coefficients, by default is enabled.
1578 Enabling it will normalize magnitude response at DC to 0dB.
1581 How much to use filtered signal in output. Default is 1.
1582 Range is between 0 and 1.
1585 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1586 By default it is disabled.
1589 Set for which IR channel to display frequency response. By default is first channel
1590 displayed. This option is used only when @var{response} is enabled.
1593 Set video stream size. This option is used only when @var{response} is enabled.
1596 Coefficients in @code{tf} and @code{sf} format are separated by spaces and are in ascending
1599 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1600 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1603 Different coefficients and gains can be provided for every channel, in such case
1604 use '|' to separate coefficients or gains. Last provided coefficients will be
1605 used for all remaining channels.
1607 @subsection Examples
1611 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1613 aiir=k=1:z=7.957584807809675810E-1 -2.575128568908332300 3.674839853930788710 -2.57512875289799137 7.957586296317130880E-1:p=1 -2.86950072432325953 3.63022088054647218 -2.28075678147272232 6.361362326477423500E-1:f=tf:r=d
1617 Same as above but in @code{zp} format:
1619 aiir=k=0.79575848078096756:z=0.80918701+0.58773007i 0.80918701-0.58773007i 0.80884700+0.58784055i 0.80884700-0.58784055i:p=0.63892345+0.59951235i 0.63892345-0.59951235i 0.79582691+0.44198673i 0.79582691-0.44198673i:f=zp:r=s
1623 Apply 3-rd order analog normalized Butterworth low-pass filter, using analog transfer function format:
1625 aiir=z=1.3057 0 0 0:p=1.3057 2.3892 2.1860 1:f=sf:r=d
1631 The limiter prevents an input signal from rising over a desired threshold.
1632 This limiter uses lookahead technology to prevent your signal from distorting.
1633 It means that there is a small delay after the signal is processed. Keep in mind
1634 that the delay it produces is the attack time you set.
1636 The filter accepts the following options:
1640 Set input gain. Default is 1.
1643 Set output gain. Default is 1.
1646 Don't let signals above this level pass the limiter. Default is 1.
1649 The limiter will reach its attenuation level in this amount of time in
1650 milliseconds. Default is 5 milliseconds.
1653 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1654 Default is 50 milliseconds.
1657 When gain reduction is always needed ASC takes care of releasing to an
1658 average reduction level rather than reaching a reduction of 0 in the release
1662 Select how much the release time is affected by ASC, 0 means nearly no changes
1663 in release time while 1 produces higher release times.
1666 Auto level output signal. Default is enabled.
1667 This normalizes audio back to 0dB if enabled.
1670 Depending on picked setting it is recommended to upsample input 2x or 4x times
1671 with @ref{aresample} before applying this filter.
1675 Apply a two-pole all-pass filter with central frequency (in Hz)
1676 @var{frequency}, and filter-width @var{width}.
1677 An all-pass filter changes the audio's frequency to phase relationship
1678 without changing its frequency to amplitude relationship.
1680 The filter accepts the following options:
1684 Set frequency in Hz.
1687 Set method to specify band-width of filter.
1702 Specify the band-width of a filter in width_type units.
1705 How much to use filtered signal in output. Default is 1.
1706 Range is between 0 and 1.
1709 Specify which channels to filter, by default all available are filtered.
1712 Normalize biquad coefficients, by default is disabled.
1713 Enabling it will normalize magnitude response at DC to 0dB.
1716 Set the filter order, can be 1 or 2. Default is 2.
1719 Set transform type of IIR filter.
1728 Set precison of filtering.
1731 Pick automatic sample format depending on surround filters.
1733 Always use signed 16-bit.
1735 Always use signed 32-bit.
1737 Always use float 32-bit.
1739 Always use float 64-bit.
1743 @subsection Commands
1745 This filter supports the following commands:
1748 Change allpass frequency.
1749 Syntax for the command is : "@var{frequency}"
1752 Change allpass width_type.
1753 Syntax for the command is : "@var{width_type}"
1756 Change allpass width.
1757 Syntax for the command is : "@var{width}"
1761 Syntax for the command is : "@var{mix}"
1768 The filter accepts the following options:
1772 Set the number of loops. Setting this value to -1 will result in infinite loops.
1776 Set maximal number of samples. Default is 0.
1779 Set first sample of loop. Default is 0.
1785 Merge two or more audio streams into a single multi-channel stream.
1787 The filter accepts the following options:
1792 Set the number of inputs. Default is 2.
1796 If the channel layouts of the inputs are disjoint, and therefore compatible,
1797 the channel layout of the output will be set accordingly and the channels
1798 will be reordered as necessary. If the channel layouts of the inputs are not
1799 disjoint, the output will have all the channels of the first input then all
1800 the channels of the second input, in that order, and the channel layout of
1801 the output will be the default value corresponding to the total number of
1804 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1805 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1806 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1807 first input, b1 is the first channel of the second input).
1809 On the other hand, if both input are in stereo, the output channels will be
1810 in the default order: a1, a2, b1, b2, and the channel layout will be
1811 arbitrarily set to 4.0, which may or may not be the expected value.
1813 All inputs must have the same sample rate, and format.
1815 If inputs do not have the same duration, the output will stop with the
1818 @subsection Examples
1822 Merge two mono files into a stereo stream:
1824 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1828 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1830 ffmpeg -i input.mkv -filter_complex "[0:1][0:2][0:3][0:4][0:5][0:6] amerge=inputs=6" -c:a pcm_s16le output.mkv
1836 Mixes multiple audio inputs into a single output.
1838 Note that this filter only supports float samples (the @var{amerge}
1839 and @var{pan} audio filters support many formats). If the @var{amix}
1840 input has integer samples then @ref{aresample} will be automatically
1841 inserted to perform the conversion to float samples.
1845 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1847 will mix 3 input audio streams to a single output with the same duration as the
1848 first input and a dropout transition time of 3 seconds.
1850 It accepts the following parameters:
1854 The number of inputs. If unspecified, it defaults to 2.
1857 How to determine the end-of-stream.
1861 The duration of the longest input. (default)
1864 The duration of the shortest input.
1867 The duration of the first input.
1871 @item dropout_transition
1872 The transition time, in seconds, for volume renormalization when an input
1873 stream ends. The default value is 2 seconds.
1876 Specify weight of each input audio stream as sequence.
1877 Each weight is separated by space. By default all inputs have same weight.
1880 @subsection Commands
1882 This filter supports the following commands:
1885 Syntax is same as option with same name.
1890 Multiply first audio stream with second audio stream and store result
1891 in output audio stream. Multiplication is done by multiplying each
1892 sample from first stream with sample at same position from second stream.
1894 With this element-wise multiplication one can create amplitude fades and
1895 amplitude modulations.
1897 @section anequalizer
1899 High-order parametric multiband equalizer for each channel.
1901 It accepts the following parameters:
1905 This option string is in format:
1906 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1907 Each equalizer band is separated by '|'.
1911 Set channel number to which equalization will be applied.
1912 If input doesn't have that channel the entry is ignored.
1915 Set central frequency for band.
1916 If input doesn't have that frequency the entry is ignored.
1919 Set band width in Hertz.
1922 Set band gain in dB.
1925 Set filter type for band, optional, can be:
1929 Butterworth, this is default.
1940 With this option activated frequency response of anequalizer is displayed
1944 Set video stream size. Only useful if curves option is activated.
1947 Set max gain that will be displayed. Only useful if curves option is activated.
1948 Setting this to a reasonable value makes it possible to display gain which is derived from
1949 neighbour bands which are too close to each other and thus produce higher gain
1950 when both are activated.
1953 Set frequency scale used to draw frequency response in video output.
1954 Can be linear or logarithmic. Default is logarithmic.
1957 Set color for each channel curve which is going to be displayed in video stream.
1958 This is list of color names separated by space or by '|'.
1959 Unrecognised or missing colors will be replaced by white color.
1962 @subsection Examples
1966 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1967 for first 2 channels using Chebyshev type 1 filter:
1969 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1973 @subsection Commands
1975 This filter supports the following commands:
1978 Alter existing filter parameters.
1979 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1981 @var{fN} is existing filter number, starting from 0, if no such filter is available
1983 @var{freq} set new frequency parameter.
1984 @var{width} set new width parameter in Hertz.
1985 @var{gain} set new gain parameter in dB.
1987 Full filter invocation with asendcmd may look like this:
1988 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1993 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1995 Each sample is adjusted by looking for other samples with similar contexts. This
1996 context similarity is defined by comparing their surrounding patches of size
1997 @option{p}. Patches are searched in an area of @option{r} around the sample.
1999 The filter accepts the following options:
2003 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
2006 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
2007 Default value is 2 milliseconds.
2010 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
2011 Default value is 6 milliseconds.
2014 Set the output mode.
2016 It accepts the following values:
2019 Pass input unchanged.
2022 Pass noise filtered out.
2027 Default value is @var{o}.
2031 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
2034 @subsection Commands
2036 This filter supports the all above options as @ref{commands}.
2039 Apply Normalized Least-Mean-Squares algorithm to the first audio stream using the second audio stream.
2041 This adaptive filter is used to mimic a desired filter by finding the filter coefficients that
2042 relate to producing the least mean square of the error signal (difference between the desired,
2043 2nd input audio stream and the actual signal, the 1st input audio stream).
2045 A description of the accepted options follows.
2058 Set the filter leakage.
2061 It accepts the following values:
2070 Pass filtered samples.
2073 Pass difference between desired and filtered samples.
2075 Default value is @var{o}.
2079 @subsection Examples
2083 One of many usages of this filter is noise reduction, input audio is filtered
2084 with same samples that are delayed by fixed amount, one such example for stereo audio is:
2086 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
2090 @subsection Commands
2092 This filter supports the same commands as options, excluding option @code{order}.
2096 Pass the audio source unchanged to the output.
2100 Pad the end of an audio stream with silence.
2102 This can be used together with @command{ffmpeg} @option{-shortest} to
2103 extend audio streams to the same length as the video stream.
2105 A description of the accepted options follows.
2109 Set silence packet size. Default value is 4096.
2112 Set the number of samples of silence to add to the end. After the
2113 value is reached, the stream is terminated. This option is mutually
2114 exclusive with @option{whole_len}.
2117 Set the minimum total number of samples in the output audio stream. If
2118 the value is longer than the input audio length, silence is added to
2119 the end, until the value is reached. This option is mutually exclusive
2120 with @option{pad_len}.
2123 Specify the duration of samples of silence to add. See
2124 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2125 for the accepted syntax. Used only if set to non-zero value.
2128 Specify the minimum total duration in the output audio stream. See
2129 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2130 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
2131 the input audio length, silence is added to the end, until the value is reached.
2132 This option is mutually exclusive with @option{pad_dur}
2135 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
2136 nor @option{whole_dur} option is set, the filter will add silence to the end of
2137 the input stream indefinitely.
2139 @subsection Examples
2143 Add 1024 samples of silence to the end of the input:
2149 Make sure the audio output will contain at least 10000 samples, pad
2150 the input with silence if required:
2152 apad=whole_len=10000
2156 Use @command{ffmpeg} to pad the audio input with silence, so that the
2157 video stream will always result the shortest and will be converted
2158 until the end in the output file when using the @option{shortest}
2161 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
2166 Add a phasing effect to the input audio.
2168 A phaser filter creates series of peaks and troughs in the frequency spectrum.
2169 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
2171 A description of the accepted parameters follows.
2175 Set input gain. Default is 0.4.
2178 Set output gain. Default is 0.74
2181 Set delay in milliseconds. Default is 3.0.
2184 Set decay. Default is 0.4.
2187 Set modulation speed in Hz. Default is 0.5.
2190 Set modulation type. Default is triangular.
2192 It accepts the following values:
2199 @section aphaseshift
2200 Apply phase shift to input audio samples.
2202 The filter accepts the following options:
2206 Specify phase shift. Allowed range is from -1.0 to 1.0.
2207 Default value is 0.0.
2210 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
2211 Default value is 1.0.
2214 @subsection Commands
2216 This filter supports the all above options as @ref{commands}.
2220 Audio pulsator is something between an autopanner and a tremolo.
2221 But it can produce funny stereo effects as well. Pulsator changes the volume
2222 of the left and right channel based on a LFO (low frequency oscillator) with
2223 different waveforms and shifted phases.
2224 This filter have the ability to define an offset between left and right
2225 channel. An offset of 0 means that both LFO shapes match each other.
2226 The left and right channel are altered equally - a conventional tremolo.
2227 An offset of 50% means that the shape of the right channel is exactly shifted
2228 in phase (or moved backwards about half of the frequency) - pulsator acts as
2229 an autopanner. At 1 both curves match again. Every setting in between moves the
2230 phase shift gapless between all stages and produces some "bypassing" sounds with
2231 sine and triangle waveforms. The more you set the offset near 1 (starting from
2232 the 0.5) the faster the signal passes from the left to the right speaker.
2234 The filter accepts the following options:
2238 Set input gain. By default it is 1. Range is [0.015625 - 64].
2241 Set output gain. By default it is 1. Range is [0.015625 - 64].
2244 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
2245 sawup or sawdown. Default is sine.
2248 Set modulation. Define how much of original signal is affected by the LFO.
2251 Set left channel offset. Default is 0. Allowed range is [0 - 1].
2254 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
2257 Set pulse width. Default is 1. Allowed range is [0 - 2].
2260 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
2263 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
2267 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
2271 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
2272 if timing is set to hz.
2278 Resample the input audio to the specified parameters, using the
2279 libswresample library. If none are specified then the filter will
2280 automatically convert between its input and output.
2282 This filter is also able to stretch/squeeze the audio data to make it match
2283 the timestamps or to inject silence / cut out audio to make it match the
2284 timestamps, do a combination of both or do neither.
2286 The filter accepts the syntax
2287 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
2288 expresses a sample rate and @var{resampler_options} is a list of
2289 @var{key}=@var{value} pairs, separated by ":". See the
2290 @ref{Resampler Options,,"Resampler Options" section in the
2291 ffmpeg-resampler(1) manual,ffmpeg-resampler}
2292 for the complete list of supported options.
2294 @subsection Examples
2298 Resample the input audio to 44100Hz:
2304 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2305 samples per second compensation:
2307 aresample=async=1000
2313 Reverse an audio clip.
2315 Warning: This filter requires memory to buffer the entire clip, so trimming
2318 @subsection Examples
2322 Take the first 5 seconds of a clip, and reverse it.
2324 atrim=end=5,areverse
2330 Reduce noise from speech using Recurrent Neural Networks.
2332 This filter accepts the following options:
2336 Set train model file to load. This option is always required.
2339 Set how much to mix filtered samples into final output.
2340 Allowed range is from -1 to 1. Default value is 1.
2341 Negative values are special, they set how much to keep filtered noise
2342 in the final filter output. Set this option to -1 to hear actual
2343 noise removed from input signal.
2346 @section asetnsamples
2348 Set the number of samples per each output audio frame.
2350 The last output packet may contain a different number of samples, as
2351 the filter will flush all the remaining samples when the input audio
2354 The filter accepts the following options:
2358 @item nb_out_samples, n
2359 Set the number of frames per each output audio frame. The number is
2360 intended as the number of samples @emph{per each channel}.
2361 Default value is 1024.
2364 If set to 1, the filter will pad the last audio frame with zeroes, so
2365 that the last frame will contain the same number of samples as the
2366 previous ones. Default value is 1.
2369 For example, to set the number of per-frame samples to 1234 and
2370 disable padding for the last frame, use:
2372 asetnsamples=n=1234:p=0
2377 Set the sample rate without altering the PCM data.
2378 This will result in a change of speed and pitch.
2380 The filter accepts the following options:
2383 @item sample_rate, r
2384 Set the output sample rate. Default is 44100 Hz.
2389 Show a line containing various information for each input audio frame.
2390 The input audio is not modified.
2392 The shown line contains a sequence of key/value pairs of the form
2393 @var{key}:@var{value}.
2395 The following values are shown in the output:
2399 The (sequential) number of the input frame, starting from 0.
2402 The presentation timestamp of the input frame, in time base units; the time base
2403 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2406 The presentation timestamp of the input frame in seconds.
2409 position of the frame in the input stream, -1 if this information in
2410 unavailable and/or meaningless (for example in case of synthetic audio)
2419 The sample rate for the audio frame.
2422 The number of samples (per channel) in the frame.
2425 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2426 audio, the data is treated as if all the planes were concatenated.
2428 @item plane_checksums
2429 A list of Adler-32 checksums for each data plane.
2433 Apply audio soft clipping.
2435 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2436 along a smooth curve, rather than the abrupt shape of hard-clipping.
2438 This filter accepts the following options:
2442 Set type of soft-clipping.
2444 It accepts the following values:
2458 Set threshold from where to start clipping. Default value is 0dB or 1.
2461 Set gain applied to output. Default value is 0dB or 1.
2464 Set additional parameter which controls sigmoid function.
2467 Set oversampling factor.
2470 @subsection Commands
2472 This filter supports the all above options as @ref{commands}.
2475 Automatic Speech Recognition
2477 This filter uses PocketSphinx for speech recognition. To enable
2478 compilation of this filter, you need to configure FFmpeg with
2479 @code{--enable-pocketsphinx}.
2481 It accepts the following options:
2485 Set sampling rate of input audio. Defaults is @code{16000}.
2486 This need to match speech models, otherwise one will get poor results.
2489 Set dictionary containing acoustic model files.
2492 Set pronunciation dictionary.
2495 Set language model file.
2498 Set language model set.
2501 Set which language model to use.
2504 Set output for log messages.
2507 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2512 Display time domain statistical information about the audio channels.
2513 Statistics are calculated and displayed for each audio channel and,
2514 where applicable, an overall figure is also given.
2516 It accepts the following option:
2519 Short window length in seconds, used for peak and trough RMS measurement.
2520 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2524 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2525 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2528 Available keys for each channel are:
2574 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2575 this @code{lavfi.astats.Overall.Peak_count}.
2577 For description what each key means read below.
2580 Set number of frame after which stats are going to be recalculated.
2581 Default is disabled.
2583 @item measure_perchannel
2584 Select the entries which need to be measured per channel. The metadata keys can
2585 be used as flags, default is @option{all} which measures everything.
2586 @option{none} disables all per channel measurement.
2588 @item measure_overall
2589 Select the entries which need to be measured overall. The metadata keys can
2590 be used as flags, default is @option{all} which measures everything.
2591 @option{none} disables all overall measurement.
2595 A description of each shown parameter follows:
2599 Mean amplitude displacement from zero.
2602 Minimal sample level.
2605 Maximal sample level.
2607 @item Min difference
2608 Minimal difference between two consecutive samples.
2610 @item Max difference
2611 Maximal difference between two consecutive samples.
2613 @item Mean difference
2614 Mean difference between two consecutive samples.
2615 The average of each difference between two consecutive samples.
2617 @item RMS difference
2618 Root Mean Square difference between two consecutive samples.
2622 Standard peak and RMS level measured in dBFS.
2626 Peak and trough values for RMS level measured over a short window.
2629 Standard ratio of peak to RMS level (note: not in dB).
2632 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2633 (i.e. either @var{Min level} or @var{Max level}).
2636 Number of occasions (not the number of samples) that the signal attained either
2637 @var{Min level} or @var{Max level}.
2639 @item Noise floor dB
2640 Minimum local peak measured in dBFS over a short window.
2642 @item Noise floor count
2643 Number of occasions (not the number of samples) that the signal attained
2647 Overall bit depth of audio. Number of bits used for each sample.
2650 Measured dynamic range of audio in dB.
2652 @item Zero crossings
2653 Number of points where the waveform crosses the zero level axis.
2655 @item Zero crossings rate
2656 Rate of Zero crossings and number of audio samples.
2660 Boost subwoofer frequencies.
2662 The filter accepts the following options:
2666 Set dry gain, how much of original signal is kept. Allowed range is from 0 to 1.
2667 Default value is 0.7.
2670 Set wet gain, how much of filtered signal is kept. Allowed range is from 0 to 1.
2671 Default value is 0.7.
2674 Set delay line decay gain value. Allowed range is from 0 to 1.
2675 Default value is 0.7.
2678 Set delay line feedback gain value. Allowed range is from 0 to 1.
2679 Default value is 0.9.
2682 Set cutoff frequency in Hertz. Allowed range is 50 to 900.
2683 Default value is 100.
2686 Set slope amount for cutoff frequency. Allowed range is 0.0001 to 1.
2687 Default value is 0.5.
2690 Set delay. Allowed range is from 1 to 100.
2691 Default value is 20.
2694 @subsection Commands
2696 This filter supports the all above options as @ref{commands}.
2699 Cut subwoofer frequencies.
2701 This filter allows to set custom, steeper
2702 roll off than highpass filter, and thus is able to more attenuate
2703 frequency content in stop-band.
2705 The filter accepts the following options:
2709 Set cutoff frequency in Hertz. Allowed range is 2 to 200.
2710 Default value is 20.
2713 Set filter order. Available values are from 3 to 20.
2714 Default value is 10.
2717 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2720 @subsection Commands
2722 This filter supports the all above options as @ref{commands}.
2725 Cut super frequencies.
2727 The filter accepts the following options:
2731 Set cutoff frequency in Hertz. Allowed range is 20000 to 192000.
2732 Default value is 20000.
2735 Set filter order. Available values are from 3 to 20.
2736 Default value is 10.
2739 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2742 @subsection Commands
2744 This filter supports the all above options as @ref{commands}.
2747 Apply high order Butterworth band-pass filter.
2749 The filter accepts the following options:
2753 Set center frequency in Hertz. Allowed range is 2 to 999999.
2754 Default value is 1000.
2757 Set filter order. Available values are from 4 to 20.
2761 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2764 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2767 @subsection Commands
2769 This filter supports the all above options as @ref{commands}.
2772 Apply high order Butterworth band-stop filter.
2774 The filter accepts the following options:
2778 Set center frequency in Hertz. Allowed range is 2 to 999999.
2779 Default value is 1000.
2782 Set filter order. Available values are from 4 to 20.
2786 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2789 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2792 @subsection Commands
2794 This filter supports the all above options as @ref{commands}.
2800 The filter accepts exactly one parameter, the audio tempo. If not
2801 specified then the filter will assume nominal 1.0 tempo. Tempo must
2802 be in the [0.5, 100.0] range.
2804 Note that tempo greater than 2 will skip some samples rather than
2805 blend them in. If for any reason this is a concern it is always
2806 possible to daisy-chain several instances of atempo to achieve the
2807 desired product tempo.
2809 @subsection Examples
2813 Slow down audio to 80% tempo:
2819 To speed up audio to 300% tempo:
2825 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2827 atempo=sqrt(3),atempo=sqrt(3)
2831 @subsection Commands
2833 This filter supports the following commands:
2836 Change filter tempo scale factor.
2837 Syntax for the command is : "@var{tempo}"
2842 Trim the input so that the output contains one continuous subpart of the input.
2844 It accepts the following parameters:
2847 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2848 sample with the timestamp @var{start} will be the first sample in the output.
2851 Specify time of the first audio sample that will be dropped, i.e. the
2852 audio sample immediately preceding the one with the timestamp @var{end} will be
2853 the last sample in the output.
2856 Same as @var{start}, except this option sets the start timestamp in samples
2860 Same as @var{end}, except this option sets the end timestamp in samples instead
2864 The maximum duration of the output in seconds.
2867 The number of the first sample that should be output.
2870 The number of the first sample that should be dropped.
2873 @option{start}, @option{end}, and @option{duration} are expressed as time
2874 duration specifications; see
2875 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2877 Note that the first two sets of the start/end options and the @option{duration}
2878 option look at the frame timestamp, while the _sample options simply count the
2879 samples that pass through the filter. So start/end_pts and start/end_sample will
2880 give different results when the timestamps are wrong, inexact or do not start at
2881 zero. Also note that this filter does not modify the timestamps. If you wish
2882 to have the output timestamps start at zero, insert the asetpts filter after the
2885 If multiple start or end options are set, this filter tries to be greedy and
2886 keep all samples that match at least one of the specified constraints. To keep
2887 only the part that matches all the constraints at once, chain multiple atrim
2890 The defaults are such that all the input is kept. So it is possible to set e.g.
2891 just the end values to keep everything before the specified time.
2896 Drop everything except the second minute of input:
2898 ffmpeg -i INPUT -af atrim=60:120
2902 Keep only the first 1000 samples:
2904 ffmpeg -i INPUT -af atrim=end_sample=1000
2909 @section axcorrelate
2910 Calculate normalized cross-correlation between two input audio streams.
2912 Resulted samples are always between -1 and 1 inclusive.
2913 If result is 1 it means two input samples are highly correlated in that selected segment.
2914 Result 0 means they are not correlated at all.
2915 If result is -1 it means two input samples are out of phase, which means they cancel each
2918 The filter accepts the following options:
2922 Set size of segment over which cross-correlation is calculated.
2923 Default is 256. Allowed range is from 2 to 131072.
2926 Set algorithm for cross-correlation. Can be @code{slow} or @code{fast}.
2927 Default is @code{slow}. Fast algorithm assumes mean values over any given segment
2928 are always zero and thus need much less calculations to make.
2929 This is generally not true, but is valid for typical audio streams.
2932 @subsection Examples
2936 Calculate correlation between channels in stereo audio stream:
2938 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
2944 Apply a two-pole Butterworth band-pass filter with central
2945 frequency @var{frequency}, and (3dB-point) band-width width.
2946 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2947 instead of the default: constant 0dB peak gain.
2948 The filter roll off at 6dB per octave (20dB per decade).
2950 The filter accepts the following options:
2954 Set the filter's central frequency. Default is @code{3000}.
2957 Constant skirt gain if set to 1. Defaults to 0.
2960 Set method to specify band-width of filter.
2975 Specify the band-width of a filter in width_type units.
2978 How much to use filtered signal in output. Default is 1.
2979 Range is between 0 and 1.
2982 Specify which channels to filter, by default all available are filtered.
2985 Normalize biquad coefficients, by default is disabled.
2986 Enabling it will normalize magnitude response at DC to 0dB.
2989 Set transform type of IIR filter.
2998 Set precison of filtering.
3001 Pick automatic sample format depending on surround filters.
3003 Always use signed 16-bit.
3005 Always use signed 32-bit.
3007 Always use float 32-bit.
3009 Always use float 64-bit.
3013 @subsection Commands
3015 This filter supports the following commands:
3018 Change bandpass frequency.
3019 Syntax for the command is : "@var{frequency}"
3022 Change bandpass width_type.
3023 Syntax for the command is : "@var{width_type}"
3026 Change bandpass width.
3027 Syntax for the command is : "@var{width}"
3030 Change bandpass mix.
3031 Syntax for the command is : "@var{mix}"
3036 Apply a two-pole Butterworth band-reject filter with central
3037 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
3038 The filter roll off at 6dB per octave (20dB per decade).
3040 The filter accepts the following options:
3044 Set the filter's central frequency. Default is @code{3000}.
3047 Set method to specify band-width of filter.
3062 Specify the band-width of a filter in width_type units.
3065 How much to use filtered signal in output. Default is 1.
3066 Range is between 0 and 1.
3069 Specify which channels to filter, by default all available are filtered.
3072 Normalize biquad coefficients, by default is disabled.
3073 Enabling it will normalize magnitude response at DC to 0dB.
3076 Set transform type of IIR filter.
3085 Set precison of filtering.
3088 Pick automatic sample format depending on surround filters.
3090 Always use signed 16-bit.
3092 Always use signed 32-bit.
3094 Always use float 32-bit.
3096 Always use float 64-bit.
3100 @subsection Commands
3102 This filter supports the following commands:
3105 Change bandreject frequency.
3106 Syntax for the command is : "@var{frequency}"
3109 Change bandreject width_type.
3110 Syntax for the command is : "@var{width_type}"
3113 Change bandreject width.
3114 Syntax for the command is : "@var{width}"
3117 Change bandreject mix.
3118 Syntax for the command is : "@var{mix}"
3121 @section bass, lowshelf
3123 Boost or cut the bass (lower) frequencies of the audio using a two-pole
3124 shelving filter with a response similar to that of a standard
3125 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
3127 The filter accepts the following options:
3131 Give the gain at 0 Hz. Its useful range is about -20
3132 (for a large cut) to +20 (for a large boost).
3133 Beware of clipping when using a positive gain.
3136 Set the filter's central frequency and so can be used
3137 to extend or reduce the frequency range to be boosted or cut.
3138 The default value is @code{100} Hz.
3141 Set method to specify band-width of filter.
3156 Determine how steep is the filter's shelf transition.
3159 Set number of poles. Default is 2.
3162 How much to use filtered signal in output. Default is 1.
3163 Range is between 0 and 1.
3166 Specify which channels to filter, by default all available are filtered.
3169 Normalize biquad coefficients, by default is disabled.
3170 Enabling it will normalize magnitude response at DC to 0dB.
3173 Set transform type of IIR filter.
3182 Set precison of filtering.
3185 Pick automatic sample format depending on surround filters.
3187 Always use signed 16-bit.
3189 Always use signed 32-bit.
3191 Always use float 32-bit.
3193 Always use float 64-bit.
3197 @subsection Commands
3199 This filter supports the following commands:
3202 Change bass frequency.
3203 Syntax for the command is : "@var{frequency}"
3206 Change bass width_type.
3207 Syntax for the command is : "@var{width_type}"
3211 Syntax for the command is : "@var{width}"
3215 Syntax for the command is : "@var{gain}"
3219 Syntax for the command is : "@var{mix}"
3224 Apply a biquad IIR filter with the given coefficients.
3225 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
3226 are the numerator and denominator coefficients respectively.
3227 and @var{channels}, @var{c} specify which channels to filter, by default all
3228 available are filtered.
3230 @subsection Commands
3232 This filter supports the following commands:
3240 Change biquad parameter.
3241 Syntax for the command is : "@var{value}"
3244 How much to use filtered signal in output. Default is 1.
3245 Range is between 0 and 1.
3248 Specify which channels to filter, by default all available are filtered.
3251 Normalize biquad coefficients, by default is disabled.
3252 Enabling it will normalize magnitude response at DC to 0dB.
3255 Set transform type of IIR filter.
3264 Set precison of filtering.
3267 Pick automatic sample format depending on surround filters.
3269 Always use signed 16-bit.
3271 Always use signed 32-bit.
3273 Always use float 32-bit.
3275 Always use float 64-bit.
3280 Bauer stereo to binaural transformation, which improves headphone listening of
3281 stereo audio records.
3283 To enable compilation of this filter you need to configure FFmpeg with
3284 @code{--enable-libbs2b}.
3286 It accepts the following parameters:
3290 Pre-defined crossfeed level.
3294 Default level (fcut=700, feed=50).
3297 Chu Moy circuit (fcut=700, feed=60).
3300 Jan Meier circuit (fcut=650, feed=95).
3305 Cut frequency (in Hz).
3314 Remap input channels to new locations.
3316 It accepts the following parameters:
3319 Map channels from input to output. The argument is a '|'-separated list of
3320 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
3321 @var{in_channel} form. @var{in_channel} can be either the name of the input
3322 channel (e.g. FL for front left) or its index in the input channel layout.
3323 @var{out_channel} is the name of the output channel or its index in the output
3324 channel layout. If @var{out_channel} is not given then it is implicitly an
3325 index, starting with zero and increasing by one for each mapping.
3327 @item channel_layout
3328 The channel layout of the output stream.
3331 If no mapping is present, the filter will implicitly map input channels to
3332 output channels, preserving indices.
3334 @subsection Examples
3338 For example, assuming a 5.1+downmix input MOV file,
3340 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
3342 will create an output WAV file tagged as stereo from the downmix channels of
3346 To fix a 5.1 WAV improperly encoded in AAC's native channel order
3348 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
3352 @section channelsplit
3354 Split each channel from an input audio stream into a separate output stream.
3356 It accepts the following parameters:
3358 @item channel_layout
3359 The channel layout of the input stream. The default is "stereo".
3361 A channel layout describing the channels to be extracted as separate output streams
3362 or "all" to extract each input channel as a separate stream. The default is "all".
3364 Choosing channels not present in channel layout in the input will result in an error.
3367 @subsection Examples
3371 For example, assuming a stereo input MP3 file,
3373 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
3375 will create an output Matroska file with two audio streams, one containing only
3376 the left channel and the other the right channel.
3379 Split a 5.1 WAV file into per-channel files:
3381 ffmpeg -i in.wav -filter_complex
3382 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
3383 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
3384 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
3389 Extract only LFE from a 5.1 WAV file:
3391 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
3392 -map '[LFE]' lfe.wav
3397 Add a chorus effect to the audio.
3399 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
3401 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
3402 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
3403 The modulation depth defines the range the modulated delay is played before or after
3404 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
3405 sound tuned around the original one, like in a chorus where some vocals are slightly
3408 It accepts the following parameters:
3411 Set input gain. Default is 0.4.
3414 Set output gain. Default is 0.4.
3417 Set delays. A typical delay is around 40ms to 60ms.
3429 @subsection Examples
3435 chorus=0.7:0.9:55:0.4:0.25:2
3441 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
3445 Fuller sounding chorus with three delays:
3447 chorus=0.5:0.9:50|60|40:0.4|0.32|0.3:0.25|0.4|0.3:2|2.3|1.3
3452 Compress or expand the audio's dynamic range.
3454 It accepts the following parameters:
3460 A list of times in seconds for each channel over which the instantaneous level
3461 of the input signal is averaged to determine its volume. @var{attacks} refers to
3462 increase of volume and @var{decays} refers to decrease of volume. For most
3463 situations, the attack time (response to the audio getting louder) should be
3464 shorter than the decay time, because the human ear is more sensitive to sudden
3465 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
3466 a typical value for decay is 0.8 seconds.
3467 If specified number of attacks & decays is lower than number of channels, the last
3468 set attack/decay will be used for all remaining channels.
3471 A list of points for the transfer function, specified in dB relative to the
3472 maximum possible signal amplitude. Each key points list must be defined using
3473 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
3474 @code{x0/y0 x1/y1 x2/y2 ....}
3476 The input values must be in strictly increasing order but the transfer function
3477 does not have to be monotonically rising. The point @code{0/0} is assumed but
3478 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
3479 function are @code{-70/-70|-60/-20|1/0}.
3482 Set the curve radius in dB for all joints. It defaults to 0.01.
3485 Set the additional gain in dB to be applied at all points on the transfer
3486 function. This allows for easy adjustment of the overall gain.
3490 Set an initial volume, in dB, to be assumed for each channel when filtering
3491 starts. This permits the user to supply a nominal level initially, so that, for
3492 example, a very large gain is not applied to initial signal levels before the
3493 companding has begun to operate. A typical value for audio which is initially
3494 quiet is -90 dB. It defaults to 0.
3497 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
3498 delayed before being fed to the volume adjuster. Specifying a delay
3499 approximately equal to the attack/decay times allows the filter to effectively
3500 operate in predictive rather than reactive mode. It defaults to 0.
3504 @subsection Examples
3508 Make music with both quiet and loud passages suitable for listening to in a
3511 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
3514 Another example for audio with whisper and explosion parts:
3516 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
3520 A noise gate for when the noise is at a lower level than the signal:
3522 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
3526 Here is another noise gate, this time for when the noise is at a higher level
3527 than the signal (making it, in some ways, similar to squelch):
3529 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3533 2:1 compression starting at -6dB:
3535 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3539 2:1 compression starting at -9dB:
3541 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3545 2:1 compression starting at -12dB:
3547 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3551 2:1 compression starting at -18dB:
3553 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3557 3:1 compression starting at -15dB:
3559 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3565 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3571 compand=attacks=0:points=-80/-169|-54/-80|-49.5/-64.6|-41.1/-41.1|-25.8/-15|-10.8/-4.5|0/0|20/8.3
3575 Hard limiter at -6dB:
3577 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3581 Hard limiter at -12dB:
3583 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3587 Hard noise gate at -35 dB:
3589 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3595 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3599 @section compensationdelay
3601 Compensation Delay Line is a metric based delay to compensate differing
3602 positions of microphones or speakers.
3604 For example, you have recorded guitar with two microphones placed in
3605 different locations. Because the front of sound wave has fixed speed in
3606 normal conditions, the phasing of microphones can vary and depends on
3607 their location and interposition. The best sound mix can be achieved when
3608 these microphones are in phase (synchronized). Note that a distance of
3609 ~30 cm between microphones makes one microphone capture the signal in
3610 antiphase to the other microphone. That makes the final mix sound moody.
3611 This filter helps to solve phasing problems by adding different delays
3612 to each microphone track and make them synchronized.
3614 The best result can be reached when you take one track as base and
3615 synchronize other tracks one by one with it.
3616 Remember that synchronization/delay tolerance depends on sample rate, too.
3617 Higher sample rates will give more tolerance.
3619 The filter accepts the following parameters:
3623 Set millimeters distance. This is compensation distance for fine tuning.
3627 Set cm distance. This is compensation distance for tightening distance setup.
3631 Set meters distance. This is compensation distance for hard distance setup.
3635 Set dry amount. Amount of unprocessed (dry) signal.
3639 Set wet amount. Amount of processed (wet) signal.
3643 Set temperature in degrees Celsius. This is the temperature of the environment.
3648 Apply headphone crossfeed filter.
3650 Crossfeed is the process of blending the left and right channels of stereo
3652 It is mainly used to reduce extreme stereo separation of low frequencies.
3654 The intent is to produce more speaker like sound to the listener.
3656 The filter accepts the following options:
3660 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3661 This sets gain of low shelf filter for side part of stereo image.
3662 Default is -6dB. Max allowed is -30db when strength is set to 1.
3665 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3666 This sets cut off frequency of low shelf filter. Default is cut off near
3667 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3670 Set curve slope of low shelf filter. Default is 0.5.
3671 Allowed range is from 0.01 to 1.
3674 Set input gain. Default is 0.9.
3677 Set output gain. Default is 1.
3680 @subsection Commands
3682 This filter supports the all above options as @ref{commands}.
3684 @section crystalizer
3685 Simple algorithm for audio noise sharpening.
3687 This filter linearly increases differences betweeen each audio sample.
3689 The filter accepts the following options:
3693 Sets the intensity of effect (default: 2.0). Must be in range between -10.0 to 0
3694 (unchanged sound) to 10.0 (maximum effect).
3695 To inverse filtering use negative value.
3698 Enable clipping. By default is enabled.
3701 @subsection Commands
3703 This filter supports the all above options as @ref{commands}.
3706 Apply a DC shift to the audio.
3708 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3709 in the recording chain) from the audio. The effect of a DC offset is reduced
3710 headroom and hence volume. The @ref{astats} filter can be used to determine if
3711 a signal has a DC offset.
3715 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3719 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3720 used to prevent clipping.
3725 Apply de-essing to the audio samples.
3729 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3733 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3737 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3741 Set the output mode.
3743 It accepts the following values:
3746 Pass input unchanged.
3749 Pass ess filtered out.
3754 Default value is @var{o}.
3760 Measure audio dynamic range.
3762 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3763 is found in transition material. And anything less that 8 have very poor dynamics
3764 and is very compressed.
3766 The filter accepts the following options:
3770 Set window length in seconds used to split audio into segments of equal length.
3771 Default is 3 seconds.
3775 Dynamic Audio Normalizer.
3777 This filter applies a certain amount of gain to the input audio in order
3778 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3779 contrast to more "simple" normalization algorithms, the Dynamic Audio
3780 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3781 This allows for applying extra gain to the "quiet" sections of the audio
3782 while avoiding distortions or clipping the "loud" sections. In other words:
3783 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3784 sections, in the sense that the volume of each section is brought to the
3785 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3786 this goal *without* applying "dynamic range compressing". It will retain 100%
3787 of the dynamic range *within* each section of the audio file.
3791 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3792 Default is 500 milliseconds.
3793 The Dynamic Audio Normalizer processes the input audio in small chunks,
3794 referred to as frames. This is required, because a peak magnitude has no
3795 meaning for just a single sample value. Instead, we need to determine the
3796 peak magnitude for a contiguous sequence of sample values. While a "standard"
3797 normalizer would simply use the peak magnitude of the complete file, the
3798 Dynamic Audio Normalizer determines the peak magnitude individually for each
3799 frame. The length of a frame is specified in milliseconds. By default, the
3800 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3801 been found to give good results with most files.
3802 Note that the exact frame length, in number of samples, will be determined
3803 automatically, based on the sampling rate of the individual input audio file.
3806 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3807 number. Default is 31.
3808 Probably the most important parameter of the Dynamic Audio Normalizer is the
3809 @code{window size} of the Gaussian smoothing filter. The filter's window size
3810 is specified in frames, centered around the current frame. For the sake of
3811 simplicity, this must be an odd number. Consequently, the default value of 31
3812 takes into account the current frame, as well as the 15 preceding frames and
3813 the 15 subsequent frames. Using a larger window results in a stronger
3814 smoothing effect and thus in less gain variation, i.e. slower gain
3815 adaptation. Conversely, using a smaller window results in a weaker smoothing
3816 effect and thus in more gain variation, i.e. faster gain adaptation.
3817 In other words, the more you increase this value, the more the Dynamic Audio
3818 Normalizer will behave like a "traditional" normalization filter. On the
3819 contrary, the more you decrease this value, the more the Dynamic Audio
3820 Normalizer will behave like a dynamic range compressor.
3823 Set the target peak value. This specifies the highest permissible magnitude
3824 level for the normalized audio input. This filter will try to approach the
3825 target peak magnitude as closely as possible, but at the same time it also
3826 makes sure that the normalized signal will never exceed the peak magnitude.
3827 A frame's maximum local gain factor is imposed directly by the target peak
3828 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3829 It is not recommended to go above this value.
3832 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3833 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3834 factor for each input frame, i.e. the maximum gain factor that does not
3835 result in clipping or distortion. The maximum gain factor is determined by
3836 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3837 additionally bounds the frame's maximum gain factor by a predetermined
3838 (global) maximum gain factor. This is done in order to avoid excessive gain
3839 factors in "silent" or almost silent frames. By default, the maximum gain
3840 factor is 10.0, For most inputs the default value should be sufficient and
3841 it usually is not recommended to increase this value. Though, for input
3842 with an extremely low overall volume level, it may be necessary to allow even
3843 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3844 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3845 Instead, a "sigmoid" threshold function will be applied. This way, the
3846 gain factors will smoothly approach the threshold value, but never exceed that
3850 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3851 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3852 This means that the maximum local gain factor for each frame is defined
3853 (only) by the frame's highest magnitude sample. This way, the samples can
3854 be amplified as much as possible without exceeding the maximum signal
3855 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3856 Normalizer can also take into account the frame's root mean square,
3857 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3858 determine the power of a time-varying signal. It is therefore considered
3859 that the RMS is a better approximation of the "perceived loudness" than
3860 just looking at the signal's peak magnitude. Consequently, by adjusting all
3861 frames to a constant RMS value, a uniform "perceived loudness" can be
3862 established. If a target RMS value has been specified, a frame's local gain
3863 factor is defined as the factor that would result in exactly that RMS value.
3864 Note, however, that the maximum local gain factor is still restricted by the
3865 frame's highest magnitude sample, in order to prevent clipping.
3868 Enable channels coupling. By default is enabled.
3869 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3870 amount. This means the same gain factor will be applied to all channels, i.e.
3871 the maximum possible gain factor is determined by the "loudest" channel.
3872 However, in some recordings, it may happen that the volume of the different
3873 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3874 In this case, this option can be used to disable the channel coupling. This way,
3875 the gain factor will be determined independently for each channel, depending
3876 only on the individual channel's highest magnitude sample. This allows for
3877 harmonizing the volume of the different channels.
3880 Enable DC bias correction. By default is disabled.
3881 An audio signal (in the time domain) is a sequence of sample values.
3882 In the Dynamic Audio Normalizer these sample values are represented in the
3883 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3884 audio signal, or "waveform", should be centered around the zero point.
3885 That means if we calculate the mean value of all samples in a file, or in a
3886 single frame, then the result should be 0.0 or at least very close to that
3887 value. If, however, there is a significant deviation of the mean value from
3888 0.0, in either positive or negative direction, this is referred to as a
3889 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3890 Audio Normalizer provides optional DC bias correction.
3891 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3892 the mean value, or "DC correction" offset, of each input frame and subtract
3893 that value from all of the frame's sample values which ensures those samples
3894 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3895 boundaries, the DC correction offset values will be interpolated smoothly
3896 between neighbouring frames.
3898 @item altboundary, b
3899 Enable alternative boundary mode. By default is disabled.
3900 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3901 around each frame. This includes the preceding frames as well as the
3902 subsequent frames. However, for the "boundary" frames, located at the very
3903 beginning and at the very end of the audio file, not all neighbouring
3904 frames are available. In particular, for the first few frames in the audio
3905 file, the preceding frames are not known. And, similarly, for the last few
3906 frames in the audio file, the subsequent frames are not known. Thus, the
3907 question arises which gain factors should be assumed for the missing frames
3908 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3909 to deal with this situation. The default boundary mode assumes a gain factor
3910 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3911 "fade out" at the beginning and at the end of the input, respectively.
3914 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3915 By default, the Dynamic Audio Normalizer does not apply "traditional"
3916 compression. This means that signal peaks will not be pruned and thus the
3917 full dynamic range will be retained within each local neighbourhood. However,
3918 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3919 normalization algorithm with a more "traditional" compression.
3920 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3921 (thresholding) function. If (and only if) the compression feature is enabled,
3922 all input frames will be processed by a soft knee thresholding function prior
3923 to the actual normalization process. Put simply, the thresholding function is
3924 going to prune all samples whose magnitude exceeds a certain threshold value.
3925 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3926 value. Instead, the threshold value will be adjusted for each individual
3928 In general, smaller parameters result in stronger compression, and vice versa.
3929 Values below 3.0 are not recommended, because audible distortion may appear.
3932 Set the target threshold value. This specifies the lowest permissible
3933 magnitude level for the audio input which will be normalized.
3934 If input frame volume is above this value frame will be normalized.
3935 Otherwise frame may not be normalized at all. The default value is set
3936 to 0, which means all input frames will be normalized.
3937 This option is mostly useful if digital noise is not wanted to be amplified.
3940 @subsection Commands
3942 This filter supports the all above options as @ref{commands}.
3946 Make audio easier to listen to on headphones.
3948 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3949 so that when listened to on headphones the stereo image is moved from
3950 inside your head (standard for headphones) to outside and in front of
3951 the listener (standard for speakers).
3957 Apply a two-pole peaking equalisation (EQ) filter. With this
3958 filter, the signal-level at and around a selected frequency can
3959 be increased or decreased, whilst (unlike bandpass and bandreject
3960 filters) that at all other frequencies is unchanged.
3962 In order to produce complex equalisation curves, this filter can
3963 be given several times, each with a different central frequency.
3965 The filter accepts the following options:
3969 Set the filter's central frequency in Hz.
3972 Set method to specify band-width of filter.
3987 Specify the band-width of a filter in width_type units.
3990 Set the required gain or attenuation in dB.
3991 Beware of clipping when using a positive gain.
3994 How much to use filtered signal in output. Default is 1.
3995 Range is between 0 and 1.
3998 Specify which channels to filter, by default all available are filtered.
4001 Normalize biquad coefficients, by default is disabled.
4002 Enabling it will normalize magnitude response at DC to 0dB.
4005 Set transform type of IIR filter.
4014 Set precison of filtering.
4017 Pick automatic sample format depending on surround filters.
4019 Always use signed 16-bit.
4021 Always use signed 32-bit.
4023 Always use float 32-bit.
4025 Always use float 64-bit.
4029 @subsection Examples
4032 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
4034 equalizer=f=1000:t=h:width=200:g=-10
4038 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
4040 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
4044 @subsection Commands
4046 This filter supports the following commands:
4049 Change equalizer frequency.
4050 Syntax for the command is : "@var{frequency}"
4053 Change equalizer width_type.
4054 Syntax for the command is : "@var{width_type}"
4057 Change equalizer width.
4058 Syntax for the command is : "@var{width}"
4061 Change equalizer gain.
4062 Syntax for the command is : "@var{gain}"
4065 Change equalizer mix.
4066 Syntax for the command is : "@var{mix}"
4069 @section extrastereo
4071 Linearly increases the difference between left and right channels which
4072 adds some sort of "live" effect to playback.
4074 The filter accepts the following options:
4078 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
4079 (average of both channels), with 1.0 sound will be unchanged, with
4080 -1.0 left and right channels will be swapped.
4083 Enable clipping. By default is enabled.
4086 @subsection Commands
4088 This filter supports the all above options as @ref{commands}.
4090 @section firequalizer
4091 Apply FIR Equalization using arbitrary frequency response.
4093 The filter accepts the following option:
4097 Set gain curve equation (in dB). The expression can contain variables:
4100 the evaluated frequency
4104 channel number, set to 0 when multichannels evaluation is disabled
4106 channel id, see libavutil/channel_layout.h, set to the first channel id when
4107 multichannels evaluation is disabled
4111 channel_layout, see libavutil/channel_layout.h
4116 @item gain_interpolate(f)
4117 interpolate gain on frequency f based on gain_entry
4118 @item cubic_interpolate(f)
4119 same as gain_interpolate, but smoother
4121 This option is also available as command. Default is @code{gain_interpolate(f)}.
4124 Set gain entry for gain_interpolate function. The expression can
4128 store gain entry at frequency f with value g
4130 This option is also available as command.
4133 Set filter delay in seconds. Higher value means more accurate.
4134 Default is @code{0.01}.
4137 Set filter accuracy in Hz. Lower value means more accurate.
4138 Default is @code{5}.
4141 Set window function. Acceptable values are:
4144 rectangular window, useful when gain curve is already smooth
4146 hann window (default)
4152 3-terms continuous 1st derivative nuttall window
4154 minimum 3-terms discontinuous nuttall window
4156 4-terms continuous 1st derivative nuttall window
4158 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
4160 blackman-harris window
4166 If enabled, use fixed number of audio samples. This improves speed when
4167 filtering with large delay. Default is disabled.
4170 Enable multichannels evaluation on gain. Default is disabled.
4173 Enable zero phase mode by subtracting timestamp to compensate delay.
4174 Default is disabled.
4177 Set scale used by gain. Acceptable values are:
4180 linear frequency, linear gain
4182 linear frequency, logarithmic (in dB) gain (default)
4184 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
4186 logarithmic frequency, logarithmic gain
4190 Set file for dumping, suitable for gnuplot.
4193 Set scale for dumpfile. Acceptable values are same with scale option.
4197 Enable 2-channel convolution using complex FFT. This improves speed significantly.
4198 Default is disabled.
4201 Enable minimum phase impulse response. Default is disabled.
4204 @subsection Examples
4209 firequalizer=gain='if(lt(f,1000), 0, -INF)'
4212 lowpass at 1000 Hz with gain_entry:
4214 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
4217 custom equalization:
4219 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
4222 higher delay with zero phase to compensate delay:
4224 firequalizer=delay=0.1:fixed=on:zero_phase=on
4227 lowpass on left channel, highpass on right channel:
4229 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
4230 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
4235 Apply a flanging effect to the audio.
4237 The filter accepts the following options:
4241 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
4244 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
4247 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
4251 Set percentage of delayed signal mixed with original. Range from 0 to 100.
4252 Default value is 71.
4255 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
4258 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
4259 Default value is @var{sinusoidal}.
4262 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
4263 Default value is 25.
4266 Set delay-line interpolation, @var{linear} or @var{quadratic}.
4267 Default is @var{linear}.
4271 Apply Haas effect to audio.
4273 Note that this makes most sense to apply on mono signals.
4274 With this filter applied to mono signals it give some directionality and
4275 stretches its stereo image.
4277 The filter accepts the following options:
4281 Set input level. By default is @var{1}, or 0dB
4284 Set output level. By default is @var{1}, or 0dB.
4287 Set gain applied to side part of signal. By default is @var{1}.
4290 Set kind of middle source. Can be one of the following:
4300 Pick middle part signal of stereo image.
4303 Pick side part signal of stereo image.
4307 Change middle phase. By default is disabled.
4310 Set left channel delay. By default is @var{2.05} milliseconds.
4313 Set left channel balance. By default is @var{-1}.
4316 Set left channel gain. By default is @var{1}.
4319 Change left phase. By default is disabled.
4322 Set right channel delay. By defaults is @var{2.12} milliseconds.
4325 Set right channel balance. By default is @var{1}.
4328 Set right channel gain. By default is @var{1}.
4331 Change right phase. By default is enabled.
4336 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
4337 embedded HDCD codes is expanded into a 20-bit PCM stream.
4339 The filter supports the Peak Extend and Low-level Gain Adjustment features
4340 of HDCD, and detects the Transient Filter flag.
4343 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
4346 When using the filter with wav, note the default encoding for wav is 16-bit,
4347 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
4348 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
4350 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
4351 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
4354 The filter accepts the following options:
4357 @item disable_autoconvert
4358 Disable any automatic format conversion or resampling in the filter graph.
4360 @item process_stereo
4361 Process the stereo channels together. If target_gain does not match between
4362 channels, consider it invalid and use the last valid target_gain.
4365 Set the code detect timer period in ms.
4368 Always extend peaks above -3dBFS even if PE isn't signaled.
4371 Replace audio with a solid tone and adjust the amplitude to signal some
4372 specific aspect of the decoding process. The output file can be loaded in
4373 an audio editor alongside the original to aid analysis.
4375 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
4382 Gain adjustment level at each sample
4384 Samples where peak extend occurs
4386 Samples where the code detect timer is active
4388 Samples where the target gain does not match between channels
4394 Apply head-related transfer functions (HRTFs) to create virtual
4395 loudspeakers around the user for binaural listening via headphones.
4396 The HRIRs are provided via additional streams, for each channel
4397 one stereo input stream is needed.
4399 The filter accepts the following options:
4403 Set mapping of input streams for convolution.
4404 The argument is a '|'-separated list of channel names in order as they
4405 are given as additional stream inputs for filter.
4406 This also specify number of input streams. Number of input streams
4407 must be not less than number of channels in first stream plus one.
4410 Set gain applied to audio. Value is in dB. Default is 0.
4413 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4414 processing audio in time domain which is slow.
4415 @var{freq} is processing audio in frequency domain which is fast.
4416 Default is @var{freq}.
4419 Set custom gain for LFE channels. Value is in dB. Default is 0.
4422 Set size of frame in number of samples which will be processed at once.
4423 Default value is @var{1024}. Allowed range is from 1024 to 96000.
4426 Set format of hrir stream.
4427 Default value is @var{stereo}. Alternative value is @var{multich}.
4428 If value is set to @var{stereo}, number of additional streams should
4429 be greater or equal to number of input channels in first input stream.
4430 Also each additional stream should have stereo number of channels.
4431 If value is set to @var{multich}, number of additional streams should
4432 be exactly one. Also number of input channels of additional stream
4433 should be equal or greater than twice number of channels of first input
4437 @subsection Examples
4441 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4442 each amovie filter use stereo file with IR coefficients as input.
4443 The files give coefficients for each position of virtual loudspeaker:
4446 -filter_complex "amovie=azi_270_ele_0_DFC.wav[sr];amovie=azi_90_ele_0_DFC.wav[sl];amovie=azi_225_ele_0_DFC.wav[br];amovie=azi_135_ele_0_DFC.wav[bl];amovie=azi_0_ele_0_DFC.wav,asplit[fc][lfe];amovie=azi_35_ele_0_DFC.wav[fl];amovie=azi_325_ele_0_DFC.wav[fr];[0:a][fl][fr][fc][lfe][bl][br][sl][sr]headphone=FL|FR|FC|LFE|BL|BR|SL|SR"
4451 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4452 but now in @var{multich} @var{hrir} format.
4454 ffmpeg -i input.wav -filter_complex "amovie=minp.wav[hrirs];[0:a][hrirs]headphone=map=FL|FR|FC|LFE|BL|BR|SL|SR:hrir=multich"
4461 Apply a high-pass filter with 3dB point frequency.
4462 The filter can be either single-pole, or double-pole (the default).
4463 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4465 The filter accepts the following options:
4469 Set frequency in Hz. Default is 3000.
4472 Set number of poles. Default is 2.
4475 Set method to specify band-width of filter.
4490 Specify the band-width of a filter in width_type units.
4491 Applies only to double-pole filter.
4492 The default is 0.707q and gives a Butterworth response.
4495 How much to use filtered signal in output. Default is 1.
4496 Range is between 0 and 1.
4499 Specify which channels to filter, by default all available are filtered.
4502 Normalize biquad coefficients, by default is disabled.
4503 Enabling it will normalize magnitude response at DC to 0dB.
4506 Set transform type of IIR filter.
4515 Set precison of filtering.
4518 Pick automatic sample format depending on surround filters.
4520 Always use signed 16-bit.
4522 Always use signed 32-bit.
4524 Always use float 32-bit.
4526 Always use float 64-bit.
4530 @subsection Commands
4532 This filter supports the following commands:
4535 Change highpass frequency.
4536 Syntax for the command is : "@var{frequency}"
4539 Change highpass width_type.
4540 Syntax for the command is : "@var{width_type}"
4543 Change highpass width.
4544 Syntax for the command is : "@var{width}"
4547 Change highpass mix.
4548 Syntax for the command is : "@var{mix}"
4553 Join multiple input streams into one multi-channel stream.
4555 It accepts the following parameters:
4559 The number of input streams. It defaults to 2.
4561 @item channel_layout
4562 The desired output channel layout. It defaults to stereo.
4565 Map channels from inputs to output. The argument is a '|'-separated list of
4566 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4567 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4568 can be either the name of the input channel (e.g. FL for front left) or its
4569 index in the specified input stream. @var{out_channel} is the name of the output
4573 The filter will attempt to guess the mappings when they are not specified
4574 explicitly. It does so by first trying to find an unused matching input channel
4575 and if that fails it picks the first unused input channel.
4577 Join 3 inputs (with properly set channel layouts):
4579 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4582 Build a 5.1 output from 6 single-channel streams:
4584 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4585 'join=inputs=6:channel_layout=5.1:map=0.0-FL|1.0-FR|2.0-FC|3.0-SL|4.0-SR|5.0-LFE'
4591 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4593 To enable compilation of this filter you need to configure FFmpeg with
4594 @code{--enable-ladspa}.
4598 Specifies the name of LADSPA plugin library to load. If the environment
4599 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4600 each one of the directories specified by the colon separated list in
4601 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4602 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4603 @file{/usr/lib/ladspa/}.
4606 Specifies the plugin within the library. Some libraries contain only
4607 one plugin, but others contain many of them. If this is not set filter
4608 will list all available plugins within the specified library.
4611 Set the '|' separated list of controls which are zero or more floating point
4612 values that determine the behavior of the loaded plugin (for example delay,
4614 Controls need to be defined using the following syntax:
4615 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4616 @var{valuei} is the value set on the @var{i}-th control.
4617 Alternatively they can be also defined using the following syntax:
4618 @var{value0}|@var{value1}|@var{value2}|..., where
4619 @var{valuei} is the value set on the @var{i}-th control.
4620 If @option{controls} is set to @code{help}, all available controls and
4621 their valid ranges are printed.
4623 @item sample_rate, s
4624 Specify the sample rate, default to 44100. Only used if plugin have
4628 Set the number of samples per channel per each output frame, default
4629 is 1024. Only used if plugin have zero inputs.
4632 Set the minimum duration of the sourced audio. See
4633 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4634 for the accepted syntax.
4635 Note that the resulting duration may be greater than the specified duration,
4636 as the generated audio is always cut at the end of a complete frame.
4637 If not specified, or the expressed duration is negative, the audio is
4638 supposed to be generated forever.
4639 Only used if plugin have zero inputs.
4642 Enable latency compensation, by default is disabled.
4643 Only used if plugin have inputs.
4646 @subsection Examples
4650 List all available plugins within amp (LADSPA example plugin) library:
4656 List all available controls and their valid ranges for @code{vcf_notch}
4657 plugin from @code{VCF} library:
4659 ladspa=f=vcf:p=vcf_notch:c=help
4663 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4666 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4670 Add reverberation to the audio using TAP-plugins
4671 (Tom's Audio Processing plugins):
4673 ladspa=file=tap_reverb:tap_reverb
4677 Generate white noise, with 0.2 amplitude:
4679 ladspa=file=cmt:noise_source_white:c=c0=.2
4683 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4684 @code{C* Audio Plugin Suite} (CAPS) library:
4686 ladspa=file=caps:Click:c=c1=20'
4690 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4692 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4696 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4697 @code{SWH Plugins} collection:
4699 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4703 Attenuate low frequencies using Multiband EQ from Steve Harris
4704 @code{SWH Plugins} collection:
4706 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4710 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4713 ladspa=caps:Narrower
4717 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4719 ladspa=caps:White:.2
4723 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4725 ladspa=caps:Fractal:c=c1=1
4729 Dynamic volume normalization using @code{VLevel} plugin:
4731 ladspa=vlevel-ladspa:vlevel_mono
4735 @subsection Commands
4737 This filter supports the following commands:
4740 Modify the @var{N}-th control value.
4742 If the specified value is not valid, it is ignored and prior one is kept.
4747 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4748 Support for both single pass (livestreams, files) and double pass (files) modes.
4749 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4750 detect true peaks, the audio stream will be upsampled to 192 kHz.
4751 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4753 The filter accepts the following options:
4757 Set integrated loudness target.
4758 Range is -70.0 - -5.0. Default value is -24.0.
4761 Set loudness range target.
4762 Range is 1.0 - 20.0. Default value is 7.0.
4765 Set maximum true peak.
4766 Range is -9.0 - +0.0. Default value is -2.0.
4768 @item measured_I, measured_i
4769 Measured IL of input file.
4770 Range is -99.0 - +0.0.
4772 @item measured_LRA, measured_lra
4773 Measured LRA of input file.
4774 Range is 0.0 - 99.0.
4776 @item measured_TP, measured_tp
4777 Measured true peak of input file.
4778 Range is -99.0 - +99.0.
4780 @item measured_thresh
4781 Measured threshold of input file.
4782 Range is -99.0 - +0.0.
4785 Set offset gain. Gain is applied before the true-peak limiter.
4786 Range is -99.0 - +99.0. Default is +0.0.
4789 Normalize by linearly scaling the source audio.
4790 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4791 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4792 be lower than source LRA and the change in integrated loudness shouldn't
4793 result in a true peak which exceeds the target TP. If any of these
4794 conditions aren't met, normalization mode will revert to @var{dynamic}.
4795 Options are @code{true} or @code{false}. Default is @code{true}.
4798 Treat mono input files as "dual-mono". If a mono file is intended for playback
4799 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4800 If set to @code{true}, this option will compensate for this effect.
4801 Multi-channel input files are not affected by this option.
4802 Options are true or false. Default is false.
4805 Set print format for stats. Options are summary, json, or none.
4806 Default value is none.
4811 Apply a low-pass filter with 3dB point frequency.
4812 The filter can be either single-pole or double-pole (the default).
4813 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4815 The filter accepts the following options:
4819 Set frequency in Hz. Default is 500.
4822 Set number of poles. Default is 2.
4825 Set method to specify band-width of filter.
4840 Specify the band-width of a filter in width_type units.
4841 Applies only to double-pole filter.
4842 The default is 0.707q and gives a Butterworth response.
4845 How much to use filtered signal in output. Default is 1.
4846 Range is between 0 and 1.
4849 Specify which channels to filter, by default all available are filtered.
4852 Normalize biquad coefficients, by default is disabled.
4853 Enabling it will normalize magnitude response at DC to 0dB.
4856 Set transform type of IIR filter.
4865 Set precison of filtering.
4868 Pick automatic sample format depending on surround filters.
4870 Always use signed 16-bit.
4872 Always use signed 32-bit.
4874 Always use float 32-bit.
4876 Always use float 64-bit.
4880 @subsection Examples
4883 Lowpass only LFE channel, it LFE is not present it does nothing:
4889 @subsection Commands
4891 This filter supports the following commands:
4894 Change lowpass frequency.
4895 Syntax for the command is : "@var{frequency}"
4898 Change lowpass width_type.
4899 Syntax for the command is : "@var{width_type}"
4902 Change lowpass width.
4903 Syntax for the command is : "@var{width}"
4907 Syntax for the command is : "@var{mix}"
4912 Load a LV2 (LADSPA Version 2) plugin.
4914 To enable compilation of this filter you need to configure FFmpeg with
4915 @code{--enable-lv2}.
4919 Specifies the plugin URI. You may need to escape ':'.
4922 Set the '|' separated list of controls which are zero or more floating point
4923 values that determine the behavior of the loaded plugin (for example delay,
4925 If @option{controls} is set to @code{help}, all available controls and
4926 their valid ranges are printed.
4928 @item sample_rate, s
4929 Specify the sample rate, default to 44100. Only used if plugin have
4933 Set the number of samples per channel per each output frame, default
4934 is 1024. Only used if plugin have zero inputs.
4937 Set the minimum duration of the sourced audio. See
4938 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4939 for the accepted syntax.
4940 Note that the resulting duration may be greater than the specified duration,
4941 as the generated audio is always cut at the end of a complete frame.
4942 If not specified, or the expressed duration is negative, the audio is
4943 supposed to be generated forever.
4944 Only used if plugin have zero inputs.
4947 @subsection Examples
4951 Apply bass enhancer plugin from Calf:
4953 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4957 Apply vinyl plugin from Calf:
4959 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4963 Apply bit crusher plugin from ArtyFX:
4965 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4970 Multiband Compress or expand the audio's dynamic range.
4972 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4973 This is akin to the crossover of a loudspeaker, and results in flat frequency
4974 response when absent compander action.
4976 It accepts the following parameters:
4980 This option syntax is:
4981 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4982 For explanation of each item refer to compand filter documentation.
4988 Mix channels with specific gain levels. The filter accepts the output
4989 channel layout followed by a set of channels definitions.
4991 This filter is also designed to efficiently remap the channels of an audio
4994 The filter accepts parameters of the form:
4995 "@var{l}|@var{outdef}|@var{outdef}|..."
4999 output channel layout or number of channels
5002 output channel specification, of the form:
5003 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
5006 output channel to define, either a channel name (FL, FR, etc.) or a channel
5007 number (c0, c1, etc.)
5010 multiplicative coefficient for the channel, 1 leaving the volume unchanged
5013 input channel to use, see out_name for details; it is not possible to mix
5014 named and numbered input channels
5017 If the `=' in a channel specification is replaced by `<', then the gains for
5018 that specification will be renormalized so that the total is 1, thus
5019 avoiding clipping noise.
5021 @subsection Mixing examples
5023 For example, if you want to down-mix from stereo to mono, but with a bigger
5024 factor for the left channel:
5026 pan=1c|c0=0.9*c0+0.1*c1
5029 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
5030 7-channels surround:
5032 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
5035 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
5036 that should be preferred (see "-ac" option) unless you have very specific
5039 @subsection Remapping examples
5041 The channel remapping will be effective if, and only if:
5044 @item gain coefficients are zeroes or ones,
5045 @item only one input per channel output,
5048 If all these conditions are satisfied, the filter will notify the user ("Pure
5049 channel mapping detected"), and use an optimized and lossless method to do the
5052 For example, if you have a 5.1 source and want a stereo audio stream by
5053 dropping the extra channels:
5055 pan="stereo| c0=FL | c1=FR"
5058 Given the same source, you can also switch front left and front right channels
5059 and keep the input channel layout:
5061 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
5064 If the input is a stereo audio stream, you can mute the front left channel (and
5065 still keep the stereo channel layout) with:
5070 Still with a stereo audio stream input, you can copy the right channel in both
5071 front left and right:
5073 pan="stereo| c0=FR | c1=FR"
5078 ReplayGain scanner filter. This filter takes an audio stream as an input and
5079 outputs it unchanged.
5080 At end of filtering it displays @code{track_gain} and @code{track_peak}.
5084 Convert the audio sample format, sample rate and channel layout. It is
5085 not meant to be used directly.
5088 Apply time-stretching and pitch-shifting with librubberband.
5090 To enable compilation of this filter, you need to configure FFmpeg with
5091 @code{--enable-librubberband}.
5093 The filter accepts the following options:
5097 Set tempo scale factor.
5100 Set pitch scale factor.
5103 Set transients detector.
5104 Possible values are:
5113 Possible values are:
5122 Possible values are:
5129 Set processing window size.
5130 Possible values are:
5139 Possible values are:
5146 Enable formant preservation when shift pitching.
5147 Possible values are:
5155 Possible values are:
5164 Possible values are:
5171 @subsection Commands
5173 This filter supports the following commands:
5176 Change filter tempo scale factor.
5177 Syntax for the command is : "@var{tempo}"
5180 Change filter pitch scale factor.
5181 Syntax for the command is : "@var{pitch}"
5184 @section sidechaincompress
5186 This filter acts like normal compressor but has the ability to compress
5187 detected signal using second input signal.
5188 It needs two input streams and returns one output stream.
5189 First input stream will be processed depending on second stream signal.
5190 The filtered signal then can be filtered with other filters in later stages of
5191 processing. See @ref{pan} and @ref{amerge} filter.
5193 The filter accepts the following options:
5197 Set input gain. Default is 1. Range is between 0.015625 and 64.
5200 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
5201 Default is @code{downward}.
5204 If a signal of second stream raises above this level it will affect the gain
5205 reduction of first stream.
5206 By default is 0.125. Range is between 0.00097563 and 1.
5209 Set a ratio about which the signal is reduced. 1:2 means that if the level
5210 raised 4dB above the threshold, it will be only 2dB above after the reduction.
5211 Default is 2. Range is between 1 and 20.
5214 Amount of milliseconds the signal has to rise above the threshold before gain
5215 reduction starts. Default is 20. Range is between 0.01 and 2000.
5218 Amount of milliseconds the signal has to fall below the threshold before
5219 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
5222 Set the amount by how much signal will be amplified after processing.
5223 Default is 1. Range is from 1 to 64.
5226 Curve the sharp knee around the threshold to enter gain reduction more softly.
5227 Default is 2.82843. Range is between 1 and 8.
5230 Choose if the @code{average} level between all channels of side-chain stream
5231 or the louder(@code{maximum}) channel of side-chain stream affects the
5232 reduction. Default is @code{average}.
5235 Should the exact signal be taken in case of @code{peak} or an RMS one in case
5236 of @code{rms}. Default is @code{rms} which is mainly smoother.
5239 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
5242 How much to use compressed signal in output. Default is 1.
5243 Range is between 0 and 1.
5246 @subsection Commands
5248 This filter supports the all above options as @ref{commands}.
5250 @subsection Examples
5254 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
5255 depending on the signal of 2nd input and later compressed signal to be
5256 merged with 2nd input:
5258 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
5262 @section sidechaingate
5264 A sidechain gate acts like a normal (wideband) gate but has the ability to
5265 filter the detected signal before sending it to the gain reduction stage.
5266 Normally a gate uses the full range signal to detect a level above the
5268 For example: If you cut all lower frequencies from your sidechain signal
5269 the gate will decrease the volume of your track only if not enough highs
5270 appear. With this technique you are able to reduce the resonation of a
5271 natural drum or remove "rumbling" of muted strokes from a heavily distorted
5273 It needs two input streams and returns one output stream.
5274 First input stream will be processed depending on second stream signal.
5276 The filter accepts the following options:
5280 Set input level before filtering.
5281 Default is 1. Allowed range is from 0.015625 to 64.
5284 Set the mode of operation. Can be @code{upward} or @code{downward}.
5285 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
5286 will be amplified, expanding dynamic range in upward direction.
5287 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
5290 Set the level of gain reduction when the signal is below the threshold.
5291 Default is 0.06125. Allowed range is from 0 to 1.
5292 Setting this to 0 disables reduction and then filter behaves like expander.
5295 If a signal rises above this level the gain reduction is released.
5296 Default is 0.125. Allowed range is from 0 to 1.
5299 Set a ratio about which the signal is reduced.
5300 Default is 2. Allowed range is from 1 to 9000.
5303 Amount of milliseconds the signal has to rise above the threshold before gain
5305 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
5308 Amount of milliseconds the signal has to fall below the threshold before the
5309 reduction is increased again. Default is 250 milliseconds.
5310 Allowed range is from 0.01 to 9000.
5313 Set amount of amplification of signal after processing.
5314 Default is 1. Allowed range is from 1 to 64.
5317 Curve the sharp knee around the threshold to enter gain reduction more softly.
5318 Default is 2.828427125. Allowed range is from 1 to 8.
5321 Choose if exact signal should be taken for detection or an RMS like one.
5322 Default is rms. Can be peak or rms.
5325 Choose if the average level between all channels or the louder channel affects
5327 Default is average. Can be average or maximum.
5330 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
5333 @subsection Commands
5335 This filter supports the all above options as @ref{commands}.
5337 @section silencedetect
5339 Detect silence in an audio stream.
5341 This filter logs a message when it detects that the input audio volume is less
5342 or equal to a noise tolerance value for a duration greater or equal to the
5343 minimum detected noise duration.
5345 The printed times and duration are expressed in seconds. The
5346 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
5347 is set on the first frame whose timestamp equals or exceeds the detection
5348 duration and it contains the timestamp of the first frame of the silence.
5350 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
5351 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
5352 keys are set on the first frame after the silence. If @option{mono} is
5353 enabled, and each channel is evaluated separately, the @code{.X}
5354 suffixed keys are used, and @code{X} corresponds to the channel number.
5356 The filter accepts the following options:
5360 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
5361 specified value) or amplitude ratio. Default is -60dB, or 0.001.
5364 Set silence duration until notification (default is 2 seconds). See
5365 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5366 for the accepted syntax.
5369 Process each channel separately, instead of combined. By default is disabled.
5372 @subsection Examples
5376 Detect 5 seconds of silence with -50dB noise tolerance:
5378 silencedetect=n=-50dB:d=5
5382 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
5383 tolerance in @file{silence.mp3}:
5385 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
5389 @section silenceremove
5391 Remove silence from the beginning, middle or end of the audio.
5393 The filter accepts the following options:
5397 This value is used to indicate if audio should be trimmed at beginning of
5398 the audio. A value of zero indicates no silence should be trimmed from the
5399 beginning. When specifying a non-zero value, it trims audio up until it
5400 finds non-silence. Normally, when trimming silence from beginning of audio
5401 the @var{start_periods} will be @code{1} but it can be increased to higher
5402 values to trim all audio up to specific count of non-silence periods.
5403 Default value is @code{0}.
5405 @item start_duration
5406 Specify the amount of time that non-silence must be detected before it stops
5407 trimming audio. By increasing the duration, bursts of noises can be treated
5408 as silence and trimmed off. Default value is @code{0}.
5410 @item start_threshold
5411 This indicates what sample value should be treated as silence. For digital
5412 audio, a value of @code{0} may be fine but for audio recorded from analog,
5413 you may wish to increase the value to account for background noise.
5414 Can be specified in dB (in case "dB" is appended to the specified value)
5415 or amplitude ratio. Default value is @code{0}.
5418 Specify max duration of silence at beginning that will be kept after
5419 trimming. Default is 0, which is equal to trimming all samples detected
5423 Specify mode of detection of silence end in start of multi-channel audio.
5424 Can be @var{any} or @var{all}. Default is @var{any}.
5425 With @var{any}, any sample that is detected as non-silence will cause
5426 stopped trimming of silence.
5427 With @var{all}, only if all channels are detected as non-silence will cause
5428 stopped trimming of silence.
5431 Set the count for trimming silence from the end of audio.
5432 To remove silence from the middle of a file, specify a @var{stop_periods}
5433 that is negative. This value is then treated as a positive value and is
5434 used to indicate the effect should restart processing as specified by
5435 @var{start_periods}, making it suitable for removing periods of silence
5436 in the middle of the audio.
5437 Default value is @code{0}.
5440 Specify a duration of silence that must exist before audio is not copied any
5441 more. By specifying a higher duration, silence that is wanted can be left in
5443 Default value is @code{0}.
5445 @item stop_threshold
5446 This is the same as @option{start_threshold} but for trimming silence from
5448 Can be specified in dB (in case "dB" is appended to the specified value)
5449 or amplitude ratio. Default value is @code{0}.
5452 Specify max duration of silence at end that will be kept after
5453 trimming. Default is 0, which is equal to trimming all samples detected
5457 Specify mode of detection of silence start in end of multi-channel audio.
5458 Can be @var{any} or @var{all}. Default is @var{any}.
5459 With @var{any}, any sample that is detected as non-silence will cause
5460 stopped trimming of silence.
5461 With @var{all}, only if all channels are detected as non-silence will cause
5462 stopped trimming of silence.
5465 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
5466 and works better with digital silence which is exactly 0.
5467 Default value is @code{rms}.
5470 Set duration in number of seconds used to calculate size of window in number
5471 of samples for detecting silence.
5472 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
5475 @subsection Examples
5479 The following example shows how this filter can be used to start a recording
5480 that does not contain the delay at the start which usually occurs between
5481 pressing the record button and the start of the performance:
5483 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
5487 Trim all silence encountered from beginning to end where there is more than 1
5488 second of silence in audio:
5490 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
5494 Trim all digital silence samples, using peak detection, from beginning to end
5495 where there is more than 0 samples of digital silence in audio and digital
5496 silence is detected in all channels at same positions in stream:
5498 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
5504 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
5505 loudspeakers around the user for binaural listening via headphones (audio
5506 formats up to 9 channels supported).
5507 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
5508 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
5509 Austrian Academy of Sciences.
5511 To enable compilation of this filter you need to configure FFmpeg with
5512 @code{--enable-libmysofa}.
5514 The filter accepts the following options:
5518 Set the SOFA file used for rendering.
5521 Set gain applied to audio. Value is in dB. Default is 0.
5524 Set rotation of virtual loudspeakers in deg. Default is 0.
5527 Set elevation of virtual speakers in deg. Default is 0.
5530 Set distance in meters between loudspeakers and the listener with near-field
5531 HRTFs. Default is 1.
5534 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
5535 processing audio in time domain which is slow.
5536 @var{freq} is processing audio in frequency domain which is fast.
5537 Default is @var{freq}.
5540 Set custom positions of virtual loudspeakers. Syntax for this option is:
5541 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
5542 Each virtual loudspeaker is described with short channel name following with
5543 azimuth and elevation in degrees.
5544 Each virtual loudspeaker description is separated by '|'.
5545 For example to override front left and front right channel positions use:
5546 'speakers=FL 45 15|FR 345 15'.
5547 Descriptions with unrecognised channel names are ignored.
5550 Set custom gain for LFE channels. Value is in dB. Default is 0.
5553 Set custom frame size in number of samples. Default is 1024.
5554 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5555 is set to @var{freq}.
5558 Should all IRs be normalized upon importing SOFA file.
5559 By default is enabled.
5562 Should nearest IRs be interpolated with neighbor IRs if exact position
5563 does not match. By default is disabled.
5566 Minphase all IRs upon loading of SOFA file. By default is disabled.
5569 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5572 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5575 @subsection Examples
5579 Using ClubFritz6 sofa file:
5581 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5585 Using ClubFritz12 sofa file and bigger radius with small rotation:
5587 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5591 Similar as above but with custom speaker positions for front left, front right, back left and back right
5592 and also with custom gain:
5594 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5601 This filter expands or compresses each half-cycle of audio samples
5602 (local set of samples all above or all below zero and between two nearest zero crossings) depending
5603 on threshold value, so audio reaches target peak value under conditions controlled by below options.
5605 The filter accepts the following options:
5609 Set the expansion target peak value. This specifies the highest allowed absolute amplitude
5610 level for the normalized audio input. Default value is 0.95. Allowed range is from 0.0 to 1.0.
5613 Set the maximum expansion factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5614 This option controls maximum local half-cycle of samples expansion. The maximum expansion
5615 would be such that local peak value reaches target peak value but never to surpass it and that
5616 ratio between new and previous peak value does not surpass this option value.
5618 @item compression, c
5619 Set the maximum compression factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5620 This option controls maximum local half-cycle of samples compression. This option is used
5621 only if @option{threshold} option is set to value greater than 0.0, then in such cases
5622 when local peak is lower or same as value set by @option{threshold} all samples belonging to
5623 that peak's half-cycle will be compressed by current compression factor.
5626 Set the threshold value. Default value is 0.0. Allowed range is from 0.0 to 1.0.
5627 This option specifies which half-cycles of samples will be compressed and which will be expanded.
5628 Any half-cycle samples with their local peak value below or same as this option value will be
5629 compressed by current compression factor, otherwise, if greater than threshold value they will be
5630 expanded with expansion factor so that it could reach peak target value but never surpass it.
5633 Set the expansion raising amount per each half-cycle of samples. Default value is 0.001.
5634 Allowed range is from 0.0 to 1.0. This controls how fast expansion factor is raised per
5635 each new half-cycle until it reaches @option{expansion} value.
5636 Setting this options too high may lead to distortions.
5639 Set the compression raising amount per each half-cycle of samples. Default value is 0.001.
5640 Allowed range is from 0.0 to 1.0. This controls how fast compression factor is raised per
5641 each new half-cycle until it reaches @option{compression} value.
5644 Specify which channels to filter, by default all available channels are filtered.
5647 Enable inverted filtering, by default is disabled. This inverts interpretation of @option{threshold}
5648 option. When enabled any half-cycle of samples with their local peak value below or same as
5649 @option{threshold} option will be expanded otherwise it will be compressed.
5652 Link channels when calculating gain applied to each filtered channel sample, by default is disabled.
5653 When disabled each filtered channel gain calculation is independent, otherwise when this option
5654 is enabled the minimum of all possible gains for each filtered channel is used.
5657 @subsection Commands
5659 This filter supports the all above options as @ref{commands}.
5661 @section stereotools
5663 This filter has some handy utilities to manage stereo signals, for converting
5664 M/S stereo recordings to L/R signal while having control over the parameters
5665 or spreading the stereo image of master track.
5667 The filter accepts the following options:
5671 Set input level before filtering for both channels. Defaults is 1.
5672 Allowed range is from 0.015625 to 64.
5675 Set output level after filtering for both channels. Defaults is 1.
5676 Allowed range is from 0.015625 to 64.
5679 Set input balance between both channels. Default is 0.
5680 Allowed range is from -1 to 1.
5683 Set output balance between both channels. Default is 0.
5684 Allowed range is from -1 to 1.
5687 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5688 clipping. Disabled by default.
5691 Mute the left channel. Disabled by default.
5694 Mute the right channel. Disabled by default.
5697 Change the phase of the left channel. Disabled by default.
5700 Change the phase of the right channel. Disabled by default.
5703 Set stereo mode. Available values are:
5707 Left/Right to Left/Right, this is default.
5710 Left/Right to Mid/Side.
5713 Mid/Side to Left/Right.
5716 Left/Right to Left/Left.
5719 Left/Right to Right/Right.
5722 Left/Right to Left + Right.
5725 Left/Right to Right/Left.
5728 Mid/Side to Left/Left.
5731 Mid/Side to Right/Right.
5734 Mid/Side to Right/Left.
5737 Left/Right to Left - Right.
5741 Set level of side signal. Default is 1.
5742 Allowed range is from 0.015625 to 64.
5745 Set balance of side signal. Default is 0.
5746 Allowed range is from -1 to 1.
5749 Set level of the middle signal. Default is 1.
5750 Allowed range is from 0.015625 to 64.
5753 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5756 Set stereo base between mono and inversed channels. Default is 0.
5757 Allowed range is from -1 to 1.
5760 Set delay in milliseconds how much to delay left from right channel and
5761 vice versa. Default is 0. Allowed range is from -20 to 20.
5764 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5767 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5769 @item bmode_in, bmode_out
5770 Set balance mode for balance_in/balance_out option.
5772 Can be one of the following:
5776 Classic balance mode. Attenuate one channel at time.
5777 Gain is raised up to 1.
5780 Similar as classic mode above but gain is raised up to 2.
5783 Equal power distribution, from -6dB to +6dB range.
5787 @subsection Commands
5789 This filter supports the all above options as @ref{commands}.
5791 @subsection Examples
5795 Apply karaoke like effect:
5797 stereotools=mlev=0.015625
5801 Convert M/S signal to L/R:
5803 "stereotools=mode=ms>lr"
5807 @section stereowiden
5809 This filter enhance the stereo effect by suppressing signal common to both
5810 channels and by delaying the signal of left into right and vice versa,
5811 thereby widening the stereo effect.
5813 The filter accepts the following options:
5817 Time in milliseconds of the delay of left signal into right and vice versa.
5818 Default is 20 milliseconds.
5821 Amount of gain in delayed signal into right and vice versa. Gives a delay
5822 effect of left signal in right output and vice versa which gives widening
5823 effect. Default is 0.3.
5826 Cross feed of left into right with inverted phase. This helps in suppressing
5827 the mono. If the value is 1 it will cancel all the signal common to both
5828 channels. Default is 0.3.
5831 Set level of input signal of original channel. Default is 0.8.
5834 @subsection Commands
5836 This filter supports the all above options except @code{delay} as @ref{commands}.
5838 @section superequalizer
5839 Apply 18 band equalizer.
5841 The filter accepts the following options:
5848 Set 131Hz band gain.
5850 Set 185Hz band gain.
5852 Set 262Hz band gain.
5854 Set 370Hz band gain.
5856 Set 523Hz band gain.
5858 Set 740Hz band gain.
5860 Set 1047Hz band gain.
5862 Set 1480Hz band gain.
5864 Set 2093Hz band gain.
5866 Set 2960Hz band gain.
5868 Set 4186Hz band gain.
5870 Set 5920Hz band gain.
5872 Set 8372Hz band gain.
5874 Set 11840Hz band gain.
5876 Set 16744Hz band gain.
5878 Set 20000Hz band gain.
5882 Apply audio surround upmix filter.
5884 This filter allows to produce multichannel output from audio stream.
5886 The filter accepts the following options:
5890 Set output channel layout. By default, this is @var{5.1}.
5892 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5893 for the required syntax.
5896 Set input channel layout. By default, this is @var{stereo}.
5898 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5899 for the required syntax.
5902 Set input volume level. By default, this is @var{1}.
5905 Set output volume level. By default, this is @var{1}.
5908 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5911 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5914 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5917 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5918 In @var{add} mode, LFE channel is created from input audio and added to output.
5919 In @var{sub} mode, LFE channel is created from input audio and added to output but
5920 also all non-LFE output channels are subtracted with output LFE channel.
5923 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5924 Default is @var{90}.
5927 Set front center input volume. By default, this is @var{1}.
5930 Set front center output volume. By default, this is @var{1}.
5933 Set front left input volume. By default, this is @var{1}.
5936 Set front left output volume. By default, this is @var{1}.
5939 Set front right input volume. By default, this is @var{1}.
5942 Set front right output volume. By default, this is @var{1}.
5945 Set side left input volume. By default, this is @var{1}.
5948 Set side left output volume. By default, this is @var{1}.
5951 Set side right input volume. By default, this is @var{1}.
5954 Set side right output volume. By default, this is @var{1}.
5957 Set back left input volume. By default, this is @var{1}.
5960 Set back left output volume. By default, this is @var{1}.
5963 Set back right input volume. By default, this is @var{1}.
5966 Set back right output volume. By default, this is @var{1}.
5969 Set back center input volume. By default, this is @var{1}.
5972 Set back center output volume. By default, this is @var{1}.
5975 Set LFE input volume. By default, this is @var{1}.
5978 Set LFE output volume. By default, this is @var{1}.
5981 Set spread usage of stereo image across X axis for all channels.
5984 Set spread usage of stereo image across Y axis for all channels.
5986 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5987 Set spread usage of stereo image across X axis for each channel.
5989 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5990 Set spread usage of stereo image across Y axis for each channel.
5993 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5996 Set window function.
5998 It accepts the following values:
6021 Default is @code{hann}.
6024 Set window overlap. If set to 1, the recommended overlap for selected
6025 window function will be picked. Default is @code{0.5}.
6028 @section treble, highshelf
6030 Boost or cut treble (upper) frequencies of the audio using a two-pole
6031 shelving filter with a response similar to that of a standard
6032 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
6034 The filter accepts the following options:
6038 Give the gain at whichever is the lower of ~22 kHz and the
6039 Nyquist frequency. Its useful range is about -20 (for a large cut)
6040 to +20 (for a large boost). Beware of clipping when using a positive gain.
6043 Set the filter's central frequency and so can be used
6044 to extend or reduce the frequency range to be boosted or cut.
6045 The default value is @code{3000} Hz.
6048 Set method to specify band-width of filter.
6063 Determine how steep is the filter's shelf transition.
6066 Set number of poles. Default is 2.
6069 How much to use filtered signal in output. Default is 1.
6070 Range is between 0 and 1.
6073 Specify which channels to filter, by default all available are filtered.
6076 Normalize biquad coefficients, by default is disabled.
6077 Enabling it will normalize magnitude response at DC to 0dB.
6080 Set transform type of IIR filter.
6089 Set precison of filtering.
6092 Pick automatic sample format depending on surround filters.
6094 Always use signed 16-bit.
6096 Always use signed 32-bit.
6098 Always use float 32-bit.
6100 Always use float 64-bit.
6104 @subsection Commands
6106 This filter supports the following commands:
6109 Change treble frequency.
6110 Syntax for the command is : "@var{frequency}"
6113 Change treble width_type.
6114 Syntax for the command is : "@var{width_type}"
6117 Change treble width.
6118 Syntax for the command is : "@var{width}"
6122 Syntax for the command is : "@var{gain}"
6126 Syntax for the command is : "@var{mix}"
6131 Sinusoidal amplitude modulation.
6133 The filter accepts the following options:
6137 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
6138 (20 Hz or lower) will result in a tremolo effect.
6139 This filter may also be used as a ring modulator by specifying
6140 a modulation frequency higher than 20 Hz.
6141 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6144 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6145 Default value is 0.5.
6150 Sinusoidal phase modulation.
6152 The filter accepts the following options:
6156 Modulation frequency in Hertz.
6157 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6160 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6161 Default value is 0.5.
6166 Adjust the input audio volume.
6168 It accepts the following parameters:
6172 Set audio volume expression.
6174 Output values are clipped to the maximum value.
6176 The output audio volume is given by the relation:
6178 @var{output_volume} = @var{volume} * @var{input_volume}
6181 The default value for @var{volume} is "1.0".
6184 This parameter represents the mathematical precision.
6186 It determines which input sample formats will be allowed, which affects the
6187 precision of the volume scaling.
6191 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
6193 32-bit floating-point; this limits input sample format to FLT. (default)
6195 64-bit floating-point; this limits input sample format to DBL.
6199 Choose the behaviour on encountering ReplayGain side data in input frames.
6203 Remove ReplayGain side data, ignoring its contents (the default).
6206 Ignore ReplayGain side data, but leave it in the frame.
6209 Prefer the track gain, if present.
6212 Prefer the album gain, if present.
6215 @item replaygain_preamp
6216 Pre-amplification gain in dB to apply to the selected replaygain gain.
6218 Default value for @var{replaygain_preamp} is 0.0.
6220 @item replaygain_noclip
6221 Prevent clipping by limiting the gain applied.
6223 Default value for @var{replaygain_noclip} is 1.
6226 Set when the volume expression is evaluated.
6228 It accepts the following values:
6231 only evaluate expression once during the filter initialization, or
6232 when the @samp{volume} command is sent
6235 evaluate expression for each incoming frame
6238 Default value is @samp{once}.
6241 The volume expression can contain the following parameters.
6245 frame number (starting at zero)
6248 @item nb_consumed_samples
6249 number of samples consumed by the filter
6251 number of samples in the current frame
6253 original frame position in the file
6259 PTS at start of stream
6261 time at start of stream
6267 last set volume value
6270 Note that when @option{eval} is set to @samp{once} only the
6271 @var{sample_rate} and @var{tb} variables are available, all other
6272 variables will evaluate to NAN.
6274 @subsection Commands
6276 This filter supports the following commands:
6279 Modify the volume expression.
6280 The command accepts the same syntax of the corresponding option.
6282 If the specified expression is not valid, it is kept at its current
6286 @subsection Examples
6290 Halve the input audio volume:
6294 volume=volume=-6.0206dB
6297 In all the above example the named key for @option{volume} can be
6298 omitted, for example like in:
6304 Increase input audio power by 6 decibels using fixed-point precision:
6306 volume=volume=6dB:precision=fixed
6310 Fade volume after time 10 with an annihilation period of 5 seconds:
6312 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
6316 @section volumedetect
6318 Detect the volume of the input video.
6320 The filter has no parameters. The input is not modified. Statistics about
6321 the volume will be printed in the log when the input stream end is reached.
6323 In particular it will show the mean volume (root mean square), maximum
6324 volume (on a per-sample basis), and the beginning of a histogram of the
6325 registered volume values (from the maximum value to a cumulated 1/1000 of
6328 All volumes are in decibels relative to the maximum PCM value.
6330 @subsection Examples
6332 Here is an excerpt of the output:
6334 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
6335 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
6336 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
6337 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
6338 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
6339 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
6340 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
6341 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
6342 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
6348 The mean square energy is approximately -27 dB, or 10^-2.7.
6350 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
6352 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
6355 In other words, raising the volume by +4 dB does not cause any clipping,
6356 raising it by +5 dB causes clipping for 6 samples, etc.
6358 @c man end AUDIO FILTERS
6360 @chapter Audio Sources
6361 @c man begin AUDIO SOURCES
6363 Below is a description of the currently available audio sources.
6367 Buffer audio frames, and make them available to the filter chain.
6369 This source is mainly intended for a programmatic use, in particular
6370 through the interface defined in @file{libavfilter/buffersrc.h}.
6372 It accepts the following parameters:
6376 The timebase which will be used for timestamps of submitted frames. It must be
6377 either a floating-point number or in @var{numerator}/@var{denominator} form.
6380 The sample rate of the incoming audio buffers.
6383 The sample format of the incoming audio buffers.
6384 Either a sample format name or its corresponding integer representation from
6385 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
6387 @item channel_layout
6388 The channel layout of the incoming audio buffers.
6389 Either a channel layout name from channel_layout_map in
6390 @file{libavutil/channel_layout.c} or its corresponding integer representation
6391 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
6394 The number of channels of the incoming audio buffers.
6395 If both @var{channels} and @var{channel_layout} are specified, then they
6400 @subsection Examples
6403 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
6406 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
6407 Since the sample format with name "s16p" corresponds to the number
6408 6 and the "stereo" channel layout corresponds to the value 0x3, this is
6411 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
6416 Generate an audio signal specified by an expression.
6418 This source accepts in input one or more expressions (one for each
6419 channel), which are evaluated and used to generate a corresponding
6422 This source accepts the following options:
6426 Set the '|'-separated expressions list for each separate channel. In case the
6427 @option{channel_layout} option is not specified, the selected channel layout
6428 depends on the number of provided expressions. Otherwise the last
6429 specified expression is applied to the remaining output channels.
6431 @item channel_layout, c
6432 Set the channel layout. The number of channels in the specified layout
6433 must be equal to the number of specified expressions.
6436 Set the minimum duration of the sourced audio. See
6437 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6438 for the accepted syntax.
6439 Note that the resulting duration may be greater than the specified
6440 duration, as the generated audio is always cut at the end of a
6443 If not specified, or the expressed duration is negative, the audio is
6444 supposed to be generated forever.
6447 Set the number of samples per channel per each output frame,
6450 @item sample_rate, s
6451 Specify the sample rate, default to 44100.
6454 Each expression in @var{exprs} can contain the following constants:
6458 number of the evaluated sample, starting from 0
6461 time of the evaluated sample expressed in seconds, starting from 0
6468 @subsection Examples
6478 Generate a sin signal with frequency of 440 Hz, set sample rate to
6481 aevalsrc="sin(440*2*PI*t):s=8000"
6485 Generate a two channels signal, specify the channel layout (Front
6486 Center + Back Center) explicitly:
6488 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
6492 Generate white noise:
6494 aevalsrc="-2+random(0)"
6498 Generate an amplitude modulated signal:
6500 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
6504 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
6506 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
6513 Generate a FIR coefficients using frequency sampling method.
6515 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6517 The filter accepts the following options:
6521 Set number of filter coefficents in output audio stream.
6522 Default value is 1025.
6525 Set frequency points from where magnitude and phase are set.
6526 This must be in non decreasing order, and first element must be 0, while last element
6527 must be 1. Elements are separated by white spaces.
6530 Set magnitude value for every frequency point set by @option{frequency}.
6531 Number of values must be same as number of frequency points.
6532 Values are separated by white spaces.
6535 Set phase value for every frequency point set by @option{frequency}.
6536 Number of values must be same as number of frequency points.
6537 Values are separated by white spaces.
6539 @item sample_rate, r
6540 Set sample rate, default is 44100.
6543 Set number of samples per each frame. Default is 1024.
6546 Set window function. Default is blackman.
6551 The null audio source, return unprocessed audio frames. It is mainly useful
6552 as a template and to be employed in analysis / debugging tools, or as
6553 the source for filters which ignore the input data (for example the sox
6556 This source accepts the following options:
6560 @item channel_layout, cl
6562 Specifies the channel layout, and can be either an integer or a string
6563 representing a channel layout. The default value of @var{channel_layout}
6566 Check the channel_layout_map definition in
6567 @file{libavutil/channel_layout.c} for the mapping between strings and
6568 channel layout values.
6570 @item sample_rate, r
6571 Specifies the sample rate, and defaults to 44100.
6574 Set the number of samples per requested frames.
6577 Set the duration of the sourced audio. See
6578 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6579 for the accepted syntax.
6581 If not specified, or the expressed duration is negative, the audio is
6582 supposed to be generated forever.
6585 @subsection Examples
6589 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
6591 anullsrc=r=48000:cl=4
6595 Do the same operation with a more obvious syntax:
6597 anullsrc=r=48000:cl=mono
6601 All the parameters need to be explicitly defined.
6605 Synthesize a voice utterance using the libflite library.
6607 To enable compilation of this filter you need to configure FFmpeg with
6608 @code{--enable-libflite}.
6610 Note that versions of the flite library prior to 2.0 are not thread-safe.
6612 The filter accepts the following options:
6617 If set to 1, list the names of the available voices and exit
6618 immediately. Default value is 0.
6621 Set the maximum number of samples per frame. Default value is 512.
6624 Set the filename containing the text to speak.
6627 Set the text to speak.
6630 Set the voice to use for the speech synthesis. Default value is
6631 @code{kal}. See also the @var{list_voices} option.
6634 @subsection Examples
6638 Read from file @file{speech.txt}, and synthesize the text using the
6639 standard flite voice:
6641 flite=textfile=speech.txt
6645 Read the specified text selecting the @code{slt} voice:
6647 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6651 Input text to ffmpeg:
6653 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6657 Make @file{ffplay} speak the specified text, using @code{flite} and
6658 the @code{lavfi} device:
6660 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
6664 For more information about libflite, check:
6665 @url{http://www.festvox.org/flite/}
6669 Generate a noise audio signal.
6671 The filter accepts the following options:
6674 @item sample_rate, r
6675 Specify the sample rate. Default value is 48000 Hz.
6678 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
6682 Specify the duration of the generated audio stream. Not specifying this option
6683 results in noise with an infinite length.
6685 @item color, colour, c
6686 Specify the color of noise. Available noise colors are white, pink, brown,
6687 blue, violet and velvet. Default color is white.
6690 Specify a value used to seed the PRNG.
6693 Set the number of samples per each output frame, default is 1024.
6696 @subsection Examples
6701 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6703 anoisesrc=d=60:c=pink:r=44100:a=0.5
6709 Generate odd-tap Hilbert transform FIR coefficients.
6711 The resulting stream can be used with @ref{afir} filter for phase-shifting
6712 the signal by 90 degrees.
6714 This is used in many matrix coding schemes and for analytic signal generation.
6715 The process is often written as a multiplication by i (or j), the imaginary unit.
6717 The filter accepts the following options:
6721 @item sample_rate, s
6722 Set sample rate, default is 44100.
6725 Set length of FIR filter, default is 22051.
6728 Set number of samples per each frame.
6731 Set window function to be used when generating FIR coefficients.
6736 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6738 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6740 The filter accepts the following options:
6743 @item sample_rate, r
6744 Set sample rate, default is 44100.
6747 Set number of samples per each frame. Default is 1024.
6750 Set high-pass frequency. Default is 0.
6753 Set low-pass frequency. Default is 0.
6754 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6755 is higher than 0 then filter will create band-pass filter coefficients,
6756 otherwise band-reject filter coefficients.
6759 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6762 Set Kaiser window beta.
6765 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6768 Enable rounding, by default is disabled.
6771 Set number of taps for high-pass filter.
6774 Set number of taps for low-pass filter.
6779 Generate an audio signal made of a sine wave with amplitude 1/8.
6781 The audio signal is bit-exact.
6783 The filter accepts the following options:
6788 Set the carrier frequency. Default is 440 Hz.
6790 @item beep_factor, b
6791 Enable a periodic beep every second with frequency @var{beep_factor} times
6792 the carrier frequency. Default is 0, meaning the beep is disabled.
6794 @item sample_rate, r
6795 Specify the sample rate, default is 44100.
6798 Specify the duration of the generated audio stream.
6800 @item samples_per_frame
6801 Set the number of samples per output frame.
6803 The expression can contain the following constants:
6807 The (sequential) number of the output audio frame, starting from 0.
6810 The PTS (Presentation TimeStamp) of the output audio frame,
6811 expressed in @var{TB} units.
6814 The PTS of the output audio frame, expressed in seconds.
6817 The timebase of the output audio frames.
6820 Default is @code{1024}.
6823 @subsection Examples
6828 Generate a simple 440 Hz sine wave:
6834 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6838 sine=frequency=220:beep_factor=4:duration=5
6842 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6845 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6849 @c man end AUDIO SOURCES
6851 @chapter Audio Sinks
6852 @c man begin AUDIO SINKS
6854 Below is a description of the currently available audio sinks.
6856 @section abuffersink
6858 Buffer audio frames, and make them available to the end of filter chain.
6860 This sink is mainly intended for programmatic use, in particular
6861 through the interface defined in @file{libavfilter/buffersink.h}
6862 or the options system.
6864 It accepts a pointer to an AVABufferSinkContext structure, which
6865 defines the incoming buffers' formats, to be passed as the opaque
6866 parameter to @code{avfilter_init_filter} for initialization.
6869 Null audio sink; do absolutely nothing with the input audio. It is
6870 mainly useful as a template and for use in analysis / debugging
6873 @c man end AUDIO SINKS
6875 @chapter Video Filters
6876 @c man begin VIDEO FILTERS
6878 When you configure your FFmpeg build, you can disable any of the
6879 existing filters using @code{--disable-filters}.
6880 The configure output will show the video filters included in your
6883 Below is a description of the currently available video filters.
6887 Mark a region of interest in a video frame.
6889 The frame data is passed through unchanged, but metadata is attached
6890 to the frame indicating regions of interest which can affect the
6891 behaviour of later encoding. Multiple regions can be marked by
6892 applying the filter multiple times.
6896 Region distance in pixels from the left edge of the frame.
6898 Region distance in pixels from the top edge of the frame.
6900 Region width in pixels.
6902 Region height in pixels.
6904 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6905 and may contain the following variables:
6908 Width of the input frame.
6910 Height of the input frame.
6914 Quantisation offset to apply within the region.
6916 This must be a real value in the range -1 to +1. A value of zero
6917 indicates no quality change. A negative value asks for better quality
6918 (less quantisation), while a positive value asks for worse quality
6919 (greater quantisation).
6921 The range is calibrated so that the extreme values indicate the
6922 largest possible offset - if the rest of the frame is encoded with the
6923 worst possible quality, an offset of -1 indicates that this region
6924 should be encoded with the best possible quality anyway. Intermediate
6925 values are then interpolated in some codec-dependent way.
6927 For example, in 10-bit H.264 the quantisation parameter varies between
6928 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6929 this region should be encoded with a QP around one-tenth of the full
6930 range better than the rest of the frame. So, if most of the frame
6931 were to be encoded with a QP of around 30, this region would get a QP
6932 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6933 An extreme value of -1 would indicate that this region should be
6934 encoded with the best possible quality regardless of the treatment of
6935 the rest of the frame - that is, should be encoded at a QP of -12.
6937 If set to true, remove any existing regions of interest marked on the
6938 frame before adding the new one.
6941 @subsection Examples
6945 Mark the centre quarter of the frame as interesting.
6947 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6950 Mark the 100-pixel-wide region on the left edge of the frame as very
6951 uninteresting (to be encoded at much lower quality than the rest of
6954 addroi=0:0:100:ih:+1/5
6958 @section alphaextract
6960 Extract the alpha component from the input as a grayscale video. This
6961 is especially useful with the @var{alphamerge} filter.
6965 Add or replace the alpha component of the primary input with the
6966 grayscale value of a second input. This is intended for use with
6967 @var{alphaextract} to allow the transmission or storage of frame
6968 sequences that have alpha in a format that doesn't support an alpha
6971 For example, to reconstruct full frames from a normal YUV-encoded video
6972 and a separate video created with @var{alphaextract}, you might use:
6974 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6979 Amplify differences between current pixel and pixels of adjacent frames in
6980 same pixel location.
6982 This filter accepts the following options:
6986 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6987 For example radius of 3 will instruct filter to calculate average of 7 frames.
6990 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6993 Set threshold for difference amplification. Any difference greater or equal to
6994 this value will not alter source pixel. Default is 10.
6995 Allowed range is from 0 to 65535.
6998 Set tolerance for difference amplification. Any difference lower to
6999 this value will not alter source pixel. Default is 0.
7000 Allowed range is from 0 to 65535.
7003 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
7004 This option controls maximum possible value that will decrease source pixel value.
7007 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
7008 This option controls maximum possible value that will increase source pixel value.
7011 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
7014 @subsection Commands
7016 This filter supports the following @ref{commands} that corresponds to option of same name:
7028 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
7029 and libavformat to work. On the other hand, it is limited to ASS (Advanced
7030 Substation Alpha) subtitles files.
7032 This filter accepts the following option in addition to the common options from
7033 the @ref{subtitles} filter:
7037 Set the shaping engine
7039 Available values are:
7042 The default libass shaping engine, which is the best available.
7044 Fast, font-agnostic shaper that can do only substitutions
7046 Slower shaper using OpenType for substitutions and positioning
7049 The default is @code{auto}.
7053 Apply an Adaptive Temporal Averaging Denoiser to the video input.
7055 The filter accepts the following options:
7059 Set threshold A for 1st plane. Default is 0.02.
7060 Valid range is 0 to 0.3.
7063 Set threshold B for 1st plane. Default is 0.04.
7064 Valid range is 0 to 5.
7067 Set threshold A for 2nd plane. Default is 0.02.
7068 Valid range is 0 to 0.3.
7071 Set threshold B for 2nd plane. Default is 0.04.
7072 Valid range is 0 to 5.
7075 Set threshold A for 3rd plane. Default is 0.02.
7076 Valid range is 0 to 0.3.
7079 Set threshold B for 3rd plane. Default is 0.04.
7080 Valid range is 0 to 5.
7082 Threshold A is designed to react on abrupt changes in the input signal and
7083 threshold B is designed to react on continuous changes in the input signal.
7086 Set number of frames filter will use for averaging. Default is 9. Must be odd
7087 number in range [5, 129].
7090 Set what planes of frame filter will use for averaging. Default is all.
7093 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
7094 Alternatively can be set to @code{s} serial.
7096 Parallel can be faster then serial, while other way around is never true.
7097 Parallel will abort early on first change being greater then thresholds, while serial
7098 will continue processing other side of frames if they are equal or below thresholds.
7101 @subsection Commands
7102 This filter supports same @ref{commands} as options except option @code{s}.
7103 The command accepts the same syntax of the corresponding option.
7107 Apply average blur filter.
7109 The filter accepts the following options:
7113 Set horizontal radius size.
7116 Set which planes to filter. By default all planes are filtered.
7119 Set vertical radius size, if zero it will be same as @code{sizeX}.
7120 Default is @code{0}.
7123 @subsection Commands
7124 This filter supports same commands as options.
7125 The command accepts the same syntax of the corresponding option.
7127 If the specified expression is not valid, it is kept at its current
7132 Compute the bounding box for the non-black pixels in the input frame
7135 This filter computes the bounding box containing all the pixels with a
7136 luminance value greater than the minimum allowed value.
7137 The parameters describing the bounding box are printed on the filter
7140 The filter accepts the following option:
7144 Set the minimal luminance value. Default is @code{16}.
7148 Apply bilateral filter, spatial smoothing while preserving edges.
7150 The filter accepts the following options:
7153 Set sigma of gaussian function to calculate spatial weight.
7154 Allowed range is 0 to 512. Default is 0.1.
7157 Set sigma of gaussian function to calculate range weight.
7158 Allowed range is 0 to 1. Default is 0.1.
7161 Set planes to filter. Default is first only.
7164 @subsection Commands
7166 This filter supports the all above options as @ref{commands}.
7168 @section bitplanenoise
7170 Show and measure bit plane noise.
7172 The filter accepts the following options:
7176 Set which plane to analyze. Default is @code{1}.
7179 Filter out noisy pixels from @code{bitplane} set above.
7180 Default is disabled.
7183 @section blackdetect
7185 Detect video intervals that are (almost) completely black. Can be
7186 useful to detect chapter transitions, commercials, or invalid
7189 The filter outputs its detection analysis to both the log as well as
7190 frame metadata. If a black segment of at least the specified minimum
7191 duration is found, a line with the start and end timestamps as well
7192 as duration is printed to the log with level @code{info}. In addition,
7193 a log line with level @code{debug} is printed per frame showing the
7194 black amount detected for that frame.
7196 The filter also attaches metadata to the first frame of a black
7197 segment with key @code{lavfi.black_start} and to the first frame
7198 after the black segment ends with key @code{lavfi.black_end}. The
7199 value is the frame's timestamp. This metadata is added regardless
7200 of the minimum duration specified.
7202 The filter accepts the following options:
7205 @item black_min_duration, d
7206 Set the minimum detected black duration expressed in seconds. It must
7207 be a non-negative floating point number.
7209 Default value is 2.0.
7211 @item picture_black_ratio_th, pic_th
7212 Set the threshold for considering a picture "black".
7213 Express the minimum value for the ratio:
7215 @var{nb_black_pixels} / @var{nb_pixels}
7218 for which a picture is considered black.
7219 Default value is 0.98.
7221 @item pixel_black_th, pix_th
7222 Set the threshold for considering a pixel "black".
7224 The threshold expresses the maximum pixel luminance value for which a
7225 pixel is considered "black". The provided value is scaled according to
7226 the following equation:
7228 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
7231 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
7232 the input video format, the range is [0-255] for YUV full-range
7233 formats and [16-235] for YUV non full-range formats.
7235 Default value is 0.10.
7238 The following example sets the maximum pixel threshold to the minimum
7239 value, and detects only black intervals of 2 or more seconds:
7241 blackdetect=d=2:pix_th=0.00
7246 Detect frames that are (almost) completely black. Can be useful to
7247 detect chapter transitions or commercials. Output lines consist of
7248 the frame number of the detected frame, the percentage of blackness,
7249 the position in the file if known or -1 and the timestamp in seconds.
7251 In order to display the output lines, you need to set the loglevel at
7252 least to the AV_LOG_INFO value.
7254 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
7255 The value represents the percentage of pixels in the picture that
7256 are below the threshold value.
7258 It accepts the following parameters:
7263 The percentage of the pixels that have to be below the threshold; it defaults to
7266 @item threshold, thresh
7267 The threshold below which a pixel value is considered black; it defaults to
7275 Blend two video frames into each other.
7277 The @code{blend} filter takes two input streams and outputs one
7278 stream, the first input is the "top" layer and second input is
7279 "bottom" layer. By default, the output terminates when the longest input terminates.
7281 The @code{tblend} (time blend) filter takes two consecutive frames
7282 from one single stream, and outputs the result obtained by blending
7283 the new frame on top of the old frame.
7285 A description of the accepted options follows.
7293 Set blend mode for specific pixel component or all pixel components in case
7294 of @var{all_mode}. Default value is @code{normal}.
7296 Available values for component modes are:
7338 Set blend opacity for specific pixel component or all pixel components in case
7339 of @var{all_opacity}. Only used in combination with pixel component blend modes.
7346 Set blend expression for specific pixel component or all pixel components in case
7347 of @var{all_expr}. Note that related mode options will be ignored if those are set.
7349 The expressions can use the following variables:
7353 The sequential number of the filtered frame, starting from @code{0}.
7357 the coordinates of the current sample
7361 the width and height of currently filtered plane
7365 Width and height scale for the plane being filtered. It is the
7366 ratio between the dimensions of the current plane to the luma plane,
7367 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
7368 the luma plane and @code{0.5,0.5} for the chroma planes.
7371 Time of the current frame, expressed in seconds.
7374 Value of pixel component at current location for first video frame (top layer).
7377 Value of pixel component at current location for second video frame (bottom layer).
7381 The @code{blend} filter also supports the @ref{framesync} options.
7383 @subsection Examples
7387 Apply transition from bottom layer to top layer in first 10 seconds:
7389 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
7393 Apply linear horizontal transition from top layer to bottom layer:
7395 blend=all_expr='A*(X/W)+B*(1-X/W)'
7399 Apply 1x1 checkerboard effect:
7401 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
7405 Apply uncover left effect:
7407 blend=all_expr='if(gte(N*SW+X,W),A,B)'
7411 Apply uncover down effect:
7413 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
7417 Apply uncover up-left effect:
7419 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
7423 Split diagonally video and shows top and bottom layer on each side:
7425 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
7429 Display differences between the current and the previous frame:
7431 tblend=all_mode=grainextract
7437 Denoise frames using Block-Matching 3D algorithm.
7439 The filter accepts the following options.
7443 Set denoising strength. Default value is 1.
7444 Allowed range is from 0 to 999.9.
7445 The denoising algorithm is very sensitive to sigma, so adjust it
7446 according to the source.
7449 Set local patch size. This sets dimensions in 2D.
7452 Set sliding step for processing blocks. Default value is 4.
7453 Allowed range is from 1 to 64.
7454 Smaller values allows processing more reference blocks and is slower.
7457 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
7458 When set to 1, no block matching is done. Larger values allows more blocks
7460 Allowed range is from 1 to 256.
7463 Set radius for search block matching. Default is 9.
7464 Allowed range is from 1 to INT32_MAX.
7467 Set step between two search locations for block matching. Default is 1.
7468 Allowed range is from 1 to 64. Smaller is slower.
7471 Set threshold of mean square error for block matching. Valid range is 0 to
7475 Set thresholding parameter for hard thresholding in 3D transformed domain.
7476 Larger values results in stronger hard-thresholding filtering in frequency
7480 Set filtering estimation mode. Can be @code{basic} or @code{final}.
7481 Default is @code{basic}.
7484 If enabled, filter will use 2nd stream for block matching.
7485 Default is disabled for @code{basic} value of @var{estim} option,
7486 and always enabled if value of @var{estim} is @code{final}.
7489 Set planes to filter. Default is all available except alpha.
7492 @subsection Examples
7496 Basic filtering with bm3d:
7498 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
7502 Same as above, but filtering only luma:
7504 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
7508 Same as above, but with both estimation modes:
7510 split[a][b],[a]bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic[a],[b][a]bm3d=sigma=3:block=4:bstep=2:group=16:estim=final:ref=1
7514 Same as above, but prefilter with @ref{nlmeans} filter instead:
7516 split[a][b],[a]nlmeans=s=3:r=7:p=3[a],[b][a]bm3d=sigma=3:block=4:bstep=2:group=16:estim=final:ref=1
7522 Apply a boxblur algorithm to the input video.
7524 It accepts the following parameters:
7528 @item luma_radius, lr
7529 @item luma_power, lp
7530 @item chroma_radius, cr
7531 @item chroma_power, cp
7532 @item alpha_radius, ar
7533 @item alpha_power, ap
7537 A description of the accepted options follows.
7540 @item luma_radius, lr
7541 @item chroma_radius, cr
7542 @item alpha_radius, ar
7543 Set an expression for the box radius in pixels used for blurring the
7544 corresponding input plane.
7546 The radius value must be a non-negative number, and must not be
7547 greater than the value of the expression @code{min(w,h)/2} for the
7548 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
7551 Default value for @option{luma_radius} is "2". If not specified,
7552 @option{chroma_radius} and @option{alpha_radius} default to the
7553 corresponding value set for @option{luma_radius}.
7555 The expressions can contain the following constants:
7559 The input width and height in pixels.
7563 The input chroma image width and height in pixels.
7567 The horizontal and vertical chroma subsample values. For example, for the
7568 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
7571 @item luma_power, lp
7572 @item chroma_power, cp
7573 @item alpha_power, ap
7574 Specify how many times the boxblur filter is applied to the
7575 corresponding plane.
7577 Default value for @option{luma_power} is 2. If not specified,
7578 @option{chroma_power} and @option{alpha_power} default to the
7579 corresponding value set for @option{luma_power}.
7581 A value of 0 will disable the effect.
7584 @subsection Examples
7588 Apply a boxblur filter with the luma, chroma, and alpha radii
7591 boxblur=luma_radius=2:luma_power=1
7596 Set the luma radius to 2, and alpha and chroma radius to 0:
7598 boxblur=2:1:cr=0:ar=0
7602 Set the luma and chroma radii to a fraction of the video dimension:
7604 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
7610 Deinterlace the input video ("bwdif" stands for "Bob Weaver
7611 Deinterlacing Filter").
7613 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
7614 interpolation algorithms.
7615 It accepts the following parameters:
7619 The interlacing mode to adopt. It accepts one of the following values:
7623 Output one frame for each frame.
7625 Output one frame for each field.
7628 The default value is @code{send_field}.
7631 The picture field parity assumed for the input interlaced video. It accepts one
7632 of the following values:
7636 Assume the top field is first.
7638 Assume the bottom field is first.
7640 Enable automatic detection of field parity.
7643 The default value is @code{auto}.
7644 If the interlacing is unknown or the decoder does not export this information,
7645 top field first will be assumed.
7648 Specify which frames to deinterlace. Accepts one of the following
7653 Deinterlace all frames.
7655 Only deinterlace frames marked as interlaced.
7658 The default value is @code{all}.
7663 Apply Contrast Adaptive Sharpen filter to video stream.
7665 The filter accepts the following options:
7669 Set the sharpening strength. Default value is 0.
7672 Set planes to filter. Default value is to filter all
7673 planes except alpha plane.
7676 @subsection Commands
7677 This filter supports same @ref{commands} as options.
7680 Remove all color information for all colors except for certain one.
7682 The filter accepts the following options:
7686 The color which will not be replaced with neutral chroma.
7689 Similarity percentage with the above color.
7690 0.01 matches only the exact key color, while 1.0 matches everything.
7694 0.0 makes pixels either fully gray, or not gray at all.
7695 Higher values result in more preserved color.
7698 Signals that the color passed is already in YUV instead of RGB.
7700 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7701 This can be used to pass exact YUV values as hexadecimal numbers.
7704 @subsection Commands
7705 This filter supports same @ref{commands} as options.
7706 The command accepts the same syntax of the corresponding option.
7708 If the specified expression is not valid, it is kept at its current
7712 YUV colorspace color/chroma keying.
7714 The filter accepts the following options:
7718 The color which will be replaced with transparency.
7721 Similarity percentage with the key color.
7723 0.01 matches only the exact key color, while 1.0 matches everything.
7728 0.0 makes pixels either fully transparent, or not transparent at all.
7730 Higher values result in semi-transparent pixels, with a higher transparency
7731 the more similar the pixels color is to the key color.
7734 Signals that the color passed is already in YUV instead of RGB.
7736 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7737 This can be used to pass exact YUV values as hexadecimal numbers.
7740 @subsection Commands
7741 This filter supports same @ref{commands} as options.
7742 The command accepts the same syntax of the corresponding option.
7744 If the specified expression is not valid, it is kept at its current
7747 @subsection Examples
7751 Make every green pixel in the input image transparent:
7753 ffmpeg -i input.png -vf chromakey=green out.png
7757 Overlay a greenscreen-video on top of a static black background.
7759 ffmpeg -f lavfi -i color=c=black:s=1280x720 -i video.mp4 -shortest -filter_complex "[1:v]chromakey=0x70de77:0.1:0.2[ckout];[0:v][ckout]overlay[out]" -map "[out]" output.mkv
7764 Reduce chrominance noise.
7766 The filter accepts the following options:
7770 Set threshold for averaging chrominance values.
7771 Sum of absolute difference of Y, U and V pixel components of current
7772 pixel and neighbour pixels lower than this threshold will be used in
7773 averaging. Luma component is left unchanged and is copied to output.
7774 Default value is 30. Allowed range is from 1 to 200.
7777 Set horizontal radius of rectangle used for averaging.
7778 Allowed range is from 1 to 100. Default value is 5.
7781 Set vertical radius of rectangle used for averaging.
7782 Allowed range is from 1 to 100. Default value is 5.
7785 Set horizontal step when averaging. Default value is 1.
7786 Allowed range is from 1 to 50.
7787 Mostly useful to speed-up filtering.
7790 Set vertical step when averaging. Default value is 1.
7791 Allowed range is from 1 to 50.
7792 Mostly useful to speed-up filtering.
7795 Set Y threshold for averaging chrominance values.
7796 Set finer control for max allowed difference between Y components
7797 of current pixel and neigbour pixels.
7798 Default value is 200. Allowed range is from 1 to 200.
7801 Set U threshold for averaging chrominance values.
7802 Set finer control for max allowed difference between U components
7803 of current pixel and neigbour pixels.
7804 Default value is 200. Allowed range is from 1 to 200.
7807 Set V threshold for averaging chrominance values.
7808 Set finer control for max allowed difference between V components
7809 of current pixel and neigbour pixels.
7810 Default value is 200. Allowed range is from 1 to 200.
7813 @subsection Commands
7814 This filter supports same @ref{commands} as options.
7815 The command accepts the same syntax of the corresponding option.
7817 @section chromashift
7818 Shift chroma pixels horizontally and/or vertically.
7820 The filter accepts the following options:
7823 Set amount to shift chroma-blue horizontally.
7825 Set amount to shift chroma-blue vertically.
7827 Set amount to shift chroma-red horizontally.
7829 Set amount to shift chroma-red vertically.
7831 Set edge mode, can be @var{smear}, default, or @var{warp}.
7834 @subsection Commands
7836 This filter supports the all above options as @ref{commands}.
7840 Display CIE color diagram with pixels overlaid onto it.
7842 The filter accepts the following options:
7857 @item uhdtv, rec2020
7871 Set what gamuts to draw.
7873 See @code{system} option for available values.
7876 Set ciescope size, by default set to 512.
7879 Set intensity used to map input pixel values to CIE diagram.
7882 Set contrast used to draw tongue colors that are out of active color system gamut.
7885 Correct gamma displayed on scope, by default enabled.
7888 Show white point on CIE diagram, by default disabled.
7891 Set input gamma. Used only with XYZ input color space.
7896 Visualize information exported by some codecs.
7898 Some codecs can export information through frames using side-data or other
7899 means. For example, some MPEG based codecs export motion vectors through the
7900 @var{export_mvs} flag in the codec @option{flags2} option.
7902 The filter accepts the following option:
7906 Set motion vectors to visualize.
7908 Available flags for @var{mv} are:
7912 forward predicted MVs of P-frames
7914 forward predicted MVs of B-frames
7916 backward predicted MVs of B-frames
7920 Display quantization parameters using the chroma planes.
7923 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7925 Available flags for @var{mv_type} are:
7929 forward predicted MVs
7931 backward predicted MVs
7934 @item frame_type, ft
7935 Set frame type to visualize motion vectors of.
7937 Available flags for @var{frame_type} are:
7941 intra-coded frames (I-frames)
7943 predicted frames (P-frames)
7945 bi-directionally predicted frames (B-frames)
7949 @subsection Examples
7953 Visualize forward predicted MVs of all frames using @command{ffplay}:
7955 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7959 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7961 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7965 @section colorbalance
7966 Modify intensity of primary colors (red, green and blue) of input frames.
7968 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7969 regions for the red-cyan, green-magenta or blue-yellow balance.
7971 A positive adjustment value shifts the balance towards the primary color, a negative
7972 value towards the complementary color.
7974 The filter accepts the following options:
7980 Adjust red, green and blue shadows (darkest pixels).
7985 Adjust red, green and blue midtones (medium pixels).
7990 Adjust red, green and blue highlights (brightest pixels).
7992 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7995 Preserve lightness when changing color balance. Default is disabled.
7998 @subsection Examples
8002 Add red color cast to shadows:
8008 @subsection Commands
8010 This filter supports the all above options as @ref{commands}.
8012 @section colorchannelmixer
8014 Adjust video input frames by re-mixing color channels.
8016 This filter modifies a color channel by adding the values associated to
8017 the other channels of the same pixels. For example if the value to
8018 modify is red, the output value will be:
8020 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
8023 The filter accepts the following options:
8030 Adjust contribution of input red, green, blue and alpha channels for output red channel.
8031 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
8037 Adjust contribution of input red, green, blue and alpha channels for output green channel.
8038 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
8044 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
8045 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
8051 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
8052 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
8054 Allowed ranges for options are @code{[-2.0, 2.0]}.
8057 @subsection Examples
8061 Convert source to grayscale:
8063 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
8066 Simulate sepia tones:
8068 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
8072 @subsection Commands
8074 This filter supports the all above options as @ref{commands}.
8077 RGB colorspace color keying.
8079 The filter accepts the following options:
8083 The color which will be replaced with transparency.
8086 Similarity percentage with the key color.
8088 0.01 matches only the exact key color, while 1.0 matches everything.
8093 0.0 makes pixels either fully transparent, or not transparent at all.
8095 Higher values result in semi-transparent pixels, with a higher transparency
8096 the more similar the pixels color is to the key color.
8099 @subsection Examples
8103 Make every green pixel in the input image transparent:
8105 ffmpeg -i input.png -vf colorkey=green out.png
8109 Overlay a greenscreen-video on top of a static background image.
8111 ffmpeg -i background.png -i video.mp4 -filter_complex "[1:v]colorkey=0x3BBD1E:0.3:0.2[ckout];[0:v][ckout]overlay[out]" -map "[out]" output.flv
8115 @subsection Commands
8116 This filter supports same @ref{commands} as options.
8117 The command accepts the same syntax of the corresponding option.
8119 If the specified expression is not valid, it is kept at its current
8123 Remove all color information for all RGB colors except for certain one.
8125 The filter accepts the following options:
8129 The color which will not be replaced with neutral gray.
8132 Similarity percentage with the above color.
8133 0.01 matches only the exact key color, while 1.0 matches everything.
8136 Blend percentage. 0.0 makes pixels fully gray.
8137 Higher values result in more preserved color.
8140 @subsection Commands
8141 This filter supports same @ref{commands} as options.
8142 The command accepts the same syntax of the corresponding option.
8144 If the specified expression is not valid, it is kept at its current
8147 @section colorlevels
8149 Adjust video input frames using levels.
8151 The filter accepts the following options:
8158 Adjust red, green, blue and alpha input black point.
8159 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
8165 Adjust red, green, blue and alpha input white point.
8166 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
8168 Input levels are used to lighten highlights (bright tones), darken shadows
8169 (dark tones), change the balance of bright and dark tones.
8175 Adjust red, green, blue and alpha output black point.
8176 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
8182 Adjust red, green, blue and alpha output white point.
8183 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
8185 Output levels allows manual selection of a constrained output level range.
8188 @subsection Examples
8192 Make video output darker:
8194 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
8200 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
8204 Make video output lighter:
8206 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
8210 Increase brightness:
8212 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
8216 @subsection Commands
8218 This filter supports the all above options as @ref{commands}.
8220 @section colormatrix
8222 Convert color matrix.
8224 The filter accepts the following options:
8229 Specify the source and destination color matrix. Both values must be
8232 The accepted values are:
8260 For example to convert from BT.601 to SMPTE-240M, use the command:
8262 colormatrix=bt601:smpte240m
8267 Convert colorspace, transfer characteristics or color primaries.
8268 Input video needs to have an even size.
8270 The filter accepts the following options:
8275 Specify all color properties at once.
8277 The accepted values are:
8307 Specify output colorspace.
8309 The accepted values are:
8318 BT.470BG or BT.601-6 625
8321 SMPTE-170M or BT.601-6 525
8330 BT.2020 with non-constant luminance
8336 Specify output transfer characteristics.
8338 The accepted values are:
8350 Constant gamma of 2.2
8353 Constant gamma of 2.8
8356 SMPTE-170M, BT.601-6 625 or BT.601-6 525
8374 BT.2020 for 10-bits content
8377 BT.2020 for 12-bits content
8383 Specify output color primaries.
8385 The accepted values are:
8394 BT.470BG or BT.601-6 625
8397 SMPTE-170M or BT.601-6 525
8421 Specify output color range.
8423 The accepted values are:
8426 TV (restricted) range
8429 MPEG (restricted) range
8440 Specify output color format.
8442 The accepted values are:
8445 YUV 4:2:0 planar 8-bits
8448 YUV 4:2:0 planar 10-bits
8451 YUV 4:2:0 planar 12-bits
8454 YUV 4:2:2 planar 8-bits
8457 YUV 4:2:2 planar 10-bits
8460 YUV 4:2:2 planar 12-bits
8463 YUV 4:4:4 planar 8-bits
8466 YUV 4:4:4 planar 10-bits
8469 YUV 4:4:4 planar 12-bits
8474 Do a fast conversion, which skips gamma/primary correction. This will take
8475 significantly less CPU, but will be mathematically incorrect. To get output
8476 compatible with that produced by the colormatrix filter, use fast=1.
8479 Specify dithering mode.
8481 The accepted values are:
8487 Floyd-Steinberg dithering
8491 Whitepoint adaptation mode.
8493 The accepted values are:
8496 Bradford whitepoint adaptation
8499 von Kries whitepoint adaptation
8502 identity whitepoint adaptation (i.e. no whitepoint adaptation)
8506 Override all input properties at once. Same accepted values as @ref{all}.
8509 Override input colorspace. Same accepted values as @ref{space}.
8512 Override input color primaries. Same accepted values as @ref{primaries}.
8515 Override input transfer characteristics. Same accepted values as @ref{trc}.
8518 Override input color range. Same accepted values as @ref{range}.
8522 The filter converts the transfer characteristics, color space and color
8523 primaries to the specified user values. The output value, if not specified,
8524 is set to a default value based on the "all" property. If that property is
8525 also not specified, the filter will log an error. The output color range and
8526 format default to the same value as the input color range and format. The
8527 input transfer characteristics, color space, color primaries and color range
8528 should be set on the input data. If any of these are missing, the filter will
8529 log an error and no conversion will take place.
8531 For example to convert the input to SMPTE-240M, use the command:
8533 colorspace=smpte240m
8536 @section convolution
8538 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
8540 The filter accepts the following options:
8547 Set matrix for each plane.
8548 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
8549 and from 1 to 49 odd number of signed integers in @var{row} mode.
8555 Set multiplier for calculated value for each plane.
8556 If unset or 0, it will be sum of all matrix elements.
8562 Set bias for each plane. This value is added to the result of the multiplication.
8563 Useful for making the overall image brighter or darker. Default is 0.0.
8569 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
8570 Default is @var{square}.
8573 @subsection Commands
8575 This filter supports the all above options as @ref{commands}.
8577 @subsection Examples
8583 convolution="0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0"
8589 convolution="1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1/9:1/9:1/9:1/9"
8595 convolution="0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:5:1:1:1:0:128:128:128"
8601 convolution="0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:5:5:5:1:0:128:128:128"
8605 Apply laplacian edge detector which includes diagonals:
8607 convolution="1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:5:5:5:1:0:128:128:0"
8613 convolution="-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2"
8619 Apply 2D convolution of video stream in frequency domain using second stream
8622 The filter accepts the following options:
8626 Set which planes to process.
8629 Set which impulse video frames will be processed, can be @var{first}
8630 or @var{all}. Default is @var{all}.
8633 The @code{convolve} filter also supports the @ref{framesync} options.
8637 Copy the input video source unchanged to the output. This is mainly useful for
8642 Video filtering on GPU using Apple's CoreImage API on OSX.
8644 Hardware acceleration is based on an OpenGL context. Usually, this means it is
8645 processed by video hardware. However, software-based OpenGL implementations
8646 exist which means there is no guarantee for hardware processing. It depends on
8649 There are many filters and image generators provided by Apple that come with a
8650 large variety of options. The filter has to be referenced by its name along
8653 The coreimage filter accepts the following options:
8656 List all available filters and generators along with all their respective
8657 options as well as possible minimum and maximum values along with the default
8664 Specify all filters by their respective name and options.
8665 Use @var{list_filters} to determine all valid filter names and options.
8666 Numerical options are specified by a float value and are automatically clamped
8667 to their respective value range. Vector and color options have to be specified
8668 by a list of space separated float values. Character escaping has to be done.
8669 A special option name @code{default} is available to use default options for a
8672 It is required to specify either @code{default} or at least one of the filter options.
8673 All omitted options are used with their default values.
8674 The syntax of the filter string is as follows:
8676 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
8680 Specify a rectangle where the output of the filter chain is copied into the
8681 input image. It is given by a list of space separated float values:
8683 output_rect=x\ y\ width\ height
8685 If not given, the output rectangle equals the dimensions of the input image.
8686 The output rectangle is automatically cropped at the borders of the input
8687 image. Negative values are valid for each component.
8689 output_rect=25\ 25\ 100\ 100
8693 Several filters can be chained for successive processing without GPU-HOST
8694 transfers allowing for fast processing of complex filter chains.
8695 Currently, only filters with zero (generators) or exactly one (filters) input
8696 image and one output image are supported. Also, transition filters are not yet
8699 Some filters generate output images with additional padding depending on the
8700 respective filter kernel. The padding is automatically removed to ensure the
8701 filter output has the same size as the input image.
8703 For image generators, the size of the output image is determined by the
8704 previous output image of the filter chain or the input image of the whole
8705 filterchain, respectively. The generators do not use the pixel information of
8706 this image to generate their output. However, the generated output is
8707 blended onto this image, resulting in partial or complete coverage of the
8710 The @ref{coreimagesrc} video source can be used for generating input images
8711 which are directly fed into the filter chain. By using it, providing input
8712 images by another video source or an input video is not required.
8714 @subsection Examples
8719 List all filters available:
8721 coreimage=list_filters=true
8725 Use the CIBoxBlur filter with default options to blur an image:
8727 coreimage=filter=CIBoxBlur@@default
8731 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8732 its center at 100x100 and a radius of 50 pixels:
8734 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8738 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8739 given as complete and escaped command-line for Apple's standard bash shell:
8741 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8747 Cover a rectangular object
8749 It accepts the following options:
8753 Filepath of the optional cover image, needs to be in yuv420.
8758 It accepts the following values:
8761 cover it by the supplied image
8763 cover it by interpolating the surrounding pixels
8766 Default value is @var{blur}.
8769 @subsection Examples
8773 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8775 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8781 Crop the input video to given dimensions.
8783 It accepts the following parameters:
8787 The width of the output video. It defaults to @code{iw}.
8788 This expression is evaluated only once during the filter
8789 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8792 The height of the output video. It defaults to @code{ih}.
8793 This expression is evaluated only once during the filter
8794 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8797 The horizontal position, in the input video, of the left edge of the output
8798 video. It defaults to @code{(in_w-out_w)/2}.
8799 This expression is evaluated per-frame.
8802 The vertical position, in the input video, of the top edge of the output video.
8803 It defaults to @code{(in_h-out_h)/2}.
8804 This expression is evaluated per-frame.
8807 If set to 1 will force the output display aspect ratio
8808 to be the same of the input, by changing the output sample aspect
8809 ratio. It defaults to 0.
8812 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8813 width/height/x/y as specified and will not be rounded to nearest smaller value.
8817 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8818 expressions containing the following constants:
8823 The computed values for @var{x} and @var{y}. They are evaluated for
8828 The input width and height.
8832 These are the same as @var{in_w} and @var{in_h}.
8836 The output (cropped) width and height.
8840 These are the same as @var{out_w} and @var{out_h}.
8843 same as @var{iw} / @var{ih}
8846 input sample aspect ratio
8849 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8853 horizontal and vertical chroma subsample values. For example for the
8854 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8857 The number of the input frame, starting from 0.
8860 the position in the file of the input frame, NAN if unknown
8863 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8867 The expression for @var{out_w} may depend on the value of @var{out_h},
8868 and the expression for @var{out_h} may depend on @var{out_w}, but they
8869 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8870 evaluated after @var{out_w} and @var{out_h}.
8872 The @var{x} and @var{y} parameters specify the expressions for the
8873 position of the top-left corner of the output (non-cropped) area. They
8874 are evaluated for each frame. If the evaluated value is not valid, it
8875 is approximated to the nearest valid value.
8877 The expression for @var{x} may depend on @var{y}, and the expression
8878 for @var{y} may depend on @var{x}.
8880 @subsection Examples
8884 Crop area with size 100x100 at position (12,34).
8889 Using named options, the example above becomes:
8891 crop=w=100:h=100:x=12:y=34
8895 Crop the central input area with size 100x100:
8901 Crop the central input area with size 2/3 of the input video:
8903 crop=2/3*in_w:2/3*in_h
8907 Crop the input video central square:
8914 Delimit the rectangle with the top-left corner placed at position
8915 100:100 and the right-bottom corner corresponding to the right-bottom
8916 corner of the input image.
8918 crop=in_w-100:in_h-100:100:100
8922 Crop 10 pixels from the left and right borders, and 20 pixels from
8923 the top and bottom borders
8925 crop=in_w-2*10:in_h-2*20
8929 Keep only the bottom right quarter of the input image:
8931 crop=in_w/2:in_h/2:in_w/2:in_h/2
8935 Crop height for getting Greek harmony:
8937 crop=in_w:1/PHI*in_w
8941 Apply trembling effect:
8943 crop=in_w/2:in_h/2:(in_w-out_w)/2+((in_w-out_w)/2)*sin(n/10):(in_h-out_h)/2 +((in_h-out_h)/2)*sin(n/7)
8947 Apply erratic camera effect depending on timestamp:
8949 crop=in_w/2:in_h/2:(in_w-out_w)/2+((in_w-out_w)/2)*sin(t*10):(in_h-out_h)/2 +((in_h-out_h)/2)*sin(t*13)"
8953 Set x depending on the value of y:
8955 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8959 @subsection Commands
8961 This filter supports the following commands:
8967 Set width/height of the output video and the horizontal/vertical position
8969 The command accepts the same syntax of the corresponding option.
8971 If the specified expression is not valid, it is kept at its current
8977 Auto-detect the crop size.
8979 It calculates the necessary cropping parameters and prints the
8980 recommended parameters via the logging system. The detected dimensions
8981 correspond to the non-black area of the input video.
8983 It accepts the following parameters:
8988 Set higher black value threshold, which can be optionally specified
8989 from nothing (0) to everything (255 for 8-bit based formats). An intensity
8990 value greater to the set value is considered non-black. It defaults to 24.
8991 You can also specify a value between 0.0 and 1.0 which will be scaled depending
8992 on the bitdepth of the pixel format.
8995 The value which the width/height should be divisible by. It defaults to
8996 16. The offset is automatically adjusted to center the video. Use 2 to
8997 get only even dimensions (needed for 4:2:2 video). 16 is best when
8998 encoding to most video codecs.
9001 Set the number of initial frames for which evaluation is skipped.
9002 Default is 2. Range is 0 to INT_MAX.
9004 @item reset_count, reset
9005 Set the counter that determines after how many frames cropdetect will
9006 reset the previously detected largest video area and start over to
9007 detect the current optimal crop area. Default value is 0.
9009 This can be useful when channel logos distort the video area. 0
9010 indicates 'never reset', and returns the largest area encountered during
9017 Delay video filtering until a given wallclock timestamp. The filter first
9018 passes on @option{preroll} amount of frames, then it buffers at most
9019 @option{buffer} amount of frames and waits for the cue. After reaching the cue
9020 it forwards the buffered frames and also any subsequent frames coming in its
9023 The filter can be used synchronize the output of multiple ffmpeg processes for
9024 realtime output devices like decklink. By putting the delay in the filtering
9025 chain and pre-buffering frames the process can pass on data to output almost
9026 immediately after the target wallclock timestamp is reached.
9028 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
9034 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
9037 The duration of content to pass on as preroll expressed in seconds. Default is 0.
9040 The maximum duration of content to buffer before waiting for the cue expressed
9041 in seconds. Default is 0.
9048 Apply color adjustments using curves.
9050 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
9051 component (red, green and blue) has its values defined by @var{N} key points
9052 tied from each other using a smooth curve. The x-axis represents the pixel
9053 values from the input frame, and the y-axis the new pixel values to be set for
9056 By default, a component curve is defined by the two points @var{(0;0)} and
9057 @var{(1;1)}. This creates a straight line where each original pixel value is
9058 "adjusted" to its own value, which means no change to the image.
9060 The filter allows you to redefine these two points and add some more. A new
9061 curve (using a natural cubic spline interpolation) will be define to pass
9062 smoothly through all these new coordinates. The new defined points needs to be
9063 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
9064 be in the @var{[0;1]} interval. If the computed curves happened to go outside
9065 the vector spaces, the values will be clipped accordingly.
9067 The filter accepts the following options:
9071 Select one of the available color presets. This option can be used in addition
9072 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
9073 options takes priority on the preset values.
9074 Available presets are:
9077 @item color_negative
9080 @item increase_contrast
9082 @item linear_contrast
9083 @item medium_contrast
9085 @item strong_contrast
9088 Default is @code{none}.
9090 Set the master key points. These points will define a second pass mapping. It
9091 is sometimes called a "luminance" or "value" mapping. It can be used with
9092 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
9093 post-processing LUT.
9095 Set the key points for the red component.
9097 Set the key points for the green component.
9099 Set the key points for the blue component.
9101 Set the key points for all components (not including master).
9102 Can be used in addition to the other key points component
9103 options. In this case, the unset component(s) will fallback on this
9104 @option{all} setting.
9106 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
9108 Save Gnuplot script of the curves in specified file.
9111 To avoid some filtergraph syntax conflicts, each key points list need to be
9112 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
9114 @subsection Examples
9118 Increase slightly the middle level of blue:
9120 curves=blue='0/0 0.5/0.58 1/1'
9126 curves=r='0/0.11 .42/.51 1/0.95':g='0/0 0.50/0.48 1/1':b='0/0.22 .49/.44 1/0.8'
9128 Here we obtain the following coordinates for each components:
9131 @code{(0;0.11) (0.42;0.51) (1;0.95)}
9133 @code{(0;0) (0.50;0.48) (1;1)}
9135 @code{(0;0.22) (0.49;0.44) (1;0.80)}
9139 The previous example can also be achieved with the associated built-in preset:
9141 curves=preset=vintage
9151 Use a Photoshop preset and redefine the points of the green component:
9153 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
9157 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
9158 and @command{gnuplot}:
9160 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
9161 gnuplot -p /tmp/curves.plt
9167 Video data analysis filter.
9169 This filter shows hexadecimal pixel values of part of video.
9171 The filter accepts the following options:
9175 Set output video size.
9178 Set x offset from where to pick pixels.
9181 Set y offset from where to pick pixels.
9184 Set scope mode, can be one of the following:
9187 Draw hexadecimal pixel values with white color on black background.
9190 Draw hexadecimal pixel values with input video pixel color on black
9194 Draw hexadecimal pixel values on color background picked from input video,
9195 the text color is picked in such way so its always visible.
9199 Draw rows and columns numbers on left and top of video.
9202 Set background opacity.
9205 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
9209 Apply Directional blur filter.
9211 The filter accepts the following options:
9215 Set angle of directional blur. Default is @code{45}.
9218 Set radius of directional blur. Default is @code{5}.
9221 Set which planes to filter. By default all planes are filtered.
9224 @subsection Commands
9225 This filter supports same @ref{commands} as options.
9226 The command accepts the same syntax of the corresponding option.
9228 If the specified expression is not valid, it is kept at its current
9233 Denoise frames using 2D DCT (frequency domain filtering).
9235 This filter is not designed for real time.
9237 The filter accepts the following options:
9241 Set the noise sigma constant.
9243 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
9244 coefficient (absolute value) below this threshold with be dropped.
9246 If you need a more advanced filtering, see @option{expr}.
9248 Default is @code{0}.
9251 Set number overlapping pixels for each block. Since the filter can be slow, you
9252 may want to reduce this value, at the cost of a less effective filter and the
9253 risk of various artefacts.
9255 If the overlapping value doesn't permit processing the whole input width or
9256 height, a warning will be displayed and according borders won't be denoised.
9258 Default value is @var{blocksize}-1, which is the best possible setting.
9261 Set the coefficient factor expression.
9263 For each coefficient of a DCT block, this expression will be evaluated as a
9264 multiplier value for the coefficient.
9266 If this is option is set, the @option{sigma} option will be ignored.
9268 The absolute value of the coefficient can be accessed through the @var{c}
9272 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
9273 @var{blocksize}, which is the width and height of the processed blocks.
9275 The default value is @var{3} (8x8) and can be raised to @var{4} for a
9276 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
9277 on the speed processing. Also, a larger block size does not necessarily means a
9281 @subsection Examples
9283 Apply a denoise with a @option{sigma} of @code{4.5}:
9288 The same operation can be achieved using the expression system:
9290 dctdnoiz=e='gte(c, 4.5*3)'
9293 Violent denoise using a block size of @code{16x16}:
9300 Remove banding artifacts from input video.
9301 It works by replacing banded pixels with average value of referenced pixels.
9303 The filter accepts the following options:
9310 Set banding detection threshold for each plane. Default is 0.02.
9311 Valid range is 0.00003 to 0.5.
9312 If difference between current pixel and reference pixel is less than threshold,
9313 it will be considered as banded.
9316 Banding detection range in pixels. Default is 16. If positive, random number
9317 in range 0 to set value will be used. If negative, exact absolute value
9319 The range defines square of four pixels around current pixel.
9322 Set direction in radians from which four pixel will be compared. If positive,
9323 random direction from 0 to set direction will be picked. If negative, exact of
9324 absolute value will be picked. For example direction 0, -PI or -2*PI radians
9325 will pick only pixels on same row and -PI/2 will pick only pixels on same
9329 If enabled, current pixel is compared with average value of all four
9330 surrounding pixels. The default is enabled. If disabled current pixel is
9331 compared with all four surrounding pixels. The pixel is considered banded
9332 if only all four differences with surrounding pixels are less than threshold.
9335 If enabled, current pixel is changed if and only if all pixel components are banded,
9336 e.g. banding detection threshold is triggered for all color components.
9337 The default is disabled.
9342 Remove blocking artifacts from input video.
9344 The filter accepts the following options:
9348 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
9349 This controls what kind of deblocking is applied.
9352 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
9358 Set blocking detection thresholds. Allowed range is 0 to 1.
9359 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
9360 Using higher threshold gives more deblocking strength.
9361 Setting @var{alpha} controls threshold detection at exact edge of block.
9362 Remaining options controls threshold detection near the edge. Each one for
9363 below/above or left/right. Setting any of those to @var{0} disables
9367 Set planes to filter. Default is to filter all available planes.
9370 @subsection Examples
9374 Deblock using weak filter and block size of 4 pixels.
9376 deblock=filter=weak:block=4
9380 Deblock using strong filter, block size of 4 pixels and custom thresholds for
9381 deblocking more edges.
9383 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
9387 Similar as above, but filter only first plane.
9389 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
9393 Similar as above, but filter only second and third plane.
9395 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
9402 Drop duplicated frames at regular intervals.
9404 The filter accepts the following options:
9408 Set the number of frames from which one will be dropped. Setting this to
9409 @var{N} means one frame in every batch of @var{N} frames will be dropped.
9410 Default is @code{5}.
9413 Set the threshold for duplicate detection. If the difference metric for a frame
9414 is less than or equal to this value, then it is declared as duplicate. Default
9418 Set scene change threshold. Default is @code{15}.
9422 Set the size of the x and y-axis blocks used during metric calculations.
9423 Larger blocks give better noise suppression, but also give worse detection of
9424 small movements. Must be a power of two. Default is @code{32}.
9427 Mark main input as a pre-processed input and activate clean source input
9428 stream. This allows the input to be pre-processed with various filters to help
9429 the metrics calculation while keeping the frame selection lossless. When set to
9430 @code{1}, the first stream is for the pre-processed input, and the second
9431 stream is the clean source from where the kept frames are chosen. Default is
9435 Set whether or not chroma is considered in the metric calculations. Default is
9441 Apply 2D deconvolution of video stream in frequency domain using second stream
9444 The filter accepts the following options:
9448 Set which planes to process.
9451 Set which impulse video frames will be processed, can be @var{first}
9452 or @var{all}. Default is @var{all}.
9455 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
9456 and height are not same and not power of 2 or if stream prior to convolving
9460 The @code{deconvolve} filter also supports the @ref{framesync} options.
9464 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
9466 It accepts the following options:
9470 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
9471 @var{rainbows} for cross-color reduction.
9474 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
9477 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
9480 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
9483 Set temporal chroma threshold. Lower values increases reduction of cross-color.
9488 Apply deflate effect to the video.
9490 This filter replaces the pixel by the local(3x3) average by taking into account
9491 only values lower than the pixel.
9493 It accepts the following options:
9500 Limit the maximum change for each plane, default is 65535.
9501 If 0, plane will remain unchanged.
9504 @subsection Commands
9506 This filter supports the all above options as @ref{commands}.
9510 Remove temporal frame luminance variations.
9512 It accepts the following options:
9516 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
9519 Set averaging mode to smooth temporal luminance variations.
9521 Available values are:
9546 Do not actually modify frame. Useful when one only wants metadata.
9551 Remove judder produced by partially interlaced telecined content.
9553 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
9554 source was partially telecined content then the output of @code{pullup,dejudder}
9555 will have a variable frame rate. May change the recorded frame rate of the
9556 container. Aside from that change, this filter will not affect constant frame
9559 The option available in this filter is:
9563 Specify the length of the window over which the judder repeats.
9565 Accepts any integer greater than 1. Useful values are:
9569 If the original was telecined from 24 to 30 fps (Film to NTSC).
9572 If the original was telecined from 25 to 30 fps (PAL to NTSC).
9575 If a mixture of the two.
9578 The default is @samp{4}.
9583 Suppress a TV station logo by a simple interpolation of the surrounding
9584 pixels. Just set a rectangle covering the logo and watch it disappear
9585 (and sometimes something even uglier appear - your mileage may vary).
9587 It accepts the following parameters:
9592 Specify the top left corner coordinates of the logo. They must be
9597 Specify the width and height of the logo to clear. They must be
9601 Specify the thickness of the fuzzy edge of the rectangle (added to
9602 @var{w} and @var{h}). The default value is 1. This option is
9603 deprecated, setting higher values should no longer be necessary and
9607 When set to 1, a green rectangle is drawn on the screen to simplify
9608 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
9609 The default value is 0.
9611 The rectangle is drawn on the outermost pixels which will be (partly)
9612 replaced with interpolated values. The values of the next pixels
9613 immediately outside this rectangle in each direction will be used to
9614 compute the interpolated pixel values inside the rectangle.
9618 @subsection Examples
9622 Set a rectangle covering the area with top left corner coordinates 0,0
9623 and size 100x77, and a band of size 10:
9625 delogo=x=0:y=0:w=100:h=77:band=10
9633 Remove the rain in the input image/video by applying the derain methods based on
9634 convolutional neural networks. Supported models:
9638 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
9639 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
9642 Training as well as model generation scripts are provided in
9643 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
9645 Native model files (.model) can be generated from TensorFlow model
9646 files (.pb) by using tools/python/convert.py
9648 The filter accepts the following options:
9652 Specify which filter to use. This option accepts the following values:
9656 Derain filter. To conduct derain filter, you need to use a derain model.
9659 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
9661 Default value is @samp{derain}.
9664 Specify which DNN backend to use for model loading and execution. This option accepts
9665 the following values:
9669 Native implementation of DNN loading and execution.
9672 TensorFlow backend. To enable this backend you
9673 need to install the TensorFlow for C library (see
9674 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9675 @code{--enable-libtensorflow}
9677 Default value is @samp{native}.
9680 Set path to model file specifying network architecture and its parameters.
9681 Note that different backends use different file formats. TensorFlow and native
9682 backend can load files for only its format.
9685 It can also be finished with @ref{dnn_processing} filter.
9689 Attempt to fix small changes in horizontal and/or vertical shift. This
9690 filter helps remove camera shake from hand-holding a camera, bumping a
9691 tripod, moving on a vehicle, etc.
9693 The filter accepts the following options:
9701 Specify a rectangular area where to limit the search for motion
9703 If desired the search for motion vectors can be limited to a
9704 rectangular area of the frame defined by its top left corner, width
9705 and height. These parameters have the same meaning as the drawbox
9706 filter which can be used to visualise the position of the bounding
9709 This is useful when simultaneous movement of subjects within the frame
9710 might be confused for camera motion by the motion vector search.
9712 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
9713 then the full frame is used. This allows later options to be set
9714 without specifying the bounding box for the motion vector search.
9716 Default - search the whole frame.
9720 Specify the maximum extent of movement in x and y directions in the
9721 range 0-64 pixels. Default 16.
9724 Specify how to generate pixels to fill blanks at the edge of the
9725 frame. Available values are:
9728 Fill zeroes at blank locations
9730 Original image at blank locations
9732 Extruded edge value at blank locations
9734 Mirrored edge at blank locations
9736 Default value is @samp{mirror}.
9739 Specify the blocksize to use for motion search. Range 4-128 pixels,
9743 Specify the contrast threshold for blocks. Only blocks with more than
9744 the specified contrast (difference between darkest and lightest
9745 pixels) will be considered. Range 1-255, default 125.
9748 Specify the search strategy. Available values are:
9751 Set exhaustive search
9753 Set less exhaustive search.
9755 Default value is @samp{exhaustive}.
9758 If set then a detailed log of the motion search is written to the
9765 Remove unwanted contamination of foreground colors, caused by reflected color of
9766 greenscreen or bluescreen.
9768 This filter accepts the following options:
9772 Set what type of despill to use.
9775 Set how spillmap will be generated.
9778 Set how much to get rid of still remaining spill.
9781 Controls amount of red in spill area.
9784 Controls amount of green in spill area.
9785 Should be -1 for greenscreen.
9788 Controls amount of blue in spill area.
9789 Should be -1 for bluescreen.
9792 Controls brightness of spill area, preserving colors.
9795 Modify alpha from generated spillmap.
9798 @subsection Commands
9800 This filter supports the all above options as @ref{commands}.
9804 Apply an exact inverse of the telecine operation. It requires a predefined
9805 pattern specified using the pattern option which must be the same as that passed
9806 to the telecine filter.
9808 This filter accepts the following options:
9817 The default value is @code{top}.
9821 A string of numbers representing the pulldown pattern you wish to apply.
9822 The default value is @code{23}.
9825 A number representing position of the first frame with respect to the telecine
9826 pattern. This is to be used if the stream is cut. The default value is @code{0}.
9831 Apply dilation effect to the video.
9833 This filter replaces the pixel by the local(3x3) maximum.
9835 It accepts the following options:
9842 Limit the maximum change for each plane, default is 65535.
9843 If 0, plane will remain unchanged.
9846 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9849 Flags to local 3x3 coordinates maps like this:
9856 @subsection Commands
9858 This filter supports the all above options as @ref{commands}.
9862 Displace pixels as indicated by second and third input stream.
9864 It takes three input streams and outputs one stream, the first input is the
9865 source, and second and third input are displacement maps.
9867 The second input specifies how much to displace pixels along the
9868 x-axis, while the third input specifies how much to displace pixels
9870 If one of displacement map streams terminates, last frame from that
9871 displacement map will be used.
9873 Note that once generated, displacements maps can be reused over and over again.
9875 A description of the accepted options follows.
9879 Set displace behavior for pixels that are out of range.
9881 Available values are:
9884 Missing pixels are replaced by black pixels.
9887 Adjacent pixels will spread out to replace missing pixels.
9890 Out of range pixels are wrapped so they point to pixels of other side.
9893 Out of range pixels will be replaced with mirrored pixels.
9895 Default is @samp{smear}.
9899 @subsection Examples
9903 Add ripple effect to rgb input of video size hd720:
9905 ffmpeg -i INPUT -f lavfi -i nullsrc=s=hd720,lutrgb=128:128:128 -f lavfi -i nullsrc=s=hd720,geq='r=128+30*sin(2*PI*X/400+T):g=128+30*sin(2*PI*X/400+T):b=128+30*sin(2*PI*X/400+T)' -lavfi '[0][1][2]displace' OUTPUT
9909 Add wave effect to rgb input of video size hd720:
9911 ffmpeg -i INPUT -f lavfi -i nullsrc=hd720,geq='r=128+80*(sin(sqrt((X-W/2)*(X-W/2)+(Y-H/2)*(Y-H/2))/220*2*PI+T)):g=128+80*(sin(sqrt((X-W/2)*(X-W/2)+(Y-H/2)*(Y-H/2))/220*2*PI+T)):b=128+80*(sin(sqrt((X-W/2)*(X-W/2)+(Y-H/2)*(Y-H/2))/220*2*PI+T))' -lavfi '[1]split[x][y],[0][x][y]displace' OUTPUT
9915 @anchor{dnn_processing}
9916 @section dnn_processing
9918 Do image processing with deep neural networks. It works together with another filter
9919 which converts the pixel format of the Frame to what the dnn network requires.
9921 The filter accepts the following options:
9925 Specify which DNN backend to use for model loading and execution. This option accepts
9926 the following values:
9930 Native implementation of DNN loading and execution.
9933 TensorFlow backend. To enable this backend you
9934 need to install the TensorFlow for C library (see
9935 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9936 @code{--enable-libtensorflow}
9939 OpenVINO backend. To enable this backend you
9940 need to build and install the OpenVINO for C library (see
9941 @url{https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md}) and configure FFmpeg with
9942 @code{--enable-libopenvino} (--extra-cflags=-I... --extra-ldflags=-L... might
9943 be needed if the header files and libraries are not installed into system path)
9947 Default value is @samp{native}.
9950 Set path to model file specifying network architecture and its parameters.
9951 Note that different backends use different file formats. TensorFlow, OpenVINO and native
9952 backend can load files for only its format.
9954 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9957 Set the input name of the dnn network.
9960 Set the output name of the dnn network.
9964 @subsection Examples
9968 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
9970 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
9974 Halve the pixel value of the frame with format gray32f:
9976 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
9980 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
9982 ./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
9986 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
9988 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
9995 Draw a colored box on the input image.
9997 It accepts the following parameters:
10002 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
10006 The expressions which specify the width and height of the box; if 0 they are interpreted as
10007 the input width and height. It defaults to 0.
10010 Specify the color of the box to write. For the general syntax of this option,
10011 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10012 value @code{invert} is used, the box edge color is the same as the
10013 video with inverted luma.
10016 The expression which sets the thickness of the box edge.
10017 A value of @code{fill} will create a filled box. Default value is @code{3}.
10019 See below for the list of accepted constants.
10022 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
10023 will overwrite the video's color and alpha pixels.
10024 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
10027 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10028 following constants:
10032 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10036 horizontal and vertical chroma subsample values. For example for the
10037 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10041 The input width and height.
10044 The input sample aspect ratio.
10048 The x and y offset coordinates where the box is drawn.
10052 The width and height of the drawn box.
10055 The thickness of the drawn box.
10057 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10058 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10062 @subsection Examples
10066 Draw a black box around the edge of the input image:
10072 Draw a box with color red and an opacity of 50%:
10074 drawbox=10:20:200:60:red@@0.5
10077 The previous example can be specified as:
10079 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
10083 Fill the box with pink color:
10085 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
10089 Draw a 2-pixel red 2.40:1 mask:
10091 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
10095 @subsection Commands
10096 This filter supports same commands as options.
10097 The command accepts the same syntax of the corresponding option.
10099 If the specified expression is not valid, it is kept at its current
10104 Draw a graph using input video metadata.
10106 It accepts the following parameters:
10110 Set 1st frame metadata key from which metadata values will be used to draw a graph.
10113 Set 1st foreground color expression.
10116 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
10119 Set 2nd foreground color expression.
10122 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
10125 Set 3rd foreground color expression.
10128 Set 4th frame metadata key from which metadata values will be used to draw a graph.
10131 Set 4th foreground color expression.
10134 Set minimal value of metadata value.
10137 Set maximal value of metadata value.
10140 Set graph background color. Default is white.
10145 Available values for mode is:
10152 Default is @code{line}.
10157 Available values for slide is:
10160 Draw new frame when right border is reached.
10163 Replace old columns with new ones.
10166 Scroll from right to left.
10169 Scroll from left to right.
10172 Draw single picture.
10175 Default is @code{frame}.
10178 Set size of graph video. For the syntax of this option, check the
10179 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
10180 The default value is @code{900x256}.
10183 Set the output frame rate. Default value is @code{25}.
10185 The foreground color expressions can use the following variables:
10188 Minimal value of metadata value.
10191 Maximal value of metadata value.
10194 Current metadata key value.
10197 The color is defined as 0xAABBGGRR.
10200 Example using metadata from @ref{signalstats} filter:
10202 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
10205 Example using metadata from @ref{ebur128} filter:
10207 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
10212 Draw a grid on the input image.
10214 It accepts the following parameters:
10219 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
10223 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
10224 input width and height, respectively, minus @code{thickness}, so image gets
10225 framed. Default to 0.
10228 Specify the color of the grid. For the general syntax of this option,
10229 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10230 value @code{invert} is used, the grid color is the same as the
10231 video with inverted luma.
10234 The expression which sets the thickness of the grid line. Default value is @code{1}.
10236 See below for the list of accepted constants.
10239 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
10240 will overwrite the video's color and alpha pixels.
10241 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
10244 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10245 following constants:
10249 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10253 horizontal and vertical chroma subsample values. For example for the
10254 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10258 The input grid cell width and height.
10261 The input sample aspect ratio.
10265 The x and y coordinates of some point of grid intersection (meant to configure offset).
10269 The width and height of the drawn cell.
10272 The thickness of the drawn cell.
10274 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10275 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10279 @subsection Examples
10283 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
10285 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
10289 Draw a white 3x3 grid with an opacity of 50%:
10291 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
10295 @subsection Commands
10296 This filter supports same commands as options.
10297 The command accepts the same syntax of the corresponding option.
10299 If the specified expression is not valid, it is kept at its current
10305 Draw a text string or text from a specified file on top of a video, using the
10306 libfreetype library.
10308 To enable compilation of this filter, you need to configure FFmpeg with
10309 @code{--enable-libfreetype}.
10310 To enable default font fallback and the @var{font} option you need to
10311 configure FFmpeg with @code{--enable-libfontconfig}.
10312 To enable the @var{text_shaping} option, you need to configure FFmpeg with
10313 @code{--enable-libfribidi}.
10317 It accepts the following parameters:
10322 Used to draw a box around text using the background color.
10323 The value must be either 1 (enable) or 0 (disable).
10324 The default value of @var{box} is 0.
10327 Set the width of the border to be drawn around the box using @var{boxcolor}.
10328 The default value of @var{boxborderw} is 0.
10331 The color to be used for drawing box around text. For the syntax of this
10332 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10334 The default value of @var{boxcolor} is "white".
10337 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
10338 The default value of @var{line_spacing} is 0.
10341 Set the width of the border to be drawn around the text using @var{bordercolor}.
10342 The default value of @var{borderw} is 0.
10345 Set the color to be used for drawing border around text. For the syntax of this
10346 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10348 The default value of @var{bordercolor} is "black".
10351 Select how the @var{text} is expanded. Can be either @code{none},
10352 @code{strftime} (deprecated) or
10353 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
10357 Set a start time for the count. Value is in microseconds. Only applied
10358 in the deprecated strftime expansion mode. To emulate in normal expansion
10359 mode use the @code{pts} function, supplying the start time (in seconds)
10360 as the second argument.
10363 If true, check and fix text coords to avoid clipping.
10366 The color to be used for drawing fonts. For the syntax of this option, check
10367 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10369 The default value of @var{fontcolor} is "black".
10371 @item fontcolor_expr
10372 String which is expanded the same way as @var{text} to obtain dynamic
10373 @var{fontcolor} value. By default this option has empty value and is not
10374 processed. When this option is set, it overrides @var{fontcolor} option.
10377 The font family to be used for drawing text. By default Sans.
10380 The font file to be used for drawing text. The path must be included.
10381 This parameter is mandatory if the fontconfig support is disabled.
10384 Draw the text applying alpha blending. The value can
10385 be a number between 0.0 and 1.0.
10386 The expression accepts the same variables @var{x, y} as well.
10387 The default value is 1.
10388 Please see @var{fontcolor_expr}.
10391 The font size to be used for drawing text.
10392 The default value of @var{fontsize} is 16.
10395 If set to 1, attempt to shape the text (for example, reverse the order of
10396 right-to-left text and join Arabic characters) before drawing it.
10397 Otherwise, just draw the text exactly as given.
10398 By default 1 (if supported).
10400 @item ft_load_flags
10401 The flags to be used for loading the fonts.
10403 The flags map the corresponding flags supported by libfreetype, and are
10404 a combination of the following values:
10411 @item vertical_layout
10412 @item force_autohint
10415 @item ignore_global_advance_width
10417 @item ignore_transform
10419 @item linear_design
10423 Default value is "default".
10425 For more information consult the documentation for the FT_LOAD_*
10429 The color to be used for drawing a shadow behind the drawn text. For the
10430 syntax of this option, check the @ref{color syntax,,"Color" section in the
10431 ffmpeg-utils manual,ffmpeg-utils}.
10433 The default value of @var{shadowcolor} is "black".
10437 The x and y offsets for the text shadow position with respect to the
10438 position of the text. They can be either positive or negative
10439 values. The default value for both is "0".
10442 The starting frame number for the n/frame_num variable. The default value
10446 The size in number of spaces to use for rendering the tab.
10447 Default value is 4.
10450 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
10451 format. It can be used with or without text parameter. @var{timecode_rate}
10452 option must be specified.
10454 @item timecode_rate, rate, r
10455 Set the timecode frame rate (timecode only). Value will be rounded to nearest
10456 integer. Minimum value is "1".
10457 Drop-frame timecode is supported for frame rates 30 & 60.
10460 If set to 1, the output of the timecode option will wrap around at 24 hours.
10461 Default is 0 (disabled).
10464 The text string to be drawn. The text must be a sequence of UTF-8
10465 encoded characters.
10466 This parameter is mandatory if no file is specified with the parameter
10470 A text file containing text to be drawn. The text must be a sequence
10471 of UTF-8 encoded characters.
10473 This parameter is mandatory if no text string is specified with the
10474 parameter @var{text}.
10476 If both @var{text} and @var{textfile} are specified, an error is thrown.
10479 If set to 1, the @var{textfile} will be reloaded before each frame.
10480 Be sure to update it atomically, or it may be read partially, or even fail.
10484 The expressions which specify the offsets where text will be drawn
10485 within the video frame. They are relative to the top/left border of the
10488 The default value of @var{x} and @var{y} is "0".
10490 See below for the list of accepted constants and functions.
10493 The parameters for @var{x} and @var{y} are expressions containing the
10494 following constants and functions:
10498 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
10502 horizontal and vertical chroma subsample values. For example for the
10503 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10506 the height of each text line
10514 @item max_glyph_a, ascent
10515 the maximum distance from the baseline to the highest/upper grid
10516 coordinate used to place a glyph outline point, for all the rendered
10518 It is a positive value, due to the grid's orientation with the Y axis
10521 @item max_glyph_d, descent
10522 the maximum distance from the baseline to the lowest grid coordinate
10523 used to place a glyph outline point, for all the rendered glyphs.
10524 This is a negative value, due to the grid's orientation, with the Y axis
10528 maximum glyph height, that is the maximum height for all the glyphs
10529 contained in the rendered text, it is equivalent to @var{ascent} -
10533 maximum glyph width, that is the maximum width for all the glyphs
10534 contained in the rendered text
10537 the number of input frame, starting from 0
10539 @item rand(min, max)
10540 return a random number included between @var{min} and @var{max}
10543 The input sample aspect ratio.
10546 timestamp expressed in seconds, NAN if the input timestamp is unknown
10549 the height of the rendered text
10552 the width of the rendered text
10556 the x and y offset coordinates where the text is drawn.
10558 These parameters allow the @var{x} and @var{y} expressions to refer
10559 to each other, so you can for example specify @code{y=x/dar}.
10562 A one character description of the current frame's picture type.
10565 The current packet's position in the input file or stream
10566 (in bytes, from the start of the input). A value of -1 indicates
10567 this info is not available.
10570 The current packet's duration, in seconds.
10573 The current packet's size (in bytes).
10576 @anchor{drawtext_expansion}
10577 @subsection Text expansion
10579 If @option{expansion} is set to @code{strftime},
10580 the filter recognizes strftime() sequences in the provided text and
10581 expands them accordingly. Check the documentation of strftime(). This
10582 feature is deprecated.
10584 If @option{expansion} is set to @code{none}, the text is printed verbatim.
10586 If @option{expansion} is set to @code{normal} (which is the default),
10587 the following expansion mechanism is used.
10589 The backslash character @samp{\}, followed by any character, always expands to
10590 the second character.
10592 Sequences of the form @code{%@{...@}} are expanded. The text between the
10593 braces is a function name, possibly followed by arguments separated by ':'.
10594 If the arguments contain special characters or delimiters (':' or '@}'),
10595 they should be escaped.
10597 Note that they probably must also be escaped as the value for the
10598 @option{text} option in the filter argument string and as the filter
10599 argument in the filtergraph description, and possibly also for the shell,
10600 that makes up to four levels of escaping; using a text file avoids these
10603 The following functions are available:
10608 The expression evaluation result.
10610 It must take one argument specifying the expression to be evaluated,
10611 which accepts the same constants and functions as the @var{x} and
10612 @var{y} values. Note that not all constants should be used, for
10613 example the text size is not known when evaluating the expression, so
10614 the constants @var{text_w} and @var{text_h} will have an undefined
10617 @item expr_int_format, eif
10618 Evaluate the expression's value and output as formatted integer.
10620 The first argument is the expression to be evaluated, just as for the @var{expr} function.
10621 The second argument specifies the output format. Allowed values are @samp{x},
10622 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
10623 @code{printf} function.
10624 The third parameter is optional and sets the number of positions taken by the output.
10625 It can be used to add padding with zeros from the left.
10628 The time at which the filter is running, expressed in UTC.
10629 It can accept an argument: a strftime() format string.
10632 The time at which the filter is running, expressed in the local time zone.
10633 It can accept an argument: a strftime() format string.
10636 Frame metadata. Takes one or two arguments.
10638 The first argument is mandatory and specifies the metadata key.
10640 The second argument is optional and specifies a default value, used when the
10641 metadata key is not found or empty.
10643 Available metadata can be identified by inspecting entries
10644 starting with TAG included within each frame section
10645 printed by running @code{ffprobe -show_frames}.
10647 String metadata generated in filters leading to
10648 the drawtext filter are also available.
10651 The frame number, starting from 0.
10654 A one character description of the current picture type.
10657 The timestamp of the current frame.
10658 It can take up to three arguments.
10660 The first argument is the format of the timestamp; it defaults to @code{flt}
10661 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
10662 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
10663 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
10664 @code{localtime} stands for the timestamp of the frame formatted as
10665 local time zone time.
10667 The second argument is an offset added to the timestamp.
10669 If the format is set to @code{hms}, a third argument @code{24HH} may be
10670 supplied to present the hour part of the formatted timestamp in 24h format
10673 If the format is set to @code{localtime} or @code{gmtime},
10674 a third argument may be supplied: a strftime() format string.
10675 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
10678 @subsection Commands
10680 This filter supports altering parameters via commands:
10683 Alter existing filter parameters.
10685 Syntax for the argument is the same as for filter invocation, e.g.
10688 fontsize=56:fontcolor=green:text='Hello World'
10691 Full filter invocation with sendcmd would look like this:
10694 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
10698 If the entire argument can't be parsed or applied as valid values then the filter will
10699 continue with its existing parameters.
10701 @subsection Examples
10705 Draw "Test Text" with font FreeSerif, using the default values for the
10706 optional parameters.
10709 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
10713 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
10714 and y=50 (counting from the top-left corner of the screen), text is
10715 yellow with a red box around it. Both the text and the box have an
10719 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
10720 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
10723 Note that the double quotes are not necessary if spaces are not used
10724 within the parameter list.
10727 Show the text at the center of the video frame:
10729 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
10733 Show the text at a random position, switching to a new position every 30 seconds:
10735 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)"
10739 Show a text line sliding from right to left in the last row of the video
10740 frame. The file @file{LONG_LINE} is assumed to contain a single line
10743 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
10747 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
10749 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
10753 Draw a single green letter "g", at the center of the input video.
10754 The glyph baseline is placed at half screen height.
10756 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
10760 Show text for 1 second every 3 seconds:
10762 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
10766 Use fontconfig to set the font. Note that the colons need to be escaped.
10768 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
10772 Draw "Test Text" with font size dependent on height of the video.
10774 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
10778 Print the date of a real-time encoding (see strftime(3)):
10780 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
10784 Show text fading in and out (appearing/disappearing):
10787 DS=1.0 # display start
10788 DE=10.0 # display end
10789 FID=1.5 # fade in duration
10790 FOD=5 # fade out duration
10791 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 @}"
10795 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
10796 and the @option{fontsize} value are included in the @option{y} offset.
10798 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
10799 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
10803 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
10804 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
10805 must have option @option{-export_path_metadata 1} for the special metadata fields
10806 to be available for filters.
10808 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
10813 For more information about libfreetype, check:
10814 @url{http://www.freetype.org/}.
10816 For more information about fontconfig, check:
10817 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
10819 For more information about libfribidi, check:
10820 @url{http://fribidi.org/}.
10822 @section edgedetect
10824 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
10826 The filter accepts the following options:
10831 Set low and high threshold values used by the Canny thresholding
10834 The high threshold selects the "strong" edge pixels, which are then
10835 connected through 8-connectivity with the "weak" edge pixels selected
10836 by the low threshold.
10838 @var{low} and @var{high} threshold values must be chosen in the range
10839 [0,1], and @var{low} should be lesser or equal to @var{high}.
10841 Default value for @var{low} is @code{20/255}, and default value for @var{high}
10845 Define the drawing mode.
10849 Draw white/gray wires on black background.
10852 Mix the colors to create a paint/cartoon effect.
10855 Apply Canny edge detector on all selected planes.
10857 Default value is @var{wires}.
10860 Select planes for filtering. By default all available planes are filtered.
10863 @subsection Examples
10867 Standard edge detection with custom values for the hysteresis thresholding:
10869 edgedetect=low=0.1:high=0.4
10873 Painting effect without thresholding:
10875 edgedetect=mode=colormix:high=0
10881 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10883 For each input image, the filter will compute the optimal mapping from
10884 the input to the output given the codebook length, that is the number
10885 of distinct output colors.
10887 This filter accepts the following options.
10890 @item codebook_length, l
10891 Set codebook length. The value must be a positive integer, and
10892 represents the number of distinct output colors. Default value is 256.
10895 Set the maximum number of iterations to apply for computing the optimal
10896 mapping. The higher the value the better the result and the higher the
10897 computation time. Default value is 1.
10900 Set a random seed, must be an integer included between 0 and
10901 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10902 will try to use a good random seed on a best effort basis.
10905 Set pal8 output pixel format. This option does not work with codebook
10906 length greater than 256.
10911 Measure graylevel entropy in histogram of color channels of video frames.
10913 It accepts the following parameters:
10917 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10919 @var{diff} mode measures entropy of histogram delta values, absolute differences
10920 between neighbour histogram values.
10924 Set brightness, contrast, saturation and approximate gamma adjustment.
10926 The filter accepts the following options:
10930 Set the contrast expression. The value must be a float value in range
10931 @code{-1000.0} to @code{1000.0}. The default value is "1".
10934 Set the brightness expression. The value must be a float value in
10935 range @code{-1.0} to @code{1.0}. The default value is "0".
10938 Set the saturation expression. The value must be a float in
10939 range @code{0.0} to @code{3.0}. The default value is "1".
10942 Set the gamma expression. The value must be a float in range
10943 @code{0.1} to @code{10.0}. The default value is "1".
10946 Set the gamma expression for red. The value must be a float in
10947 range @code{0.1} to @code{10.0}. The default value is "1".
10950 Set the gamma expression for green. The value must be a float in range
10951 @code{0.1} to @code{10.0}. The default value is "1".
10954 Set the gamma expression for blue. The value must be a float in range
10955 @code{0.1} to @code{10.0}. The default value is "1".
10958 Set the gamma weight expression. It can be used to reduce the effect
10959 of a high gamma value on bright image areas, e.g. keep them from
10960 getting overamplified and just plain white. The value must be a float
10961 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
10962 gamma correction all the way down while @code{1.0} leaves it at its
10963 full strength. Default is "1".
10966 Set when the expressions for brightness, contrast, saturation and
10967 gamma expressions are evaluated.
10969 It accepts the following values:
10972 only evaluate expressions once during the filter initialization or
10973 when a command is processed
10976 evaluate expressions for each incoming frame
10979 Default value is @samp{init}.
10982 The expressions accept the following parameters:
10985 frame count of the input frame starting from 0
10988 byte position of the corresponding packet in the input file, NAN if
10992 frame rate of the input video, NAN if the input frame rate is unknown
10995 timestamp expressed in seconds, NAN if the input timestamp is unknown
10998 @subsection Commands
10999 The filter supports the following commands:
11003 Set the contrast expression.
11006 Set the brightness expression.
11009 Set the saturation expression.
11012 Set the gamma expression.
11015 Set the gamma_r expression.
11018 Set gamma_g expression.
11021 Set gamma_b expression.
11024 Set gamma_weight expression.
11026 The command accepts the same syntax of the corresponding option.
11028 If the specified expression is not valid, it is kept at its current
11035 Apply erosion effect to the video.
11037 This filter replaces the pixel by the local(3x3) minimum.
11039 It accepts the following options:
11046 Limit the maximum change for each plane, default is 65535.
11047 If 0, plane will remain unchanged.
11050 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
11053 Flags to local 3x3 coordinates maps like this:
11060 @subsection Commands
11062 This filter supports the all above options as @ref{commands}.
11064 @section extractplanes
11066 Extract color channel components from input video stream into
11067 separate grayscale video streams.
11069 The filter accepts the following option:
11073 Set plane(s) to extract.
11075 Available values for planes are:
11086 Choosing planes not available in the input will result in an error.
11087 That means you cannot select @code{r}, @code{g}, @code{b} planes
11088 with @code{y}, @code{u}, @code{v} planes at same time.
11091 @subsection Examples
11095 Extract luma, u and v color channel component from input video frame
11096 into 3 grayscale outputs:
11098 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
11104 Apply a fade-in/out effect to the input video.
11106 It accepts the following parameters:
11110 The effect type can be either "in" for a fade-in, or "out" for a fade-out
11112 Default is @code{in}.
11114 @item start_frame, s
11115 Specify the number of the frame to start applying the fade
11116 effect at. Default is 0.
11119 The number of frames that the fade effect lasts. At the end of the
11120 fade-in effect, the output video will have the same intensity as the input video.
11121 At the end of the fade-out transition, the output video will be filled with the
11122 selected @option{color}.
11126 If set to 1, fade only alpha channel, if one exists on the input.
11127 Default value is 0.
11129 @item start_time, st
11130 Specify the timestamp (in seconds) of the frame to start to apply the fade
11131 effect. If both start_frame and start_time are specified, the fade will start at
11132 whichever comes last. Default is 0.
11135 The number of seconds for which the fade effect has to last. At the end of the
11136 fade-in effect the output video will have the same intensity as the input video,
11137 at the end of the fade-out transition the output video will be filled with the
11138 selected @option{color}.
11139 If both duration and nb_frames are specified, duration is used. Default is 0
11140 (nb_frames is used by default).
11143 Specify the color of the fade. Default is "black".
11146 @subsection Examples
11150 Fade in the first 30 frames of video:
11155 The command above is equivalent to:
11161 Fade out the last 45 frames of a 200-frame video:
11164 fade=type=out:start_frame=155:nb_frames=45
11168 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
11170 fade=in:0:25, fade=out:975:25
11174 Make the first 5 frames yellow, then fade in from frame 5-24:
11176 fade=in:5:20:color=yellow
11180 Fade in alpha over first 25 frames of video:
11182 fade=in:0:25:alpha=1
11186 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
11188 fade=t=in:st=5.5:d=0.5
11194 Denoise frames using 3D FFT (frequency domain filtering).
11196 The filter accepts the following options:
11200 Set the noise sigma constant. This sets denoising strength.
11201 Default value is 1. Allowed range is from 0 to 30.
11202 Using very high sigma with low overlap may give blocking artifacts.
11205 Set amount of denoising. By default all detected noise is reduced.
11206 Default value is 1. Allowed range is from 0 to 1.
11209 Set size of block, Default is 4, can be 3, 4, 5 or 6.
11210 Actual size of block in pixels is 2 to power of @var{block}, so by default
11211 block size in pixels is 2^4 which is 16.
11214 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
11217 Set number of previous frames to use for denoising. By default is set to 0.
11220 Set number of next frames to to use for denoising. By default is set to 0.
11223 Set planes which will be filtered, by default are all available filtered
11228 Apply arbitrary expressions to samples in frequency domain
11232 Adjust the dc value (gain) of the luma plane of the image. The filter
11233 accepts an integer value in range @code{0} to @code{1000}. The default
11234 value is set to @code{0}.
11237 Adjust the dc value (gain) of the 1st chroma plane of the image. The
11238 filter accepts an integer value in range @code{0} to @code{1000}. The
11239 default value is set to @code{0}.
11242 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
11243 filter accepts an integer value in range @code{0} to @code{1000}. The
11244 default value is set to @code{0}.
11247 Set the frequency domain weight expression for the luma plane.
11250 Set the frequency domain weight expression for the 1st chroma plane.
11253 Set the frequency domain weight expression for the 2nd chroma plane.
11256 Set when the expressions are evaluated.
11258 It accepts the following values:
11261 Only evaluate expressions once during the filter initialization.
11264 Evaluate expressions for each incoming frame.
11267 Default value is @samp{init}.
11269 The filter accepts the following variables:
11272 The coordinates of the current sample.
11276 The width and height of the image.
11279 The number of input frame, starting from 0.
11282 @subsection Examples
11288 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
11294 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
11300 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
11306 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
11313 Extract a single field from an interlaced image using stride
11314 arithmetic to avoid wasting CPU time. The output frames are marked as
11317 The filter accepts the following options:
11321 Specify whether to extract the top (if the value is @code{0} or
11322 @code{top}) or the bottom field (if the value is @code{1} or
11328 Create new frames by copying the top and bottom fields from surrounding frames
11329 supplied as numbers by the hint file.
11333 Set file containing hints: absolute/relative frame numbers.
11335 There must be one line for each frame in a clip. Each line must contain two
11336 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
11337 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
11338 is current frame number for @code{absolute} mode or out of [-1, 1] range
11339 for @code{relative} mode. First number tells from which frame to pick up top
11340 field and second number tells from which frame to pick up bottom field.
11342 If optionally followed by @code{+} output frame will be marked as interlaced,
11343 else if followed by @code{-} output frame will be marked as progressive, else
11344 it will be marked same as input frame.
11345 If optionally followed by @code{t} output frame will use only top field, or in
11346 case of @code{b} it will use only bottom field.
11347 If line starts with @code{#} or @code{;} that line is skipped.
11350 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
11353 Example of first several lines of @code{hint} file for @code{relative} mode:
11355 0,0 - # first frame
11356 1,0 - # second frame, use third's frame top field and second's frame bottom field
11357 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
11372 @section fieldmatch
11374 Field matching filter for inverse telecine. It is meant to reconstruct the
11375 progressive frames from a telecined stream. The filter does not drop duplicated
11376 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
11377 followed by a decimation filter such as @ref{decimate} in the filtergraph.
11379 The separation of the field matching and the decimation is notably motivated by
11380 the possibility of inserting a de-interlacing filter fallback between the two.
11381 If the source has mixed telecined and real interlaced content,
11382 @code{fieldmatch} will not be able to match fields for the interlaced parts.
11383 But these remaining combed frames will be marked as interlaced, and thus can be
11384 de-interlaced by a later filter such as @ref{yadif} before decimation.
11386 In addition to the various configuration options, @code{fieldmatch} can take an
11387 optional second stream, activated through the @option{ppsrc} option. If
11388 enabled, the frames reconstruction will be based on the fields and frames from
11389 this second stream. This allows the first input to be pre-processed in order to
11390 help the various algorithms of the filter, while keeping the output lossless
11391 (assuming the fields are matched properly). Typically, a field-aware denoiser,
11392 or brightness/contrast adjustments can help.
11394 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
11395 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
11396 which @code{fieldmatch} is based on. While the semantic and usage are very
11397 close, some behaviour and options names can differ.
11399 The @ref{decimate} filter currently only works for constant frame rate input.
11400 If your input has mixed telecined (30fps) and progressive content with a lower
11401 framerate like 24fps use the following filterchain to produce the necessary cfr
11402 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
11404 The filter accepts the following options:
11408 Specify the assumed field order of the input stream. Available values are:
11412 Auto detect parity (use FFmpeg's internal parity value).
11414 Assume bottom field first.
11416 Assume top field first.
11419 Note that it is sometimes recommended not to trust the parity announced by the
11422 Default value is @var{auto}.
11425 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
11426 sense that it won't risk creating jerkiness due to duplicate frames when
11427 possible, but if there are bad edits or blended fields it will end up
11428 outputting combed frames when a good match might actually exist. On the other
11429 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
11430 but will almost always find a good frame if there is one. The other values are
11431 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
11432 jerkiness and creating duplicate frames versus finding good matches in sections
11433 with bad edits, orphaned fields, blended fields, etc.
11435 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
11437 Available values are:
11441 2-way matching (p/c)
11443 2-way matching, and trying 3rd match if still combed (p/c + n)
11445 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
11447 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
11448 still combed (p/c + n + u/b)
11450 3-way matching (p/c/n)
11452 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
11453 detected as combed (p/c/n + u/b)
11456 The parenthesis at the end indicate the matches that would be used for that
11457 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
11460 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
11463 Default value is @var{pc_n}.
11466 Mark the main input stream as a pre-processed input, and enable the secondary
11467 input stream as the clean source to pick the fields from. See the filter
11468 introduction for more details. It is similar to the @option{clip2} feature from
11471 Default value is @code{0} (disabled).
11474 Set the field to match from. It is recommended to set this to the same value as
11475 @option{order} unless you experience matching failures with that setting. In
11476 certain circumstances changing the field that is used to match from can have a
11477 large impact on matching performance. Available values are:
11481 Automatic (same value as @option{order}).
11483 Match from the bottom field.
11485 Match from the top field.
11488 Default value is @var{auto}.
11491 Set whether or not chroma is included during the match comparisons. In most
11492 cases it is recommended to leave this enabled. You should set this to @code{0}
11493 only if your clip has bad chroma problems such as heavy rainbowing or other
11494 artifacts. Setting this to @code{0} could also be used to speed things up at
11495 the cost of some accuracy.
11497 Default value is @code{1}.
11501 These define an exclusion band which excludes the lines between @option{y0} and
11502 @option{y1} from being included in the field matching decision. An exclusion
11503 band can be used to ignore subtitles, a logo, or other things that may
11504 interfere with the matching. @option{y0} sets the starting scan line and
11505 @option{y1} sets the ending line; all lines in between @option{y0} and
11506 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
11507 @option{y0} and @option{y1} to the same value will disable the feature.
11508 @option{y0} and @option{y1} defaults to @code{0}.
11511 Set the scene change detection threshold as a percentage of maximum change on
11512 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
11513 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
11514 @option{scthresh} is @code{[0.0, 100.0]}.
11516 Default value is @code{12.0}.
11519 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
11520 account the combed scores of matches when deciding what match to use as the
11521 final match. Available values are:
11525 No final matching based on combed scores.
11527 Combed scores are only used when a scene change is detected.
11529 Use combed scores all the time.
11532 Default is @var{sc}.
11535 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
11536 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
11537 Available values are:
11541 No forced calculation.
11543 Force p/c/n calculations.
11545 Force p/c/n/u/b calculations.
11548 Default value is @var{none}.
11551 This is the area combing threshold used for combed frame detection. This
11552 essentially controls how "strong" or "visible" combing must be to be detected.
11553 Larger values mean combing must be more visible and smaller values mean combing
11554 can be less visible or strong and still be detected. Valid settings are from
11555 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
11556 be detected as combed). This is basically a pixel difference value. A good
11557 range is @code{[8, 12]}.
11559 Default value is @code{9}.
11562 Sets whether or not chroma is considered in the combed frame decision. Only
11563 disable this if your source has chroma problems (rainbowing, etc.) that are
11564 causing problems for the combed frame detection with chroma enabled. Actually,
11565 using @option{chroma}=@var{0} is usually more reliable, except for the case
11566 where there is chroma only combing in the source.
11568 Default value is @code{0}.
11572 Respectively set the x-axis and y-axis size of the window used during combed
11573 frame detection. This has to do with the size of the area in which
11574 @option{combpel} pixels are required to be detected as combed for a frame to be
11575 declared combed. See the @option{combpel} parameter description for more info.
11576 Possible values are any number that is a power of 2 starting at 4 and going up
11579 Default value is @code{16}.
11582 The number of combed pixels inside any of the @option{blocky} by
11583 @option{blockx} size blocks on the frame for the frame to be detected as
11584 combed. While @option{cthresh} controls how "visible" the combing must be, this
11585 setting controls "how much" combing there must be in any localized area (a
11586 window defined by the @option{blockx} and @option{blocky} settings) on the
11587 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
11588 which point no frames will ever be detected as combed). This setting is known
11589 as @option{MI} in TFM/VFM vocabulary.
11591 Default value is @code{80}.
11594 @anchor{p/c/n/u/b meaning}
11595 @subsection p/c/n/u/b meaning
11597 @subsubsection p/c/n
11599 We assume the following telecined stream:
11602 Top fields: 1 2 2 3 4
11603 Bottom fields: 1 2 3 4 4
11606 The numbers correspond to the progressive frame the fields relate to. Here, the
11607 first two frames are progressive, the 3rd and 4th are combed, and so on.
11609 When @code{fieldmatch} is configured to run a matching from bottom
11610 (@option{field}=@var{bottom}) this is how this input stream get transformed:
11615 B 1 2 3 4 4 <-- matching reference
11624 As a result of the field matching, we can see that some frames get duplicated.
11625 To perform a complete inverse telecine, you need to rely on a decimation filter
11626 after this operation. See for instance the @ref{decimate} filter.
11628 The same operation now matching from top fields (@option{field}=@var{top})
11633 T 1 2 2 3 4 <-- matching reference
11643 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
11644 basically, they refer to the frame and field of the opposite parity:
11647 @item @var{p} matches the field of the opposite parity in the previous frame
11648 @item @var{c} matches the field of the opposite parity in the current frame
11649 @item @var{n} matches the field of the opposite parity in the next frame
11654 The @var{u} and @var{b} matching are a bit special in the sense that they match
11655 from the opposite parity flag. In the following examples, we assume that we are
11656 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
11657 'x' is placed above and below each matched fields.
11659 With bottom matching (@option{field}=@var{bottom}):
11664 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11665 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11673 With top matching (@option{field}=@var{top}):
11678 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11679 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11687 @subsection Examples
11689 Simple IVTC of a top field first telecined stream:
11691 fieldmatch=order=tff:combmatch=none, decimate
11694 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
11696 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
11699 @section fieldorder
11701 Transform the field order of the input video.
11703 It accepts the following parameters:
11708 The output field order. Valid values are @var{tff} for top field first or @var{bff}
11709 for bottom field first.
11712 The default value is @samp{tff}.
11714 The transformation is done by shifting the picture content up or down
11715 by one line, and filling the remaining line with appropriate picture content.
11716 This method is consistent with most broadcast field order converters.
11718 If the input video is not flagged as being interlaced, or it is already
11719 flagged as being of the required output field order, then this filter does
11720 not alter the incoming video.
11722 It is very useful when converting to or from PAL DV material,
11723 which is bottom field first.
11727 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
11730 @section fifo, afifo
11732 Buffer input images and send them when they are requested.
11734 It is mainly useful when auto-inserted by the libavfilter
11737 It does not take parameters.
11739 @section fillborders
11741 Fill borders of the input video, without changing video stream dimensions.
11742 Sometimes video can have garbage at the four edges and you may not want to
11743 crop video input to keep size multiple of some number.
11745 This filter accepts the following options:
11749 Number of pixels to fill from left border.
11752 Number of pixels to fill from right border.
11755 Number of pixels to fill from top border.
11758 Number of pixels to fill from bottom border.
11763 It accepts the following values:
11766 fill pixels using outermost pixels
11769 fill pixels using mirroring (half sample symmetric)
11772 fill pixels with constant value
11775 fill pixels using reflecting (whole sample symmetric)
11778 fill pixels using wrapping
11781 fade pixels to constant value
11784 Default is @var{smear}.
11787 Set color for pixels in fixed or fade mode. Default is @var{black}.
11790 @subsection Commands
11791 This filter supports same @ref{commands} as options.
11792 The command accepts the same syntax of the corresponding option.
11794 If the specified expression is not valid, it is kept at its current
11799 Find a rectangular object
11801 It accepts the following options:
11805 Filepath of the object image, needs to be in gray8.
11808 Detection threshold, default is 0.5.
11811 Number of mipmaps, default is 3.
11813 @item xmin, ymin, xmax, ymax
11814 Specifies the rectangle in which to search.
11817 @subsection Examples
11821 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
11823 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
11829 Flood area with values of same pixel components with another values.
11831 It accepts the following options:
11834 Set pixel x coordinate.
11837 Set pixel y coordinate.
11840 Set source #0 component value.
11843 Set source #1 component value.
11846 Set source #2 component value.
11849 Set source #3 component value.
11852 Set destination #0 component value.
11855 Set destination #1 component value.
11858 Set destination #2 component value.
11861 Set destination #3 component value.
11867 Convert the input video to one of the specified pixel formats.
11868 Libavfilter will try to pick one that is suitable as input to
11871 It accepts the following parameters:
11875 A '|'-separated list of pixel format names, such as
11876 "pix_fmts=yuv420p|monow|rgb24".
11880 @subsection Examples
11884 Convert the input video to the @var{yuv420p} format
11886 format=pix_fmts=yuv420p
11889 Convert the input video to any of the formats in the list
11891 format=pix_fmts=yuv420p|yuv444p|yuv410p
11898 Convert the video to specified constant frame rate by duplicating or dropping
11899 frames as necessary.
11901 It accepts the following parameters:
11905 The desired output frame rate. The default is @code{25}.
11908 Assume the first PTS should be the given value, in seconds. This allows for
11909 padding/trimming at the start of stream. By default, no assumption is made
11910 about the first frame's expected PTS, so no padding or trimming is done.
11911 For example, this could be set to 0 to pad the beginning with duplicates of
11912 the first frame if a video stream starts after the audio stream or to trim any
11913 frames with a negative PTS.
11916 Timestamp (PTS) rounding method.
11918 Possible values are:
11925 round towards -infinity
11927 round towards +infinity
11931 The default is @code{near}.
11934 Action performed when reading the last frame.
11936 Possible values are:
11939 Use same timestamp rounding method as used for other frames.
11941 Pass through last frame if input duration has not been reached yet.
11943 The default is @code{round}.
11947 Alternatively, the options can be specified as a flat string:
11948 @var{fps}[:@var{start_time}[:@var{round}]].
11950 See also the @ref{setpts} filter.
11952 @subsection Examples
11956 A typical usage in order to set the fps to 25:
11962 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
11964 fps=fps=film:round=near
11970 Pack two different video streams into a stereoscopic video, setting proper
11971 metadata on supported codecs. The two views should have the same size and
11972 framerate and processing will stop when the shorter video ends. Please note
11973 that you may conveniently adjust view properties with the @ref{scale} and
11976 It accepts the following parameters:
11980 The desired packing format. Supported values are:
11985 The views are next to each other (default).
11988 The views are on top of each other.
11991 The views are packed by line.
11994 The views are packed by column.
11997 The views are temporally interleaved.
12006 # Convert left and right views into a frame-sequential video
12007 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
12009 # Convert views into a side-by-side video with the same output resolution as the input
12010 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
12015 Change the frame rate by interpolating new video output frames from the source
12018 This filter is not designed to function correctly with interlaced media. If
12019 you wish to change the frame rate of interlaced media then you are required
12020 to deinterlace before this filter and re-interlace after this filter.
12022 A description of the accepted options follows.
12026 Specify the output frames per second. This option can also be specified
12027 as a value alone. The default is @code{50}.
12030 Specify the start of a range where the output frame will be created as a
12031 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12032 the default is @code{15}.
12035 Specify the end of a range where the output frame will be created as a
12036 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12037 the default is @code{240}.
12040 Specify the level at which a scene change is detected as a value between
12041 0 and 100 to indicate a new scene; a low value reflects a low
12042 probability for the current frame to introduce a new scene, while a higher
12043 value means the current frame is more likely to be one.
12044 The default is @code{8.2}.
12047 Specify flags influencing the filter process.
12049 Available value for @var{flags} is:
12052 @item scene_change_detect, scd
12053 Enable scene change detection using the value of the option @var{scene}.
12054 This flag is enabled by default.
12060 Select one frame every N-th frame.
12062 This filter accepts the following option:
12065 Select frame after every @code{step} frames.
12066 Allowed values are positive integers higher than 0. Default value is @code{1}.
12069 @section freezedetect
12071 Detect frozen video.
12073 This filter logs a message and sets frame metadata when it detects that the
12074 input video has no significant change in content during a specified duration.
12075 Video freeze detection calculates the mean average absolute difference of all
12076 the components of video frames and compares it to a noise floor.
12078 The printed times and duration are expressed in seconds. The
12079 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
12080 whose timestamp equals or exceeds the detection duration and it contains the
12081 timestamp of the first frame of the freeze. The
12082 @code{lavfi.freezedetect.freeze_duration} and
12083 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
12086 The filter accepts the following options:
12090 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
12091 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
12095 Set freeze duration until notification (default is 2 seconds).
12098 @section freezeframes
12100 Freeze video frames.
12102 This filter freezes video frames using frame from 2nd input.
12104 The filter accepts the following options:
12108 Set number of first frame from which to start freeze.
12111 Set number of last frame from which to end freeze.
12114 Set number of frame from 2nd input which will be used instead of replaced frames.
12120 Apply a frei0r effect to the input video.
12122 To enable the compilation of this filter, you need to install the frei0r
12123 header and configure FFmpeg with @code{--enable-frei0r}.
12125 It accepts the following parameters:
12130 The name of the frei0r effect to load. If the environment variable
12131 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
12132 directories specified by the colon-separated list in @env{FREI0R_PATH}.
12133 Otherwise, the standard frei0r paths are searched, in this order:
12134 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
12135 @file{/usr/lib/frei0r-1/}.
12137 @item filter_params
12138 A '|'-separated list of parameters to pass to the frei0r effect.
12142 A frei0r effect parameter can be a boolean (its value is either
12143 "y" or "n"), a double, a color (specified as
12144 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
12145 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
12146 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
12147 a position (specified as @var{X}/@var{Y}, where
12148 @var{X} and @var{Y} are floating point numbers) and/or a string.
12150 The number and types of parameters depend on the loaded effect. If an
12151 effect parameter is not specified, the default value is set.
12153 @subsection Examples
12157 Apply the distort0r effect, setting the first two double parameters:
12159 frei0r=filter_name=distort0r:filter_params=0.5|0.01
12163 Apply the colordistance effect, taking a color as the first parameter:
12165 frei0r=colordistance:0.2/0.3/0.4
12166 frei0r=colordistance:violet
12167 frei0r=colordistance:0x112233
12171 Apply the perspective effect, specifying the top left and top right image
12174 frei0r=perspective:0.2/0.2|0.8/0.2
12178 For more information, see
12179 @url{http://frei0r.dyne.org}
12181 @subsection Commands
12183 This filter supports the @option{filter_params} option as @ref{commands}.
12187 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
12189 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
12190 processing filter, one of them is performed once per block, not per pixel.
12191 This allows for much higher speed.
12193 The filter accepts the following options:
12197 Set quality. This option defines the number of levels for averaging. It accepts
12198 an integer in the range 4-5. Default value is @code{4}.
12201 Force a constant quantization parameter. It accepts an integer in range 0-63.
12202 If not set, the filter will use the QP from the video stream (if available).
12205 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
12206 more details but also more artifacts, while higher values make the image smoother
12207 but also blurrier. Default value is @code{0} − PSNR optimal.
12209 @item use_bframe_qp
12210 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
12211 option may cause flicker since the B-Frames have often larger QP. Default is
12212 @code{0} (not enabled).
12218 Apply Gaussian blur filter.
12220 The filter accepts the following options:
12224 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
12227 Set number of steps for Gaussian approximation. Default is @code{1}.
12230 Set which planes to filter. By default all planes are filtered.
12233 Set vertical sigma, if negative it will be same as @code{sigma}.
12234 Default is @code{-1}.
12237 @subsection Commands
12238 This filter supports same commands as options.
12239 The command accepts the same syntax of the corresponding option.
12241 If the specified expression is not valid, it is kept at its current
12246 Apply generic equation to each pixel.
12248 The filter accepts the following options:
12251 @item lum_expr, lum
12252 Set the luminance expression.
12254 Set the chrominance blue expression.
12256 Set the chrominance red expression.
12257 @item alpha_expr, a
12258 Set the alpha expression.
12260 Set the red expression.
12261 @item green_expr, g
12262 Set the green expression.
12264 Set the blue expression.
12267 The colorspace is selected according to the specified options. If one
12268 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
12269 options is specified, the filter will automatically select a YCbCr
12270 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
12271 @option{blue_expr} options is specified, it will select an RGB
12274 If one of the chrominance expression is not defined, it falls back on the other
12275 one. If no alpha expression is specified it will evaluate to opaque value.
12276 If none of chrominance expressions are specified, they will evaluate
12277 to the luminance expression.
12279 The expressions can use the following variables and functions:
12283 The sequential number of the filtered frame, starting from @code{0}.
12287 The coordinates of the current sample.
12291 The width and height of the image.
12295 Width and height scale depending on the currently filtered plane. It is the
12296 ratio between the corresponding luma plane number of pixels and the current
12297 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
12298 @code{0.5,0.5} for chroma planes.
12301 Time of the current frame, expressed in seconds.
12304 Return the value of the pixel at location (@var{x},@var{y}) of the current
12308 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
12312 Return the value of the pixel at location (@var{x},@var{y}) of the
12313 blue-difference chroma plane. Return 0 if there is no such plane.
12316 Return the value of the pixel at location (@var{x},@var{y}) of the
12317 red-difference chroma plane. Return 0 if there is no such plane.
12322 Return the value of the pixel at location (@var{x},@var{y}) of the
12323 red/green/blue component. Return 0 if there is no such component.
12326 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
12327 plane. Return 0 if there is no such plane.
12329 @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)
12330 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
12331 sums of samples within a rectangle. See the functions without the sum postfix.
12333 @item interpolation
12334 Set one of interpolation methods:
12339 Default is bilinear.
12342 For functions, if @var{x} and @var{y} are outside the area, the value will be
12343 automatically clipped to the closer edge.
12345 Please note that this filter can use multiple threads in which case each slice
12346 will have its own expression state. If you want to use only a single expression
12347 state because your expressions depend on previous state then you should limit
12348 the number of filter threads to 1.
12350 @subsection Examples
12354 Flip the image horizontally:
12360 Generate a bidimensional sine wave, with angle @code{PI/3} and a
12361 wavelength of 100 pixels:
12363 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
12367 Generate a fancy enigmatic moving light:
12369 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
12373 Generate a quick emboss effect:
12375 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
12379 Modify RGB components depending on pixel position:
12381 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
12385 Create a radial gradient that is the same size as the input (also see
12386 the @ref{vignette} filter):
12388 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
12394 Fix the banding artifacts that are sometimes introduced into nearly flat
12395 regions by truncation to 8-bit color depth.
12396 Interpolate the gradients that should go where the bands are, and
12399 It is designed for playback only. Do not use it prior to
12400 lossy compression, because compression tends to lose the dither and
12401 bring back the bands.
12403 It accepts the following parameters:
12408 The maximum amount by which the filter will change any one pixel. This is also
12409 the threshold for detecting nearly flat regions. Acceptable values range from
12410 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
12414 The neighborhood to fit the gradient to. A larger radius makes for smoother
12415 gradients, but also prevents the filter from modifying the pixels near detailed
12416 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
12417 values will be clipped to the valid range.
12421 Alternatively, the options can be specified as a flat string:
12422 @var{strength}[:@var{radius}]
12424 @subsection Examples
12428 Apply the filter with a @code{3.5} strength and radius of @code{8}:
12434 Specify radius, omitting the strength (which will fall-back to the default
12442 @anchor{graphmonitor}
12443 @section graphmonitor
12444 Show various filtergraph stats.
12446 With this filter one can debug complete filtergraph.
12447 Especially issues with links filling with queued frames.
12449 The filter accepts the following options:
12453 Set video output size. Default is @var{hd720}.
12456 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
12459 Set output mode, can be @var{fulll} or @var{compact}.
12460 In @var{compact} mode only filters with some queued frames have displayed stats.
12463 Set flags which enable which stats are shown in video.
12465 Available values for flags are:
12468 Display number of queued frames in each link.
12470 @item frame_count_in
12471 Display number of frames taken from filter.
12473 @item frame_count_out
12474 Display number of frames given out from filter.
12477 Display current filtered frame pts.
12480 Display current filtered frame time.
12483 Display time base for filter link.
12486 Display used format for filter link.
12489 Display video size or number of audio channels in case of audio used by filter link.
12492 Display video frame rate or sample rate in case of audio used by filter link.
12495 Display link output status.
12499 Set upper limit for video rate of output stream, Default value is @var{25}.
12500 This guarantee that output video frame rate will not be higher than this value.
12504 A color constancy variation filter which estimates scene illumination via grey edge algorithm
12505 and corrects the scene colors accordingly.
12507 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
12509 The filter accepts the following options:
12513 The order of differentiation to be applied on the scene. Must be chosen in the range
12514 [0,2] and default value is 1.
12517 The Minkowski parameter to be used for calculating the Minkowski distance. Must
12518 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
12519 max value instead of calculating Minkowski distance.
12522 The standard deviation of Gaussian blur to be applied on the scene. Must be
12523 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
12524 can't be equal to 0 if @var{difford} is greater than 0.
12527 @subsection Examples
12533 greyedge=difford=1:minknorm=5:sigma=2
12539 greyedge=difford=1:minknorm=0:sigma=2
12547 Apply a Hald CLUT to a video stream.
12549 First input is the video stream to process, and second one is the Hald CLUT.
12550 The Hald CLUT input can be a simple picture or a complete video stream.
12552 The filter accepts the following options:
12556 Force termination when the shortest input terminates. Default is @code{0}.
12558 Continue applying the last CLUT after the end of the stream. A value of
12559 @code{0} disable the filter after the last frame of the CLUT is reached.
12560 Default is @code{1}.
12563 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
12564 filters share the same internals).
12566 This filter also supports the @ref{framesync} options.
12568 More information about the Hald CLUT can be found on Eskil Steenberg's website
12569 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
12571 @subsection Workflow examples
12573 @subsubsection Hald CLUT video stream
12575 Generate an identity Hald CLUT stream altered with various effects:
12577 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
12580 Note: make sure you use a lossless codec.
12582 Then use it with @code{haldclut} to apply it on some random stream:
12584 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
12587 The Hald CLUT will be applied to the 10 first seconds (duration of
12588 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
12589 to the remaining frames of the @code{mandelbrot} stream.
12591 @subsubsection Hald CLUT with preview
12593 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
12594 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
12595 biggest possible square starting at the top left of the picture. The remaining
12596 padding pixels (bottom or right) will be ignored. This area can be used to add
12597 a preview of the Hald CLUT.
12599 Typically, the following generated Hald CLUT will be supported by the
12600 @code{haldclut} filter:
12603 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
12604 pad=iw+320 [padded_clut];
12605 smptebars=s=320x256, split [a][b];
12606 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
12607 [main][b] overlay=W-320" -frames:v 1 clut.png
12610 It contains the original and a preview of the effect of the CLUT: SMPTE color
12611 bars are displayed on the right-top, and below the same color bars processed by
12614 Then, the effect of this Hald CLUT can be visualized with:
12616 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
12621 Flip the input video horizontally.
12623 For example, to horizontally flip the input video with @command{ffmpeg}:
12625 ffmpeg -i in.avi -vf "hflip" out.avi
12629 This filter applies a global color histogram equalization on a
12632 It can be used to correct video that has a compressed range of pixel
12633 intensities. The filter redistributes the pixel intensities to
12634 equalize their distribution across the intensity range. It may be
12635 viewed as an "automatically adjusting contrast filter". This filter is
12636 useful only for correcting degraded or poorly captured source
12639 The filter accepts the following options:
12643 Determine the amount of equalization to be applied. As the strength
12644 is reduced, the distribution of pixel intensities more-and-more
12645 approaches that of the input frame. The value must be a float number
12646 in the range [0,1] and defaults to 0.200.
12649 Set the maximum intensity that can generated and scale the output
12650 values appropriately. The strength should be set as desired and then
12651 the intensity can be limited if needed to avoid washing-out. The value
12652 must be a float number in the range [0,1] and defaults to 0.210.
12655 Set the antibanding level. If enabled the filter will randomly vary
12656 the luminance of output pixels by a small amount to avoid banding of
12657 the histogram. Possible values are @code{none}, @code{weak} or
12658 @code{strong}. It defaults to @code{none}.
12664 Compute and draw a color distribution histogram for the input video.
12666 The computed histogram is a representation of the color component
12667 distribution in an image.
12669 Standard histogram displays the color components distribution in an image.
12670 Displays color graph for each color component. Shows distribution of
12671 the Y, U, V, A or R, G, B components, depending on input format, in the
12672 current frame. Below each graph a color component scale meter is shown.
12674 The filter accepts the following options:
12678 Set height of level. Default value is @code{200}.
12679 Allowed range is [50, 2048].
12682 Set height of color scale. Default value is @code{12}.
12683 Allowed range is [0, 40].
12687 It accepts the following values:
12690 Per color component graphs are placed below each other.
12693 Per color component graphs are placed side by side.
12696 Presents information identical to that in the @code{parade}, except
12697 that the graphs representing color components are superimposed directly
12700 Default is @code{stack}.
12703 Set mode. Can be either @code{linear}, or @code{logarithmic}.
12704 Default is @code{linear}.
12707 Set what color components to display.
12708 Default is @code{7}.
12711 Set foreground opacity. Default is @code{0.7}.
12714 Set background opacity. Default is @code{0.5}.
12717 @subsection Examples
12722 Calculate and draw histogram:
12724 ffplay -i input -vf histogram
12732 This is a high precision/quality 3d denoise filter. It aims to reduce
12733 image noise, producing smooth images and making still images really
12734 still. It should enhance compressibility.
12736 It accepts the following optional parameters:
12740 A non-negative floating point number which specifies spatial luma strength.
12741 It defaults to 4.0.
12743 @item chroma_spatial
12744 A non-negative floating point number which specifies spatial chroma strength.
12745 It defaults to 3.0*@var{luma_spatial}/4.0.
12748 A floating point number which specifies luma temporal strength. It defaults to
12749 6.0*@var{luma_spatial}/4.0.
12752 A floating point number which specifies chroma temporal strength. It defaults to
12753 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
12756 @subsection Commands
12757 This filter supports same @ref{commands} as options.
12758 The command accepts the same syntax of the corresponding option.
12760 If the specified expression is not valid, it is kept at its current
12763 @anchor{hwdownload}
12764 @section hwdownload
12766 Download hardware frames to system memory.
12768 The input must be in hardware frames, and the output a non-hardware format.
12769 Not all formats will be supported on the output - it may be necessary to insert
12770 an additional @option{format} filter immediately following in the graph to get
12771 the output in a supported format.
12775 Map hardware frames to system memory or to another device.
12777 This filter has several different modes of operation; which one is used depends
12778 on the input and output formats:
12781 Hardware frame input, normal frame output
12783 Map the input frames to system memory and pass them to the output. If the
12784 original hardware frame is later required (for example, after overlaying
12785 something else on part of it), the @option{hwmap} filter can be used again
12786 in the next mode to retrieve it.
12788 Normal frame input, hardware frame output
12790 If the input is actually a software-mapped hardware frame, then unmap it -
12791 that is, return the original hardware frame.
12793 Otherwise, a device must be provided. Create new hardware surfaces on that
12794 device for the output, then map them back to the software format at the input
12795 and give those frames to the preceding filter. This will then act like the
12796 @option{hwupload} filter, but may be able to avoid an additional copy when
12797 the input is already in a compatible format.
12799 Hardware frame input and output
12801 A device must be supplied for the output, either directly or with the
12802 @option{derive_device} option. The input and output devices must be of
12803 different types and compatible - the exact meaning of this is
12804 system-dependent, but typically it means that they must refer to the same
12805 underlying hardware context (for example, refer to the same graphics card).
12807 If the input frames were originally created on the output device, then unmap
12808 to retrieve the original frames.
12810 Otherwise, map the frames to the output device - create new hardware frames
12811 on the output corresponding to the frames on the input.
12814 The following additional parameters are accepted:
12818 Set the frame mapping mode. Some combination of:
12821 The mapped frame should be readable.
12823 The mapped frame should be writeable.
12825 The mapping will always overwrite the entire frame.
12827 This may improve performance in some cases, as the original contents of the
12828 frame need not be loaded.
12830 The mapping must not involve any copying.
12832 Indirect mappings to copies of frames are created in some cases where either
12833 direct mapping is not possible or it would have unexpected properties.
12834 Setting this flag ensures that the mapping is direct and will fail if that is
12837 Defaults to @var{read+write} if not specified.
12839 @item derive_device @var{type}
12840 Rather than using the device supplied at initialisation, instead derive a new
12841 device of type @var{type} from the device the input frames exist on.
12844 In a hardware to hardware mapping, map in reverse - create frames in the sink
12845 and map them back to the source. This may be necessary in some cases where
12846 a mapping in one direction is required but only the opposite direction is
12847 supported by the devices being used.
12849 This option is dangerous - it may break the preceding filter in undefined
12850 ways if there are any additional constraints on that filter's output.
12851 Do not use it without fully understanding the implications of its use.
12857 Upload system memory frames to hardware surfaces.
12859 The device to upload to must be supplied when the filter is initialised. If
12860 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
12861 option or with the @option{derive_device} option. The input and output devices
12862 must be of different types and compatible - the exact meaning of this is
12863 system-dependent, but typically it means that they must refer to the same
12864 underlying hardware context (for example, refer to the same graphics card).
12866 The following additional parameters are accepted:
12869 @item derive_device @var{type}
12870 Rather than using the device supplied at initialisation, instead derive a new
12871 device of type @var{type} from the device the input frames exist on.
12874 @anchor{hwupload_cuda}
12875 @section hwupload_cuda
12877 Upload system memory frames to a CUDA device.
12879 It accepts the following optional parameters:
12883 The number of the CUDA device to use
12888 Apply a high-quality magnification filter designed for pixel art. This filter
12889 was originally created by Maxim Stepin.
12891 It accepts the following option:
12895 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
12896 @code{hq3x} and @code{4} for @code{hq4x}.
12897 Default is @code{3}.
12901 Stack input videos horizontally.
12903 All streams must be of same pixel format and of same height.
12905 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
12906 to create same output.
12908 The filter accepts the following option:
12912 Set number of input streams. Default is 2.
12915 If set to 1, force the output to terminate when the shortest input
12916 terminates. Default value is 0.
12921 Modify the hue and/or the saturation of the input.
12923 It accepts the following parameters:
12927 Specify the hue angle as a number of degrees. It accepts an expression,
12928 and defaults to "0".
12931 Specify the saturation in the [-10,10] range. It accepts an expression and
12935 Specify the hue angle as a number of radians. It accepts an
12936 expression, and defaults to "0".
12939 Specify the brightness in the [-10,10] range. It accepts an expression and
12943 @option{h} and @option{H} are mutually exclusive, and can't be
12944 specified at the same time.
12946 The @option{b}, @option{h}, @option{H} and @option{s} option values are
12947 expressions containing the following constants:
12951 frame count of the input frame starting from 0
12954 presentation timestamp of the input frame expressed in time base units
12957 frame rate of the input video, NAN if the input frame rate is unknown
12960 timestamp expressed in seconds, NAN if the input timestamp is unknown
12963 time base of the input video
12966 @subsection Examples
12970 Set the hue to 90 degrees and the saturation to 1.0:
12976 Same command but expressing the hue in radians:
12982 Rotate hue and make the saturation swing between 0
12983 and 2 over a period of 1 second:
12985 hue="H=2*PI*t: s=sin(2*PI*t)+1"
12989 Apply a 3 seconds saturation fade-in effect starting at 0:
12991 hue="s=min(t/3\,1)"
12994 The general fade-in expression can be written as:
12996 hue="s=min(0\, max((t-START)/DURATION\, 1))"
13000 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
13002 hue="s=max(0\, min(1\, (8-t)/3))"
13005 The general fade-out expression can be written as:
13007 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
13012 @subsection Commands
13014 This filter supports the following commands:
13020 Modify the hue and/or the saturation and/or brightness of the input video.
13021 The command accepts the same syntax of the corresponding option.
13023 If the specified expression is not valid, it is kept at its current
13027 @section hysteresis
13029 Grow first stream into second stream by connecting components.
13030 This makes it possible to build more robust edge masks.
13032 This filter accepts the following options:
13036 Set which planes will be processed as bitmap, unprocessed planes will be
13037 copied from first stream.
13038 By default value 0xf, all planes will be processed.
13041 Set threshold which is used in filtering. If pixel component value is higher than
13042 this value filter algorithm for connecting components is activated.
13043 By default value is 0.
13046 The @code{hysteresis} filter also supports the @ref{framesync} options.
13050 Detect video interlacing type.
13052 This filter tries to detect if the input frames are interlaced, progressive,
13053 top or bottom field first. It will also try to detect fields that are
13054 repeated between adjacent frames (a sign of telecine).
13056 Single frame detection considers only immediately adjacent frames when classifying each frame.
13057 Multiple frame detection incorporates the classification history of previous frames.
13059 The filter will log these metadata values:
13062 @item single.current_frame
13063 Detected type of current frame using single-frame detection. One of:
13064 ``tff'' (top field first), ``bff'' (bottom field first),
13065 ``progressive'', or ``undetermined''
13068 Cumulative number of frames detected as top field first using single-frame detection.
13071 Cumulative number of frames detected as top field first using multiple-frame detection.
13074 Cumulative number of frames detected as bottom field first using single-frame detection.
13076 @item multiple.current_frame
13077 Detected type of current frame using multiple-frame detection. One of:
13078 ``tff'' (top field first), ``bff'' (bottom field first),
13079 ``progressive'', or ``undetermined''
13082 Cumulative number of frames detected as bottom field first using multiple-frame detection.
13084 @item single.progressive
13085 Cumulative number of frames detected as progressive using single-frame detection.
13087 @item multiple.progressive
13088 Cumulative number of frames detected as progressive using multiple-frame detection.
13090 @item single.undetermined
13091 Cumulative number of frames that could not be classified using single-frame detection.
13093 @item multiple.undetermined
13094 Cumulative number of frames that could not be classified using multiple-frame detection.
13096 @item repeated.current_frame
13097 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
13099 @item repeated.neither
13100 Cumulative number of frames with no repeated field.
13103 Cumulative number of frames with the top field repeated from the previous frame's top field.
13105 @item repeated.bottom
13106 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
13109 The filter accepts the following options:
13113 Set interlacing threshold.
13115 Set progressive threshold.
13117 Threshold for repeated field detection.
13119 Number of frames after which a given frame's contribution to the
13120 statistics is halved (i.e., it contributes only 0.5 to its
13121 classification). The default of 0 means that all frames seen are given
13122 full weight of 1.0 forever.
13123 @item analyze_interlaced_flag
13124 When this is not 0 then idet will use the specified number of frames to determine
13125 if the interlaced flag is accurate, it will not count undetermined frames.
13126 If the flag is found to be accurate it will be used without any further
13127 computations, if it is found to be inaccurate it will be cleared without any
13128 further computations. This allows inserting the idet filter as a low computational
13129 method to clean up the interlaced flag
13134 Deinterleave or interleave fields.
13136 This filter allows one to process interlaced images fields without
13137 deinterlacing them. Deinterleaving splits the input frame into 2
13138 fields (so called half pictures). Odd lines are moved to the top
13139 half of the output image, even lines to the bottom half.
13140 You can process (filter) them independently and then re-interleave them.
13142 The filter accepts the following options:
13146 @item chroma_mode, c
13147 @item alpha_mode, a
13148 Available values for @var{luma_mode}, @var{chroma_mode} and
13149 @var{alpha_mode} are:
13155 @item deinterleave, d
13156 Deinterleave fields, placing one above the other.
13158 @item interleave, i
13159 Interleave fields. Reverse the effect of deinterleaving.
13161 Default value is @code{none}.
13163 @item luma_swap, ls
13164 @item chroma_swap, cs
13165 @item alpha_swap, as
13166 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
13169 @subsection Commands
13171 This filter supports the all above options as @ref{commands}.
13175 Apply inflate effect to the video.
13177 This filter replaces the pixel by the local(3x3) average by taking into account
13178 only values higher than the pixel.
13180 It accepts the following options:
13187 Limit the maximum change for each plane, default is 65535.
13188 If 0, plane will remain unchanged.
13191 @subsection Commands
13193 This filter supports the all above options as @ref{commands}.
13197 Simple interlacing filter from progressive contents. This interleaves upper (or
13198 lower) lines from odd frames with lower (or upper) lines from even frames,
13199 halving the frame rate and preserving image height.
13202 Original Original New Frame
13203 Frame 'j' Frame 'j+1' (tff)
13204 ========== =========== ==================
13205 Line 0 --------------------> Frame 'j' Line 0
13206 Line 1 Line 1 ----> Frame 'j+1' Line 1
13207 Line 2 ---------------------> Frame 'j' Line 2
13208 Line 3 Line 3 ----> Frame 'j+1' Line 3
13210 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
13213 It accepts the following optional parameters:
13217 This determines whether the interlaced frame is taken from the even
13218 (tff - default) or odd (bff) lines of the progressive frame.
13221 Vertical lowpass filter to avoid twitter interlacing and
13222 reduce moire patterns.
13226 Disable vertical lowpass filter
13229 Enable linear filter (default)
13232 Enable complex filter. This will slightly less reduce twitter and moire
13233 but better retain detail and subjective sharpness impression.
13240 Deinterlace input video by applying Donald Graft's adaptive kernel
13241 deinterling. Work on interlaced parts of a video to produce
13242 progressive frames.
13244 The description of the accepted parameters follows.
13248 Set the threshold which affects the filter's tolerance when
13249 determining if a pixel line must be processed. It must be an integer
13250 in the range [0,255] and defaults to 10. A value of 0 will result in
13251 applying the process on every pixels.
13254 Paint pixels exceeding the threshold value to white if set to 1.
13258 Set the fields order. Swap fields if set to 1, leave fields alone if
13262 Enable additional sharpening if set to 1. Default is 0.
13265 Enable twoway sharpening if set to 1. Default is 0.
13268 @subsection Examples
13272 Apply default values:
13274 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
13278 Enable additional sharpening:
13284 Paint processed pixels in white:
13292 Slowly update darker pixels.
13294 This filter makes short flashes of light appear longer.
13295 This filter accepts the following options:
13299 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
13302 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
13305 @section lenscorrection
13307 Correct radial lens distortion
13309 This filter can be used to correct for radial distortion as can result from the use
13310 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
13311 one can use tools available for example as part of opencv or simply trial-and-error.
13312 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
13313 and extract the k1 and k2 coefficients from the resulting matrix.
13315 Note that effectively the same filter is available in the open-source tools Krita and
13316 Digikam from the KDE project.
13318 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
13319 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
13320 brightness distribution, so you may want to use both filters together in certain
13321 cases, though you will have to take care of ordering, i.e. whether vignetting should
13322 be applied before or after lens correction.
13324 @subsection Options
13326 The filter accepts the following options:
13330 Relative x-coordinate of the focal point of the image, and thereby the center of the
13331 distortion. This value has a range [0,1] and is expressed as fractions of the image
13332 width. Default is 0.5.
13334 Relative y-coordinate of the focal point of the image, and thereby the center of the
13335 distortion. This value has a range [0,1] and is expressed as fractions of the image
13336 height. Default is 0.5.
13338 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
13339 no correction. Default is 0.
13341 Coefficient of the double quadratic correction term. This value has a range [-1,1].
13342 0 means no correction. Default is 0.
13345 The formula that generates the correction is:
13347 @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)
13349 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
13350 distances from the focal point in the source and target images, respectively.
13354 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
13356 The @code{lensfun} filter requires the camera make, camera model, and lens model
13357 to apply the lens correction. The filter will load the lensfun database and
13358 query it to find the corresponding camera and lens entries in the database. As
13359 long as these entries can be found with the given options, the filter can
13360 perform corrections on frames. Note that incomplete strings will result in the
13361 filter choosing the best match with the given options, and the filter will
13362 output the chosen camera and lens models (logged with level "info"). You must
13363 provide the make, camera model, and lens model as they are required.
13365 The filter accepts the following options:
13369 The make of the camera (for example, "Canon"). This option is required.
13372 The model of the camera (for example, "Canon EOS 100D"). This option is
13376 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
13377 option is required.
13380 The type of correction to apply. The following values are valid options:
13384 Enables fixing lens vignetting.
13387 Enables fixing lens geometry. This is the default.
13390 Enables fixing chromatic aberrations.
13393 Enables fixing lens vignetting and lens geometry.
13396 Enables fixing lens vignetting and chromatic aberrations.
13399 Enables fixing both lens geometry and chromatic aberrations.
13402 Enables all possible corrections.
13406 The focal length of the image/video (zoom; expected constant for video). For
13407 example, a 18--55mm lens has focal length range of [18--55], so a value in that
13408 range should be chosen when using that lens. Default 18.
13411 The aperture of the image/video (expected constant for video). Note that
13412 aperture is only used for vignetting correction. Default 3.5.
13414 @item focus_distance
13415 The focus distance of the image/video (expected constant for video). Note that
13416 focus distance is only used for vignetting and only slightly affects the
13417 vignetting correction process. If unknown, leave it at the default value (which
13421 The scale factor which is applied after transformation. After correction the
13422 video is no longer necessarily rectangular. This parameter controls how much of
13423 the resulting image is visible. The value 0 means that a value will be chosen
13424 automatically such that there is little or no unmapped area in the output
13425 image. 1.0 means that no additional scaling is done. Lower values may result
13426 in more of the corrected image being visible, while higher values may avoid
13427 unmapped areas in the output.
13429 @item target_geometry
13430 The target geometry of the output image/video. The following values are valid
13434 @item rectilinear (default)
13437 @item equirectangular
13438 @item fisheye_orthographic
13439 @item fisheye_stereographic
13440 @item fisheye_equisolid
13441 @item fisheye_thoby
13444 Apply the reverse of image correction (instead of correcting distortion, apply
13447 @item interpolation
13448 The type of interpolation used when correcting distortion. The following values
13453 @item linear (default)
13458 @subsection Examples
13462 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
13463 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
13467 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
13471 Apply the same as before, but only for the first 5 seconds of video.
13474 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
13481 Obtain the VMAF (Video Multi-Method Assessment Fusion)
13482 score between two input videos.
13484 The obtained VMAF score is printed through the logging system.
13486 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
13487 After installing the library it can be enabled using:
13488 @code{./configure --enable-libvmaf}.
13489 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
13491 The filter has following options:
13495 Set the model path which is to be used for SVM.
13496 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
13499 Set the file path to be used to store logs.
13502 Set the format of the log file (csv, json or xml).
13504 @item enable_transform
13505 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
13506 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
13507 Default value: @code{false}
13510 Invokes the phone model which will generate VMAF scores higher than in the
13511 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
13512 Default value: @code{false}
13515 Enables computing psnr along with vmaf.
13516 Default value: @code{false}
13519 Enables computing ssim along with vmaf.
13520 Default value: @code{false}
13523 Enables computing ms_ssim along with vmaf.
13524 Default value: @code{false}
13527 Set the pool method to be used for computing vmaf.
13528 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
13531 Set number of threads to be used when computing vmaf.
13532 Default value: @code{0}, which makes use of all available logical processors.
13535 Set interval for frame subsampling used when computing vmaf.
13536 Default value: @code{1}
13538 @item enable_conf_interval
13539 Enables confidence interval.
13540 Default value: @code{false}
13543 This filter also supports the @ref{framesync} options.
13545 @subsection Examples
13548 On the below examples the input file @file{main.mpg} being processed is
13549 compared with the reference file @file{ref.mpg}.
13552 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
13556 Example with options:
13558 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
13562 Example with options and different containers:
13564 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 -
13570 Limits the pixel components values to the specified range [min, max].
13572 The filter accepts the following options:
13576 Lower bound. Defaults to the lowest allowed value for the input.
13579 Upper bound. Defaults to the highest allowed value for the input.
13582 Specify which planes will be processed. Defaults to all available.
13585 @subsection Commands
13587 This filter supports the all above options as @ref{commands}.
13593 The filter accepts the following options:
13597 Set the number of loops. Setting this value to -1 will result in infinite loops.
13601 Set maximal size in number of frames. Default is 0.
13604 Set first frame of loop. Default is 0.
13607 @subsection Examples
13611 Loop single first frame infinitely:
13613 loop=loop=-1:size=1:start=0
13617 Loop single first frame 10 times:
13619 loop=loop=10:size=1:start=0
13623 Loop 10 first frames 5 times:
13625 loop=loop=5:size=10:start=0
13631 Apply a 1D LUT to an input video.
13633 The filter accepts the following options:
13637 Set the 1D LUT file name.
13639 Currently supported formats:
13648 Select interpolation mode.
13650 Available values are:
13654 Use values from the nearest defined point.
13656 Interpolate values using the linear interpolation.
13658 Interpolate values using the cosine interpolation.
13660 Interpolate values using the cubic interpolation.
13662 Interpolate values using the spline interpolation.
13669 Apply a 3D LUT to an input video.
13671 The filter accepts the following options:
13675 Set the 3D LUT file name.
13677 Currently supported formats:
13691 Select interpolation mode.
13693 Available values are:
13697 Use values from the nearest defined point.
13699 Interpolate values using the 8 points defining a cube.
13701 Interpolate values using a tetrahedron.
13707 Turn certain luma values into transparency.
13709 The filter accepts the following options:
13713 Set the luma which will be used as base for transparency.
13714 Default value is @code{0}.
13717 Set the range of luma values to be keyed out.
13718 Default value is @code{0.01}.
13721 Set the range of softness. Default value is @code{0}.
13722 Use this to control gradual transition from zero to full transparency.
13725 @subsection Commands
13726 This filter supports same @ref{commands} as options.
13727 The command accepts the same syntax of the corresponding option.
13729 If the specified expression is not valid, it is kept at its current
13732 @section lut, lutrgb, lutyuv
13734 Compute a look-up table for binding each pixel component input value
13735 to an output value, and apply it to the input video.
13737 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
13738 to an RGB input video.
13740 These filters accept the following parameters:
13743 set first pixel component expression
13745 set second pixel component expression
13747 set third pixel component expression
13749 set fourth pixel component expression, corresponds to the alpha component
13752 set red component expression
13754 set green component expression
13756 set blue component expression
13758 alpha component expression
13761 set Y/luminance component expression
13763 set U/Cb component expression
13765 set V/Cr component expression
13768 Each of them specifies the expression to use for computing the lookup table for
13769 the corresponding pixel component values.
13771 The exact component associated to each of the @var{c*} options depends on the
13774 The @var{lut} filter requires either YUV or RGB pixel formats in input,
13775 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
13777 The expressions can contain the following constants and functions:
13782 The input width and height.
13785 The input value for the pixel component.
13788 The input value, clipped to the @var{minval}-@var{maxval} range.
13791 The maximum value for the pixel component.
13794 The minimum value for the pixel component.
13797 The negated value for the pixel component value, clipped to the
13798 @var{minval}-@var{maxval} range; it corresponds to the expression
13799 "maxval-clipval+minval".
13802 The computed value in @var{val}, clipped to the
13803 @var{minval}-@var{maxval} range.
13805 @item gammaval(gamma)
13806 The computed gamma correction value of the pixel component value,
13807 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
13809 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
13813 All expressions default to "val".
13815 @subsection Examples
13819 Negate input video:
13821 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
13822 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
13825 The above is the same as:
13827 lutrgb="r=negval:g=negval:b=negval"
13828 lutyuv="y=negval:u=negval:v=negval"
13838 Remove chroma components, turning the video into a graytone image:
13840 lutyuv="u=128:v=128"
13844 Apply a luma burning effect:
13850 Remove green and blue components:
13856 Set a constant alpha channel value on input:
13858 format=rgba,lutrgb=a="maxval-minval/2"
13862 Correct luminance gamma by a factor of 0.5:
13864 lutyuv=y=gammaval(0.5)
13868 Discard least significant bits of luma:
13870 lutyuv=y='bitand(val, 128+64+32)'
13874 Technicolor like effect:
13876 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
13880 @section lut2, tlut2
13882 The @code{lut2} filter takes two input streams and outputs one
13885 The @code{tlut2} (time lut2) filter takes two consecutive frames
13886 from one single stream.
13888 This filter accepts the following parameters:
13891 set first pixel component expression
13893 set second pixel component expression
13895 set third pixel component expression
13897 set fourth pixel component expression, corresponds to the alpha component
13900 set output bit depth, only available for @code{lut2} filter. By default is 0,
13901 which means bit depth is automatically picked from first input format.
13904 The @code{lut2} filter also supports the @ref{framesync} options.
13906 Each of them specifies the expression to use for computing the lookup table for
13907 the corresponding pixel component values.
13909 The exact component associated to each of the @var{c*} options depends on the
13912 The expressions can contain the following constants:
13917 The input width and height.
13920 The first input value for the pixel component.
13923 The second input value for the pixel component.
13926 The first input video bit depth.
13929 The second input video bit depth.
13932 All expressions default to "x".
13934 @subsection Examples
13938 Highlight differences between two RGB video streams:
13940 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)'
13944 Highlight differences between two YUV video streams:
13946 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)'
13950 Show max difference between two video streams:
13952 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)))'
13956 @section maskedclamp
13958 Clamp the first input stream with the second input and third input stream.
13960 Returns the value of first stream to be between second input
13961 stream - @code{undershoot} and third input stream + @code{overshoot}.
13963 This filter accepts the following options:
13966 Default value is @code{0}.
13969 Default value is @code{0}.
13972 Set which planes will be processed as bitmap, unprocessed planes will be
13973 copied from first stream.
13974 By default value 0xf, all planes will be processed.
13977 @subsection Commands
13979 This filter supports the all above options as @ref{commands}.
13983 Merge the second and third input stream into output stream using absolute differences
13984 between second input stream and first input stream and absolute difference between
13985 third input stream and first input stream. The picked value will be from second input
13986 stream if second absolute difference is greater than first one or from third input stream
13989 This filter accepts the following options:
13992 Set which planes will be processed as bitmap, unprocessed planes will be
13993 copied from first stream.
13994 By default value 0xf, all planes will be processed.
13997 @subsection Commands
13999 This filter supports the all above options as @ref{commands}.
14001 @section maskedmerge
14003 Merge the first input stream with the second input stream using per pixel
14004 weights in the third input stream.
14006 A value of 0 in the third stream pixel component means that pixel component
14007 from first stream is returned unchanged, while maximum value (eg. 255 for
14008 8-bit videos) means that pixel component from second stream is returned
14009 unchanged. Intermediate values define the amount of merging between both
14010 input stream's pixel components.
14012 This filter accepts the following options:
14015 Set which planes will be processed as bitmap, unprocessed planes will be
14016 copied from first stream.
14017 By default value 0xf, all planes will be processed.
14022 Merge the second and third input stream into output stream using absolute differences
14023 between second input stream and first input stream and absolute difference between
14024 third input stream and first input stream. The picked value will be from second input
14025 stream if second absolute difference is less than first one or from third input stream
14028 This filter accepts the following options:
14031 Set which planes will be processed as bitmap, unprocessed planes will be
14032 copied from first stream.
14033 By default value 0xf, all planes will be processed.
14036 @subsection Commands
14038 This filter supports the all above options as @ref{commands}.
14040 @section maskedthreshold
14041 Pick pixels comparing absolute difference of two video streams with fixed
14044 If absolute difference between pixel component of first and second video
14045 stream is equal or lower than user supplied threshold than pixel component
14046 from first video stream is picked, otherwise pixel component from second
14047 video stream is picked.
14049 This filter accepts the following options:
14052 Set threshold used when picking pixels from absolute difference from two input
14056 Set which planes will be processed as bitmap, unprocessed planes will be
14057 copied from second stream.
14058 By default value 0xf, all planes will be processed.
14061 @subsection Commands
14063 This filter supports the all above options as @ref{commands}.
14066 Create mask from input video.
14068 For example it is useful to create motion masks after @code{tblend} filter.
14070 This filter accepts the following options:
14074 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
14077 Set high threshold. Any pixel component higher than this value will be set to max value
14078 allowed for current pixel format.
14081 Set planes to filter, by default all available planes are filtered.
14084 Fill all frame pixels with this value.
14087 Set max average pixel value for frame. If sum of all pixel components is higher that this
14088 average, output frame will be completely filled with value set by @var{fill} option.
14089 Typically useful for scene changes when used in combination with @code{tblend} filter.
14094 Apply motion-compensation deinterlacing.
14096 It needs one field per frame as input and must thus be used together
14097 with yadif=1/3 or equivalent.
14099 This filter accepts the following options:
14102 Set the deinterlacing mode.
14104 It accepts one of the following values:
14109 use iterative motion estimation
14111 like @samp{slow}, but use multiple reference frames.
14113 Default value is @samp{fast}.
14116 Set the picture field parity assumed for the input video. It must be
14117 one of the following values:
14121 assume top field first
14123 assume bottom field first
14126 Default value is @samp{bff}.
14129 Set per-block quantization parameter (QP) used by the internal
14132 Higher values should result in a smoother motion vector field but less
14133 optimal individual vectors. Default value is 1.
14138 Pick median pixel from certain rectangle defined by radius.
14140 This filter accepts the following options:
14144 Set horizontal radius size. Default value is @code{1}.
14145 Allowed range is integer from 1 to 127.
14148 Set which planes to process. Default is @code{15}, which is all available planes.
14151 Set vertical radius size. Default value is @code{0}.
14152 Allowed range is integer from 0 to 127.
14153 If it is 0, value will be picked from horizontal @code{radius} option.
14156 Set median percentile. Default value is @code{0.5}.
14157 Default value of @code{0.5} will pick always median values, while @code{0} will pick
14158 minimum values, and @code{1} maximum values.
14161 @subsection Commands
14162 This filter supports same @ref{commands} as options.
14163 The command accepts the same syntax of the corresponding option.
14165 If the specified expression is not valid, it is kept at its current
14168 @section mergeplanes
14170 Merge color channel components from several video streams.
14172 The filter accepts up to 4 input streams, and merge selected input
14173 planes to the output video.
14175 This filter accepts the following options:
14178 Set input to output plane mapping. Default is @code{0}.
14180 The mappings is specified as a bitmap. It should be specified as a
14181 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
14182 mapping for the first plane of the output stream. 'A' sets the number of
14183 the input stream to use (from 0 to 3), and 'a' the plane number of the
14184 corresponding input to use (from 0 to 3). The rest of the mappings is
14185 similar, 'Bb' describes the mapping for the output stream second
14186 plane, 'Cc' describes the mapping for the output stream third plane and
14187 'Dd' describes the mapping for the output stream fourth plane.
14190 Set output pixel format. Default is @code{yuva444p}.
14193 @subsection Examples
14197 Merge three gray video streams of same width and height into single video stream:
14199 [a0][a1][a2]mergeplanes=0x001020:yuv444p
14203 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
14205 [a0][a1]mergeplanes=0x00010210:yuva444p
14209 Swap Y and A plane in yuva444p stream:
14211 format=yuva444p,mergeplanes=0x03010200:yuva444p
14215 Swap U and V plane in yuv420p stream:
14217 format=yuv420p,mergeplanes=0x000201:yuv420p
14221 Cast a rgb24 clip to yuv444p:
14223 format=rgb24,mergeplanes=0x000102:yuv444p
14229 Estimate and export motion vectors using block matching algorithms.
14230 Motion vectors are stored in frame side data to be used by other filters.
14232 This filter accepts the following options:
14235 Specify the motion estimation method. Accepts one of the following values:
14239 Exhaustive search algorithm.
14241 Three step search algorithm.
14243 Two dimensional logarithmic search algorithm.
14245 New three step search algorithm.
14247 Four step search algorithm.
14249 Diamond search algorithm.
14251 Hexagon-based search algorithm.
14253 Enhanced predictive zonal search algorithm.
14255 Uneven multi-hexagon search algorithm.
14257 Default value is @samp{esa}.
14260 Macroblock size. Default @code{16}.
14263 Search parameter. Default @code{7}.
14266 @section midequalizer
14268 Apply Midway Image Equalization effect using two video streams.
14270 Midway Image Equalization adjusts a pair of images to have the same
14271 histogram, while maintaining their dynamics as much as possible. It's
14272 useful for e.g. matching exposures from a pair of stereo cameras.
14274 This filter has two inputs and one output, which must be of same pixel format, but
14275 may be of different sizes. The output of filter is first input adjusted with
14276 midway histogram of both inputs.
14278 This filter accepts the following option:
14282 Set which planes to process. Default is @code{15}, which is all available planes.
14285 @section minterpolate
14287 Convert the video to specified frame rate using motion interpolation.
14289 This filter accepts the following options:
14292 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}.
14295 Motion interpolation mode. Following values are accepted:
14298 Duplicate previous or next frame for interpolating new ones.
14300 Blend source frames. Interpolated frame is mean of previous and next frames.
14302 Motion compensated interpolation. Following options are effective when this mode is selected:
14306 Motion compensation mode. Following values are accepted:
14309 Overlapped block motion compensation.
14311 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
14313 Default mode is @samp{obmc}.
14316 Motion estimation mode. Following values are accepted:
14319 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
14321 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
14323 Default mode is @samp{bilat}.
14326 The algorithm to be used for motion estimation. Following values are accepted:
14329 Exhaustive search algorithm.
14331 Three step search algorithm.
14333 Two dimensional logarithmic search algorithm.
14335 New three step search algorithm.
14337 Four step search algorithm.
14339 Diamond search algorithm.
14341 Hexagon-based search algorithm.
14343 Enhanced predictive zonal search algorithm.
14345 Uneven multi-hexagon search algorithm.
14347 Default algorithm is @samp{epzs}.
14350 Macroblock size. Default @code{16}.
14353 Motion estimation search parameter. Default @code{32}.
14356 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).
14361 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:
14364 Disable scene change detection.
14366 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
14368 Default method is @samp{fdiff}.
14370 @item scd_threshold
14371 Scene change detection threshold. Default is @code{10.}.
14376 Mix several video input streams into one video stream.
14378 A description of the accepted options follows.
14382 The number of inputs. If unspecified, it defaults to 2.
14385 Specify weight of each input video stream as sequence.
14386 Each weight is separated by space. If number of weights
14387 is smaller than number of @var{frames} last specified
14388 weight will be used for all remaining unset weights.
14391 Specify scale, if it is set it will be multiplied with sum
14392 of each weight multiplied with pixel values to give final destination
14393 pixel value. By default @var{scale} is auto scaled to sum of weights.
14396 Specify how end of stream is determined.
14399 The duration of the longest input. (default)
14402 The duration of the shortest input.
14405 The duration of the first input.
14409 @section mpdecimate
14411 Drop frames that do not differ greatly from the previous frame in
14412 order to reduce frame rate.
14414 The main use of this filter is for very-low-bitrate encoding
14415 (e.g. streaming over dialup modem), but it could in theory be used for
14416 fixing movies that were inverse-telecined incorrectly.
14418 A description of the accepted options follows.
14422 Set the maximum number of consecutive frames which can be dropped (if
14423 positive), or the minimum interval between dropped frames (if
14424 negative). If the value is 0, the frame is dropped disregarding the
14425 number of previous sequentially dropped frames.
14427 Default value is 0.
14432 Set the dropping threshold values.
14434 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
14435 represent actual pixel value differences, so a threshold of 64
14436 corresponds to 1 unit of difference for each pixel, or the same spread
14437 out differently over the block.
14439 A frame is a candidate for dropping if no 8x8 blocks differ by more
14440 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
14441 meaning the whole image) differ by more than a threshold of @option{lo}.
14443 Default value for @option{hi} is 64*12, default value for @option{lo} is
14444 64*5, and default value for @option{frac} is 0.33.
14450 Negate (invert) the input video.
14452 It accepts the following option:
14457 With value 1, it negates the alpha component, if present. Default value is 0.
14463 Denoise frames using Non-Local Means algorithm.
14465 Each pixel is adjusted by looking for other pixels with similar contexts. This
14466 context similarity is defined by comparing their surrounding patches of size
14467 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
14470 Note that the research area defines centers for patches, which means some
14471 patches will be made of pixels outside that research area.
14473 The filter accepts the following options.
14477 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
14480 Set patch size. Default is 7. Must be odd number in range [0, 99].
14483 Same as @option{p} but for chroma planes.
14485 The default value is @var{0} and means automatic.
14488 Set research size. Default is 15. Must be odd number in range [0, 99].
14491 Same as @option{r} but for chroma planes.
14493 The default value is @var{0} and means automatic.
14498 Deinterlace video using neural network edge directed interpolation.
14500 This filter accepts the following options:
14504 Mandatory option, without binary file filter can not work.
14505 Currently file can be found here:
14506 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
14509 Set which frames to deinterlace, by default it is @code{all}.
14510 Can be @code{all} or @code{interlaced}.
14513 Set mode of operation.
14515 Can be one of the following:
14519 Use frame flags, both fields.
14521 Use frame flags, single field.
14523 Use top field only.
14525 Use bottom field only.
14527 Use both fields, top first.
14529 Use both fields, bottom first.
14533 Set which planes to process, by default filter process all frames.
14536 Set size of local neighborhood around each pixel, used by the predictor neural
14539 Can be one of the following:
14552 Set the number of neurons in predictor neural network.
14553 Can be one of the following:
14564 Controls the number of different neural network predictions that are blended
14565 together to compute the final output value. Can be @code{fast}, default or
14569 Set which set of weights to use in the predictor.
14570 Can be one of the following:
14574 weights trained to minimize absolute error
14576 weights trained to minimize squared error
14580 Controls whether or not the prescreener neural network is used to decide
14581 which pixels should be processed by the predictor neural network and which
14582 can be handled by simple cubic interpolation.
14583 The prescreener is trained to know whether cubic interpolation will be
14584 sufficient for a pixel or whether it should be predicted by the predictor nn.
14585 The computational complexity of the prescreener nn is much less than that of
14586 the predictor nn. Since most pixels can be handled by cubic interpolation,
14587 using the prescreener generally results in much faster processing.
14588 The prescreener is pretty accurate, so the difference between using it and not
14589 using it is almost always unnoticeable.
14591 Can be one of the following:
14599 Default is @code{new}.
14602 Set various debugging flags.
14607 Force libavfilter not to use any of the specified pixel formats for the
14608 input to the next filter.
14610 It accepts the following parameters:
14614 A '|'-separated list of pixel format names, such as
14615 pix_fmts=yuv420p|monow|rgb24".
14619 @subsection Examples
14623 Force libavfilter to use a format different from @var{yuv420p} for the
14624 input to the vflip filter:
14626 noformat=pix_fmts=yuv420p,vflip
14630 Convert the input video to any of the formats not contained in the list:
14632 noformat=yuv420p|yuv444p|yuv410p
14638 Add noise on video input frame.
14640 The filter accepts the following options:
14648 Set noise seed for specific pixel component or all pixel components in case
14649 of @var{all_seed}. Default value is @code{123457}.
14651 @item all_strength, alls
14652 @item c0_strength, c0s
14653 @item c1_strength, c1s
14654 @item c2_strength, c2s
14655 @item c3_strength, c3s
14656 Set noise strength for specific pixel component or all pixel components in case
14657 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
14659 @item all_flags, allf
14660 @item c0_flags, c0f
14661 @item c1_flags, c1f
14662 @item c2_flags, c2f
14663 @item c3_flags, c3f
14664 Set pixel component flags or set flags for all components if @var{all_flags}.
14665 Available values for component flags are:
14668 averaged temporal noise (smoother)
14670 mix random noise with a (semi)regular pattern
14672 temporal noise (noise pattern changes between frames)
14674 uniform noise (gaussian otherwise)
14678 @subsection Examples
14680 Add temporal and uniform noise to input video:
14682 noise=alls=20:allf=t+u
14687 Normalize RGB video (aka histogram stretching, contrast stretching).
14688 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
14690 For each channel of each frame, the filter computes the input range and maps
14691 it linearly to the user-specified output range. The output range defaults
14692 to the full dynamic range from pure black to pure white.
14694 Temporal smoothing can be used on the input range to reduce flickering (rapid
14695 changes in brightness) caused when small dark or bright objects enter or leave
14696 the scene. This is similar to the auto-exposure (automatic gain control) on a
14697 video camera, and, like a video camera, it may cause a period of over- or
14698 under-exposure of the video.
14700 The R,G,B channels can be normalized independently, which may cause some
14701 color shifting, or linked together as a single channel, which prevents
14702 color shifting. Linked normalization preserves hue. Independent normalization
14703 does not, so it can be used to remove some color casts. Independent and linked
14704 normalization can be combined in any ratio.
14706 The normalize filter accepts the following options:
14711 Colors which define the output range. The minimum input value is mapped to
14712 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
14713 The defaults are black and white respectively. Specifying white for
14714 @var{blackpt} and black for @var{whitept} will give color-inverted,
14715 normalized video. Shades of grey can be used to reduce the dynamic range
14716 (contrast). Specifying saturated colors here can create some interesting
14720 The number of previous frames to use for temporal smoothing. The input range
14721 of each channel is smoothed using a rolling average over the current frame
14722 and the @var{smoothing} previous frames. The default is 0 (no temporal
14726 Controls the ratio of independent (color shifting) channel normalization to
14727 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
14728 independent. Defaults to 1.0 (fully independent).
14731 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
14732 expensive no-op. Defaults to 1.0 (full strength).
14736 @subsection Commands
14737 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
14738 The command accepts the same syntax of the corresponding option.
14740 If the specified expression is not valid, it is kept at its current
14743 @subsection Examples
14745 Stretch video contrast to use the full dynamic range, with no temporal
14746 smoothing; may flicker depending on the source content:
14748 normalize=blackpt=black:whitept=white:smoothing=0
14751 As above, but with 50 frames of temporal smoothing; flicker should be
14752 reduced, depending on the source content:
14754 normalize=blackpt=black:whitept=white:smoothing=50
14757 As above, but with hue-preserving linked channel normalization:
14759 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
14762 As above, but with half strength:
14764 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
14767 Map the darkest input color to red, the brightest input color to cyan:
14769 normalize=blackpt=red:whitept=cyan
14774 Pass the video source unchanged to the output.
14777 Optical Character Recognition
14779 This filter uses Tesseract for optical character recognition. To enable
14780 compilation of this filter, you need to configure FFmpeg with
14781 @code{--enable-libtesseract}.
14783 It accepts the following options:
14787 Set datapath to tesseract data. Default is to use whatever was
14788 set at installation.
14791 Set language, default is "eng".
14794 Set character whitelist.
14797 Set character blacklist.
14800 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
14801 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
14805 Apply a video transform using libopencv.
14807 To enable this filter, install the libopencv library and headers and
14808 configure FFmpeg with @code{--enable-libopencv}.
14810 It accepts the following parameters:
14815 The name of the libopencv filter to apply.
14817 @item filter_params
14818 The parameters to pass to the libopencv filter. If not specified, the default
14819 values are assumed.
14823 Refer to the official libopencv documentation for more precise
14825 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
14827 Several libopencv filters are supported; see the following subsections.
14832 Dilate an image by using a specific structuring element.
14833 It corresponds to the libopencv function @code{cvDilate}.
14835 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
14837 @var{struct_el} represents a structuring element, and has the syntax:
14838 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
14840 @var{cols} and @var{rows} represent the number of columns and rows of
14841 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
14842 point, and @var{shape} the shape for the structuring element. @var{shape}
14843 must be "rect", "cross", "ellipse", or "custom".
14845 If the value for @var{shape} is "custom", it must be followed by a
14846 string of the form "=@var{filename}". The file with name
14847 @var{filename} is assumed to represent a binary image, with each
14848 printable character corresponding to a bright pixel. When a custom
14849 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
14850 or columns and rows of the read file are assumed instead.
14852 The default value for @var{struct_el} is "3x3+0x0/rect".
14854 @var{nb_iterations} specifies the number of times the transform is
14855 applied to the image, and defaults to 1.
14859 # Use the default values
14862 # Dilate using a structuring element with a 5x5 cross, iterating two times
14863 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
14865 # Read the shape from the file diamond.shape, iterating two times.
14866 # The file diamond.shape may contain a pattern of characters like this
14872 # The specified columns and rows are ignored
14873 # but the anchor point coordinates are not
14874 ocv=dilate:0x0+2x2/custom=diamond.shape|2
14879 Erode an image by using a specific structuring element.
14880 It corresponds to the libopencv function @code{cvErode}.
14882 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
14883 with the same syntax and semantics as the @ref{dilate} filter.
14887 Smooth the input video.
14889 The filter takes the following parameters:
14890 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
14892 @var{type} is the type of smooth filter to apply, and must be one of
14893 the following values: "blur", "blur_no_scale", "median", "gaussian",
14894 or "bilateral". The default value is "gaussian".
14896 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
14897 depends on the smooth type. @var{param1} and
14898 @var{param2} accept integer positive values or 0. @var{param3} and
14899 @var{param4} accept floating point values.
14901 The default value for @var{param1} is 3. The default value for the
14902 other parameters is 0.
14904 These parameters correspond to the parameters assigned to the
14905 libopencv function @code{cvSmooth}.
14907 @section oscilloscope
14909 2D Video Oscilloscope.
14911 Useful to measure spatial impulse, step responses, chroma delays, etc.
14913 It accepts the following parameters:
14917 Set scope center x position.
14920 Set scope center y position.
14923 Set scope size, relative to frame diagonal.
14926 Set scope tilt/rotation.
14932 Set trace center x position.
14935 Set trace center y position.
14938 Set trace width, relative to width of frame.
14941 Set trace height, relative to height of frame.
14944 Set which components to trace. By default it traces first three components.
14947 Draw trace grid. By default is enabled.
14950 Draw some statistics. By default is enabled.
14953 Draw scope. By default is enabled.
14956 @subsection Commands
14957 This filter supports same @ref{commands} as options.
14958 The command accepts the same syntax of the corresponding option.
14960 If the specified expression is not valid, it is kept at its current
14963 @subsection Examples
14967 Inspect full first row of video frame.
14969 oscilloscope=x=0.5:y=0:s=1
14973 Inspect full last row of video frame.
14975 oscilloscope=x=0.5:y=1:s=1
14979 Inspect full 5th line of video frame of height 1080.
14981 oscilloscope=x=0.5:y=5/1080:s=1
14985 Inspect full last column of video frame.
14987 oscilloscope=x=1:y=0.5:s=1:t=1
14995 Overlay one video on top of another.
14997 It takes two inputs and has one output. The first input is the "main"
14998 video on which the second input is overlaid.
15000 It accepts the following parameters:
15002 A description of the accepted options follows.
15007 Set the expression for the x and y coordinates of the overlaid video
15008 on the main video. Default value is "0" for both expressions. In case
15009 the expression is invalid, it is set to a huge value (meaning that the
15010 overlay will not be displayed within the output visible area).
15013 See @ref{framesync}.
15016 Set when the expressions for @option{x}, and @option{y} are evaluated.
15018 It accepts the following values:
15021 only evaluate expressions once during the filter initialization or
15022 when a command is processed
15025 evaluate expressions for each incoming frame
15028 Default value is @samp{frame}.
15031 See @ref{framesync}.
15034 Set the format for the output video.
15036 It accepts the following values:
15039 force YUV420 output
15042 force YUV420p10 output
15045 force YUV422 output
15048 force YUV422p10 output
15051 force YUV444 output
15054 force packed RGB output
15057 force planar RGB output
15060 automatically pick format
15063 Default value is @samp{yuv420}.
15066 See @ref{framesync}.
15069 Set format of alpha of the overlaid video, it can be @var{straight} or
15070 @var{premultiplied}. Default is @var{straight}.
15073 The @option{x}, and @option{y} expressions can contain the following
15079 The main input width and height.
15083 The overlay input width and height.
15087 The computed values for @var{x} and @var{y}. They are evaluated for
15092 horizontal and vertical chroma subsample values of the output
15093 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
15097 the number of input frame, starting from 0
15100 the position in the file of the input frame, NAN if unknown
15103 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
15107 This filter also supports the @ref{framesync} options.
15109 Note that the @var{n}, @var{pos}, @var{t} variables are available only
15110 when evaluation is done @emph{per frame}, and will evaluate to NAN
15111 when @option{eval} is set to @samp{init}.
15113 Be aware that frames are taken from each input video in timestamp
15114 order, hence, if their initial timestamps differ, it is a good idea
15115 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
15116 have them begin in the same zero timestamp, as the example for
15117 the @var{movie} filter does.
15119 You can chain together more overlays but you should test the
15120 efficiency of such approach.
15122 @subsection Commands
15124 This filter supports the following commands:
15128 Modify the x and y of the overlay input.
15129 The command accepts the same syntax of the corresponding option.
15131 If the specified expression is not valid, it is kept at its current
15135 @subsection Examples
15139 Draw the overlay at 10 pixels from the bottom right corner of the main
15142 overlay=main_w-overlay_w-10:main_h-overlay_h-10
15145 Using named options the example above becomes:
15147 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
15151 Insert a transparent PNG logo in the bottom left corner of the input,
15152 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
15154 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
15158 Insert 2 different transparent PNG logos (second logo on bottom
15159 right corner) using the @command{ffmpeg} tool:
15161 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
15165 Add a transparent color layer on top of the main video; @code{WxH}
15166 must specify the size of the main input to the overlay filter:
15168 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
15172 Play an original video and a filtered version (here with the deshake
15173 filter) side by side using the @command{ffplay} tool:
15175 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
15178 The above command is the same as:
15180 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
15184 Make a sliding overlay appearing from the left to the right top part of the
15185 screen starting since time 2:
15187 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
15191 Compose output by putting two input videos side to side:
15193 ffmpeg -i left.avi -i right.avi -filter_complex "
15194 nullsrc=size=200x100 [background];
15195 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
15196 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
15197 [background][left] overlay=shortest=1 [background+left];
15198 [background+left][right] overlay=shortest=1:x=100 [left+right]
15203 Mask 10-20 seconds of a video by applying the delogo filter to a section
15205 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
15206 -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]'
15211 Chain several overlays in cascade:
15213 nullsrc=s=200x200 [bg];
15214 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
15215 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
15216 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
15217 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
15218 [in3] null, [mid2] overlay=100:100 [out0]
15223 @anchor{overlay_cuda}
15224 @section overlay_cuda
15226 Overlay one video on top of another.
15228 This is the CUDA variant of the @ref{overlay} filter.
15229 It only accepts CUDA frames. The underlying input pixel formats have to match.
15231 It takes two inputs and has one output. The first input is the "main"
15232 video on which the second input is overlaid.
15234 It accepts the following parameters:
15239 Set the x and y coordinates of the overlaid video on the main video.
15240 Default value is "0" for both expressions.
15243 See @ref{framesync}.
15246 See @ref{framesync}.
15249 See @ref{framesync}.
15253 This filter also supports the @ref{framesync} options.
15257 Apply Overcomplete Wavelet denoiser.
15259 The filter accepts the following options:
15265 Larger depth values will denoise lower frequency components more, but
15266 slow down filtering.
15268 Must be an int in the range 8-16, default is @code{8}.
15270 @item luma_strength, ls
15273 Must be a double value in the range 0-1000, default is @code{1.0}.
15275 @item chroma_strength, cs
15276 Set chroma strength.
15278 Must be a double value in the range 0-1000, default is @code{1.0}.
15284 Add paddings to the input image, and place the original input at the
15285 provided @var{x}, @var{y} coordinates.
15287 It accepts the following parameters:
15292 Specify an expression for the size of the output image with the
15293 paddings added. If the value for @var{width} or @var{height} is 0, the
15294 corresponding input size is used for the output.
15296 The @var{width} expression can reference the value set by the
15297 @var{height} expression, and vice versa.
15299 The default value of @var{width} and @var{height} is 0.
15303 Specify the offsets to place the input image at within the padded area,
15304 with respect to the top/left border of the output image.
15306 The @var{x} expression can reference the value set by the @var{y}
15307 expression, and vice versa.
15309 The default value of @var{x} and @var{y} is 0.
15311 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
15312 so the input image is centered on the padded area.
15315 Specify the color of the padded area. For the syntax of this option,
15316 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
15317 manual,ffmpeg-utils}.
15319 The default value of @var{color} is "black".
15322 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
15324 It accepts the following values:
15328 Only evaluate expressions once during the filter initialization or when
15329 a command is processed.
15332 Evaluate expressions for each incoming frame.
15336 Default value is @samp{init}.
15339 Pad to aspect instead to a resolution.
15343 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
15344 options are expressions containing the following constants:
15349 The input video width and height.
15353 These are the same as @var{in_w} and @var{in_h}.
15357 The output width and height (the size of the padded area), as
15358 specified by the @var{width} and @var{height} expressions.
15362 These are the same as @var{out_w} and @var{out_h}.
15366 The x and y offsets as specified by the @var{x} and @var{y}
15367 expressions, or NAN if not yet specified.
15370 same as @var{iw} / @var{ih}
15373 input sample aspect ratio
15376 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
15380 The horizontal and vertical chroma subsample values. For example for the
15381 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15384 @subsection Examples
15388 Add paddings with the color "violet" to the input video. The output video
15389 size is 640x480, and the top-left corner of the input video is placed at
15392 pad=640:480:0:40:violet
15395 The example above is equivalent to the following command:
15397 pad=width=640:height=480:x=0:y=40:color=violet
15401 Pad the input to get an output with dimensions increased by 3/2,
15402 and put the input video at the center of the padded area:
15404 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
15408 Pad the input to get a squared output with size equal to the maximum
15409 value between the input width and height, and put the input video at
15410 the center of the padded area:
15412 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
15416 Pad the input to get a final w/h ratio of 16:9:
15418 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
15422 In case of anamorphic video, in order to set the output display aspect
15423 correctly, it is necessary to use @var{sar} in the expression,
15424 according to the relation:
15426 (ih * X / ih) * sar = output_dar
15427 X = output_dar / sar
15430 Thus the previous example needs to be modified to:
15432 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
15436 Double the output size and put the input video in the bottom-right
15437 corner of the output padded area:
15439 pad="2*iw:2*ih:ow-iw:oh-ih"
15443 @anchor{palettegen}
15444 @section palettegen
15446 Generate one palette for a whole video stream.
15448 It accepts the following options:
15452 Set the maximum number of colors to quantize in the palette.
15453 Note: the palette will still contain 256 colors; the unused palette entries
15456 @item reserve_transparent
15457 Create a palette of 255 colors maximum and reserve the last one for
15458 transparency. Reserving the transparency color is useful for GIF optimization.
15459 If not set, the maximum of colors in the palette will be 256. You probably want
15460 to disable this option for a standalone image.
15463 @item transparency_color
15464 Set the color that will be used as background for transparency.
15467 Set statistics mode.
15469 It accepts the following values:
15472 Compute full frame histograms.
15474 Compute histograms only for the part that differs from previous frame. This
15475 might be relevant to give more importance to the moving part of your input if
15476 the background is static.
15478 Compute new histogram for each frame.
15481 Default value is @var{full}.
15484 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
15485 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
15486 color quantization of the palette. This information is also visible at
15487 @var{info} logging level.
15489 @subsection Examples
15493 Generate a representative palette of a given video using @command{ffmpeg}:
15495 ffmpeg -i input.mkv -vf palettegen palette.png
15499 @section paletteuse
15501 Use a palette to downsample an input video stream.
15503 The filter takes two inputs: one video stream and a palette. The palette must
15504 be a 256 pixels image.
15506 It accepts the following options:
15510 Select dithering mode. Available algorithms are:
15513 Ordered 8x8 bayer dithering (deterministic)
15515 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
15516 Note: this dithering is sometimes considered "wrong" and is included as a
15518 @item floyd_steinberg
15519 Floyd and Steingberg dithering (error diffusion)
15521 Frankie Sierra dithering v2 (error diffusion)
15523 Frankie Sierra dithering v2 "Lite" (error diffusion)
15526 Default is @var{sierra2_4a}.
15529 When @var{bayer} dithering is selected, this option defines the scale of the
15530 pattern (how much the crosshatch pattern is visible). A low value means more
15531 visible pattern for less banding, and higher value means less visible pattern
15532 at the cost of more banding.
15534 The option must be an integer value in the range [0,5]. Default is @var{2}.
15537 If set, define the zone to process
15541 Only the changing rectangle will be reprocessed. This is similar to GIF
15542 cropping/offsetting compression mechanism. This option can be useful for speed
15543 if only a part of the image is changing, and has use cases such as limiting the
15544 scope of the error diffusal @option{dither} to the rectangle that bounds the
15545 moving scene (it leads to more deterministic output if the scene doesn't change
15546 much, and as a result less moving noise and better GIF compression).
15549 Default is @var{none}.
15552 Take new palette for each output frame.
15554 @item alpha_threshold
15555 Sets the alpha threshold for transparency. Alpha values above this threshold
15556 will be treated as completely opaque, and values below this threshold will be
15557 treated as completely transparent.
15559 The option must be an integer value in the range [0,255]. Default is @var{128}.
15562 @subsection Examples
15566 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
15567 using @command{ffmpeg}:
15569 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
15573 @section perspective
15575 Correct perspective of video not recorded perpendicular to the screen.
15577 A description of the accepted parameters follows.
15588 Set coordinates expression for top left, top right, bottom left and bottom right corners.
15589 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
15590 If the @code{sense} option is set to @code{source}, then the specified points will be sent
15591 to the corners of the destination. If the @code{sense} option is set to @code{destination},
15592 then the corners of the source will be sent to the specified coordinates.
15594 The expressions can use the following variables:
15599 the width and height of video frame.
15603 Output frame count.
15606 @item interpolation
15607 Set interpolation for perspective correction.
15609 It accepts the following values:
15615 Default value is @samp{linear}.
15618 Set interpretation of coordinate options.
15620 It accepts the following values:
15624 Send point in the source specified by the given coordinates to
15625 the corners of the destination.
15627 @item 1, destination
15629 Send the corners of the source to the point in the destination specified
15630 by the given coordinates.
15632 Default value is @samp{source}.
15636 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
15638 It accepts the following values:
15641 only evaluate expressions once during the filter initialization or
15642 when a command is processed
15645 evaluate expressions for each incoming frame
15648 Default value is @samp{init}.
15653 Delay interlaced video by one field time so that the field order changes.
15655 The intended use is to fix PAL movies that have been captured with the
15656 opposite field order to the film-to-video transfer.
15658 A description of the accepted parameters follows.
15664 It accepts the following values:
15667 Capture field order top-first, transfer bottom-first.
15668 Filter will delay the bottom field.
15671 Capture field order bottom-first, transfer top-first.
15672 Filter will delay the top field.
15675 Capture and transfer with the same field order. This mode only exists
15676 for the documentation of the other options to refer to, but if you
15677 actually select it, the filter will faithfully do nothing.
15680 Capture field order determined automatically by field flags, transfer
15682 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
15683 basis using field flags. If no field information is available,
15684 then this works just like @samp{u}.
15687 Capture unknown or varying, transfer opposite.
15688 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
15689 analyzing the images and selecting the alternative that produces best
15690 match between the fields.
15693 Capture top-first, transfer unknown or varying.
15694 Filter selects among @samp{t} and @samp{p} using image analysis.
15697 Capture bottom-first, transfer unknown or varying.
15698 Filter selects among @samp{b} and @samp{p} using image analysis.
15701 Capture determined by field flags, transfer unknown or varying.
15702 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
15703 image analysis. If no field information is available, then this works just
15704 like @samp{U}. This is the default mode.
15707 Both capture and transfer unknown or varying.
15708 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
15712 @section photosensitivity
15713 Reduce various flashes in video, so to help users with epilepsy.
15715 It accepts the following options:
15718 Set how many frames to use when filtering. Default is 30.
15721 Set detection threshold factor. Default is 1.
15725 Set how many pixels to skip when sampling frames. Default is 1.
15726 Allowed range is from 1 to 1024.
15729 Leave frames unchanged. Default is disabled.
15732 @section pixdesctest
15734 Pixel format descriptor test filter, mainly useful for internal
15735 testing. The output video should be equal to the input video.
15739 format=monow, pixdesctest
15742 can be used to test the monowhite pixel format descriptor definition.
15746 Display sample values of color channels. Mainly useful for checking color
15747 and levels. Minimum supported resolution is 640x480.
15749 The filters accept the following options:
15753 Set scope X position, relative offset on X axis.
15756 Set scope Y position, relative offset on Y axis.
15765 Set window opacity. This window also holds statistics about pixel area.
15768 Set window X position, relative offset on X axis.
15771 Set window Y position, relative offset on Y axis.
15776 Enable the specified chain of postprocessing subfilters using libpostproc. This
15777 library should be automatically selected with a GPL build (@code{--enable-gpl}).
15778 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
15779 Each subfilter and some options have a short and a long name that can be used
15780 interchangeably, i.e. dr/dering are the same.
15782 The filters accept the following options:
15786 Set postprocessing subfilters string.
15789 All subfilters share common options to determine their scope:
15793 Honor the quality commands for this subfilter.
15796 Do chrominance filtering, too (default).
15799 Do luminance filtering only (no chrominance).
15802 Do chrominance filtering only (no luminance).
15805 These options can be appended after the subfilter name, separated by a '|'.
15807 Available subfilters are:
15810 @item hb/hdeblock[|difference[|flatness]]
15811 Horizontal deblocking filter
15814 Difference factor where higher values mean more deblocking (default: @code{32}).
15816 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15819 @item vb/vdeblock[|difference[|flatness]]
15820 Vertical deblocking filter
15823 Difference factor where higher values mean more deblocking (default: @code{32}).
15825 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15828 @item ha/hadeblock[|difference[|flatness]]
15829 Accurate horizontal deblocking filter
15832 Difference factor where higher values mean more deblocking (default: @code{32}).
15834 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15837 @item va/vadeblock[|difference[|flatness]]
15838 Accurate vertical deblocking filter
15841 Difference factor where higher values mean more deblocking (default: @code{32}).
15843 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15847 The horizontal and vertical deblocking filters share the difference and
15848 flatness values so you cannot set different horizontal and vertical
15852 @item h1/x1hdeblock
15853 Experimental horizontal deblocking filter
15855 @item v1/x1vdeblock
15856 Experimental vertical deblocking filter
15861 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
15864 larger -> stronger filtering
15866 larger -> stronger filtering
15868 larger -> stronger filtering
15871 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
15874 Stretch luminance to @code{0-255}.
15877 @item lb/linblenddeint
15878 Linear blend deinterlacing filter that deinterlaces the given block by
15879 filtering all lines with a @code{(1 2 1)} filter.
15881 @item li/linipoldeint
15882 Linear interpolating deinterlacing filter that deinterlaces the given block by
15883 linearly interpolating every second line.
15885 @item ci/cubicipoldeint
15886 Cubic interpolating deinterlacing filter deinterlaces the given block by
15887 cubically interpolating every second line.
15889 @item md/mediandeint
15890 Median deinterlacing filter that deinterlaces the given block by applying a
15891 median filter to every second line.
15893 @item fd/ffmpegdeint
15894 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
15895 second line with a @code{(-1 4 2 4 -1)} filter.
15898 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
15899 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
15901 @item fq/forceQuant[|quantizer]
15902 Overrides the quantizer table from the input with the constant quantizer you
15910 Default pp filter combination (@code{hb|a,vb|a,dr|a})
15913 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
15916 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
15919 @subsection Examples
15923 Apply horizontal and vertical deblocking, deringing and automatic
15924 brightness/contrast:
15930 Apply default filters without brightness/contrast correction:
15936 Apply default filters and temporal denoiser:
15938 pp=default/tmpnoise|1|2|3
15942 Apply deblocking on luminance only, and switch vertical deblocking on or off
15943 automatically depending on available CPU time:
15950 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
15951 similar to spp = 6 with 7 point DCT, where only the center sample is
15954 The filter accepts the following options:
15958 Force a constant quantization parameter. It accepts an integer in range
15959 0 to 63. If not set, the filter will use the QP from the video stream
15963 Set thresholding mode. Available modes are:
15967 Set hard thresholding.
15969 Set soft thresholding (better de-ringing effect, but likely blurrier).
15971 Set medium thresholding (good results, default).
15975 @section premultiply
15976 Apply alpha premultiply effect to input video stream using first plane
15977 of second stream as alpha.
15979 Both streams must have same dimensions and same pixel format.
15981 The filter accepts the following option:
15985 Set which planes will be processed, unprocessed planes will be copied.
15986 By default value 0xf, all planes will be processed.
15989 Do not require 2nd input for processing, instead use alpha plane from input stream.
15993 Apply prewitt operator to input video stream.
15995 The filter accepts the following option:
15999 Set which planes will be processed, unprocessed planes will be copied.
16000 By default value 0xf, all planes will be processed.
16003 Set value which will be multiplied with filtered result.
16006 Set value which will be added to filtered result.
16009 @subsection Commands
16011 This filter supports the all above options as @ref{commands}.
16013 @section pseudocolor
16015 Alter frame colors in video with pseudocolors.
16017 This filter accepts the following options:
16021 set pixel first component expression
16024 set pixel second component expression
16027 set pixel third component expression
16030 set pixel fourth component expression, corresponds to the alpha component
16033 set component to use as base for altering colors
16036 Each of them specifies the expression to use for computing the lookup table for
16037 the corresponding pixel component values.
16039 The expressions can contain the following constants and functions:
16044 The input width and height.
16047 The input value for the pixel component.
16049 @item ymin, umin, vmin, amin
16050 The minimum allowed component value.
16052 @item ymax, umax, vmax, amax
16053 The maximum allowed component value.
16056 All expressions default to "val".
16058 @subsection Examples
16062 Change too high luma values to gradient:
16064 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'"
16070 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
16071 Ratio) between two input videos.
16073 This filter takes in input two input videos, the first input is
16074 considered the "main" source and is passed unchanged to the
16075 output. The second input is used as a "reference" video for computing
16078 Both video inputs must have the same resolution and pixel format for
16079 this filter to work correctly. Also it assumes that both inputs
16080 have the same number of frames, which are compared one by one.
16082 The obtained average PSNR is printed through the logging system.
16084 The filter stores the accumulated MSE (mean squared error) of each
16085 frame, and at the end of the processing it is averaged across all frames
16086 equally, and the following formula is applied to obtain the PSNR:
16089 PSNR = 10*log10(MAX^2/MSE)
16092 Where MAX is the average of the maximum values of each component of the
16095 The description of the accepted parameters follows.
16098 @item stats_file, f
16099 If specified the filter will use the named file to save the PSNR of
16100 each individual frame. When filename equals "-" the data is sent to
16103 @item stats_version
16104 Specifies which version of the stats file format to use. Details of
16105 each format are written below.
16106 Default value is 1.
16108 @item stats_add_max
16109 Determines whether the max value is output to the stats log.
16110 Default value is 0.
16111 Requires stats_version >= 2. If this is set and stats_version < 2,
16112 the filter will return an error.
16115 This filter also supports the @ref{framesync} options.
16117 The file printed if @var{stats_file} is selected, contains a sequence of
16118 key/value pairs of the form @var{key}:@var{value} for each compared
16121 If a @var{stats_version} greater than 1 is specified, a header line precedes
16122 the list of per-frame-pair stats, with key value pairs following the frame
16123 format with the following parameters:
16126 @item psnr_log_version
16127 The version of the log file format. Will match @var{stats_version}.
16130 A comma separated list of the per-frame-pair parameters included in
16134 A description of each shown per-frame-pair parameter follows:
16138 sequential number of the input frame, starting from 1
16141 Mean Square Error pixel-by-pixel average difference of the compared
16142 frames, averaged over all the image components.
16144 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
16145 Mean Square Error pixel-by-pixel average difference of the compared
16146 frames for the component specified by the suffix.
16148 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
16149 Peak Signal to Noise ratio of the compared frames for the component
16150 specified by the suffix.
16152 @item max_avg, max_y, max_u, max_v
16153 Maximum allowed value for each channel, and average over all
16157 @subsection Examples
16162 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
16163 [main][ref] psnr="stats_file=stats.log" [out]
16166 On this example the input file being processed is compared with the
16167 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
16168 is stored in @file{stats.log}.
16171 Another example with different containers:
16173 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 -
16180 Pulldown reversal (inverse telecine) filter, capable of handling mixed
16181 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
16184 The pullup filter is designed to take advantage of future context in making
16185 its decisions. This filter is stateless in the sense that it does not lock
16186 onto a pattern to follow, but it instead looks forward to the following
16187 fields in order to identify matches and rebuild progressive frames.
16189 To produce content with an even framerate, insert the fps filter after
16190 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
16191 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
16193 The filter accepts the following options:
16200 These options set the amount of "junk" to ignore at the left, right, top, and
16201 bottom of the image, respectively. Left and right are in units of 8 pixels,
16202 while top and bottom are in units of 2 lines.
16203 The default is 8 pixels on each side.
16206 Set the strict breaks. Setting this option to 1 will reduce the chances of
16207 filter generating an occasional mismatched frame, but it may also cause an
16208 excessive number of frames to be dropped during high motion sequences.
16209 Conversely, setting it to -1 will make filter match fields more easily.
16210 This may help processing of video where there is slight blurring between
16211 the fields, but may also cause there to be interlaced frames in the output.
16212 Default value is @code{0}.
16215 Set the metric plane to use. It accepts the following values:
16221 Use chroma blue plane.
16224 Use chroma red plane.
16227 This option may be set to use chroma plane instead of the default luma plane
16228 for doing filter's computations. This may improve accuracy on very clean
16229 source material, but more likely will decrease accuracy, especially if there
16230 is chroma noise (rainbow effect) or any grayscale video.
16231 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
16232 load and make pullup usable in realtime on slow machines.
16235 For best results (without duplicated frames in the output file) it is
16236 necessary to change the output frame rate. For example, to inverse
16237 telecine NTSC input:
16239 ffmpeg -i input -vf pullup -r 24000/1001 ...
16244 Change video quantization parameters (QP).
16246 The filter accepts the following option:
16250 Set expression for quantization parameter.
16253 The expression is evaluated through the eval API and can contain, among others,
16254 the following constants:
16258 1 if index is not 129, 0 otherwise.
16261 Sequential index starting from -129 to 128.
16264 @subsection Examples
16268 Some equation like:
16276 Flush video frames from internal cache of frames into a random order.
16277 No frame is discarded.
16278 Inspired by @ref{frei0r} nervous filter.
16282 Set size in number of frames of internal cache, in range from @code{2} to
16283 @code{512}. Default is @code{30}.
16286 Set seed for random number generator, must be an integer included between
16287 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
16288 less than @code{0}, the filter will try to use a good random seed on a
16292 @section readeia608
16294 Read closed captioning (EIA-608) information from the top lines of a video frame.
16296 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
16297 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
16298 with EIA-608 data (starting from 0). A description of each metadata value follows:
16301 @item lavfi.readeia608.X.cc
16302 The two bytes stored as EIA-608 data (printed in hexadecimal).
16304 @item lavfi.readeia608.X.line
16305 The number of the line on which the EIA-608 data was identified and read.
16308 This filter accepts the following options:
16312 Set the line to start scanning for EIA-608 data. Default is @code{0}.
16315 Set the line to end scanning for EIA-608 data. Default is @code{29}.
16318 Set the ratio of width reserved for sync code detection.
16319 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
16322 Enable checking the parity bit. In the event of a parity error, the filter will output
16323 @code{0x00} for that character. Default is false.
16326 Lowpass lines prior to further processing. Default is enabled.
16329 @subsection Commands
16331 This filter supports the all above options as @ref{commands}.
16333 @subsection Examples
16337 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
16339 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
16345 Read vertical interval timecode (VITC) information from the top lines of a
16348 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
16349 timecode value, if a valid timecode has been detected. Further metadata key
16350 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
16351 timecode data has been found or not.
16353 This filter accepts the following options:
16357 Set the maximum number of lines to scan for VITC data. If the value is set to
16358 @code{-1} the full video frame is scanned. Default is @code{45}.
16361 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
16362 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
16365 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
16366 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
16369 @subsection Examples
16373 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
16374 draw @code{--:--:--:--} as a placeholder:
16376 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
16382 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
16384 Destination pixel at position (X, Y) will be picked from source (x, y) position
16385 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
16386 value for pixel will be used for destination pixel.
16388 Xmap and Ymap input video streams must be of same dimensions. Output video stream
16389 will have Xmap/Ymap video stream dimensions.
16390 Xmap and Ymap input video streams are 16bit depth, single channel.
16394 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
16395 Default is @code{color}.
16398 Specify the color of the unmapped pixels. For the syntax of this option,
16399 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
16400 manual,ffmpeg-utils}. Default color is @code{black}.
16403 @section removegrain
16405 The removegrain filter is a spatial denoiser for progressive video.
16409 Set mode for the first plane.
16412 Set mode for the second plane.
16415 Set mode for the third plane.
16418 Set mode for the fourth plane.
16421 Range of mode is from 0 to 24. Description of each mode follows:
16425 Leave input plane unchanged. Default.
16428 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
16431 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
16434 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
16437 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
16438 This is equivalent to a median filter.
16441 Line-sensitive clipping giving the minimal change.
16444 Line-sensitive clipping, intermediate.
16447 Line-sensitive clipping, intermediate.
16450 Line-sensitive clipping, intermediate.
16453 Line-sensitive clipping on a line where the neighbours pixels are the closest.
16456 Replaces the target pixel with the closest neighbour.
16459 [1 2 1] horizontal and vertical kernel blur.
16465 Bob mode, interpolates top field from the line where the neighbours
16466 pixels are the closest.
16469 Bob mode, interpolates bottom field from the line where the neighbours
16470 pixels are the closest.
16473 Bob mode, interpolates top field. Same as 13 but with a more complicated
16474 interpolation formula.
16477 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
16478 interpolation formula.
16481 Clips the pixel with the minimum and maximum of respectively the maximum and
16482 minimum of each pair of opposite neighbour pixels.
16485 Line-sensitive clipping using opposite neighbours whose greatest distance from
16486 the current pixel is minimal.
16489 Replaces the pixel with the average of its 8 neighbours.
16492 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
16495 Clips pixels using the averages of opposite neighbour.
16498 Same as mode 21 but simpler and faster.
16501 Small edge and halo removal, but reputed useless.
16507 @section removelogo
16509 Suppress a TV station logo, using an image file to determine which
16510 pixels comprise the logo. It works by filling in the pixels that
16511 comprise the logo with neighboring pixels.
16513 The filter accepts the following options:
16517 Set the filter bitmap file, which can be any image format supported by
16518 libavformat. The width and height of the image file must match those of the
16519 video stream being processed.
16522 Pixels in the provided bitmap image with a value of zero are not
16523 considered part of the logo, non-zero pixels are considered part of
16524 the logo. If you use white (255) for the logo and black (0) for the
16525 rest, you will be safe. For making the filter bitmap, it is
16526 recommended to take a screen capture of a black frame with the logo
16527 visible, and then using a threshold filter followed by the erode
16528 filter once or twice.
16530 If needed, little splotches can be fixed manually. Remember that if
16531 logo pixels are not covered, the filter quality will be much
16532 reduced. Marking too many pixels as part of the logo does not hurt as
16533 much, but it will increase the amount of blurring needed to cover over
16534 the image and will destroy more information than necessary, and extra
16535 pixels will slow things down on a large logo.
16537 @section repeatfields
16539 This filter uses the repeat_field flag from the Video ES headers and hard repeats
16540 fields based on its value.
16544 Reverse a video clip.
16546 Warning: This filter requires memory to buffer the entire clip, so trimming
16549 @subsection Examples
16553 Take the first 5 seconds of a clip, and reverse it.
16560 Shift R/G/B/A pixels horizontally and/or vertically.
16562 The filter accepts the following options:
16565 Set amount to shift red horizontally.
16567 Set amount to shift red vertically.
16569 Set amount to shift green horizontally.
16571 Set amount to shift green vertically.
16573 Set amount to shift blue horizontally.
16575 Set amount to shift blue vertically.
16577 Set amount to shift alpha horizontally.
16579 Set amount to shift alpha vertically.
16581 Set edge mode, can be @var{smear}, default, or @var{warp}.
16584 @subsection Commands
16586 This filter supports the all above options as @ref{commands}.
16589 Apply roberts cross operator to input video stream.
16591 The filter accepts the following option:
16595 Set which planes will be processed, unprocessed planes will be copied.
16596 By default value 0xf, all planes will be processed.
16599 Set value which will be multiplied with filtered result.
16602 Set value which will be added to filtered result.
16605 @subsection Commands
16607 This filter supports the all above options as @ref{commands}.
16611 Rotate video by an arbitrary angle expressed in radians.
16613 The filter accepts the following options:
16615 A description of the optional parameters follows.
16618 Set an expression for the angle by which to rotate the input video
16619 clockwise, expressed as a number of radians. A negative value will
16620 result in a counter-clockwise rotation. By default it is set to "0".
16622 This expression is evaluated for each frame.
16625 Set the output width expression, default value is "iw".
16626 This expression is evaluated just once during configuration.
16629 Set the output height expression, default value is "ih".
16630 This expression is evaluated just once during configuration.
16633 Enable bilinear interpolation if set to 1, a value of 0 disables
16634 it. Default value is 1.
16637 Set the color used to fill the output area not covered by the rotated
16638 image. For the general syntax of this option, check the
16639 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
16640 If the special value "none" is selected then no
16641 background is printed (useful for example if the background is never shown).
16643 Default value is "black".
16646 The expressions for the angle and the output size can contain the
16647 following constants and functions:
16651 sequential number of the input frame, starting from 0. It is always NAN
16652 before the first frame is filtered.
16655 time in seconds of the input frame, it is set to 0 when the filter is
16656 configured. It is always NAN before the first frame is filtered.
16660 horizontal and vertical chroma subsample values. For example for the
16661 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16665 the input video width and height
16669 the output width and height, that is the size of the padded area as
16670 specified by the @var{width} and @var{height} expressions
16674 the minimal width/height required for completely containing the input
16675 video rotated by @var{a} radians.
16677 These are only available when computing the @option{out_w} and
16678 @option{out_h} expressions.
16681 @subsection Examples
16685 Rotate the input by PI/6 radians clockwise:
16691 Rotate the input by PI/6 radians counter-clockwise:
16697 Rotate the input by 45 degrees clockwise:
16703 Apply a constant rotation with period T, starting from an angle of PI/3:
16705 rotate=PI/3+2*PI*t/T
16709 Make the input video rotation oscillating with a period of T
16710 seconds and an amplitude of A radians:
16712 rotate=A*sin(2*PI/T*t)
16716 Rotate the video, output size is chosen so that the whole rotating
16717 input video is always completely contained in the output:
16719 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
16723 Rotate the video, reduce the output size so that no background is ever
16726 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
16730 @subsection Commands
16732 The filter supports the following commands:
16736 Set the angle expression.
16737 The command accepts the same syntax of the corresponding option.
16739 If the specified expression is not valid, it is kept at its current
16745 Apply Shape Adaptive Blur.
16747 The filter accepts the following options:
16750 @item luma_radius, lr
16751 Set luma blur filter strength, must be a value in range 0.1-4.0, default
16752 value is 1.0. A greater value will result in a more blurred image, and
16753 in slower processing.
16755 @item luma_pre_filter_radius, lpfr
16756 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
16759 @item luma_strength, ls
16760 Set luma maximum difference between pixels to still be considered, must
16761 be a value in the 0.1-100.0 range, default value is 1.0.
16763 @item chroma_radius, cr
16764 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
16765 greater value will result in a more blurred image, and in slower
16768 @item chroma_pre_filter_radius, cpfr
16769 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
16771 @item chroma_strength, cs
16772 Set chroma maximum difference between pixels to still be considered,
16773 must be a value in the -0.9-100.0 range.
16776 Each chroma option value, if not explicitly specified, is set to the
16777 corresponding luma option value.
16782 Scale (resize) the input video, using the libswscale library.
16784 The scale filter forces the output display aspect ratio to be the same
16785 of the input, by changing the output sample aspect ratio.
16787 If the input image format is different from the format requested by
16788 the next filter, the scale filter will convert the input to the
16791 @subsection Options
16792 The filter accepts the following options, or any of the options
16793 supported by the libswscale scaler.
16795 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
16796 the complete list of scaler options.
16801 Set the output video dimension expression. Default value is the input
16804 If the @var{width} or @var{w} value is 0, the input width is used for
16805 the output. If the @var{height} or @var{h} value is 0, the input height
16806 is used for the output.
16808 If one and only one of the values is -n with n >= 1, the scale filter
16809 will use a value that maintains the aspect ratio of the input image,
16810 calculated from the other specified dimension. After that it will,
16811 however, make sure that the calculated dimension is divisible by n and
16812 adjust the value if necessary.
16814 If both values are -n with n >= 1, the behavior will be identical to
16815 both values being set to 0 as previously detailed.
16817 See below for the list of accepted constants for use in the dimension
16821 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
16825 Only evaluate expressions once during the filter initialization or when a command is processed.
16828 Evaluate expressions for each incoming frame.
16832 Default value is @samp{init}.
16836 Set the interlacing mode. It accepts the following values:
16840 Force interlaced aware scaling.
16843 Do not apply interlaced scaling.
16846 Select interlaced aware scaling depending on whether the source frames
16847 are flagged as interlaced or not.
16850 Default value is @samp{0}.
16853 Set libswscale scaling flags. See
16854 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16855 complete list of values. If not explicitly specified the filter applies
16859 @item param0, param1
16860 Set libswscale input parameters for scaling algorithms that need them. See
16861 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16862 complete documentation. If not explicitly specified the filter applies
16868 Set the video size. For the syntax of this option, check the
16869 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16871 @item in_color_matrix
16872 @item out_color_matrix
16873 Set in/output YCbCr color space type.
16875 This allows the autodetected value to be overridden as well as allows forcing
16876 a specific value used for the output and encoder.
16878 If not specified, the color space type depends on the pixel format.
16884 Choose automatically.
16887 Format conforming to International Telecommunication Union (ITU)
16888 Recommendation BT.709.
16891 Set color space conforming to the United States Federal Communications
16892 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
16897 Set color space conforming to:
16901 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
16904 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
16907 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
16912 Set color space conforming to SMPTE ST 240:1999.
16915 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
16920 Set in/output YCbCr sample range.
16922 This allows the autodetected value to be overridden as well as allows forcing
16923 a specific value used for the output and encoder. If not specified, the
16924 range depends on the pixel format. Possible values:
16928 Choose automatically.
16931 Set full range (0-255 in case of 8-bit luma).
16933 @item mpeg/limited/tv
16934 Set "MPEG" range (16-235 in case of 8-bit luma).
16937 @item force_original_aspect_ratio
16938 Enable decreasing or increasing output video width or height if necessary to
16939 keep the original aspect ratio. Possible values:
16943 Scale the video as specified and disable this feature.
16946 The output video dimensions will automatically be decreased if needed.
16949 The output video dimensions will automatically be increased if needed.
16953 One useful instance of this option is that when you know a specific device's
16954 maximum allowed resolution, you can use this to limit the output video to
16955 that, while retaining the aspect ratio. For example, device A allows
16956 1280x720 playback, and your video is 1920x800. Using this option (set it to
16957 decrease) and specifying 1280x720 to the command line makes the output
16960 Please note that this is a different thing than specifying -1 for @option{w}
16961 or @option{h}, you still need to specify the output resolution for this option
16964 @item force_divisible_by
16965 Ensures that both the output dimensions, width and height, are divisible by the
16966 given integer when used together with @option{force_original_aspect_ratio}. This
16967 works similar to using @code{-n} in the @option{w} and @option{h} options.
16969 This option respects the value set for @option{force_original_aspect_ratio},
16970 increasing or decreasing the resolution accordingly. The video's aspect ratio
16971 may be slightly modified.
16973 This option can be handy if you need to have a video fit within or exceed
16974 a defined resolution using @option{force_original_aspect_ratio} but also have
16975 encoder restrictions on width or height divisibility.
16979 The values of the @option{w} and @option{h} options are expressions
16980 containing the following constants:
16985 The input width and height
16989 These are the same as @var{in_w} and @var{in_h}.
16993 The output (scaled) width and height
16997 These are the same as @var{out_w} and @var{out_h}
17000 The same as @var{iw} / @var{ih}
17003 input sample aspect ratio
17006 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
17010 horizontal and vertical input chroma subsample values. For example for the
17011 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17015 horizontal and vertical output chroma subsample values. For example for the
17016 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17019 The (sequential) number of the input frame, starting from 0.
17020 Only available with @code{eval=frame}.
17023 The presentation timestamp of the input frame, expressed as a number of
17024 seconds. Only available with @code{eval=frame}.
17027 The position (byte offset) of the frame in the input stream, or NaN if
17028 this information is unavailable and/or meaningless (for example in case of synthetic video).
17029 Only available with @code{eval=frame}.
17032 @subsection Examples
17036 Scale the input video to a size of 200x100
17041 This is equivalent to:
17052 Specify a size abbreviation for the output size:
17057 which can also be written as:
17063 Scale the input to 2x:
17065 scale=w=2*iw:h=2*ih
17069 The above is the same as:
17071 scale=2*in_w:2*in_h
17075 Scale the input to 2x with forced interlaced scaling:
17077 scale=2*iw:2*ih:interl=1
17081 Scale the input to half size:
17083 scale=w=iw/2:h=ih/2
17087 Increase the width, and set the height to the same size:
17093 Seek Greek harmony:
17100 Increase the height, and set the width to 3/2 of the height:
17102 scale=w=3/2*oh:h=3/5*ih
17106 Increase the size, making the size a multiple of the chroma
17109 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
17113 Increase the width to a maximum of 500 pixels,
17114 keeping the same aspect ratio as the input:
17116 scale=w='min(500\, iw*3/2):h=-1'
17120 Make pixels square by combining scale and setsar:
17122 scale='trunc(ih*dar):ih',setsar=1/1
17126 Make pixels square by combining scale and setsar,
17127 making sure the resulting resolution is even (required by some codecs):
17129 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
17133 @subsection Commands
17135 This filter supports the following commands:
17139 Set the output video dimension expression.
17140 The command accepts the same syntax of the corresponding option.
17142 If the specified expression is not valid, it is kept at its current
17148 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
17149 format conversion on CUDA video frames. Setting the output width and height
17150 works in the same way as for the @var{scale} filter.
17152 The following additional options are accepted:
17155 The pixel format of the output CUDA frames. If set to the string "same" (the
17156 default), the input format will be kept. Note that automatic format negotiation
17157 and conversion is not yet supported for hardware frames
17160 The interpolation algorithm used for resizing. One of the following:
17167 @item cubic2p_bspline
17168 2-parameter cubic (B=1, C=0)
17170 @item cubic2p_catmullrom
17171 2-parameter cubic (B=0, C=1/2)
17173 @item cubic2p_b05c03
17174 2-parameter cubic (B=1/2, C=3/10)
17182 @item force_original_aspect_ratio
17183 Enable decreasing or increasing output video width or height if necessary to
17184 keep the original aspect ratio. Possible values:
17188 Scale the video as specified and disable this feature.
17191 The output video dimensions will automatically be decreased if needed.
17194 The output video dimensions will automatically be increased if needed.
17198 One useful instance of this option is that when you know a specific device's
17199 maximum allowed resolution, you can use this to limit the output video to
17200 that, while retaining the aspect ratio. For example, device A allows
17201 1280x720 playback, and your video is 1920x800. Using this option (set it to
17202 decrease) and specifying 1280x720 to the command line makes the output
17205 Please note that this is a different thing than specifying -1 for @option{w}
17206 or @option{h}, you still need to specify the output resolution for this option
17209 @item force_divisible_by
17210 Ensures that both the output dimensions, width and height, are divisible by the
17211 given integer when used together with @option{force_original_aspect_ratio}. This
17212 works similar to using @code{-n} in the @option{w} and @option{h} options.
17214 This option respects the value set for @option{force_original_aspect_ratio},
17215 increasing or decreasing the resolution accordingly. The video's aspect ratio
17216 may be slightly modified.
17218 This option can be handy if you need to have a video fit within or exceed
17219 a defined resolution using @option{force_original_aspect_ratio} but also have
17220 encoder restrictions on width or height divisibility.
17226 Scale (resize) the input video, based on a reference video.
17228 See the scale filter for available options, scale2ref supports the same but
17229 uses the reference video instead of the main input as basis. scale2ref also
17230 supports the following additional constants for the @option{w} and
17231 @option{h} options:
17236 The main input video's width and height
17239 The same as @var{main_w} / @var{main_h}
17242 The main input video's sample aspect ratio
17244 @item main_dar, mdar
17245 The main input video's display aspect ratio. Calculated from
17246 @code{(main_w / main_h) * main_sar}.
17250 The main input video's horizontal and vertical chroma subsample values.
17251 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
17255 The (sequential) number of the main input frame, starting from 0.
17256 Only available with @code{eval=frame}.
17259 The presentation timestamp of the main input frame, expressed as a number of
17260 seconds. Only available with @code{eval=frame}.
17263 The position (byte offset) of the frame in the main input stream, or NaN if
17264 this information is unavailable and/or meaningless (for example in case of synthetic video).
17265 Only available with @code{eval=frame}.
17268 @subsection Examples
17272 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
17274 'scale2ref[b][a];[a][b]overlay'
17278 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
17280 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
17284 @subsection Commands
17286 This filter supports the following commands:
17290 Set the output video dimension expression.
17291 The command accepts the same syntax of the corresponding option.
17293 If the specified expression is not valid, it is kept at its current
17298 Scroll input video horizontally and/or vertically by constant speed.
17300 The filter accepts the following options:
17302 @item horizontal, h
17303 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
17304 Negative values changes scrolling direction.
17307 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
17308 Negative values changes scrolling direction.
17311 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
17314 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
17317 @subsection Commands
17319 This filter supports the following @ref{commands}:
17321 @item horizontal, h
17322 Set the horizontal scrolling speed.
17324 Set the vertical scrolling speed.
17330 Detect video scene change.
17332 This filter sets frame metadata with mafd between frame, the scene score, and
17333 forward the frame to the next filter, so they can use these metadata to detect
17334 scene change or others.
17336 In addition, this filter logs a message and sets frame metadata when it detects
17337 a scene change by @option{threshold}.
17339 @code{lavfi.scd.mafd} metadata keys are set with mafd for every frame.
17341 @code{lavfi.scd.score} metadata keys are set with scene change score for every frame
17342 to detect scene change.
17344 @code{lavfi.scd.time} metadata keys are set with current filtered frame time which
17345 detect scene change with @option{threshold}.
17347 The filter accepts the following options:
17351 Set the scene change detection threshold as a percentage of maximum change. Good
17352 values are in the @code{[8.0, 14.0]} range. The range for @option{threshold} is
17355 Default value is @code{10.}.
17358 Set the flag to pass scene change frames to the next filter. Default value is @code{0}
17359 You can enable it if you want to get snapshot of scene change frames only.
17362 @anchor{selectivecolor}
17363 @section selectivecolor
17365 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
17366 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
17367 by the "purity" of the color (that is, how saturated it already is).
17369 This filter is similar to the Adobe Photoshop Selective Color tool.
17371 The filter accepts the following options:
17374 @item correction_method
17375 Select color correction method.
17377 Available values are:
17380 Specified adjustments are applied "as-is" (added/subtracted to original pixel
17383 Specified adjustments are relative to the original component value.
17385 Default is @code{absolute}.
17387 Adjustments for red pixels (pixels where the red component is the maximum)
17389 Adjustments for yellow pixels (pixels where the blue component is the minimum)
17391 Adjustments for green pixels (pixels where the green component is the maximum)
17393 Adjustments for cyan pixels (pixels where the red component is the minimum)
17395 Adjustments for blue pixels (pixels where the blue component is the maximum)
17397 Adjustments for magenta pixels (pixels where the green component is the minimum)
17399 Adjustments for white pixels (pixels where all components are greater than 128)
17401 Adjustments for all pixels except pure black and pure white
17403 Adjustments for black pixels (pixels where all components are lesser than 128)
17405 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
17408 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
17409 4 space separated floating point adjustment values in the [-1,1] range,
17410 respectively to adjust the amount of cyan, magenta, yellow and black for the
17411 pixels of its range.
17413 @subsection Examples
17417 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
17418 increase magenta by 27% in blue areas:
17420 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
17424 Use a Photoshop selective color preset:
17426 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
17430 @anchor{separatefields}
17431 @section separatefields
17433 The @code{separatefields} takes a frame-based video input and splits
17434 each frame into its components fields, producing a new half height clip
17435 with twice the frame rate and twice the frame count.
17437 This filter use field-dominance information in frame to decide which
17438 of each pair of fields to place first in the output.
17439 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
17441 @section setdar, setsar
17443 The @code{setdar} filter sets the Display Aspect Ratio for the filter
17446 This is done by changing the specified Sample (aka Pixel) Aspect
17447 Ratio, according to the following equation:
17449 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
17452 Keep in mind that the @code{setdar} filter does not modify the pixel
17453 dimensions of the video frame. Also, the display aspect ratio set by
17454 this filter may be changed by later filters in the filterchain,
17455 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
17458 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
17459 the filter output video.
17461 Note that as a consequence of the application of this filter, the
17462 output display aspect ratio will change according to the equation
17465 Keep in mind that the sample aspect ratio set by the @code{setsar}
17466 filter may be changed by later filters in the filterchain, e.g. if
17467 another "setsar" or a "setdar" filter is applied.
17469 It accepts the following parameters:
17472 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
17473 Set the aspect ratio used by the filter.
17475 The parameter can be a floating point number string, an expression, or
17476 a string of the form @var{num}:@var{den}, where @var{num} and
17477 @var{den} are the numerator and denominator of the aspect ratio. If
17478 the parameter is not specified, it is assumed the value "0".
17479 In case the form "@var{num}:@var{den}" is used, the @code{:} character
17483 Set the maximum integer value to use for expressing numerator and
17484 denominator when reducing the expressed aspect ratio to a rational.
17485 Default value is @code{100}.
17489 The parameter @var{sar} is an expression containing
17490 the following constants:
17494 These are approximated values for the mathematical constants e
17495 (Euler's number), pi (Greek pi), and phi (the golden ratio).
17498 The input width and height.
17501 These are the same as @var{w} / @var{h}.
17504 The input sample aspect ratio.
17507 The input display aspect ratio. It is the same as
17508 (@var{w} / @var{h}) * @var{sar}.
17511 Horizontal and vertical chroma subsample values. For example, for the
17512 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17515 @subsection Examples
17520 To change the display aspect ratio to 16:9, specify one of the following:
17527 To change the sample aspect ratio to 10:11, specify:
17533 To set a display aspect ratio of 16:9, and specify a maximum integer value of
17534 1000 in the aspect ratio reduction, use the command:
17536 setdar=ratio=16/9:max=1000
17544 Force field for the output video frame.
17546 The @code{setfield} filter marks the interlace type field for the
17547 output frames. It does not change the input frame, but only sets the
17548 corresponding property, which affects how the frame is treated by
17549 following filters (e.g. @code{fieldorder} or @code{yadif}).
17551 The filter accepts the following options:
17556 Available values are:
17560 Keep the same field property.
17563 Mark the frame as bottom-field-first.
17566 Mark the frame as top-field-first.
17569 Mark the frame as progressive.
17576 Force frame parameter for the output video frame.
17578 The @code{setparams} filter marks interlace and color range for the
17579 output frames. It does not change the input frame, but only sets the
17580 corresponding property, which affects how the frame is treated by
17585 Available values are:
17589 Keep the same field property (default).
17592 Mark the frame as bottom-field-first.
17595 Mark the frame as top-field-first.
17598 Mark the frame as progressive.
17602 Available values are:
17606 Keep the same color range property (default).
17608 @item unspecified, unknown
17609 Mark the frame as unspecified color range.
17611 @item limited, tv, mpeg
17612 Mark the frame as limited range.
17614 @item full, pc, jpeg
17615 Mark the frame as full range.
17618 @item color_primaries
17619 Set the color primaries.
17620 Available values are:
17624 Keep the same color primaries property (default).
17641 Set the color transfer.
17642 Available values are:
17646 Keep the same color trc property (default).
17668 Set the colorspace.
17669 Available values are:
17673 Keep the same colorspace property (default).
17686 @item chroma-derived-nc
17687 @item chroma-derived-c
17694 Show a line containing various information for each input video frame.
17695 The input video is not modified.
17697 This filter supports the following options:
17701 Calculate checksums of each plane. By default enabled.
17704 The shown line contains a sequence of key/value pairs of the form
17705 @var{key}:@var{value}.
17707 The following values are shown in the output:
17711 The (sequential) number of the input frame, starting from 0.
17714 The Presentation TimeStamp of the input frame, expressed as a number of
17715 time base units. The time base unit depends on the filter input pad.
17718 The Presentation TimeStamp of the input frame, expressed as a number of
17722 The position of the frame in the input stream, or -1 if this information is
17723 unavailable and/or meaningless (for example in case of synthetic video).
17726 The pixel format name.
17729 The sample aspect ratio of the input frame, expressed in the form
17730 @var{num}/@var{den}.
17733 The size of the input frame. For the syntax of this option, check the
17734 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17737 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
17738 for bottom field first).
17741 This is 1 if the frame is a key frame, 0 otherwise.
17744 The picture type of the input frame ("I" for an I-frame, "P" for a
17745 P-frame, "B" for a B-frame, or "?" for an unknown type).
17746 Also refer to the documentation of the @code{AVPictureType} enum and of
17747 the @code{av_get_picture_type_char} function defined in
17748 @file{libavutil/avutil.h}.
17751 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
17753 @item plane_checksum
17754 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
17755 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
17758 The mean value of pixels in each plane of the input frame, expressed in the form
17759 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
17762 The standard deviation of pixel values in each plane of the input frame, expressed
17763 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
17767 @section showpalette
17769 Displays the 256 colors palette of each frame. This filter is only relevant for
17770 @var{pal8} pixel format frames.
17772 It accepts the following option:
17776 Set the size of the box used to represent one palette color entry. Default is
17777 @code{30} (for a @code{30x30} pixel box).
17780 @section shuffleframes
17782 Reorder and/or duplicate and/or drop video frames.
17784 It accepts the following parameters:
17788 Set the destination indexes of input frames.
17789 This is space or '|' separated list of indexes that maps input frames to output
17790 frames. Number of indexes also sets maximal value that each index may have.
17791 '-1' index have special meaning and that is to drop frame.
17794 The first frame has the index 0. The default is to keep the input unchanged.
17796 @subsection Examples
17800 Swap second and third frame of every three frames of the input:
17802 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
17806 Swap 10th and 1st frame of every ten frames of the input:
17808 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
17812 @section shufflepixels
17814 Reorder pixels in video frames.
17816 This filter accepts the following options:
17820 Set shuffle direction. Can be forward or inverse direction.
17821 Default direction is forward.
17824 Set shuffle mode. Can be horizontal, vertical or block mode.
17828 Set shuffle block_size. In case of horizontal shuffle mode only width
17829 part of size is used, and in case of vertical shuffle mode only height
17830 part of size is used.
17833 Set random seed used with shuffling pixels. Mainly useful to set to be able
17834 to reverse filtering process to get original input.
17835 For example, to reverse forward shuffle you need to use same parameters
17836 and exact same seed and to set direction to inverse.
17839 @section shuffleplanes
17841 Reorder and/or duplicate video planes.
17843 It accepts the following parameters:
17848 The index of the input plane to be used as the first output plane.
17851 The index of the input plane to be used as the second output plane.
17854 The index of the input plane to be used as the third output plane.
17857 The index of the input plane to be used as the fourth output plane.
17861 The first plane has the index 0. The default is to keep the input unchanged.
17863 @subsection Examples
17867 Swap the second and third planes of the input:
17869 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
17873 @anchor{signalstats}
17874 @section signalstats
17875 Evaluate various visual metrics that assist in determining issues associated
17876 with the digitization of analog video media.
17878 By default the filter will log these metadata values:
17882 Display the minimal Y value contained within the input frame. Expressed in
17886 Display the Y value at the 10% percentile within the input frame. Expressed in
17890 Display the average Y value within the input frame. Expressed in range of
17894 Display the Y value at the 90% percentile within the input frame. Expressed in
17898 Display the maximum Y value contained within the input frame. Expressed in
17902 Display the minimal U value contained within the input frame. Expressed in
17906 Display the U value at the 10% percentile within the input frame. Expressed in
17910 Display the average U value within the input frame. Expressed in range of
17914 Display the U value at the 90% percentile within the input frame. Expressed in
17918 Display the maximum U value contained within the input frame. Expressed in
17922 Display the minimal V value contained within the input frame. Expressed in
17926 Display the V value at the 10% percentile within the input frame. Expressed in
17930 Display the average V value within the input frame. Expressed in range of
17934 Display the V value at the 90% percentile within the input frame. Expressed in
17938 Display the maximum V value contained within the input frame. Expressed in
17942 Display the minimal saturation value contained within the input frame.
17943 Expressed in range of [0-~181.02].
17946 Display the saturation value at the 10% percentile within the input frame.
17947 Expressed in range of [0-~181.02].
17950 Display the average saturation value within the input frame. Expressed in range
17954 Display the saturation value at the 90% percentile within the input frame.
17955 Expressed in range of [0-~181.02].
17958 Display the maximum saturation value contained within the input frame.
17959 Expressed in range of [0-~181.02].
17962 Display the median value for hue within the input frame. Expressed in range of
17966 Display the average value for hue within the input frame. Expressed in range of
17970 Display the average of sample value difference between all values of the Y
17971 plane in the current frame and corresponding values of the previous input frame.
17972 Expressed in range of [0-255].
17975 Display the average of sample value difference between all values of the U
17976 plane in the current frame and corresponding values of the previous input frame.
17977 Expressed in range of [0-255].
17980 Display the average of sample value difference between all values of the V
17981 plane in the current frame and corresponding values of the previous input frame.
17982 Expressed in range of [0-255].
17985 Display bit depth of Y plane in current frame.
17986 Expressed in range of [0-16].
17989 Display bit depth of U plane in current frame.
17990 Expressed in range of [0-16].
17993 Display bit depth of V plane in current frame.
17994 Expressed in range of [0-16].
17997 The filter accepts the following options:
18003 @option{stat} specify an additional form of image analysis.
18004 @option{out} output video with the specified type of pixel highlighted.
18006 Both options accept the following values:
18010 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
18011 unlike the neighboring pixels of the same field. Examples of temporal outliers
18012 include the results of video dropouts, head clogs, or tape tracking issues.
18015 Identify @var{vertical line repetition}. Vertical line repetition includes
18016 similar rows of pixels within a frame. In born-digital video vertical line
18017 repetition is common, but this pattern is uncommon in video digitized from an
18018 analog source. When it occurs in video that results from the digitization of an
18019 analog source it can indicate concealment from a dropout compensator.
18022 Identify pixels that fall outside of legal broadcast range.
18026 Set the highlight color for the @option{out} option. The default color is
18030 @subsection Examples
18034 Output data of various video metrics:
18036 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
18040 Output specific data about the minimum and maximum values of the Y plane per frame:
18042 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
18046 Playback video while highlighting pixels that are outside of broadcast range in red.
18048 ffplay example.mov -vf signalstats="out=brng:color=red"
18052 Playback video with signalstats metadata drawn over the frame.
18054 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
18057 The contents of signalstat_drawtext.txt used in the command are:
18060 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
18061 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
18062 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
18063 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
18071 Calculates the MPEG-7 Video Signature. The filter can handle more than one
18072 input. In this case the matching between the inputs can be calculated additionally.
18073 The filter always passes through the first input. The signature of each stream can
18074 be written into a file.
18076 It accepts the following options:
18080 Enable or disable the matching process.
18082 Available values are:
18086 Disable the calculation of a matching (default).
18088 Calculate the matching for the whole video and output whether the whole video
18089 matches or only parts.
18091 Calculate only until a matching is found or the video ends. Should be faster in
18096 Set the number of inputs. The option value must be a non negative integer.
18097 Default value is 1.
18100 Set the path to which the output is written. If there is more than one input,
18101 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
18102 integer), that will be replaced with the input number. If no filename is
18103 specified, no output will be written. This is the default.
18106 Choose the output format.
18108 Available values are:
18112 Use the specified binary representation (default).
18114 Use the specified xml representation.
18118 Set threshold to detect one word as similar. The option value must be an integer
18119 greater than zero. The default value is 9000.
18122 Set threshold to detect all words as similar. The option value must be an integer
18123 greater than zero. The default value is 60000.
18126 Set threshold to detect frames as similar. The option value must be an integer
18127 greater than zero. The default value is 116.
18130 Set the minimum length of a sequence in frames to recognize it as matching
18131 sequence. The option value must be a non negative integer value.
18132 The default value is 0.
18135 Set the minimum relation, that matching frames to all frames must have.
18136 The option value must be a double value between 0 and 1. The default value is 0.5.
18139 @subsection Examples
18143 To calculate the signature of an input video and store it in signature.bin:
18145 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
18149 To detect whether two videos match and store the signatures in XML format in
18150 signature0.xml and signature1.xml:
18152 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 -
18160 Blur the input video without impacting the outlines.
18162 It accepts the following options:
18165 @item luma_radius, lr
18166 Set the luma radius. The option value must be a float number in
18167 the range [0.1,5.0] that specifies the variance of the gaussian filter
18168 used to blur the image (slower if larger). Default value is 1.0.
18170 @item luma_strength, ls
18171 Set the luma strength. The option value must be a float number
18172 in the range [-1.0,1.0] that configures the blurring. A value included
18173 in [0.0,1.0] will blur the image whereas a value included in
18174 [-1.0,0.0] will sharpen the image. Default value is 1.0.
18176 @item luma_threshold, lt
18177 Set the luma threshold used as a coefficient to determine
18178 whether a pixel should be blurred or not. The option value must be an
18179 integer in the range [-30,30]. A value of 0 will filter all the image,
18180 a value included in [0,30] will filter flat areas and a value included
18181 in [-30,0] will filter edges. Default value is 0.
18183 @item chroma_radius, cr
18184 Set the chroma radius. The option value must be a float number in
18185 the range [0.1,5.0] that specifies the variance of the gaussian filter
18186 used to blur the image (slower if larger). Default value is @option{luma_radius}.
18188 @item chroma_strength, cs
18189 Set the chroma strength. The option value must be a float number
18190 in the range [-1.0,1.0] that configures the blurring. A value included
18191 in [0.0,1.0] will blur the image whereas a value included in
18192 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
18194 @item chroma_threshold, ct
18195 Set the chroma threshold used as a coefficient to determine
18196 whether a pixel should be blurred or not. The option value must be an
18197 integer in the range [-30,30]. A value of 0 will filter all the image,
18198 a value included in [0,30] will filter flat areas and a value included
18199 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
18202 If a chroma option is not explicitly set, the corresponding luma value
18206 Apply sobel operator to input video stream.
18208 The filter accepts the following option:
18212 Set which planes will be processed, unprocessed planes will be copied.
18213 By default value 0xf, all planes will be processed.
18216 Set value which will be multiplied with filtered result.
18219 Set value which will be added to filtered result.
18222 @subsection Commands
18224 This filter supports the all above options as @ref{commands}.
18229 Apply a simple postprocessing filter that compresses and decompresses the image
18230 at several (or - in the case of @option{quality} level @code{6} - all) shifts
18231 and average the results.
18233 The filter accepts the following options:
18237 Set quality. This option defines the number of levels for averaging. It accepts
18238 an integer in the range 0-6. If set to @code{0}, the filter will have no
18239 effect. A value of @code{6} means the higher quality. For each increment of
18240 that value the speed drops by a factor of approximately 2. Default value is
18244 Force a constant quantization parameter. If not set, the filter will use the QP
18245 from the video stream (if available).
18248 Set thresholding mode. Available modes are:
18252 Set hard thresholding (default).
18254 Set soft thresholding (better de-ringing effect, but likely blurrier).
18257 @item use_bframe_qp
18258 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
18259 option may cause flicker since the B-Frames have often larger QP. Default is
18260 @code{0} (not enabled).
18263 @subsection Commands
18265 This filter supports the following commands:
18267 @item quality, level
18268 Set quality level. The value @code{max} can be used to set the maximum level,
18269 currently @code{6}.
18275 Scale the input by applying one of the super-resolution methods based on
18276 convolutional neural networks. Supported models:
18280 Super-Resolution Convolutional Neural Network model (SRCNN).
18281 See @url{https://arxiv.org/abs/1501.00092}.
18284 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
18285 See @url{https://arxiv.org/abs/1609.05158}.
18288 Training scripts as well as scripts for model file (.pb) saving can be found at
18289 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
18290 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
18292 Native model files (.model) can be generated from TensorFlow model
18293 files (.pb) by using tools/python/convert.py
18295 The filter accepts the following options:
18299 Specify which DNN backend to use for model loading and execution. This option accepts
18300 the following values:
18304 Native implementation of DNN loading and execution.
18307 TensorFlow backend. To enable this backend you
18308 need to install the TensorFlow for C library (see
18309 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
18310 @code{--enable-libtensorflow}
18313 Default value is @samp{native}.
18316 Set path to model file specifying network architecture and its parameters.
18317 Note that different backends use different file formats. TensorFlow backend
18318 can load files for both formats, while native backend can load files for only
18322 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
18323 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
18324 input upscaled using bicubic upscaling with proper scale factor.
18327 This feature can also be finished with @ref{dnn_processing} filter.
18331 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
18333 This filter takes in input two input videos, the first input is
18334 considered the "main" source and is passed unchanged to the
18335 output. The second input is used as a "reference" video for computing
18338 Both video inputs must have the same resolution and pixel format for
18339 this filter to work correctly. Also it assumes that both inputs
18340 have the same number of frames, which are compared one by one.
18342 The filter stores the calculated SSIM of each frame.
18344 The description of the accepted parameters follows.
18347 @item stats_file, f
18348 If specified the filter will use the named file to save the SSIM of
18349 each individual frame. When filename equals "-" the data is sent to
18353 The file printed if @var{stats_file} is selected, contains a sequence of
18354 key/value pairs of the form @var{key}:@var{value} for each compared
18357 A description of each shown parameter follows:
18361 sequential number of the input frame, starting from 1
18363 @item Y, U, V, R, G, B
18364 SSIM of the compared frames for the component specified by the suffix.
18367 SSIM of the compared frames for the whole frame.
18370 Same as above but in dB representation.
18373 This filter also supports the @ref{framesync} options.
18375 @subsection Examples
18380 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
18381 [main][ref] ssim="stats_file=stats.log" [out]
18384 On this example the input file being processed is compared with the
18385 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
18386 is stored in @file{stats.log}.
18389 Another example with both psnr and ssim at same time:
18391 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
18395 Another example with different containers:
18397 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 -
18403 Convert between different stereoscopic image formats.
18405 The filters accept the following options:
18409 Set stereoscopic image format of input.
18411 Available values for input image formats are:
18414 side by side parallel (left eye left, right eye right)
18417 side by side crosseye (right eye left, left eye right)
18420 side by side parallel with half width resolution
18421 (left eye left, right eye right)
18424 side by side crosseye with half width resolution
18425 (right eye left, left eye right)
18429 above-below (left eye above, right eye below)
18433 above-below (right eye above, left eye below)
18437 above-below with half height resolution
18438 (left eye above, right eye below)
18442 above-below with half height resolution
18443 (right eye above, left eye below)
18446 alternating frames (left eye first, right eye second)
18449 alternating frames (right eye first, left eye second)
18452 interleaved rows (left eye has top row, right eye starts on next row)
18455 interleaved rows (right eye has top row, left eye starts on next row)
18458 interleaved columns, left eye first
18461 interleaved columns, right eye first
18463 Default value is @samp{sbsl}.
18467 Set stereoscopic image format of output.
18471 side by side parallel (left eye left, right eye right)
18474 side by side crosseye (right eye left, left eye right)
18477 side by side parallel with half width resolution
18478 (left eye left, right eye right)
18481 side by side crosseye with half width resolution
18482 (right eye left, left eye right)
18486 above-below (left eye above, right eye below)
18490 above-below (right eye above, left eye below)
18494 above-below with half height resolution
18495 (left eye above, right eye below)
18499 above-below with half height resolution
18500 (right eye above, left eye below)
18503 alternating frames (left eye first, right eye second)
18506 alternating frames (right eye first, left eye second)
18509 interleaved rows (left eye has top row, right eye starts on next row)
18512 interleaved rows (right eye has top row, left eye starts on next row)
18515 anaglyph red/blue gray
18516 (red filter on left eye, blue filter on right eye)
18519 anaglyph red/green gray
18520 (red filter on left eye, green filter on right eye)
18523 anaglyph red/cyan gray
18524 (red filter on left eye, cyan filter on right eye)
18527 anaglyph red/cyan half colored
18528 (red filter on left eye, cyan filter on right eye)
18531 anaglyph red/cyan color
18532 (red filter on left eye, cyan filter on right eye)
18535 anaglyph red/cyan color optimized with the least squares projection of dubois
18536 (red filter on left eye, cyan filter on right eye)
18539 anaglyph green/magenta gray
18540 (green filter on left eye, magenta filter on right eye)
18543 anaglyph green/magenta half colored
18544 (green filter on left eye, magenta filter on right eye)
18547 anaglyph green/magenta colored
18548 (green filter on left eye, magenta filter on right eye)
18551 anaglyph green/magenta color optimized with the least squares projection of dubois
18552 (green filter on left eye, magenta filter on right eye)
18555 anaglyph yellow/blue gray
18556 (yellow filter on left eye, blue filter on right eye)
18559 anaglyph yellow/blue half colored
18560 (yellow filter on left eye, blue filter on right eye)
18563 anaglyph yellow/blue colored
18564 (yellow filter on left eye, blue filter on right eye)
18567 anaglyph yellow/blue color optimized with the least squares projection of dubois
18568 (yellow filter on left eye, blue filter on right eye)
18571 mono output (left eye only)
18574 mono output (right eye only)
18577 checkerboard, left eye first
18580 checkerboard, right eye first
18583 interleaved columns, left eye first
18586 interleaved columns, right eye first
18592 Default value is @samp{arcd}.
18595 @subsection Examples
18599 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
18605 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
18611 @section streamselect, astreamselect
18612 Select video or audio streams.
18614 The filter accepts the following options:
18618 Set number of inputs. Default is 2.
18621 Set input indexes to remap to outputs.
18624 @subsection Commands
18626 The @code{streamselect} and @code{astreamselect} filter supports the following
18631 Set input indexes to remap to outputs.
18634 @subsection Examples
18638 Select first 5 seconds 1st stream and rest of time 2nd stream:
18640 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
18644 Same as above, but for audio:
18646 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
18653 Draw subtitles on top of input video using the libass library.
18655 To enable compilation of this filter you need to configure FFmpeg with
18656 @code{--enable-libass}. This filter also requires a build with libavcodec and
18657 libavformat to convert the passed subtitles file to ASS (Advanced Substation
18658 Alpha) subtitles format.
18660 The filter accepts the following options:
18664 Set the filename of the subtitle file to read. It must be specified.
18666 @item original_size
18667 Specify the size of the original video, the video for which the ASS file
18668 was composed. For the syntax of this option, check the
18669 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18670 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
18671 correctly scale the fonts if the aspect ratio has been changed.
18674 Set a directory path containing fonts that can be used by the filter.
18675 These fonts will be used in addition to whatever the font provider uses.
18678 Process alpha channel, by default alpha channel is untouched.
18681 Set subtitles input character encoding. @code{subtitles} filter only. Only
18682 useful if not UTF-8.
18684 @item stream_index, si
18685 Set subtitles stream index. @code{subtitles} filter only.
18688 Override default style or script info parameters of the subtitles. It accepts a
18689 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
18692 If the first key is not specified, it is assumed that the first value
18693 specifies the @option{filename}.
18695 For example, to render the file @file{sub.srt} on top of the input
18696 video, use the command:
18701 which is equivalent to:
18703 subtitles=filename=sub.srt
18706 To render the default subtitles stream from file @file{video.mkv}, use:
18708 subtitles=video.mkv
18711 To render the second subtitles stream from that file, use:
18713 subtitles=video.mkv:si=1
18716 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
18717 @code{DejaVu Serif}, use:
18719 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
18722 @section super2xsai
18724 Scale the input by 2x and smooth using the Super2xSaI (Scale and
18725 Interpolate) pixel art scaling algorithm.
18727 Useful for enlarging pixel art images without reducing sharpness.
18731 Swap two rectangular objects in video.
18733 This filter accepts the following options:
18743 Set 1st rect x coordinate.
18746 Set 1st rect y coordinate.
18749 Set 2nd rect x coordinate.
18752 Set 2nd rect y coordinate.
18754 All expressions are evaluated once for each frame.
18757 The all options are expressions containing the following constants:
18762 The input width and height.
18765 same as @var{w} / @var{h}
18768 input sample aspect ratio
18771 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
18774 The number of the input frame, starting from 0.
18777 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
18780 the position in the file of the input frame, NAN if unknown
18787 Blend successive video frames.
18793 Apply telecine process to the video.
18795 This filter accepts the following options:
18804 The default value is @code{top}.
18808 A string of numbers representing the pulldown pattern you wish to apply.
18809 The default value is @code{23}.
18813 Some typical patterns:
18818 24p: 2332 (preferred)
18825 24p: 222222222223 ("Euro pulldown")
18830 @section thistogram
18832 Compute and draw a color distribution histogram for the input video across time.
18834 Unlike @ref{histogram} video filter which only shows histogram of single input frame
18835 at certain time, this filter shows also past histograms of number of frames defined
18836 by @code{width} option.
18838 The computed histogram is a representation of the color component
18839 distribution in an image.
18841 The filter accepts the following options:
18845 Set width of single color component output. Default value is @code{0}.
18846 Value of @code{0} means width will be picked from input video.
18847 This also set number of passed histograms to keep.
18848 Allowed range is [0, 8192].
18850 @item display_mode, d
18852 It accepts the following values:
18855 Per color component graphs are placed below each other.
18858 Per color component graphs are placed side by side.
18861 Presents information identical to that in the @code{parade}, except
18862 that the graphs representing color components are superimposed directly
18865 Default is @code{stack}.
18867 @item levels_mode, m
18868 Set mode. Can be either @code{linear}, or @code{logarithmic}.
18869 Default is @code{linear}.
18871 @item components, c
18872 Set what color components to display.
18873 Default is @code{7}.
18876 Set background opacity. Default is @code{0.9}.
18879 Show envelope. Default is disabled.
18882 Set envelope color. Default is @code{gold}.
18887 Available values for slide is:
18890 Draw new frame when right border is reached.
18893 Replace old columns with new ones.
18896 Scroll from right to left.
18899 Scroll from left to right.
18902 Draw single picture.
18905 Default is @code{replace}.
18910 Apply threshold effect to video stream.
18912 This filter needs four video streams to perform thresholding.
18913 First stream is stream we are filtering.
18914 Second stream is holding threshold values, third stream is holding min values,
18915 and last, fourth stream is holding max values.
18917 The filter accepts the following option:
18921 Set which planes will be processed, unprocessed planes will be copied.
18922 By default value 0xf, all planes will be processed.
18925 For example if first stream pixel's component value is less then threshold value
18926 of pixel component from 2nd threshold stream, third stream value will picked,
18927 otherwise fourth stream pixel component value will be picked.
18929 Using color source filter one can perform various types of thresholding:
18931 @subsection Examples
18935 Binary threshold, using gray color as threshold:
18937 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
18941 Inverted binary threshold, using gray color as threshold:
18943 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
18947 Truncate binary threshold, using gray color as threshold:
18949 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
18953 Threshold to zero, using gray color as threshold:
18955 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
18959 Inverted threshold to zero, using gray color as threshold:
18961 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
18966 Select the most representative frame in a given sequence of consecutive frames.
18968 The filter accepts the following options:
18972 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
18973 will pick one of them, and then handle the next batch of @var{n} frames until
18974 the end. Default is @code{100}.
18977 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
18978 value will result in a higher memory usage, so a high value is not recommended.
18980 @subsection Examples
18984 Extract one picture each 50 frames:
18990 Complete example of a thumbnail creation with @command{ffmpeg}:
18992 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
18999 Tile several successive frames together.
19001 The @ref{untile} filter can do the reverse.
19003 The filter accepts the following options:
19008 Set the grid size (i.e. the number of lines and columns). For the syntax of
19009 this option, check the
19010 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19013 Set the maximum number of frames to render in the given area. It must be less
19014 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
19015 the area will be used.
19018 Set the outer border margin in pixels.
19021 Set the inner border thickness (i.e. the number of pixels between frames). For
19022 more advanced padding options (such as having different values for the edges),
19023 refer to the pad video filter.
19026 Specify the color of the unused area. For the syntax of this option, check the
19027 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
19028 The default value of @var{color} is "black".
19031 Set the number of frames to overlap when tiling several successive frames together.
19032 The value must be between @code{0} and @var{nb_frames - 1}.
19035 Set the number of frames to initially be empty before displaying first output frame.
19036 This controls how soon will one get first output frame.
19037 The value must be between @code{0} and @var{nb_frames - 1}.
19040 @subsection Examples
19044 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
19046 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
19048 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
19049 duplicating each output frame to accommodate the originally detected frame
19053 Display @code{5} pictures in an area of @code{3x2} frames,
19054 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
19055 mixed flat and named options:
19057 tile=3x2:nb_frames=5:padding=7:margin=2
19061 @section tinterlace
19063 Perform various types of temporal field interlacing.
19065 Frames are counted starting from 1, so the first input frame is
19068 The filter accepts the following options:
19073 Specify the mode of the interlacing. This option can also be specified
19074 as a value alone. See below for a list of values for this option.
19076 Available values are:
19080 Move odd frames into the upper field, even into the lower field,
19081 generating a double height frame at half frame rate.
19085 Frame 1 Frame 2 Frame 3 Frame 4
19087 11111 22222 33333 44444
19088 11111 22222 33333 44444
19089 11111 22222 33333 44444
19090 11111 22222 33333 44444
19104 Only output odd frames, even frames are dropped, generating a frame with
19105 unchanged height at half frame rate.
19110 Frame 1 Frame 2 Frame 3 Frame 4
19112 11111 22222 33333 44444
19113 11111 22222 33333 44444
19114 11111 22222 33333 44444
19115 11111 22222 33333 44444
19125 Only output even frames, odd frames are dropped, generating a frame with
19126 unchanged height at half frame rate.
19131 Frame 1 Frame 2 Frame 3 Frame 4
19133 11111 22222 33333 44444
19134 11111 22222 33333 44444
19135 11111 22222 33333 44444
19136 11111 22222 33333 44444
19146 Expand each frame to full height, but pad alternate lines with black,
19147 generating a frame with double height at the same input frame rate.
19152 Frame 1 Frame 2 Frame 3 Frame 4
19154 11111 22222 33333 44444
19155 11111 22222 33333 44444
19156 11111 22222 33333 44444
19157 11111 22222 33333 44444
19160 11111 ..... 33333 .....
19161 ..... 22222 ..... 44444
19162 11111 ..... 33333 .....
19163 ..... 22222 ..... 44444
19164 11111 ..... 33333 .....
19165 ..... 22222 ..... 44444
19166 11111 ..... 33333 .....
19167 ..... 22222 ..... 44444
19171 @item interleave_top, 4
19172 Interleave the upper field from odd frames with the lower field from
19173 even frames, generating a frame with unchanged height at half frame rate.
19178 Frame 1 Frame 2 Frame 3 Frame 4
19180 11111<- 22222 33333<- 44444
19181 11111 22222<- 33333 44444<-
19182 11111<- 22222 33333<- 44444
19183 11111 22222<- 33333 44444<-
19193 @item interleave_bottom, 5
19194 Interleave the lower field from odd frames with the upper field from
19195 even frames, generating a frame with unchanged height at half frame rate.
19200 Frame 1 Frame 2 Frame 3 Frame 4
19202 11111 22222<- 33333 44444<-
19203 11111<- 22222 33333<- 44444
19204 11111 22222<- 33333 44444<-
19205 11111<- 22222 33333<- 44444
19215 @item interlacex2, 6
19216 Double frame rate with unchanged height. Frames are inserted each
19217 containing the second temporal field from the previous input frame and
19218 the first temporal field from the next input frame. This mode relies on
19219 the top_field_first flag. Useful for interlaced video displays with no
19220 field synchronisation.
19225 Frame 1 Frame 2 Frame 3 Frame 4
19227 11111 22222 33333 44444
19228 11111 22222 33333 44444
19229 11111 22222 33333 44444
19230 11111 22222 33333 44444
19233 11111 22222 22222 33333 33333 44444 44444
19234 11111 11111 22222 22222 33333 33333 44444
19235 11111 22222 22222 33333 33333 44444 44444
19236 11111 11111 22222 22222 33333 33333 44444
19241 Move odd frames into the upper field, even into the lower field,
19242 generating a double height frame at same frame rate.
19247 Frame 1 Frame 2 Frame 3 Frame 4
19249 11111 22222 33333 44444
19250 11111 22222 33333 44444
19251 11111 22222 33333 44444
19252 11111 22222 33333 44444
19255 11111 33333 33333 55555
19256 22222 22222 44444 44444
19257 11111 33333 33333 55555
19258 22222 22222 44444 44444
19259 11111 33333 33333 55555
19260 22222 22222 44444 44444
19261 11111 33333 33333 55555
19262 22222 22222 44444 44444
19267 Numeric values are deprecated but are accepted for backward
19268 compatibility reasons.
19270 Default mode is @code{merge}.
19273 Specify flags influencing the filter process.
19275 Available value for @var{flags} is:
19278 @item low_pass_filter, vlpf
19279 Enable linear vertical low-pass filtering in the filter.
19280 Vertical low-pass filtering is required when creating an interlaced
19281 destination from a progressive source which contains high-frequency
19282 vertical detail. Filtering will reduce interlace 'twitter' and Moire
19285 @item complex_filter, cvlpf
19286 Enable complex vertical low-pass filtering.
19287 This will slightly less reduce interlace 'twitter' and Moire
19288 patterning but better retain detail and subjective sharpness impression.
19291 Bypass already interlaced frames, only adjust the frame rate.
19294 Vertical low-pass filtering and bypassing already interlaced frames can only be
19295 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
19300 Pick median pixels from several successive input video frames.
19302 The filter accepts the following options:
19306 Set radius of median filter.
19307 Default is 1. Allowed range is from 1 to 127.
19310 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
19313 Set median percentile. Default value is @code{0.5}.
19314 Default value of @code{0.5} will pick always median values, while @code{0} will pick
19315 minimum values, and @code{1} maximum values.
19318 @subsection Commands
19320 This filter supports all above options as @ref{commands}, excluding option @code{radius}.
19324 Mix successive video frames.
19326 A description of the accepted options follows.
19330 The number of successive frames to mix. If unspecified, it defaults to 3.
19333 Specify weight of each input video frame.
19334 Each weight is separated by space. If number of weights is smaller than
19335 number of @var{frames} last specified weight will be used for all remaining
19339 Specify scale, if it is set it will be multiplied with sum
19340 of each weight multiplied with pixel values to give final destination
19341 pixel value. By default @var{scale} is auto scaled to sum of weights.
19344 @subsection Examples
19348 Average 7 successive frames:
19350 tmix=frames=7:weights="1 1 1 1 1 1 1"
19354 Apply simple temporal convolution:
19356 tmix=frames=3:weights="-1 3 -1"
19360 Similar as above but only showing temporal differences:
19362 tmix=frames=3:weights="-1 2 -1":scale=1
19368 Tone map colors from different dynamic ranges.
19370 This filter expects data in single precision floating point, as it needs to
19371 operate on (and can output) out-of-range values. Another filter, such as
19372 @ref{zscale}, is needed to convert the resulting frame to a usable format.
19374 The tonemapping algorithms implemented only work on linear light, so input
19375 data should be linearized beforehand (and possibly correctly tagged).
19378 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
19381 @subsection Options
19382 The filter accepts the following options.
19386 Set the tone map algorithm to use.
19388 Possible values are:
19391 Do not apply any tone map, only desaturate overbright pixels.
19394 Hard-clip any out-of-range values. Use it for perfect color accuracy for
19395 in-range values, while distorting out-of-range values.
19398 Stretch the entire reference gamut to a linear multiple of the display.
19401 Fit a logarithmic transfer between the tone curves.
19404 Preserve overall image brightness with a simple curve, using nonlinear
19405 contrast, which results in flattening details and degrading color accuracy.
19408 Preserve both dark and bright details better than @var{reinhard}, at the cost
19409 of slightly darkening everything. Use it when detail preservation is more
19410 important than color and brightness accuracy.
19413 Smoothly map out-of-range values, while retaining contrast and colors for
19414 in-range material as much as possible. Use it when color accuracy is more
19415 important than detail preservation.
19421 Tune the tone mapping algorithm.
19423 This affects the following algorithms:
19429 Specifies the scale factor to use while stretching.
19433 Specifies the exponent of the function.
19437 Specify an extra linear coefficient to multiply into the signal before clipping.
19441 Specify the local contrast coefficient at the display peak.
19442 Default to 0.5, which means that in-gamut values will be about half as bright
19449 Specify the transition point from linear to mobius transform. Every value
19450 below this point is guaranteed to be mapped 1:1. The higher the value, the
19451 more accurate the result will be, at the cost of losing bright details.
19452 Default to 0.3, which due to the steep initial slope still preserves in-range
19453 colors fairly accurately.
19457 Apply desaturation for highlights that exceed this level of brightness. The
19458 higher the parameter, the more color information will be preserved. This
19459 setting helps prevent unnaturally blown-out colors for super-highlights, by
19460 (smoothly) turning into white instead. This makes images feel more natural,
19461 at the cost of reducing information about out-of-range colors.
19463 The default of 2.0 is somewhat conservative and will mostly just apply to
19464 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
19466 This option works only if the input frame has a supported color tag.
19469 Override signal/nominal/reference peak with this value. Useful when the
19470 embedded peak information in display metadata is not reliable or when tone
19471 mapping from a lower range to a higher range.
19476 Temporarily pad video frames.
19478 The filter accepts the following options:
19482 Specify number of delay frames before input video stream. Default is 0.
19485 Specify number of padding frames after input video stream.
19486 Set to -1 to pad indefinitely. Default is 0.
19489 Set kind of frames added to beginning of stream.
19490 Can be either @var{add} or @var{clone}.
19491 With @var{add} frames of solid-color are added.
19492 With @var{clone} frames are clones of first frame.
19493 Default is @var{add}.
19496 Set kind of frames added to end of stream.
19497 Can be either @var{add} or @var{clone}.
19498 With @var{add} frames of solid-color are added.
19499 With @var{clone} frames are clones of last frame.
19500 Default is @var{add}.
19502 @item start_duration, stop_duration
19503 Specify the duration of the start/stop delay. See
19504 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19505 for the accepted syntax.
19506 These options override @var{start} and @var{stop}. Default is 0.
19509 Specify the color of the padded area. For the syntax of this option,
19510 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
19511 manual,ffmpeg-utils}.
19513 The default value of @var{color} is "black".
19519 Transpose rows with columns in the input video and optionally flip it.
19521 It accepts the following parameters:
19526 Specify the transposition direction.
19528 Can assume the following values:
19530 @item 0, 4, cclock_flip
19531 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
19539 Rotate by 90 degrees clockwise, that is:
19547 Rotate by 90 degrees counterclockwise, that is:
19554 @item 3, 7, clock_flip
19555 Rotate by 90 degrees clockwise and vertically flip, that is:
19563 For values between 4-7, the transposition is only done if the input
19564 video geometry is portrait and not landscape. These values are
19565 deprecated, the @code{passthrough} option should be used instead.
19567 Numerical values are deprecated, and should be dropped in favor of
19568 symbolic constants.
19571 Do not apply the transposition if the input geometry matches the one
19572 specified by the specified value. It accepts the following values:
19575 Always apply transposition.
19577 Preserve portrait geometry (when @var{height} >= @var{width}).
19579 Preserve landscape geometry (when @var{width} >= @var{height}).
19582 Default value is @code{none}.
19585 For example to rotate by 90 degrees clockwise and preserve portrait
19588 transpose=dir=1:passthrough=portrait
19591 The command above can also be specified as:
19593 transpose=1:portrait
19596 @section transpose_npp
19598 Transpose rows with columns in the input video and optionally flip it.
19599 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
19601 It accepts the following parameters:
19606 Specify the transposition direction.
19608 Can assume the following values:
19611 Rotate by 90 degrees counterclockwise and vertically flip. (default)
19614 Rotate by 90 degrees clockwise.
19617 Rotate by 90 degrees counterclockwise.
19620 Rotate by 90 degrees clockwise and vertically flip.
19624 Do not apply the transposition if the input geometry matches the one
19625 specified by the specified value. It accepts the following values:
19628 Always apply transposition. (default)
19630 Preserve portrait geometry (when @var{height} >= @var{width}).
19632 Preserve landscape geometry (when @var{width} >= @var{height}).
19638 Trim the input so that the output contains one continuous subpart of the input.
19640 It accepts the following parameters:
19643 Specify the time of the start of the kept section, i.e. the frame with the
19644 timestamp @var{start} will be the first frame in the output.
19647 Specify the time of the first frame that will be dropped, i.e. the frame
19648 immediately preceding the one with the timestamp @var{end} will be the last
19649 frame in the output.
19652 This is the same as @var{start}, except this option sets the start timestamp
19653 in timebase units instead of seconds.
19656 This is the same as @var{end}, except this option sets the end timestamp
19657 in timebase units instead of seconds.
19660 The maximum duration of the output in seconds.
19663 The number of the first frame that should be passed to the output.
19666 The number of the first frame that should be dropped.
19669 @option{start}, @option{end}, and @option{duration} are expressed as time
19670 duration specifications; see
19671 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19672 for the accepted syntax.
19674 Note that the first two sets of the start/end options and the @option{duration}
19675 option look at the frame timestamp, while the _frame variants simply count the
19676 frames that pass through the filter. Also note that this filter does not modify
19677 the timestamps. If you wish for the output timestamps to start at zero, insert a
19678 setpts filter after the trim filter.
19680 If multiple start or end options are set, this filter tries to be greedy and
19681 keep all the frames that match at least one of the specified constraints. To keep
19682 only the part that matches all the constraints at once, chain multiple trim
19685 The defaults are such that all the input is kept. So it is possible to set e.g.
19686 just the end values to keep everything before the specified time.
19691 Drop everything except the second minute of input:
19693 ffmpeg -i INPUT -vf trim=60:120
19697 Keep only the first second:
19699 ffmpeg -i INPUT -vf trim=duration=1
19704 @section unpremultiply
19705 Apply alpha unpremultiply effect to input video stream using first plane
19706 of second stream as alpha.
19708 Both streams must have same dimensions and same pixel format.
19710 The filter accepts the following option:
19714 Set which planes will be processed, unprocessed planes will be copied.
19715 By default value 0xf, all planes will be processed.
19717 If the format has 1 or 2 components, then luma is bit 0.
19718 If the format has 3 or 4 components:
19719 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
19720 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
19721 If present, the alpha channel is always the last bit.
19724 Do not require 2nd input for processing, instead use alpha plane from input stream.
19730 Sharpen or blur the input video.
19732 It accepts the following parameters:
19735 @item luma_msize_x, lx
19736 Set the luma matrix horizontal size. It must be an odd integer between
19737 3 and 23. The default value is 5.
19739 @item luma_msize_y, ly
19740 Set the luma matrix vertical size. It must be an odd integer between 3
19741 and 23. The default value is 5.
19743 @item luma_amount, la
19744 Set the luma effect strength. It must be a floating point number, reasonable
19745 values lay between -1.5 and 1.5.
19747 Negative values will blur the input video, while positive values will
19748 sharpen it, a value of zero will disable the effect.
19750 Default value is 1.0.
19752 @item chroma_msize_x, cx
19753 Set the chroma matrix horizontal size. It must be an odd integer
19754 between 3 and 23. The default value is 5.
19756 @item chroma_msize_y, cy
19757 Set the chroma matrix vertical size. It must be an odd integer
19758 between 3 and 23. The default value is 5.
19760 @item chroma_amount, ca
19761 Set the chroma effect strength. It must be a floating point number, reasonable
19762 values lay between -1.5 and 1.5.
19764 Negative values will blur the input video, while positive values will
19765 sharpen it, a value of zero will disable the effect.
19767 Default value is 0.0.
19771 All parameters are optional and default to the equivalent of the
19772 string '5:5:1.0:5:5:0.0'.
19774 @subsection Examples
19778 Apply strong luma sharpen effect:
19780 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
19784 Apply a strong blur of both luma and chroma parameters:
19786 unsharp=7:7:-2:7:7:-2
19793 Decompose a video made of tiled images into the individual images.
19795 The frame rate of the output video is the frame rate of the input video
19796 multiplied by the number of tiles.
19798 This filter does the reverse of @ref{tile}.
19800 The filter accepts the following options:
19805 Set the grid size (i.e. the number of lines and columns). For the syntax of
19806 this option, check the
19807 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19810 @subsection Examples
19814 Produce a 1-second video from a still image file made of 25 frames stacked
19815 vertically, like an analogic film reel:
19817 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
19823 Apply ultra slow/simple postprocessing filter that compresses and decompresses
19824 the image at several (or - in the case of @option{quality} level @code{8} - all)
19825 shifts and average the results.
19827 The way this differs from the behavior of spp is that uspp actually encodes &
19828 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
19829 DCT similar to MJPEG.
19831 The filter accepts the following options:
19835 Set quality. This option defines the number of levels for averaging. It accepts
19836 an integer in the range 0-8. If set to @code{0}, the filter will have no
19837 effect. A value of @code{8} means the higher quality. For each increment of
19838 that value the speed drops by a factor of approximately 2. Default value is
19842 Force a constant quantization parameter. If not set, the filter will use the QP
19843 from the video stream (if available).
19848 Convert 360 videos between various formats.
19850 The filter accepts the following options:
19856 Set format of the input/output video.
19864 Equirectangular projection.
19869 Cubemap with 3x2/6x1/1x6 layout.
19871 Format specific options:
19876 Set padding proportion for the input/output cubemap. Values in decimals.
19883 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)
19886 Default value is @b{@samp{0}}.
19887 Maximum value is @b{@samp{0.1}}.
19891 Set fixed padding for the input/output cubemap. Values in pixels.
19893 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
19897 Set order of faces for the input/output cubemap. Choose one direction for each position.
19899 Designation of directions:
19915 Default value is @b{@samp{rludfb}}.
19919 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
19921 Designation of angles:
19924 0 degrees clockwise
19926 90 degrees clockwise
19928 180 degrees clockwise
19930 270 degrees clockwise
19933 Default value is @b{@samp{000000}}.
19937 Equi-Angular Cubemap.
19944 Format specific options:
19949 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19951 If diagonal field of view is set it overrides horizontal and vertical field of view.
19956 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19958 If diagonal field of view is set it overrides horizontal and vertical field of view.
19964 Format specific options:
19969 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19971 If diagonal field of view is set it overrides horizontal and vertical field of view.
19976 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19978 If diagonal field of view is set it overrides horizontal and vertical field of view.
19984 Facebook's 360 formats.
19987 Stereographic format.
19989 Format specific options:
19994 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19996 If diagonal field of view is set it overrides horizontal and vertical field of view.
20001 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20003 If diagonal field of view is set it overrides horizontal and vertical field of view.
20010 Ball format, gives significant distortion toward the back.
20013 Hammer-Aitoff map projection format.
20016 Sinusoidal map projection format.
20019 Fisheye projection.
20021 Format specific options:
20026 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20028 If diagonal field of view is set it overrides horizontal and vertical field of view.
20033 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20035 If diagonal field of view is set it overrides horizontal and vertical field of view.
20039 Pannini projection.
20041 Format specific options:
20044 Set output pannini parameter.
20047 Set input pannini parameter.
20051 Cylindrical projection.
20053 Format specific options:
20058 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20060 If diagonal field of view is set it overrides horizontal and vertical field of view.
20065 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20067 If diagonal field of view is set it overrides horizontal and vertical field of view.
20071 Perspective projection. @i{(output only)}
20073 Format specific options:
20076 Set perspective parameter.
20080 Tetrahedron projection.
20083 Truncated square pyramid projection.
20087 Half equirectangular projection.
20092 Format specific options:
20097 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20099 If diagonal field of view is set it overrides horizontal and vertical field of view.
20104 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20106 If diagonal field of view is set it overrides horizontal and vertical field of view.
20110 Orthographic format.
20112 Format specific options:
20117 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20119 If diagonal field of view is set it overrides horizontal and vertical field of view.
20124 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20126 If diagonal field of view is set it overrides horizontal and vertical field of view.
20130 Octahedron projection.
20134 Set interpolation method.@*
20135 @i{Note: more complex interpolation methods require much more memory to run.}
20145 Bilinear interpolation.
20147 Lagrange9 interpolation.
20150 Bicubic interpolation.
20153 Lanczos interpolation.
20156 Spline16 interpolation.
20159 Gaussian interpolation.
20161 Mitchell interpolation.
20164 Default value is @b{@samp{line}}.
20168 Set the output video resolution.
20170 Default resolution depends on formats.
20174 Set the input/output stereo format.
20185 Default value is @b{@samp{2d}} for input and output format.
20190 Set rotation for the output video. Values in degrees.
20193 Set rotation order for the output video. Choose one item for each position.
20204 Default value is @b{@samp{ypr}}.
20209 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
20213 Set if input video is flipped horizontally/vertically. Boolean values.
20216 Set if input video is transposed. Boolean value, by default disabled.
20219 Set if output video needs to be transposed. Boolean value, by default disabled.
20222 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
20225 @subsection Examples
20229 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
20231 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
20234 Extract back view of Equi-Angular Cubemap:
20236 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
20239 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
20241 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
20245 @subsection Commands
20247 This filter supports subset of above options as @ref{commands}.
20249 @section vaguedenoiser
20251 Apply a wavelet based denoiser.
20253 It transforms each frame from the video input into the wavelet domain,
20254 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
20255 the obtained coefficients. It does an inverse wavelet transform after.
20256 Due to wavelet properties, it should give a nice smoothed result, and
20257 reduced noise, without blurring picture features.
20259 This filter accepts the following options:
20263 The filtering strength. The higher, the more filtered the video will be.
20264 Hard thresholding can use a higher threshold than soft thresholding
20265 before the video looks overfiltered. Default value is 2.
20268 The filtering method the filter will use.
20270 It accepts the following values:
20273 All values under the threshold will be zeroed.
20276 All values under the threshold will be zeroed. All values above will be
20277 reduced by the threshold.
20280 Scales or nullifies coefficients - intermediary between (more) soft and
20281 (less) hard thresholding.
20284 Default is garrote.
20287 Number of times, the wavelet will decompose the picture. Picture can't
20288 be decomposed beyond a particular point (typically, 8 for a 640x480
20289 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
20292 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
20295 A list of the planes to process. By default all planes are processed.
20298 The threshold type the filter will use.
20300 It accepts the following values:
20303 Threshold used is same for all decompositions.
20306 Threshold used depends also on each decomposition coefficients.
20309 Default is universal.
20312 @section vectorscope
20314 Display 2 color component values in the two dimensional graph (which is called
20317 This filter accepts the following options:
20321 Set vectorscope mode.
20323 It accepts the following values:
20327 Gray values are displayed on graph, higher brightness means more pixels have
20328 same component color value on location in graph. This is the default mode.
20331 Gray values are displayed on graph. Surrounding pixels values which are not
20332 present in video frame are drawn in gradient of 2 color components which are
20333 set by option @code{x} and @code{y}. The 3rd color component is static.
20336 Actual color components values present in video frame are displayed on graph.
20339 Similar as color2 but higher frequency of same values @code{x} and @code{y}
20340 on graph increases value of another color component, which is luminance by
20341 default values of @code{x} and @code{y}.
20344 Actual colors present in video frame are displayed on graph. If two different
20345 colors map to same position on graph then color with higher value of component
20346 not present in graph is picked.
20349 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
20350 component picked from radial gradient.
20354 Set which color component will be represented on X-axis. Default is @code{1}.
20357 Set which color component will be represented on Y-axis. Default is @code{2}.
20360 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
20361 of color component which represents frequency of (X, Y) location in graph.
20366 No envelope, this is default.
20369 Instant envelope, even darkest single pixel will be clearly highlighted.
20372 Hold maximum and minimum values presented in graph over time. This way you
20373 can still spot out of range values without constantly looking at vectorscope.
20376 Peak and instant envelope combined together.
20380 Set what kind of graticule to draw.
20389 Set graticule opacity.
20392 Set graticule flags.
20396 Draw graticule for white point.
20399 Draw graticule for black point.
20402 Draw color points short names.
20406 Set background opacity.
20408 @item lthreshold, l
20409 Set low threshold for color component not represented on X or Y axis.
20410 Values lower than this value will be ignored. Default is 0.
20411 Note this value is multiplied with actual max possible value one pixel component
20412 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
20415 @item hthreshold, h
20416 Set high threshold for color component not represented on X or Y axis.
20417 Values higher than this value will be ignored. Default is 1.
20418 Note this value is multiplied with actual max possible value one pixel component
20419 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
20420 is 0.9 * 255 = 230.
20422 @item colorspace, c
20423 Set what kind of colorspace to use when drawing graticule.
20433 Set color tint for gray/tint vectorscope mode. By default both options are zero.
20434 This means no tint, and output will remain gray.
20437 @anchor{vidstabdetect}
20438 @section vidstabdetect
20440 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
20441 @ref{vidstabtransform} for pass 2.
20443 This filter generates a file with relative translation and rotation
20444 transform information about subsequent frames, which is then used by
20445 the @ref{vidstabtransform} filter.
20447 To enable compilation of this filter you need to configure FFmpeg with
20448 @code{--enable-libvidstab}.
20450 This filter accepts the following options:
20454 Set the path to the file used to write the transforms information.
20455 Default value is @file{transforms.trf}.
20458 Set how shaky the video is and how quick the camera is. It accepts an
20459 integer in the range 1-10, a value of 1 means little shakiness, a
20460 value of 10 means strong shakiness. Default value is 5.
20463 Set the accuracy of the detection process. It must be a value in the
20464 range 1-15. A value of 1 means low accuracy, a value of 15 means high
20465 accuracy. Default value is 15.
20468 Set stepsize of the search process. The region around minimum is
20469 scanned with 1 pixel resolution. Default value is 6.
20472 Set minimum contrast. Below this value a local measurement field is
20473 discarded. Must be a floating point value in the range 0-1. Default
20477 Set reference frame number for tripod mode.
20479 If enabled, the motion of the frames is compared to a reference frame
20480 in the filtered stream, identified by the specified number. The idea
20481 is to compensate all movements in a more-or-less static scene and keep
20482 the camera view absolutely still.
20484 If set to 0, it is disabled. The frames are counted starting from 1.
20487 Show fields and transforms in the resulting frames. It accepts an
20488 integer in the range 0-2. Default value is 0, which disables any
20492 @subsection Examples
20496 Use default values:
20502 Analyze strongly shaky movie and put the results in file
20503 @file{mytransforms.trf}:
20505 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
20509 Visualize the result of internal transformations in the resulting
20512 vidstabdetect=show=1
20516 Analyze a video with medium shakiness using @command{ffmpeg}:
20518 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
20522 @anchor{vidstabtransform}
20523 @section vidstabtransform
20525 Video stabilization/deshaking: pass 2 of 2,
20526 see @ref{vidstabdetect} for pass 1.
20528 Read a file with transform information for each frame and
20529 apply/compensate them. Together with the @ref{vidstabdetect}
20530 filter this can be used to deshake videos. See also
20531 @url{http://public.hronopik.de/vid.stab}. It is important to also use
20532 the @ref{unsharp} filter, see below.
20534 To enable compilation of this filter you need to configure FFmpeg with
20535 @code{--enable-libvidstab}.
20537 @subsection Options
20541 Set path to the file used to read the transforms. Default value is
20542 @file{transforms.trf}.
20545 Set the number of frames (value*2 + 1) used for lowpass filtering the
20546 camera movements. Default value is 10.
20548 For example a number of 10 means that 21 frames are used (10 in the
20549 past and 10 in the future) to smoothen the motion in the video. A
20550 larger value leads to a smoother video, but limits the acceleration of
20551 the camera (pan/tilt movements). 0 is a special case where a static
20552 camera is simulated.
20555 Set the camera path optimization algorithm.
20557 Accepted values are:
20560 gaussian kernel low-pass filter on camera motion (default)
20562 averaging on transformations
20566 Set maximal number of pixels to translate frames. Default value is -1,
20570 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
20571 value is -1, meaning no limit.
20574 Specify how to deal with borders that may be visible due to movement
20577 Available values are:
20580 keep image information from previous frame (default)
20582 fill the border black
20586 Invert transforms if set to 1. Default value is 0.
20589 Consider transforms as relative to previous frame if set to 1,
20590 absolute if set to 0. Default value is 0.
20593 Set percentage to zoom. A positive value will result in a zoom-in
20594 effect, a negative value in a zoom-out effect. Default value is 0 (no
20598 Set optimal zooming to avoid borders.
20600 Accepted values are:
20605 optimal static zoom value is determined (only very strong movements
20606 will lead to visible borders) (default)
20608 optimal adaptive zoom value is determined (no borders will be
20609 visible), see @option{zoomspeed}
20612 Note that the value given at zoom is added to the one calculated here.
20615 Set percent to zoom maximally each frame (enabled when
20616 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
20620 Specify type of interpolation.
20622 Available values are:
20627 linear only horizontal
20629 linear in both directions (default)
20631 cubic in both directions (slow)
20635 Enable virtual tripod mode if set to 1, which is equivalent to
20636 @code{relative=0:smoothing=0}. Default value is 0.
20638 Use also @code{tripod} option of @ref{vidstabdetect}.
20641 Increase log verbosity if set to 1. Also the detected global motions
20642 are written to the temporary file @file{global_motions.trf}. Default
20646 @subsection Examples
20650 Use @command{ffmpeg} for a typical stabilization with default values:
20652 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
20655 Note the use of the @ref{unsharp} filter which is always recommended.
20658 Zoom in a bit more and load transform data from a given file:
20660 vidstabtransform=zoom=5:input="mytransforms.trf"
20664 Smoothen the video even more:
20666 vidstabtransform=smoothing=30
20672 Flip the input video vertically.
20674 For example, to vertically flip a video with @command{ffmpeg}:
20676 ffmpeg -i in.avi -vf "vflip" out.avi
20681 Detect variable frame rate video.
20683 This filter tries to detect if the input is variable or constant frame rate.
20685 At end it will output number of frames detected as having variable delta pts,
20686 and ones with constant delta pts.
20687 If there was frames with variable delta, than it will also show min, max and
20688 average delta encountered.
20692 Boost or alter saturation.
20694 The filter accepts the following options:
20697 Set strength of boost if positive value or strength of alter if negative value.
20698 Default is 0. Allowed range is from -2 to 2.
20701 Set the red balance. Default is 1. Allowed range is from -10 to 10.
20704 Set the green balance. Default is 1. Allowed range is from -10 to 10.
20707 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
20710 Set the red luma coefficient.
20713 Set the green luma coefficient.
20716 Set the blue luma coefficient.
20719 If @code{intensity} is negative and this is set to 1, colors will change,
20720 otherwise colors will be less saturated, more towards gray.
20723 @subsection Commands
20725 This filter supports the all above options as @ref{commands}.
20730 Make or reverse a natural vignetting effect.
20732 The filter accepts the following options:
20736 Set lens angle expression as a number of radians.
20738 The value is clipped in the @code{[0,PI/2]} range.
20740 Default value: @code{"PI/5"}
20744 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
20748 Set forward/backward mode.
20750 Available modes are:
20753 The larger the distance from the central point, the darker the image becomes.
20756 The larger the distance from the central point, the brighter the image becomes.
20757 This can be used to reverse a vignette effect, though there is no automatic
20758 detection to extract the lens @option{angle} and other settings (yet). It can
20759 also be used to create a burning effect.
20762 Default value is @samp{forward}.
20765 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
20767 It accepts the following values:
20770 Evaluate expressions only once during the filter initialization.
20773 Evaluate expressions for each incoming frame. This is way slower than the
20774 @samp{init} mode since it requires all the scalers to be re-computed, but it
20775 allows advanced dynamic expressions.
20778 Default value is @samp{init}.
20781 Set dithering to reduce the circular banding effects. Default is @code{1}
20785 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
20786 Setting this value to the SAR of the input will make a rectangular vignetting
20787 following the dimensions of the video.
20789 Default is @code{1/1}.
20792 @subsection Expressions
20794 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
20795 following parameters.
20800 input width and height
20803 the number of input frame, starting from 0
20806 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
20807 @var{TB} units, NAN if undefined
20810 frame rate of the input video, NAN if the input frame rate is unknown
20813 the PTS (Presentation TimeStamp) of the filtered video frame,
20814 expressed in seconds, NAN if undefined
20817 time base of the input video
20821 @subsection Examples
20825 Apply simple strong vignetting effect:
20831 Make a flickering vignetting:
20833 vignette='PI/4+random(1)*PI/50':eval=frame
20838 @section vmafmotion
20840 Obtain the average VMAF motion score of a video.
20841 It is one of the component metrics of VMAF.
20843 The obtained average motion score is printed through the logging system.
20845 The filter accepts the following options:
20849 If specified, the filter will use the named file to save the motion score of
20850 each frame with respect to the previous frame.
20851 When filename equals "-" the data is sent to standard output.
20856 ffmpeg -i ref.mpg -vf vmafmotion -f null -
20860 Stack input videos vertically.
20862 All streams must be of same pixel format and of same width.
20864 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
20865 to create same output.
20867 The filter accepts the following options:
20871 Set number of input streams. Default is 2.
20874 If set to 1, force the output to terminate when the shortest input
20875 terminates. Default value is 0.
20880 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
20881 Deinterlacing Filter").
20883 Based on the process described by Martin Weston for BBC R&D, and
20884 implemented based on the de-interlace algorithm written by Jim
20885 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
20886 uses filter coefficients calculated by BBC R&D.
20888 This filter uses field-dominance information in frame to decide which
20889 of each pair of fields to place first in the output.
20890 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
20892 There are two sets of filter coefficients, so called "simple"
20893 and "complex". Which set of filter coefficients is used can
20894 be set by passing an optional parameter:
20898 Set the interlacing filter coefficients. Accepts one of the following values:
20902 Simple filter coefficient set.
20904 More-complex filter coefficient set.
20906 Default value is @samp{complex}.
20909 Specify which frames to deinterlace. Accepts one of the following values:
20913 Deinterlace all frames,
20915 Only deinterlace frames marked as interlaced.
20918 Default value is @samp{all}.
20922 Video waveform monitor.
20924 The waveform monitor plots color component intensity. By default luminance
20925 only. Each column of the waveform corresponds to a column of pixels in the
20928 It accepts the following options:
20932 Can be either @code{row}, or @code{column}. Default is @code{column}.
20933 In row mode, the graph on the left side represents color component value 0 and
20934 the right side represents value = 255. In column mode, the top side represents
20935 color component value = 0 and bottom side represents value = 255.
20938 Set intensity. Smaller values are useful to find out how many values of the same
20939 luminance are distributed across input rows/columns.
20940 Default value is @code{0.04}. Allowed range is [0, 1].
20943 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
20944 In mirrored mode, higher values will be represented on the left
20945 side for @code{row} mode and at the top for @code{column} mode. Default is
20946 @code{1} (mirrored).
20950 It accepts the following values:
20953 Presents information identical to that in the @code{parade}, except
20954 that the graphs representing color components are superimposed directly
20957 This display mode makes it easier to spot relative differences or similarities
20958 in overlapping areas of the color components that are supposed to be identical,
20959 such as neutral whites, grays, or blacks.
20962 Display separate graph for the color components side by side in
20963 @code{row} mode or one below the other in @code{column} mode.
20966 Display separate graph for the color components side by side in
20967 @code{column} mode or one below the other in @code{row} mode.
20969 Using this display mode makes it easy to spot color casts in the highlights
20970 and shadows of an image, by comparing the contours of the top and the bottom
20971 graphs of each waveform. Since whites, grays, and blacks are characterized
20972 by exactly equal amounts of red, green, and blue, neutral areas of the picture
20973 should display three waveforms of roughly equal width/height. If not, the
20974 correction is easy to perform by making level adjustments the three waveforms.
20976 Default is @code{stack}.
20978 @item components, c
20979 Set which color components to display. Default is 1, which means only luminance
20980 or red color component if input is in RGB colorspace. If is set for example to
20981 7 it will display all 3 (if) available color components.
20986 No envelope, this is default.
20989 Instant envelope, minimum and maximum values presented in graph will be easily
20990 visible even with small @code{step} value.
20993 Hold minimum and maximum values presented in graph across time. This way you
20994 can still spot out of range values without constantly looking at waveforms.
20997 Peak and instant envelope combined together.
21003 No filtering, this is default.
21006 Luma and chroma combined together.
21009 Similar as above, but shows difference between blue and red chroma.
21012 Similar as above, but use different colors.
21015 Similar as above, but again with different colors.
21018 Displays only chroma.
21021 Displays actual color value on waveform.
21024 Similar as above, but with luma showing frequency of chroma values.
21028 Set which graticule to display.
21032 Do not display graticule.
21035 Display green graticule showing legal broadcast ranges.
21038 Display orange graticule showing legal broadcast ranges.
21041 Display invert graticule showing legal broadcast ranges.
21045 Set graticule opacity.
21048 Set graticule flags.
21052 Draw numbers above lines. By default enabled.
21055 Draw dots instead of lines.
21059 Set scale used for displaying graticule.
21066 Default is digital.
21069 Set background opacity.
21073 Set tint for output.
21074 Only used with lowpass filter and when display is not overlay and input
21075 pixel formats are not RGB.
21078 @section weave, doubleweave
21080 The @code{weave} takes a field-based video input and join
21081 each two sequential fields into single frame, producing a new double
21082 height clip with half the frame rate and half the frame count.
21084 The @code{doubleweave} works same as @code{weave} but without
21085 halving frame rate and frame count.
21087 It accepts the following option:
21091 Set first field. Available values are:
21095 Set the frame as top-field-first.
21098 Set the frame as bottom-field-first.
21102 @subsection Examples
21106 Interlace video using @ref{select} and @ref{separatefields} filter:
21108 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
21113 Apply the xBR high-quality magnification filter which is designed for pixel
21114 art. It follows a set of edge-detection rules, see
21115 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
21117 It accepts the following option:
21121 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
21122 @code{3xBR} and @code{4} for @code{4xBR}.
21123 Default is @code{3}.
21128 Apply cross fade from one input video stream to another input video stream.
21129 The cross fade is applied for specified duration.
21131 The filter accepts the following options:
21135 Set one of available transition effects:
21183 Default transition effect is fade.
21186 Set cross fade duration in seconds.
21187 Default duration is 1 second.
21190 Set cross fade start relative to first input stream in seconds.
21191 Default offset is 0.
21194 Set expression for custom transition effect.
21196 The expressions can use the following variables and functions:
21201 The coordinates of the current sample.
21205 The width and height of the image.
21208 Progress of transition effect.
21211 Currently processed plane.
21214 Return value of first input at current location and plane.
21217 Return value of second input at current location and plane.
21223 Return the value of the pixel at location (@var{x},@var{y}) of the
21224 first/second/third/fourth component of first input.
21230 Return the value of the pixel at location (@var{x},@var{y}) of the
21231 first/second/third/fourth component of second input.
21235 @subsection Examples
21239 Cross fade from one input video to another input video, with fade transition and duration of transition
21240 of 2 seconds starting at offset of 5 seconds:
21242 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
21247 Pick median pixels from several input videos.
21249 The filter accepts the following options:
21253 Set number of inputs.
21254 Default is 3. Allowed range is from 3 to 255.
21255 If number of inputs is even number, than result will be mean value between two median values.
21258 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
21261 Set median percentile. Default value is @code{0.5}.
21262 Default value of @code{0.5} will pick always median values, while @code{0} will pick
21263 minimum values, and @code{1} maximum values.
21266 @subsection Commands
21268 This filter supports all above options as @ref{commands}, excluding option @code{inputs}.
21271 Stack video inputs into custom layout.
21273 All streams must be of same pixel format.
21275 The filter accepts the following options:
21279 Set number of input streams. Default is 2.
21282 Specify layout of inputs.
21283 This option requires the desired layout configuration to be explicitly set by the user.
21284 This sets position of each video input in output. Each input
21285 is separated by '|'.
21286 The first number represents the column, and the second number represents the row.
21287 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
21288 where X is video input from which to take width or height.
21289 Multiple values can be used when separated by '+'. In such
21290 case values are summed together.
21292 Note that if inputs are of different sizes gaps may appear, as not all of
21293 the output video frame will be filled. Similarly, videos can overlap each
21294 other if their position doesn't leave enough space for the full frame of
21297 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
21298 a layout must be set by the user.
21301 If set to 1, force the output to terminate when the shortest input
21302 terminates. Default value is 0.
21305 If set to valid color, all unused pixels will be filled with that color.
21306 By default fill is set to none, so it is disabled.
21309 @subsection Examples
21313 Display 4 inputs into 2x2 grid.
21317 input1(0, 0) | input3(w0, 0)
21318 input2(0, h0) | input4(w0, h0)
21322 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
21325 Note that if inputs are of different sizes, gaps or overlaps may occur.
21328 Display 4 inputs into 1x4 grid.
21335 input4(0, h0+h1+h2)
21339 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
21342 Note that if inputs are of different widths, unused space will appear.
21345 Display 9 inputs into 3x3 grid.
21349 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
21350 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
21351 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
21355 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
21358 Note that if inputs are of different sizes, gaps or overlaps may occur.
21361 Display 16 inputs into 4x4 grid.
21365 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
21366 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
21367 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
21368 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
21372 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|
21373 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
21376 Note that if inputs are of different sizes, gaps or overlaps may occur.
21383 Deinterlace the input video ("yadif" means "yet another deinterlacing
21386 It accepts the following parameters:
21392 The interlacing mode to adopt. It accepts one of the following values:
21395 @item 0, send_frame
21396 Output one frame for each frame.
21397 @item 1, send_field
21398 Output one frame for each field.
21399 @item 2, send_frame_nospatial
21400 Like @code{send_frame}, but it skips the spatial interlacing check.
21401 @item 3, send_field_nospatial
21402 Like @code{send_field}, but it skips the spatial interlacing check.
21405 The default value is @code{send_frame}.
21408 The picture field parity assumed for the input interlaced video. It accepts one
21409 of the following values:
21413 Assume the top field is first.
21415 Assume the bottom field is first.
21417 Enable automatic detection of field parity.
21420 The default value is @code{auto}.
21421 If the interlacing is unknown or the decoder does not export this information,
21422 top field first will be assumed.
21425 Specify which frames to deinterlace. Accepts one of the following
21430 Deinterlace all frames.
21431 @item 1, interlaced
21432 Only deinterlace frames marked as interlaced.
21435 The default value is @code{all}.
21438 @section yadif_cuda
21440 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
21441 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
21444 It accepts the following parameters:
21450 The interlacing mode to adopt. It accepts one of the following values:
21453 @item 0, send_frame
21454 Output one frame for each frame.
21455 @item 1, send_field
21456 Output one frame for each field.
21457 @item 2, send_frame_nospatial
21458 Like @code{send_frame}, but it skips the spatial interlacing check.
21459 @item 3, send_field_nospatial
21460 Like @code{send_field}, but it skips the spatial interlacing check.
21463 The default value is @code{send_frame}.
21466 The picture field parity assumed for the input interlaced video. It accepts one
21467 of the following values:
21471 Assume the top field is first.
21473 Assume the bottom field is first.
21475 Enable automatic detection of field parity.
21478 The default value is @code{auto}.
21479 If the interlacing is unknown or the decoder does not export this information,
21480 top field first will be assumed.
21483 Specify which frames to deinterlace. Accepts one of the following
21488 Deinterlace all frames.
21489 @item 1, interlaced
21490 Only deinterlace frames marked as interlaced.
21493 The default value is @code{all}.
21498 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
21499 The algorithm is described in
21500 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
21502 It accepts the following parameters:
21506 Set the window radius. Default value is 3.
21509 Set which planes to filter. Default is only the first plane.
21512 Set blur strength. Default value is 128.
21515 @subsection Commands
21516 This filter supports same @ref{commands} as options.
21520 Apply Zoom & Pan effect.
21522 This filter accepts the following options:
21526 Set the zoom expression. Range is 1-10. Default is 1.
21530 Set the x and y expression. Default is 0.
21533 Set the duration expression in number of frames.
21534 This sets for how many number of frames effect will last for
21535 single input image.
21538 Set the output image size, default is 'hd720'.
21541 Set the output frame rate, default is '25'.
21544 Each expression can contain the following constants:
21563 Output frame count.
21566 The input timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
21568 @item out_time, time, ot
21569 The output timestamp expressed in seconds.
21573 Last calculated 'x' and 'y' position from 'x' and 'y' expression
21574 for current input frame.
21578 'x' and 'y' of last output frame of previous input frame or 0 when there was
21579 not yet such frame (first input frame).
21582 Last calculated zoom from 'z' expression for current input frame.
21585 Last calculated zoom of last output frame of previous input frame.
21588 Number of output frames for current input frame. Calculated from 'd' expression
21589 for each input frame.
21592 number of output frames created for previous input frame
21595 Rational number: input width / input height
21598 sample aspect ratio
21601 display aspect ratio
21605 @subsection Examples
21609 Zoom in up to 1.5x and pan at same time to some spot near center of picture:
21611 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
21615 Zoom in up to 1.5x and pan always at center of picture:
21617 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21621 Same as above but without pausing:
21623 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21627 Zoom in 2x into center of picture only for the first second of the input video:
21629 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21636 Scale (resize) the input video, using the z.lib library:
21637 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
21638 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
21640 The zscale filter forces the output display aspect ratio to be the same
21641 as the input, by changing the output sample aspect ratio.
21643 If the input image format is different from the format requested by
21644 the next filter, the zscale filter will convert the input to the
21647 @subsection Options
21648 The filter accepts the following options.
21653 Set the output video dimension expression. Default value is the input
21656 If the @var{width} or @var{w} value is 0, the input width is used for
21657 the output. If the @var{height} or @var{h} value is 0, the input height
21658 is used for the output.
21660 If one and only one of the values is -n with n >= 1, the zscale filter
21661 will use a value that maintains the aspect ratio of the input image,
21662 calculated from the other specified dimension. After that it will,
21663 however, make sure that the calculated dimension is divisible by n and
21664 adjust the value if necessary.
21666 If both values are -n with n >= 1, the behavior will be identical to
21667 both values being set to 0 as previously detailed.
21669 See below for the list of accepted constants for use in the dimension
21673 Set the video size. For the syntax of this option, check the
21674 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21677 Set the dither type.
21679 Possible values are:
21684 @item error_diffusion
21690 Set the resize filter type.
21692 Possible values are:
21702 Default is bilinear.
21705 Set the color range.
21707 Possible values are:
21714 Default is same as input.
21717 Set the color primaries.
21719 Possible values are:
21729 Default is same as input.
21732 Set the transfer characteristics.
21734 Possible values are:
21748 Default is same as input.
21751 Set the colorspace matrix.
21753 Possible value are:
21764 Default is same as input.
21767 Set the input color range.
21769 Possible values are:
21776 Default is same as input.
21778 @item primariesin, pin
21779 Set the input color primaries.
21781 Possible values are:
21791 Default is same as input.
21793 @item transferin, tin
21794 Set the input transfer characteristics.
21796 Possible values are:
21807 Default is same as input.
21809 @item matrixin, min
21810 Set the input colorspace matrix.
21812 Possible value are:
21824 Set the output chroma location.
21826 Possible values are:
21837 @item chromalin, cin
21838 Set the input chroma location.
21840 Possible values are:
21852 Set the nominal peak luminance.
21855 The values of the @option{w} and @option{h} options are expressions
21856 containing the following constants:
21861 The input width and height
21865 These are the same as @var{in_w} and @var{in_h}.
21869 The output (scaled) width and height
21873 These are the same as @var{out_w} and @var{out_h}
21876 The same as @var{iw} / @var{ih}
21879 input sample aspect ratio
21882 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
21886 horizontal and vertical input chroma subsample values. For example for the
21887 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21891 horizontal and vertical output chroma subsample values. For example for the
21892 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21895 @subsection Commands
21897 This filter supports the following commands:
21901 Set the output video dimension expression.
21902 The command accepts the same syntax of the corresponding option.
21904 If the specified expression is not valid, it is kept at its current
21908 @c man end VIDEO FILTERS
21910 @chapter OpenCL Video Filters
21911 @c man begin OPENCL VIDEO FILTERS
21913 Below is a description of the currently available OpenCL video filters.
21915 To enable compilation of these filters you need to configure FFmpeg with
21916 @code{--enable-opencl}.
21918 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
21921 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
21922 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
21923 given device parameters.
21925 @item -filter_hw_device @var{name}
21926 Pass the hardware device called @var{name} to all filters in any filter graph.
21930 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
21934 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
21936 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
21940 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.
21942 @section avgblur_opencl
21944 Apply average blur filter.
21946 The filter accepts the following options:
21950 Set horizontal radius size.
21951 Range is @code{[1, 1024]} and default value is @code{1}.
21954 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21957 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
21960 @subsection Example
21964 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.
21966 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
21970 @section boxblur_opencl
21972 Apply a boxblur algorithm to the input video.
21974 It accepts the following parameters:
21978 @item luma_radius, lr
21979 @item luma_power, lp
21980 @item chroma_radius, cr
21981 @item chroma_power, cp
21982 @item alpha_radius, ar
21983 @item alpha_power, ap
21987 A description of the accepted options follows.
21990 @item luma_radius, lr
21991 @item chroma_radius, cr
21992 @item alpha_radius, ar
21993 Set an expression for the box radius in pixels used for blurring the
21994 corresponding input plane.
21996 The radius value must be a non-negative number, and must not be
21997 greater than the value of the expression @code{min(w,h)/2} for the
21998 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
22001 Default value for @option{luma_radius} is "2". If not specified,
22002 @option{chroma_radius} and @option{alpha_radius} default to the
22003 corresponding value set for @option{luma_radius}.
22005 The expressions can contain the following constants:
22009 The input width and height in pixels.
22013 The input chroma image width and height in pixels.
22017 The horizontal and vertical chroma subsample values. For example, for the
22018 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
22021 @item luma_power, lp
22022 @item chroma_power, cp
22023 @item alpha_power, ap
22024 Specify how many times the boxblur filter is applied to the
22025 corresponding plane.
22027 Default value for @option{luma_power} is 2. If not specified,
22028 @option{chroma_power} and @option{alpha_power} default to the
22029 corresponding value set for @option{luma_power}.
22031 A value of 0 will disable the effect.
22034 @subsection Examples
22036 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.
22040 Apply a boxblur filter with the luma, chroma, and alpha radius
22041 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.
22043 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
22044 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
22048 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.
22050 For the luma plane, a 2x2 box radius will be run once.
22052 For the chroma plane, a 4x4 box radius will be run 5 times.
22054 For the alpha plane, a 3x3 box radius will be run 7 times.
22056 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
22060 @section colorkey_opencl
22061 RGB colorspace color keying.
22063 The filter accepts the following options:
22067 The color which will be replaced with transparency.
22070 Similarity percentage with the key color.
22072 0.01 matches only the exact key color, while 1.0 matches everything.
22077 0.0 makes pixels either fully transparent, or not transparent at all.
22079 Higher values result in semi-transparent pixels, with a higher transparency
22080 the more similar the pixels color is to the key color.
22083 @subsection Examples
22087 Make every semi-green pixel in the input transparent with some slight blending:
22089 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
22093 @section convolution_opencl
22095 Apply convolution of 3x3, 5x5, 7x7 matrix.
22097 The filter accepts the following options:
22104 Set matrix for each plane.
22105 Matrix is sequence of 9, 25 or 49 signed numbers.
22106 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
22112 Set multiplier for calculated value for each plane.
22113 If unset or 0, it will be sum of all matrix elements.
22114 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
22120 Set bias for each plane. This value is added to the result of the multiplication.
22121 Useful for making the overall image brighter or darker.
22122 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
22126 @subsection Examples
22132 -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
22138 -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
22142 Apply edge enhance:
22144 -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
22150 -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
22154 Apply laplacian edge detector which includes diagonals:
22156 -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
22162 -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
22166 @section erosion_opencl
22168 Apply erosion effect to the video.
22170 This filter replaces the pixel by the local(3x3) minimum.
22172 It accepts the following options:
22179 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22180 If @code{0}, plane will remain unchanged.
22183 Flag which specifies the pixel to refer to.
22184 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22186 Flags to local 3x3 coordinates region centered on @code{x}:
22195 @subsection Example
22199 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.
22201 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22205 @section deshake_opencl
22206 Feature-point based video stabilization filter.
22208 The filter accepts the following options:
22212 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
22215 Whether or not additional debug info should be displayed, both in the processed output and in the console.
22217 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
22219 Viewing point matches in the output video is only supported for RGB input.
22221 Defaults to @code{0}.
22223 @item adaptive_crop
22224 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
22226 Defaults to @code{1}.
22228 @item refine_features
22229 Whether or not feature points should be refined at a sub-pixel level.
22231 This can be turned off for a slight performance gain at the cost of precision.
22233 Defaults to @code{1}.
22235 @item smooth_strength
22236 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
22238 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
22240 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
22242 Defaults to @code{0.0}.
22244 @item smooth_window_multiplier
22245 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
22247 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
22249 Acceptable values range from @code{0.1} to @code{10.0}.
22251 Larger values increase the amount of motion data available for determining how to smooth the camera path,
22252 potentially improving smoothness, but also increase latency and memory usage.
22254 Defaults to @code{2.0}.
22258 @subsection Examples
22262 Stabilize a video with a fixed, medium smoothing strength:
22264 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
22268 Stabilize a video with debugging (both in console and in rendered video):
22270 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
22274 @section dilation_opencl
22276 Apply dilation effect to the video.
22278 This filter replaces the pixel by the local(3x3) maximum.
22280 It accepts the following options:
22287 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22288 If @code{0}, plane will remain unchanged.
22291 Flag which specifies the pixel to refer to.
22292 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22294 Flags to local 3x3 coordinates region centered on @code{x}:
22303 @subsection Example
22307 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.
22309 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22313 @section nlmeans_opencl
22315 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
22317 @section overlay_opencl
22319 Overlay one video on top of another.
22321 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
22322 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
22324 The filter accepts the following options:
22329 Set the x coordinate of the overlaid video on the main video.
22330 Default value is @code{0}.
22333 Set the y coordinate of the overlaid video on the main video.
22334 Default value is @code{0}.
22338 @subsection Examples
22342 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
22344 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22347 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
22349 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22354 @section pad_opencl
22356 Add paddings to the input image, and place the original input at the
22357 provided @var{x}, @var{y} coordinates.
22359 It accepts the following options:
22364 Specify an expression for the size of the output image with the
22365 paddings added. If the value for @var{width} or @var{height} is 0, the
22366 corresponding input size is used for the output.
22368 The @var{width} expression can reference the value set by the
22369 @var{height} expression, and vice versa.
22371 The default value of @var{width} and @var{height} is 0.
22375 Specify the offsets to place the input image at within the padded area,
22376 with respect to the top/left border of the output image.
22378 The @var{x} expression can reference the value set by the @var{y}
22379 expression, and vice versa.
22381 The default value of @var{x} and @var{y} is 0.
22383 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
22384 so the input image is centered on the padded area.
22387 Specify the color of the padded area. For the syntax of this option,
22388 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
22389 manual,ffmpeg-utils}.
22392 Pad to an aspect instead to a resolution.
22395 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
22396 options are expressions containing the following constants:
22401 The input video width and height.
22405 These are the same as @var{in_w} and @var{in_h}.
22409 The output width and height (the size of the padded area), as
22410 specified by the @var{width} and @var{height} expressions.
22414 These are the same as @var{out_w} and @var{out_h}.
22418 The x and y offsets as specified by the @var{x} and @var{y}
22419 expressions, or NAN if not yet specified.
22422 same as @var{iw} / @var{ih}
22425 input sample aspect ratio
22428 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
22431 @section prewitt_opencl
22433 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
22435 The filter accepts the following option:
22439 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22442 Set value which will be multiplied with filtered result.
22443 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22446 Set value which will be added to filtered result.
22447 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22450 @subsection Example
22454 Apply the Prewitt operator with scale set to 2 and delta set to 10.
22456 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
22460 @anchor{program_opencl}
22461 @section program_opencl
22463 Filter video using an OpenCL program.
22468 OpenCL program source file.
22471 Kernel name in program.
22474 Number of inputs to the filter. Defaults to 1.
22477 Size of output frames. Defaults to the same as the first input.
22481 The @code{program_opencl} filter also supports the @ref{framesync} options.
22483 The program source file must contain a kernel function with the given name,
22484 which will be run once for each plane of the output. Each run on a plane
22485 gets enqueued as a separate 2D global NDRange with one work-item for each
22486 pixel to be generated. The global ID offset for each work-item is therefore
22487 the coordinates of a pixel in the destination image.
22489 The kernel function needs to take the following arguments:
22492 Destination image, @var{__write_only image2d_t}.
22494 This image will become the output; the kernel should write all of it.
22496 Frame index, @var{unsigned int}.
22498 This is a counter starting from zero and increasing by one for each frame.
22500 Source images, @var{__read_only image2d_t}.
22502 These are the most recent images on each input. The kernel may read from
22503 them to generate the output, but they can't be written to.
22510 Copy the input to the output (output must be the same size as the input).
22512 __kernel void copy(__write_only image2d_t destination,
22513 unsigned int index,
22514 __read_only image2d_t source)
22516 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
22518 int2 location = (int2)(get_global_id(0), get_global_id(1));
22520 float4 value = read_imagef(source, sampler, location);
22522 write_imagef(destination, location, value);
22527 Apply a simple transformation, rotating the input by an amount increasing
22528 with the index counter. Pixel values are linearly interpolated by the
22529 sampler, and the output need not have the same dimensions as the input.
22531 __kernel void rotate_image(__write_only image2d_t dst,
22532 unsigned int index,
22533 __read_only image2d_t src)
22535 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22536 CLK_FILTER_LINEAR);
22538 float angle = (float)index / 100.0f;
22540 float2 dst_dim = convert_float2(get_image_dim(dst));
22541 float2 src_dim = convert_float2(get_image_dim(src));
22543 float2 dst_cen = dst_dim / 2.0f;
22544 float2 src_cen = src_dim / 2.0f;
22546 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22548 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
22550 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
22551 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
22553 src_pos = src_pos * src_dim / dst_dim;
22555 float2 src_loc = src_pos + src_cen;
22557 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
22558 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
22559 write_imagef(dst, dst_loc, 0.5f);
22561 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
22566 Blend two inputs together, with the amount of each input used varying
22567 with the index counter.
22569 __kernel void blend_images(__write_only image2d_t dst,
22570 unsigned int index,
22571 __read_only image2d_t src1,
22572 __read_only image2d_t src2)
22574 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22575 CLK_FILTER_LINEAR);
22577 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
22579 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22580 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
22581 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
22583 float4 val1 = read_imagef(src1, sampler, src1_loc);
22584 float4 val2 = read_imagef(src2, sampler, src2_loc);
22586 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
22592 @section roberts_opencl
22593 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
22595 The filter accepts the following option:
22599 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22602 Set value which will be multiplied with filtered result.
22603 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22606 Set value which will be added to filtered result.
22607 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22610 @subsection Example
22614 Apply the Roberts cross operator with scale set to 2 and delta set to 10
22616 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
22620 @section sobel_opencl
22622 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
22624 The filter accepts the following option:
22628 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22631 Set value which will be multiplied with filtered result.
22632 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22635 Set value which will be added to filtered result.
22636 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22639 @subsection Example
22643 Apply sobel operator with scale set to 2 and delta set to 10
22645 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
22649 @section tonemap_opencl
22651 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
22653 It accepts the following parameters:
22657 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
22660 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
22663 Apply desaturation for highlights that exceed this level of brightness. The
22664 higher the parameter, the more color information will be preserved. This
22665 setting helps prevent unnaturally blown-out colors for super-highlights, by
22666 (smoothly) turning into white instead. This makes images feel more natural,
22667 at the cost of reducing information about out-of-range colors.
22669 The default value is 0.5, and the algorithm here is a little different from
22670 the cpu version tonemap currently. A setting of 0.0 disables this option.
22673 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
22674 is used to detect whether the scene has changed or not. If the distance between
22675 the current frame average brightness and the current running average exceeds
22676 a threshold value, we would re-calculate scene average and peak brightness.
22677 The default value is 0.2.
22680 Specify the output pixel format.
22682 Currently supported formats are:
22689 Set the output color range.
22691 Possible values are:
22697 Default is same as input.
22700 Set the output color primaries.
22702 Possible values are:
22708 Default is same as input.
22711 Set the output transfer characteristics.
22713 Possible values are:
22722 Set the output colorspace matrix.
22724 Possible value are:
22730 Default is same as input.
22734 @subsection Example
22738 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
22740 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
22744 @section unsharp_opencl
22746 Sharpen or blur the input video.
22748 It accepts the following parameters:
22751 @item luma_msize_x, lx
22752 Set the luma matrix horizontal size.
22753 Range is @code{[1, 23]} and default value is @code{5}.
22755 @item luma_msize_y, ly
22756 Set the luma matrix vertical size.
22757 Range is @code{[1, 23]} and default value is @code{5}.
22759 @item luma_amount, la
22760 Set the luma effect strength.
22761 Range is @code{[-10, 10]} and default value is @code{1.0}.
22763 Negative values will blur the input video, while positive values will
22764 sharpen it, a value of zero will disable the effect.
22766 @item chroma_msize_x, cx
22767 Set the chroma matrix horizontal size.
22768 Range is @code{[1, 23]} and default value is @code{5}.
22770 @item chroma_msize_y, cy
22771 Set the chroma matrix vertical size.
22772 Range is @code{[1, 23]} and default value is @code{5}.
22774 @item chroma_amount, ca
22775 Set the chroma effect strength.
22776 Range is @code{[-10, 10]} and default value is @code{0.0}.
22778 Negative values will blur the input video, while positive values will
22779 sharpen it, a value of zero will disable the effect.
22783 All parameters are optional and default to the equivalent of the
22784 string '5:5:1.0:5:5:0.0'.
22786 @subsection Examples
22790 Apply strong luma sharpen effect:
22792 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
22796 Apply a strong blur of both luma and chroma parameters:
22798 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
22802 @section xfade_opencl
22804 Cross fade two videos with custom transition effect by using OpenCL.
22806 It accepts the following options:
22810 Set one of possible transition effects.
22814 Select custom transition effect, the actual transition description
22815 will be picked from source and kernel options.
22827 Default transition is fade.
22831 OpenCL program source file for custom transition.
22834 Set name of kernel to use for custom transition from program source file.
22837 Set duration of video transition.
22840 Set time of start of transition relative to first video.
22843 The program source file must contain a kernel function with the given name,
22844 which will be run once for each plane of the output. Each run on a plane
22845 gets enqueued as a separate 2D global NDRange with one work-item for each
22846 pixel to be generated. The global ID offset for each work-item is therefore
22847 the coordinates of a pixel in the destination image.
22849 The kernel function needs to take the following arguments:
22852 Destination image, @var{__write_only image2d_t}.
22854 This image will become the output; the kernel should write all of it.
22857 First Source image, @var{__read_only image2d_t}.
22858 Second Source image, @var{__read_only image2d_t}.
22860 These are the most recent images on each input. The kernel may read from
22861 them to generate the output, but they can't be written to.
22864 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
22871 Apply dots curtain transition effect:
22873 __kernel void blend_images(__write_only image2d_t dst,
22874 __read_only image2d_t src1,
22875 __read_only image2d_t src2,
22878 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22879 CLK_FILTER_LINEAR);
22880 int2 p = (int2)(get_global_id(0), get_global_id(1));
22881 float2 rp = (float2)(get_global_id(0), get_global_id(1));
22882 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
22885 float2 dots = (float2)(20.0, 20.0);
22886 float2 center = (float2)(0,0);
22889 float4 val1 = read_imagef(src1, sampler, p);
22890 float4 val2 = read_imagef(src2, sampler, p);
22891 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
22893 write_imagef(dst, p, next ? val1 : val2);
22899 @c man end OPENCL VIDEO FILTERS
22901 @chapter VAAPI Video Filters
22902 @c man begin VAAPI VIDEO FILTERS
22904 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
22906 To enable compilation of these filters you need to configure FFmpeg with
22907 @code{--enable-vaapi}.
22909 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}
22911 @section tonemap_vaapi
22913 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
22914 It maps the dynamic range of HDR10 content to the SDR content.
22915 It currently only accepts HDR10 as input.
22917 It accepts the following parameters:
22921 Specify the output pixel format.
22923 Currently supported formats are:
22932 Set the output color primaries.
22934 Default is same as input.
22937 Set the output transfer characteristics.
22942 Set the output colorspace matrix.
22944 Default is same as input.
22948 @subsection Example
22952 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
22954 tonemap_vaapi=format=p010:t=bt2020-10
22958 @c man end VAAPI VIDEO FILTERS
22960 @chapter Video Sources
22961 @c man begin VIDEO SOURCES
22963 Below is a description of the currently available video sources.
22967 Buffer video frames, and make them available to the filter chain.
22969 This source is mainly intended for a programmatic use, in particular
22970 through the interface defined in @file{libavfilter/buffersrc.h}.
22972 It accepts the following parameters:
22977 Specify the size (width and height) of the buffered video frames. For the
22978 syntax of this option, check the
22979 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22982 The input video width.
22985 The input video height.
22988 A string representing the pixel format of the buffered video frames.
22989 It may be a number corresponding to a pixel format, or a pixel format
22993 Specify the timebase assumed by the timestamps of the buffered frames.
22996 Specify the frame rate expected for the video stream.
22998 @item pixel_aspect, sar
22999 The sample (pixel) aspect ratio of the input video.
23002 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
23003 to the filtergraph description to specify swscale flags for automatically
23004 inserted scalers. See @ref{Filtergraph syntax}.
23006 @item hw_frames_ctx
23007 When using a hardware pixel format, this should be a reference to an
23008 AVHWFramesContext describing input frames.
23013 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
23016 will instruct the source to accept video frames with size 320x240 and
23017 with format "yuv410p", assuming 1/24 as the timestamps timebase and
23018 square pixels (1:1 sample aspect ratio).
23019 Since the pixel format with name "yuv410p" corresponds to the number 6
23020 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
23021 this example corresponds to:
23023 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
23026 Alternatively, the options can be specified as a flat string, but this
23027 syntax is deprecated:
23029 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
23033 Create a pattern generated by an elementary cellular automaton.
23035 The initial state of the cellular automaton can be defined through the
23036 @option{filename} and @option{pattern} options. If such options are
23037 not specified an initial state is created randomly.
23039 At each new frame a new row in the video is filled with the result of
23040 the cellular automaton next generation. The behavior when the whole
23041 frame is filled is defined by the @option{scroll} option.
23043 This source accepts the following options:
23047 Read the initial cellular automaton state, i.e. the starting row, from
23048 the specified file.
23049 In the file, each non-whitespace character is considered an alive
23050 cell, a newline will terminate the row, and further characters in the
23051 file will be ignored.
23054 Read the initial cellular automaton state, i.e. the starting row, from
23055 the specified string.
23057 Each non-whitespace character in the string is considered an alive
23058 cell, a newline will terminate the row, and further characters in the
23059 string will be ignored.
23062 Set the video rate, that is the number of frames generated per second.
23065 @item random_fill_ratio, ratio
23066 Set the random fill ratio for the initial cellular automaton row. It
23067 is a floating point number value ranging from 0 to 1, defaults to
23070 This option is ignored when a file or a pattern is specified.
23072 @item random_seed, seed
23073 Set the seed for filling randomly the initial row, must be an integer
23074 included between 0 and UINT32_MAX. If not specified, or if explicitly
23075 set to -1, the filter will try to use a good random seed on a best
23079 Set the cellular automaton rule, it is a number ranging from 0 to 255.
23080 Default value is 110.
23083 Set the size of the output video. For the syntax of this option, check the
23084 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23086 If @option{filename} or @option{pattern} is specified, the size is set
23087 by default to the width of the specified initial state row, and the
23088 height is set to @var{width} * PHI.
23090 If @option{size} is set, it must contain the width of the specified
23091 pattern string, and the specified pattern will be centered in the
23094 If a filename or a pattern string is not specified, the size value
23095 defaults to "320x518" (used for a randomly generated initial state).
23098 If set to 1, scroll the output upward when all the rows in the output
23099 have been already filled. If set to 0, the new generated row will be
23100 written over the top row just after the bottom row is filled.
23103 @item start_full, full
23104 If set to 1, completely fill the output with generated rows before
23105 outputting the first frame.
23106 This is the default behavior, for disabling set the value to 0.
23109 If set to 1, stitch the left and right row edges together.
23110 This is the default behavior, for disabling set the value to 0.
23113 @subsection Examples
23117 Read the initial state from @file{pattern}, and specify an output of
23120 cellauto=f=pattern:s=200x400
23124 Generate a random initial row with a width of 200 cells, with a fill
23127 cellauto=ratio=2/3:s=200x200
23131 Create a pattern generated by rule 18 starting by a single alive cell
23132 centered on an initial row with width 100:
23134 cellauto=p=@@:s=100x400:full=0:rule=18
23138 Specify a more elaborated initial pattern:
23140 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
23145 @anchor{coreimagesrc}
23146 @section coreimagesrc
23147 Video source generated on GPU using Apple's CoreImage API on OSX.
23149 This video source is a specialized version of the @ref{coreimage} video filter.
23150 Use a core image generator at the beginning of the applied filterchain to
23151 generate the content.
23153 The coreimagesrc video source accepts the following options:
23155 @item list_generators
23156 List all available generators along with all their respective options as well as
23157 possible minimum and maximum values along with the default values.
23159 list_generators=true
23163 Specify the size of the sourced video. For the syntax of this option, check the
23164 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23165 The default value is @code{320x240}.
23168 Specify the frame rate of the sourced video, as the number of frames
23169 generated per second. It has to be a string in the format
23170 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23171 number or a valid video frame rate abbreviation. The default value is
23175 Set the sample aspect ratio of the sourced video.
23178 Set the duration of the sourced video. See
23179 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23180 for the accepted syntax.
23182 If not specified, or the expressed duration is negative, the video is
23183 supposed to be generated forever.
23186 Additionally, all options of the @ref{coreimage} video filter are accepted.
23187 A complete filterchain can be used for further processing of the
23188 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
23189 and examples for details.
23191 @subsection Examples
23196 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
23197 given as complete and escaped command-line for Apple's standard bash shell:
23199 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
23201 This example is equivalent to the QRCode example of @ref{coreimage} without the
23202 need for a nullsrc video source.
23207 Generate several gradients.
23211 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23212 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23215 Set frame rate, expressed as number of frames per second. Default
23218 @item c0, c1, c2, c3, c4, c5, c6, c7
23219 Set 8 colors. Default values for colors is to pick random one.
23221 @item x0, y0, y0, y1
23222 Set gradient line source and destination points. If negative or out of range, random ones
23226 Set number of colors to use at once. Allowed range is from 2 to 8. Default value is 2.
23229 Set seed for picking gradient line points.
23232 Set the duration of the sourced video. See
23233 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23234 for the accepted syntax.
23236 If not specified, or the expressed duration is negative, the video is
23237 supposed to be generated forever.
23240 Set speed of gradients rotation.
23244 @section mandelbrot
23246 Generate a Mandelbrot set fractal, and progressively zoom towards the
23247 point specified with @var{start_x} and @var{start_y}.
23249 This source accepts the following options:
23254 Set the terminal pts value. Default value is 400.
23257 Set the terminal scale value.
23258 Must be a floating point value. Default value is 0.3.
23261 Set the inner coloring mode, that is the algorithm used to draw the
23262 Mandelbrot fractal internal region.
23264 It shall assume one of the following values:
23269 Show time until convergence.
23271 Set color based on point closest to the origin of the iterations.
23276 Default value is @var{mincol}.
23279 Set the bailout value. Default value is 10.0.
23282 Set the maximum of iterations performed by the rendering
23283 algorithm. Default value is 7189.
23286 Set outer coloring mode.
23287 It shall assume one of following values:
23289 @item iteration_count
23290 Set iteration count mode.
23291 @item normalized_iteration_count
23292 set normalized iteration count mode.
23294 Default value is @var{normalized_iteration_count}.
23297 Set frame rate, expressed as number of frames per second. Default
23301 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23302 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23305 Set the initial scale value. Default value is 3.0.
23308 Set the initial x position. Must be a floating point value between
23309 -100 and 100. Default value is -0.743643887037158704752191506114774.
23312 Set the initial y position. Must be a floating point value between
23313 -100 and 100. Default value is -0.131825904205311970493132056385139.
23318 Generate various test patterns, as generated by the MPlayer test filter.
23320 The size of the generated video is fixed, and is 256x256.
23321 This source is useful in particular for testing encoding features.
23323 This source accepts the following options:
23328 Specify the frame rate of the sourced video, as the number of frames
23329 generated per second. It has to be a string in the format
23330 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23331 number or a valid video frame rate abbreviation. The default value is
23335 Set the duration of the sourced video. See
23336 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23337 for the accepted syntax.
23339 If not specified, or the expressed duration is negative, the video is
23340 supposed to be generated forever.
23344 Set the number or the name of the test to perform. Supported tests are:
23358 @item max_frames, m
23359 Set the maximum number of frames generated for each test, default value is 30.
23363 Default value is "all", which will cycle through the list of all tests.
23368 mptestsrc=t=dc_luma
23371 will generate a "dc_luma" test pattern.
23373 @section frei0r_src
23375 Provide a frei0r source.
23377 To enable compilation of this filter you need to install the frei0r
23378 header and configure FFmpeg with @code{--enable-frei0r}.
23380 This source accepts the following parameters:
23385 The size of the video to generate. For the syntax of this option, check the
23386 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23389 The framerate of the generated video. It may be a string of the form
23390 @var{num}/@var{den} or a frame rate abbreviation.
23393 The name to the frei0r source to load. For more information regarding frei0r and
23394 how to set the parameters, read the @ref{frei0r} section in the video filters
23397 @item filter_params
23398 A '|'-separated list of parameters to pass to the frei0r source.
23402 For example, to generate a frei0r partik0l source with size 200x200
23403 and frame rate 10 which is overlaid on the overlay filter main input:
23405 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
23410 Generate a life pattern.
23412 This source is based on a generalization of John Conway's life game.
23414 The sourced input represents a life grid, each pixel represents a cell
23415 which can be in one of two possible states, alive or dead. Every cell
23416 interacts with its eight neighbours, which are the cells that are
23417 horizontally, vertically, or diagonally adjacent.
23419 At each interaction the grid evolves according to the adopted rule,
23420 which specifies the number of neighbor alive cells which will make a
23421 cell stay alive or born. The @option{rule} option allows one to specify
23424 This source accepts the following options:
23428 Set the file from which to read the initial grid state. In the file,
23429 each non-whitespace character is considered an alive cell, and newline
23430 is used to delimit the end of each row.
23432 If this option is not specified, the initial grid is generated
23436 Set the video rate, that is the number of frames generated per second.
23439 @item random_fill_ratio, ratio
23440 Set the random fill ratio for the initial random grid. It is a
23441 floating point number value ranging from 0 to 1, defaults to 1/PHI.
23442 It is ignored when a file is specified.
23444 @item random_seed, seed
23445 Set the seed for filling the initial random grid, must be an integer
23446 included between 0 and UINT32_MAX. If not specified, or if explicitly
23447 set to -1, the filter will try to use a good random seed on a best
23453 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
23454 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
23455 @var{NS} specifies the number of alive neighbor cells which make a
23456 live cell stay alive, and @var{NB} the number of alive neighbor cells
23457 which make a dead cell to become alive (i.e. to "born").
23458 "s" and "b" can be used in place of "S" and "B", respectively.
23460 Alternatively a rule can be specified by an 18-bits integer. The 9
23461 high order bits are used to encode the next cell state if it is alive
23462 for each number of neighbor alive cells, the low order bits specify
23463 the rule for "borning" new cells. Higher order bits encode for an
23464 higher number of neighbor cells.
23465 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
23466 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
23468 Default value is "S23/B3", which is the original Conway's game of life
23469 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
23470 cells, and will born a new cell if there are three alive cells around
23474 Set the size of the output video. For the syntax of this option, check the
23475 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23477 If @option{filename} is specified, the size is set by default to the
23478 same size of the input file. If @option{size} is set, it must contain
23479 the size specified in the input file, and the initial grid defined in
23480 that file is centered in the larger resulting area.
23482 If a filename is not specified, the size value defaults to "320x240"
23483 (used for a randomly generated initial grid).
23486 If set to 1, stitch the left and right grid edges together, and the
23487 top and bottom edges also. Defaults to 1.
23490 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
23491 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
23492 value from 0 to 255.
23495 Set the color of living (or new born) cells.
23498 Set the color of dead cells. If @option{mold} is set, this is the first color
23499 used to represent a dead cell.
23502 Set mold color, for definitely dead and moldy cells.
23504 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
23505 ffmpeg-utils manual,ffmpeg-utils}.
23508 @subsection Examples
23512 Read a grid from @file{pattern}, and center it on a grid of size
23515 life=f=pattern:s=300x300
23519 Generate a random grid of size 200x200, with a fill ratio of 2/3:
23521 life=ratio=2/3:s=200x200
23525 Specify a custom rule for evolving a randomly generated grid:
23531 Full example with slow death effect (mold) using @command{ffplay}:
23533 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
23540 @anchor{haldclutsrc}
23543 @anchor{pal100bars}
23544 @anchor{rgbtestsrc}
23546 @anchor{smptehdbars}
23549 @anchor{yuvtestsrc}
23550 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
23552 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
23554 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
23556 The @code{color} source provides an uniformly colored input.
23558 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
23559 @ref{haldclut} filter.
23561 The @code{nullsrc} source returns unprocessed video frames. It is
23562 mainly useful to be employed in analysis / debugging tools, or as the
23563 source for filters which ignore the input data.
23565 The @code{pal75bars} source generates a color bars pattern, based on
23566 EBU PAL recommendations with 75% color levels.
23568 The @code{pal100bars} source generates a color bars pattern, based on
23569 EBU PAL recommendations with 100% color levels.
23571 The @code{rgbtestsrc} source generates an RGB test pattern useful for
23572 detecting RGB vs BGR issues. You should see a red, green and blue
23573 stripe from top to bottom.
23575 The @code{smptebars} source generates a color bars pattern, based on
23576 the SMPTE Engineering Guideline EG 1-1990.
23578 The @code{smptehdbars} source generates a color bars pattern, based on
23579 the SMPTE RP 219-2002.
23581 The @code{testsrc} source generates a test video pattern, showing a
23582 color pattern, a scrolling gradient and a timestamp. This is mainly
23583 intended for testing purposes.
23585 The @code{testsrc2} source is similar to testsrc, but supports more
23586 pixel formats instead of just @code{rgb24}. This allows using it as an
23587 input for other tests without requiring a format conversion.
23589 The @code{yuvtestsrc} source generates an YUV test pattern. You should
23590 see a y, cb and cr stripe from top to bottom.
23592 The sources accept the following parameters:
23597 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
23598 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
23599 pixels to be used as identity matrix for 3D lookup tables. Each component is
23600 coded on a @code{1/(N*N)} scale.
23603 Specify the color of the source, only available in the @code{color}
23604 source. For the syntax of this option, check the
23605 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
23608 Specify the size of the sourced video. For the syntax of this option, check the
23609 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23610 The default value is @code{320x240}.
23612 This option is not available with the @code{allrgb}, @code{allyuv}, and
23613 @code{haldclutsrc} filters.
23616 Specify the frame rate of the sourced video, as the number of frames
23617 generated per second. It has to be a string in the format
23618 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23619 number or a valid video frame rate abbreviation. The default value is
23623 Set the duration of the sourced video. See
23624 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23625 for the accepted syntax.
23627 If not specified, or the expressed duration is negative, the video is
23628 supposed to be generated forever.
23630 Since the frame rate is used as time base, all frames including the last one
23631 will have their full duration. If the specified duration is not a multiple
23632 of the frame duration, it will be rounded up.
23635 Set the sample aspect ratio of the sourced video.
23638 Specify the alpha (opacity) of the background, only available in the
23639 @code{testsrc2} source. The value must be between 0 (fully transparent) and
23640 255 (fully opaque, the default).
23643 Set the number of decimals to show in the timestamp, only available in the
23644 @code{testsrc} source.
23646 The displayed timestamp value will correspond to the original
23647 timestamp value multiplied by the power of 10 of the specified
23648 value. Default value is 0.
23651 @subsection Examples
23655 Generate a video with a duration of 5.3 seconds, with size
23656 176x144 and a frame rate of 10 frames per second:
23658 testsrc=duration=5.3:size=qcif:rate=10
23662 The following graph description will generate a red source
23663 with an opacity of 0.2, with size "qcif" and a frame rate of 10
23666 color=c=red@@0.2:s=qcif:r=10
23670 If the input content is to be ignored, @code{nullsrc} can be used. The
23671 following command generates noise in the luminance plane by employing
23672 the @code{geq} filter:
23674 nullsrc=s=256x256, geq=random(1)*255:128:128
23678 @subsection Commands
23680 The @code{color} source supports the following commands:
23684 Set the color of the created image. Accepts the same syntax of the
23685 corresponding @option{color} option.
23690 Generate video using an OpenCL program.
23695 OpenCL program source file.
23698 Kernel name in program.
23701 Size of frames to generate. This must be set.
23704 Pixel format to use for the generated frames. This must be set.
23707 Number of frames generated every second. Default value is '25'.
23711 For details of how the program loading works, see the @ref{program_opencl}
23718 Generate a colour ramp by setting pixel values from the position of the pixel
23719 in the output image. (Note that this will work with all pixel formats, but
23720 the generated output will not be the same.)
23722 __kernel void ramp(__write_only image2d_t dst,
23723 unsigned int index)
23725 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23728 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
23730 write_imagef(dst, loc, val);
23735 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
23737 __kernel void sierpinski_carpet(__write_only image2d_t dst,
23738 unsigned int index)
23740 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23742 float4 value = 0.0f;
23743 int x = loc.x + index;
23744 int y = loc.y + index;
23745 while (x > 0 || y > 0) {
23746 if (x % 3 == 1 && y % 3 == 1) {
23754 write_imagef(dst, loc, value);
23760 @section sierpinski
23762 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
23764 This source accepts the following options:
23768 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23769 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23772 Set frame rate, expressed as number of frames per second. Default
23776 Set seed which is used for random panning.
23779 Set max jump for single pan destination. Allowed range is from 1 to 10000.
23782 Set fractal type, can be default @code{carpet} or @code{triangle}.
23785 @c man end VIDEO SOURCES
23787 @chapter Video Sinks
23788 @c man begin VIDEO SINKS
23790 Below is a description of the currently available video sinks.
23792 @section buffersink
23794 Buffer video frames, and make them available to the end of the filter
23797 This sink is mainly intended for programmatic use, in particular
23798 through the interface defined in @file{libavfilter/buffersink.h}
23799 or the options system.
23801 It accepts a pointer to an AVBufferSinkContext structure, which
23802 defines the incoming buffers' formats, to be passed as the opaque
23803 parameter to @code{avfilter_init_filter} for initialization.
23807 Null video sink: do absolutely nothing with the input video. It is
23808 mainly useful as a template and for use in analysis / debugging
23811 @c man end VIDEO SINKS
23813 @chapter Multimedia Filters
23814 @c man begin MULTIMEDIA FILTERS
23816 Below is a description of the currently available multimedia filters.
23820 Convert input audio to a video output, displaying the audio bit scope.
23822 The filter accepts the following options:
23826 Set frame rate, expressed as number of frames per second. Default
23830 Specify the video size for the output. For the syntax of this option, check the
23831 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23832 Default value is @code{1024x256}.
23835 Specify list of colors separated by space or by '|' which will be used to
23836 draw channels. Unrecognized or missing colors will be replaced
23840 @section adrawgraph
23841 Draw a graph using input audio metadata.
23843 See @ref{drawgraph}
23845 @section agraphmonitor
23847 See @ref{graphmonitor}.
23849 @section ahistogram
23851 Convert input audio to a video output, displaying the volume histogram.
23853 The filter accepts the following options:
23857 Specify how histogram is calculated.
23859 It accepts the following values:
23862 Use single histogram for all channels.
23864 Use separate histogram for each channel.
23866 Default is @code{single}.
23869 Set frame rate, expressed as number of frames per second. Default
23873 Specify the video size for the output. For the syntax of this option, check the
23874 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23875 Default value is @code{hd720}.
23880 It accepts the following values:
23891 reverse logarithmic
23893 Default is @code{log}.
23896 Set amplitude scale.
23898 It accepts the following values:
23905 Default is @code{log}.
23908 Set how much frames to accumulate in histogram.
23909 Default is 1. Setting this to -1 accumulates all frames.
23912 Set histogram ratio of window height.
23915 Set sonogram sliding.
23917 It accepts the following values:
23920 replace old rows with new ones.
23922 scroll from top to bottom.
23924 Default is @code{replace}.
23927 @section aphasemeter
23929 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
23930 representing mean phase of current audio frame. A video output can also be produced and is
23931 enabled by default. The audio is passed through as first output.
23933 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
23934 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
23935 and @code{1} means channels are in phase.
23937 The filter accepts the following options, all related to its video output:
23941 Set the output frame rate. Default value is @code{25}.
23944 Set the video size for the output. For the syntax of this option, check the
23945 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23946 Default value is @code{800x400}.
23951 Specify the red, green, blue contrast. Default values are @code{2},
23952 @code{7} and @code{1}.
23953 Allowed range is @code{[0, 255]}.
23956 Set color which will be used for drawing median phase. If color is
23957 @code{none} which is default, no median phase value will be drawn.
23960 Enable video output. Default is enabled.
23963 @subsection phasing detection
23965 The filter also detects out of phase and mono sequences in stereo streams.
23966 It logs the sequence start, end and duration when it lasts longer or as long as the minimum set.
23968 The filter accepts the following options for this detection:
23972 Enable mono and out of phase detection. Default is disabled.
23975 Set phase tolerance for mono detection, in amplitude ratio. Default is @code{0}.
23976 Allowed range is @code{[0, 1]}.
23979 Set angle threshold for out of phase detection, in degree. Default is @code{170}.
23980 Allowed range is @code{[90, 180]}.
23983 Set mono or out of phase duration until notification, expressed in seconds. Default is @code{2}.
23986 @subsection Examples
23990 Complete example with @command{ffmpeg} to detect 1 second of mono with 0.001 phase tolerance:
23992 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
23996 @section avectorscope
23998 Convert input audio to a video output, representing the audio vector
24001 The filter is used to measure the difference between channels of stereo
24002 audio stream. A monaural signal, consisting of identical left and right
24003 signal, results in straight vertical line. Any stereo separation is visible
24004 as a deviation from this line, creating a Lissajous figure.
24005 If the straight (or deviation from it) but horizontal line appears this
24006 indicates that the left and right channels are out of phase.
24008 The filter accepts the following options:
24012 Set the vectorscope mode.
24014 Available values are:
24017 Lissajous rotated by 45 degrees.
24020 Same as above but not rotated.
24023 Shape resembling half of circle.
24026 Default value is @samp{lissajous}.
24029 Set the video size for the output. For the syntax of this option, check the
24030 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24031 Default value is @code{400x400}.
24034 Set the output frame rate. Default value is @code{25}.
24040 Specify the red, green, blue and alpha contrast. Default values are @code{40},
24041 @code{160}, @code{80} and @code{255}.
24042 Allowed range is @code{[0, 255]}.
24048 Specify the red, green, blue and alpha fade. Default values are @code{15},
24049 @code{10}, @code{5} and @code{5}.
24050 Allowed range is @code{[0, 255]}.
24053 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
24054 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
24057 Set the vectorscope drawing mode.
24059 Available values are:
24062 Draw dot for each sample.
24065 Draw line between previous and current sample.
24068 Default value is @samp{dot}.
24071 Specify amplitude scale of audio samples.
24073 Available values are:
24089 Swap left channel axis with right channel axis.
24099 Mirror only x axis.
24102 Mirror only y axis.
24110 @subsection Examples
24114 Complete example using @command{ffplay}:
24116 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
24117 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
24121 @section bench, abench
24123 Benchmark part of a filtergraph.
24125 The filter accepts the following options:
24129 Start or stop a timer.
24131 Available values are:
24134 Get the current time, set it as frame metadata (using the key
24135 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
24138 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
24139 the input frame metadata to get the time difference. Time difference, average,
24140 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
24141 @code{min}) are then printed. The timestamps are expressed in seconds.
24145 @subsection Examples
24149 Benchmark @ref{selectivecolor} filter:
24151 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
24157 Concatenate audio and video streams, joining them together one after the
24160 The filter works on segments of synchronized video and audio streams. All
24161 segments must have the same number of streams of each type, and that will
24162 also be the number of streams at output.
24164 The filter accepts the following options:
24169 Set the number of segments. Default is 2.
24172 Set the number of output video streams, that is also the number of video
24173 streams in each segment. Default is 1.
24176 Set the number of output audio streams, that is also the number of audio
24177 streams in each segment. Default is 0.
24180 Activate unsafe mode: do not fail if segments have a different format.
24184 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
24185 @var{a} audio outputs.
24187 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
24188 segment, in the same order as the outputs, then the inputs for the second
24191 Related streams do not always have exactly the same duration, for various
24192 reasons including codec frame size or sloppy authoring. For that reason,
24193 related synchronized streams (e.g. a video and its audio track) should be
24194 concatenated at once. The concat filter will use the duration of the longest
24195 stream in each segment (except the last one), and if necessary pad shorter
24196 audio streams with silence.
24198 For this filter to work correctly, all segments must start at timestamp 0.
24200 All corresponding streams must have the same parameters in all segments; the
24201 filtering system will automatically select a common pixel format for video
24202 streams, and a common sample format, sample rate and channel layout for
24203 audio streams, but other settings, such as resolution, must be converted
24204 explicitly by the user.
24206 Different frame rates are acceptable but will result in variable frame rate
24207 at output; be sure to configure the output file to handle it.
24209 @subsection Examples
24213 Concatenate an opening, an episode and an ending, all in bilingual version
24214 (video in stream 0, audio in streams 1 and 2):
24216 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
24217 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
24218 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
24219 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
24223 Concatenate two parts, handling audio and video separately, using the
24224 (a)movie sources, and adjusting the resolution:
24226 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
24227 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
24228 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
24230 Note that a desync will happen at the stitch if the audio and video streams
24231 do not have exactly the same duration in the first file.
24235 @subsection Commands
24237 This filter supports the following commands:
24240 Close the current segment and step to the next one
24246 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
24247 level. By default, it logs a message at a frequency of 10Hz with the
24248 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
24249 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
24251 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
24252 sample format is double-precision floating point. The input stream will be converted to
24253 this specification, if needed. Users may need to insert aformat and/or aresample filters
24254 after this filter to obtain the original parameters.
24256 The filter also has a video output (see the @var{video} option) with a real
24257 time graph to observe the loudness evolution. The graphic contains the logged
24258 message mentioned above, so it is not printed anymore when this option is set,
24259 unless the verbose logging is set. The main graphing area contains the
24260 short-term loudness (3 seconds of analysis), and the gauge on the right is for
24261 the momentary loudness (400 milliseconds), but can optionally be configured
24262 to instead display short-term loudness (see @var{gauge}).
24264 The green area marks a +/- 1LU target range around the target loudness
24265 (-23LUFS by default, unless modified through @var{target}).
24267 More information about the Loudness Recommendation EBU R128 on
24268 @url{http://tech.ebu.ch/loudness}.
24270 The filter accepts the following options:
24275 Activate the video output. The audio stream is passed unchanged whether this
24276 option is set or no. The video stream will be the first output stream if
24277 activated. Default is @code{0}.
24280 Set the video size. This option is for video only. For the syntax of this
24282 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24283 Default and minimum resolution is @code{640x480}.
24286 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
24287 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
24288 other integer value between this range is allowed.
24291 Set metadata injection. If set to @code{1}, the audio input will be segmented
24292 into 100ms output frames, each of them containing various loudness information
24293 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
24295 Default is @code{0}.
24298 Force the frame logging level.
24300 Available values are:
24303 information logging level
24305 verbose logging level
24308 By default, the logging level is set to @var{info}. If the @option{video} or
24309 the @option{metadata} options are set, it switches to @var{verbose}.
24314 Available modes can be cumulated (the option is a @code{flag} type). Possible
24318 Disable any peak mode (default).
24320 Enable sample-peak mode.
24322 Simple peak mode looking for the higher sample value. It logs a message
24323 for sample-peak (identified by @code{SPK}).
24325 Enable true-peak mode.
24327 If enabled, the peak lookup is done on an over-sampled version of the input
24328 stream for better peak accuracy. It logs a message for true-peak.
24329 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
24330 This mode requires a build with @code{libswresample}.
24334 Treat mono input files as "dual mono". If a mono file is intended for playback
24335 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
24336 If set to @code{true}, this option will compensate for this effect.
24337 Multi-channel input files are not affected by this option.
24340 Set a specific pan law to be used for the measurement of dual mono files.
24341 This parameter is optional, and has a default value of -3.01dB.
24344 Set a specific target level (in LUFS) used as relative zero in the visualization.
24345 This parameter is optional and has a default value of -23LUFS as specified
24346 by EBU R128. However, material published online may prefer a level of -16LUFS
24347 (e.g. for use with podcasts or video platforms).
24350 Set the value displayed by the gauge. Valid values are @code{momentary} and s
24351 @code{shortterm}. By default the momentary value will be used, but in certain
24352 scenarios it may be more useful to observe the short term value instead (e.g.
24356 Sets the display scale for the loudness. Valid parameters are @code{absolute}
24357 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
24358 video output, not the summary or continuous log output.
24361 @subsection Examples
24365 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
24367 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
24371 Run an analysis with @command{ffmpeg}:
24373 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
24377 @section interleave, ainterleave
24379 Temporally interleave frames from several inputs.
24381 @code{interleave} works with video inputs, @code{ainterleave} with audio.
24383 These filters read frames from several inputs and send the oldest
24384 queued frame to the output.
24386 Input streams must have well defined, monotonically increasing frame
24389 In order to submit one frame to output, these filters need to enqueue
24390 at least one frame for each input, so they cannot work in case one
24391 input is not yet terminated and will not receive incoming frames.
24393 For example consider the case when one input is a @code{select} filter
24394 which always drops input frames. The @code{interleave} filter will keep
24395 reading from that input, but it will never be able to send new frames
24396 to output until the input sends an end-of-stream signal.
24398 Also, depending on inputs synchronization, the filters will drop
24399 frames in case one input receives more frames than the other ones, and
24400 the queue is already filled.
24402 These filters accept the following options:
24406 Set the number of different inputs, it is 2 by default.
24409 How to determine the end-of-stream.
24413 The duration of the longest input. (default)
24416 The duration of the shortest input.
24419 The duration of the first input.
24424 @subsection Examples
24428 Interleave frames belonging to different streams using @command{ffmpeg}:
24430 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
24434 Add flickering blur effect:
24436 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
24440 @section metadata, ametadata
24442 Manipulate frame metadata.
24444 This filter accepts the following options:
24448 Set mode of operation of the filter.
24450 Can be one of the following:
24454 If both @code{value} and @code{key} is set, select frames
24455 which have such metadata. If only @code{key} is set, select
24456 every frame that has such key in metadata.
24459 Add new metadata @code{key} and @code{value}. If key is already available
24463 Modify value of already present key.
24466 If @code{value} is set, delete only keys that have such value.
24467 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
24471 Print key and its value if metadata was found. If @code{key} is not set print all
24472 metadata values available in frame.
24476 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
24479 Set metadata value which will be used. This option is mandatory for
24480 @code{modify} and @code{add} mode.
24483 Which function to use when comparing metadata value and @code{value}.
24485 Can be one of following:
24489 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
24492 Values are interpreted as strings, returns true if metadata value starts with
24493 the @code{value} option string.
24496 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
24499 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
24502 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
24505 Values are interpreted as floats, returns true if expression from option @code{expr}
24509 Values are interpreted as strings, returns true if metadata value ends with
24510 the @code{value} option string.
24514 Set expression which is used when @code{function} is set to @code{expr}.
24515 The expression is evaluated through the eval API and can contain the following
24520 Float representation of @code{value} from metadata key.
24523 Float representation of @code{value} as supplied by user in @code{value} option.
24527 If specified in @code{print} mode, output is written to the named file. Instead of
24528 plain filename any writable url can be specified. Filename ``-'' is a shorthand
24529 for standard output. If @code{file} option is not set, output is written to the log
24530 with AV_LOG_INFO loglevel.
24533 Reduces buffering in print mode when output is written to a URL set using @var{file}.
24537 @subsection Examples
24541 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
24544 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
24547 Print silencedetect output to file @file{metadata.txt}.
24549 silencedetect,ametadata=mode=print:file=metadata.txt
24552 Direct all metadata to a pipe with file descriptor 4.
24554 metadata=mode=print:file='pipe\:4'
24558 @section perms, aperms
24560 Set read/write permissions for the output frames.
24562 These filters are mainly aimed at developers to test direct path in the
24563 following filter in the filtergraph.
24565 The filters accept the following options:
24569 Select the permissions mode.
24571 It accepts the following values:
24574 Do nothing. This is the default.
24576 Set all the output frames read-only.
24578 Set all the output frames directly writable.
24580 Make the frame read-only if writable, and writable if read-only.
24582 Set each output frame read-only or writable randomly.
24586 Set the seed for the @var{random} mode, must be an integer included between
24587 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
24588 @code{-1}, the filter will try to use a good random seed on a best effort
24592 Note: in case of auto-inserted filter between the permission filter and the
24593 following one, the permission might not be received as expected in that
24594 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
24595 perms/aperms filter can avoid this problem.
24597 @section realtime, arealtime
24599 Slow down filtering to match real time approximately.
24601 These filters will pause the filtering for a variable amount of time to
24602 match the output rate with the input timestamps.
24603 They are similar to the @option{re} option to @code{ffmpeg}.
24605 They accept the following options:
24609 Time limit for the pauses. Any pause longer than that will be considered
24610 a timestamp discontinuity and reset the timer. Default is 2 seconds.
24612 Speed factor for processing. The value must be a float larger than zero.
24613 Values larger than 1.0 will result in faster than realtime processing,
24614 smaller will slow processing down. The @var{limit} is automatically adapted
24615 accordingly. Default is 1.0.
24617 A processing speed faster than what is possible without these filters cannot
24622 @section select, aselect
24624 Select frames to pass in output.
24626 This filter accepts the following options:
24631 Set expression, which is evaluated for each input frame.
24633 If the expression is evaluated to zero, the frame is discarded.
24635 If the evaluation result is negative or NaN, the frame is sent to the
24636 first output; otherwise it is sent to the output with index
24637 @code{ceil(val)-1}, assuming that the input index starts from 0.
24639 For example a value of @code{1.2} corresponds to the output with index
24640 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
24643 Set the number of outputs. The output to which to send the selected
24644 frame is based on the result of the evaluation. Default value is 1.
24647 The expression can contain the following constants:
24651 The (sequential) number of the filtered frame, starting from 0.
24654 The (sequential) number of the selected frame, starting from 0.
24656 @item prev_selected_n
24657 The sequential number of the last selected frame. It's NAN if undefined.
24660 The timebase of the input timestamps.
24663 The PTS (Presentation TimeStamp) of the filtered video frame,
24664 expressed in @var{TB} units. It's NAN if undefined.
24667 The PTS of the filtered video frame,
24668 expressed in seconds. It's NAN if undefined.
24671 The PTS of the previously filtered video frame. It's NAN if undefined.
24673 @item prev_selected_pts
24674 The PTS of the last previously filtered video frame. It's NAN if undefined.
24676 @item prev_selected_t
24677 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
24680 The PTS of the first video frame in the video. It's NAN if undefined.
24683 The time of the first video frame in the video. It's NAN if undefined.
24685 @item pict_type @emph{(video only)}
24686 The type of the filtered frame. It can assume one of the following
24698 @item interlace_type @emph{(video only)}
24699 The frame interlace type. It can assume one of the following values:
24702 The frame is progressive (not interlaced).
24704 The frame is top-field-first.
24706 The frame is bottom-field-first.
24709 @item consumed_sample_n @emph{(audio only)}
24710 the number of selected samples before the current frame
24712 @item samples_n @emph{(audio only)}
24713 the number of samples in the current frame
24715 @item sample_rate @emph{(audio only)}
24716 the input sample rate
24719 This is 1 if the filtered frame is a key-frame, 0 otherwise.
24722 the position in the file of the filtered frame, -1 if the information
24723 is not available (e.g. for synthetic video)
24725 @item scene @emph{(video only)}
24726 value between 0 and 1 to indicate a new scene; a low value reflects a low
24727 probability for the current frame to introduce a new scene, while a higher
24728 value means the current frame is more likely to be one (see the example below)
24730 @item concatdec_select
24731 The concat demuxer can select only part of a concat input file by setting an
24732 inpoint and an outpoint, but the output packets may not be entirely contained
24733 in the selected interval. By using this variable, it is possible to skip frames
24734 generated by the concat demuxer which are not exactly contained in the selected
24737 This works by comparing the frame pts against the @var{lavf.concat.start_time}
24738 and the @var{lavf.concat.duration} packet metadata values which are also
24739 present in the decoded frames.
24741 The @var{concatdec_select} variable is -1 if the frame pts is at least
24742 start_time and either the duration metadata is missing or the frame pts is less
24743 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
24746 That basically means that an input frame is selected if its pts is within the
24747 interval set by the concat demuxer.
24751 The default value of the select expression is "1".
24753 @subsection Examples
24757 Select all frames in input:
24762 The example above is the same as:
24774 Select only I-frames:
24776 select='eq(pict_type\,I)'
24780 Select one frame every 100:
24782 select='not(mod(n\,100))'
24786 Select only frames contained in the 10-20 time interval:
24788 select=between(t\,10\,20)
24792 Select only I-frames contained in the 10-20 time interval:
24794 select=between(t\,10\,20)*eq(pict_type\,I)
24798 Select frames with a minimum distance of 10 seconds:
24800 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
24804 Use aselect to select only audio frames with samples number > 100:
24806 aselect='gt(samples_n\,100)'
24810 Create a mosaic of the first scenes:
24812 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
24815 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
24819 Send even and odd frames to separate outputs, and compose them:
24821 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
24825 Select useful frames from an ffconcat file which is using inpoints and
24826 outpoints but where the source files are not intra frame only.
24828 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
24832 @section sendcmd, asendcmd
24834 Send commands to filters in the filtergraph.
24836 These filters read commands to be sent to other filters in the
24839 @code{sendcmd} must be inserted between two video filters,
24840 @code{asendcmd} must be inserted between two audio filters, but apart
24841 from that they act the same way.
24843 The specification of commands can be provided in the filter arguments
24844 with the @var{commands} option, or in a file specified by the
24845 @var{filename} option.
24847 These filters accept the following options:
24850 Set the commands to be read and sent to the other filters.
24852 Set the filename of the commands to be read and sent to the other
24856 @subsection Commands syntax
24858 A commands description consists of a sequence of interval
24859 specifications, comprising a list of commands to be executed when a
24860 particular event related to that interval occurs. The occurring event
24861 is typically the current frame time entering or leaving a given time
24864 An interval is specified by the following syntax:
24866 @var{START}[-@var{END}] @var{COMMANDS};
24869 The time interval is specified by the @var{START} and @var{END} times.
24870 @var{END} is optional and defaults to the maximum time.
24872 The current frame time is considered within the specified interval if
24873 it is included in the interval [@var{START}, @var{END}), that is when
24874 the time is greater or equal to @var{START} and is lesser than
24877 @var{COMMANDS} consists of a sequence of one or more command
24878 specifications, separated by ",", relating to that interval. The
24879 syntax of a command specification is given by:
24881 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
24884 @var{FLAGS} is optional and specifies the type of events relating to
24885 the time interval which enable sending the specified command, and must
24886 be a non-null sequence of identifier flags separated by "+" or "|" and
24887 enclosed between "[" and "]".
24889 The following flags are recognized:
24892 The command is sent when the current frame timestamp enters the
24893 specified interval. In other words, the command is sent when the
24894 previous frame timestamp was not in the given interval, and the
24898 The command is sent when the current frame timestamp leaves the
24899 specified interval. In other words, the command is sent when the
24900 previous frame timestamp was in the given interval, and the
24904 The command @var{ARG} is interpreted as expression and result of
24905 expression is passed as @var{ARG}.
24907 The expression is evaluated through the eval API and can contain the following
24912 Original position in the file of the frame, or undefined if undefined
24913 for the current frame.
24916 The presentation timestamp in input.
24919 The count of the input frame for video or audio, starting from 0.
24922 The time in seconds of the current frame.
24925 The start time in seconds of the current command interval.
24928 The end time in seconds of the current command interval.
24931 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
24936 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
24939 @var{TARGET} specifies the target of the command, usually the name of
24940 the filter class or a specific filter instance name.
24942 @var{COMMAND} specifies the name of the command for the target filter.
24944 @var{ARG} is optional and specifies the optional list of argument for
24945 the given @var{COMMAND}.
24947 Between one interval specification and another, whitespaces, or
24948 sequences of characters starting with @code{#} until the end of line,
24949 are ignored and can be used to annotate comments.
24951 A simplified BNF description of the commands specification syntax
24954 @var{COMMAND_FLAG} ::= "enter" | "leave"
24955 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
24956 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
24957 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
24958 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
24959 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
24962 @subsection Examples
24966 Specify audio tempo change at second 4:
24968 asendcmd=c='4.0 atempo tempo 1.5',atempo
24972 Target a specific filter instance:
24974 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
24978 Specify a list of drawtext and hue commands in a file.
24980 # show text in the interval 5-10
24981 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
24982 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
24984 # desaturate the image in the interval 15-20
24985 15.0-20.0 [enter] hue s 0,
24986 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
24988 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
24990 # apply an exponential saturation fade-out effect, starting from time 25
24991 25 [enter] hue s exp(25-t)
24994 A filtergraph allowing to read and process the above command list
24995 stored in a file @file{test.cmd}, can be specified with:
24997 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
25002 @section setpts, asetpts
25004 Change the PTS (presentation timestamp) of the input frames.
25006 @code{setpts} works on video frames, @code{asetpts} on audio frames.
25008 This filter accepts the following options:
25013 The expression which is evaluated for each frame to construct its timestamp.
25017 The expression is evaluated through the eval API and can contain the following
25021 @item FRAME_RATE, FR
25022 frame rate, only defined for constant frame-rate video
25025 The presentation timestamp in input
25028 The count of the input frame for video or the number of consumed samples,
25029 not including the current frame for audio, starting from 0.
25031 @item NB_CONSUMED_SAMPLES
25032 The number of consumed samples, not including the current frame (only
25035 @item NB_SAMPLES, S
25036 The number of samples in the current frame (only audio)
25038 @item SAMPLE_RATE, SR
25039 The audio sample rate.
25042 The PTS of the first frame.
25045 the time in seconds of the first frame
25048 State whether the current frame is interlaced.
25051 the time in seconds of the current frame
25054 original position in the file of the frame, or undefined if undefined
25055 for the current frame
25058 The previous input PTS.
25061 previous input time in seconds
25064 The previous output PTS.
25067 previous output time in seconds
25070 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
25074 The wallclock (RTC) time at the start of the movie in microseconds.
25077 The timebase of the input timestamps.
25081 @subsection Examples
25085 Start counting PTS from zero
25087 setpts=PTS-STARTPTS
25091 Apply fast motion effect:
25097 Apply slow motion effect:
25103 Set fixed rate of 25 frames per second:
25109 Set fixed rate 25 fps with some jitter:
25111 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
25115 Apply an offset of 10 seconds to the input PTS:
25121 Generate timestamps from a "live source" and rebase onto the current timebase:
25123 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
25127 Generate timestamps by counting samples:
25136 Force color range for the output video frame.
25138 The @code{setrange} filter marks the color range property for the
25139 output frames. It does not change the input frame, but only sets the
25140 corresponding property, which affects how the frame is treated by
25143 The filter accepts the following options:
25148 Available values are:
25152 Keep the same color range property.
25154 @item unspecified, unknown
25155 Set the color range as unspecified.
25157 @item limited, tv, mpeg
25158 Set the color range as limited.
25160 @item full, pc, jpeg
25161 Set the color range as full.
25165 @section settb, asettb
25167 Set the timebase to use for the output frames timestamps.
25168 It is mainly useful for testing timebase configuration.
25170 It accepts the following parameters:
25175 The expression which is evaluated into the output timebase.
25179 The value for @option{tb} is an arithmetic expression representing a
25180 rational. The expression can contain the constants "AVTB" (the default
25181 timebase), "intb" (the input timebase) and "sr" (the sample rate,
25182 audio only). Default value is "intb".
25184 @subsection Examples
25188 Set the timebase to 1/25:
25194 Set the timebase to 1/10:
25200 Set the timebase to 1001/1000:
25206 Set the timebase to 2*intb:
25212 Set the default timebase value:
25219 Convert input audio to a video output representing frequency spectrum
25220 logarithmically using Brown-Puckette constant Q transform algorithm with
25221 direct frequency domain coefficient calculation (but the transform itself
25222 is not really constant Q, instead the Q factor is actually variable/clamped),
25223 with musical tone scale, from E0 to D#10.
25225 The filter accepts the following options:
25229 Specify the video size for the output. It must be even. For the syntax of this option,
25230 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25231 Default value is @code{1920x1080}.
25234 Set the output frame rate. Default value is @code{25}.
25237 Set the bargraph height. It must be even. Default value is @code{-1} which
25238 computes the bargraph height automatically.
25241 Set the axis height. It must be even. Default value is @code{-1} which computes
25242 the axis height automatically.
25245 Set the sonogram height. It must be even. Default value is @code{-1} which
25246 computes the sonogram height automatically.
25249 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
25250 instead. Default value is @code{1}.
25252 @item sono_v, volume
25253 Specify the sonogram volume expression. It can contain variables:
25256 the @var{bar_v} evaluated expression
25257 @item frequency, freq, f
25258 the frequency where it is evaluated
25259 @item timeclamp, tc
25260 the value of @var{timeclamp} option
25264 @item a_weighting(f)
25265 A-weighting of equal loudness
25266 @item b_weighting(f)
25267 B-weighting of equal loudness
25268 @item c_weighting(f)
25269 C-weighting of equal loudness.
25271 Default value is @code{16}.
25273 @item bar_v, volume2
25274 Specify the bargraph volume expression. It can contain variables:
25277 the @var{sono_v} evaluated expression
25278 @item frequency, freq, f
25279 the frequency where it is evaluated
25280 @item timeclamp, tc
25281 the value of @var{timeclamp} option
25285 @item a_weighting(f)
25286 A-weighting of equal loudness
25287 @item b_weighting(f)
25288 B-weighting of equal loudness
25289 @item c_weighting(f)
25290 C-weighting of equal loudness.
25292 Default value is @code{sono_v}.
25294 @item sono_g, gamma
25295 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
25296 higher gamma makes the spectrum having more range. Default value is @code{3}.
25297 Acceptable range is @code{[1, 7]}.
25299 @item bar_g, gamma2
25300 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
25304 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
25305 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
25307 @item timeclamp, tc
25308 Specify the transform timeclamp. At low frequency, there is trade-off between
25309 accuracy in time domain and frequency domain. If timeclamp is lower,
25310 event in time domain is represented more accurately (such as fast bass drum),
25311 otherwise event in frequency domain is represented more accurately
25312 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
25315 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
25316 limits future samples by applying asymmetric windowing in time domain, useful
25317 when low latency is required. Accepted range is @code{[0, 1]}.
25320 Specify the transform base frequency. Default value is @code{20.01523126408007475},
25321 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
25324 Specify the transform end frequency. Default value is @code{20495.59681441799654},
25325 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
25328 This option is deprecated and ignored.
25331 Specify the transform length in time domain. Use this option to control accuracy
25332 trade-off between time domain and frequency domain at every frequency sample.
25333 It can contain variables:
25335 @item frequency, freq, f
25336 the frequency where it is evaluated
25337 @item timeclamp, tc
25338 the value of @var{timeclamp} option.
25340 Default value is @code{384*tc/(384+tc*f)}.
25343 Specify the transform count for every video frame. Default value is @code{6}.
25344 Acceptable range is @code{[1, 30]}.
25347 Specify the transform count for every single pixel. Default value is @code{0},
25348 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
25351 Specify font file for use with freetype to draw the axis. If not specified,
25352 use embedded font. Note that drawing with font file or embedded font is not
25353 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
25357 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
25358 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
25362 Specify font color expression. This is arithmetic expression that should return
25363 integer value 0xRRGGBB. It can contain variables:
25365 @item frequency, freq, f
25366 the frequency where it is evaluated
25367 @item timeclamp, tc
25368 the value of @var{timeclamp} option
25373 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
25374 @item r(x), g(x), b(x)
25375 red, green, and blue value of intensity x.
25377 Default value is @code{st(0, (midi(f)-59.5)/12);
25378 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
25379 r(1-ld(1)) + b(ld(1))}.
25382 Specify image file to draw the axis. This option override @var{fontfile} and
25383 @var{fontcolor} option.
25386 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
25387 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
25388 Default value is @code{1}.
25391 Set colorspace. The accepted values are:
25394 Unspecified (default)
25403 BT.470BG or BT.601-6 625
25406 SMPTE-170M or BT.601-6 525
25412 BT.2020 with non-constant luminance
25417 Set spectrogram color scheme. This is list of floating point values with format
25418 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
25419 The default is @code{1|0.5|0|0|0.5|1}.
25423 @subsection Examples
25427 Playing audio while showing the spectrum:
25429 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
25433 Same as above, but with frame rate 30 fps:
25435 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
25439 Playing at 1280x720:
25441 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
25445 Disable sonogram display:
25451 A1 and its harmonics: A1, A2, (near)E3, A3:
25453 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),
25454 asplit[a][out1]; [a] showcqt [out0]'
25458 Same as above, but with more accuracy in frequency domain:
25460 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),
25461 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
25467 bar_v=10:sono_v=bar_v*a_weighting(f)
25471 Custom gamma, now spectrum is linear to the amplitude.
25477 Custom tlength equation:
25479 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)))'
25483 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
25485 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
25489 Custom font using fontconfig:
25491 font='Courier New,Monospace,mono|bold'
25495 Custom frequency range with custom axis using image file:
25497 axisfile=myaxis.png:basefreq=40:endfreq=10000
25503 Convert input audio to video output representing the audio power spectrum.
25504 Audio amplitude is on Y-axis while frequency is on X-axis.
25506 The filter accepts the following options:
25510 Specify size of video. For the syntax of this option, check the
25511 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25512 Default is @code{1024x512}.
25516 This set how each frequency bin will be represented.
25518 It accepts the following values:
25524 Default is @code{bar}.
25527 Set amplitude scale.
25529 It accepts the following values:
25543 Default is @code{log}.
25546 Set frequency scale.
25548 It accepts the following values:
25557 Reverse logarithmic scale.
25559 Default is @code{lin}.
25562 Set window size. Allowed range is from 16 to 65536.
25564 Default is @code{2048}
25567 Set windowing function.
25569 It accepts the following values:
25592 Default is @code{hanning}.
25595 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25596 which means optimal overlap for selected window function will be picked.
25599 Set time averaging. Setting this to 0 will display current maximal peaks.
25600 Default is @code{1}, which means time averaging is disabled.
25603 Specify list of colors separated by space or by '|' which will be used to
25604 draw channel frequencies. Unrecognized or missing colors will be replaced
25608 Set channel display mode.
25610 It accepts the following values:
25615 Default is @code{combined}.
25618 Set minimum amplitude used in @code{log} amplitude scaler.
25621 Set data display mode.
25623 It accepts the following values:
25629 Default is @code{magnitude}.
25632 @section showspatial
25634 Convert stereo input audio to a video output, representing the spatial relationship
25635 between two channels.
25637 The filter accepts the following options:
25641 Specify the video size for the output. For the syntax of this option, check the
25642 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25643 Default value is @code{512x512}.
25646 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
25649 Set window function.
25651 It accepts the following values:
25676 Default value is @code{hann}.
25679 Set ratio of overlap window. Default value is @code{0.5}.
25680 When value is @code{1} overlap is set to recommended size for specific
25681 window function currently used.
25684 @anchor{showspectrum}
25685 @section showspectrum
25687 Convert input audio to a video output, representing the audio frequency
25690 The filter accepts the following options:
25694 Specify the video size for the output. For the syntax of this option, check the
25695 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25696 Default value is @code{640x512}.
25699 Specify how the spectrum should slide along the window.
25701 It accepts the following values:
25704 the samples start again on the left when they reach the right
25706 the samples scroll from right to left
25708 frames are only produced when the samples reach the right
25710 the samples scroll from left to right
25713 Default value is @code{replace}.
25716 Specify display mode.
25718 It accepts the following values:
25721 all channels are displayed in the same row
25723 all channels are displayed in separate rows
25726 Default value is @samp{combined}.
25729 Specify display color mode.
25731 It accepts the following values:
25734 each channel is displayed in a separate color
25736 each channel is displayed using the same color scheme
25738 each channel is displayed using the rainbow color scheme
25740 each channel is displayed using the moreland color scheme
25742 each channel is displayed using the nebulae color scheme
25744 each channel is displayed using the fire color scheme
25746 each channel is displayed using the fiery color scheme
25748 each channel is displayed using the fruit color scheme
25750 each channel is displayed using the cool color scheme
25752 each channel is displayed using the magma color scheme
25754 each channel is displayed using the green color scheme
25756 each channel is displayed using the viridis color scheme
25758 each channel is displayed using the plasma color scheme
25760 each channel is displayed using the cividis color scheme
25762 each channel is displayed using the terrain color scheme
25765 Default value is @samp{channel}.
25768 Specify scale used for calculating intensity color values.
25770 It accepts the following values:
25775 square root, default
25786 Default value is @samp{sqrt}.
25789 Specify frequency scale.
25791 It accepts the following values:
25799 Default value is @samp{lin}.
25802 Set saturation modifier for displayed colors. Negative values provide
25803 alternative color scheme. @code{0} is no saturation at all.
25804 Saturation must be in [-10.0, 10.0] range.
25805 Default value is @code{1}.
25808 Set window function.
25810 It accepts the following values:
25835 Default value is @code{hann}.
25838 Set orientation of time vs frequency axis. Can be @code{vertical} or
25839 @code{horizontal}. Default is @code{vertical}.
25842 Set ratio of overlap window. Default value is @code{0}.
25843 When value is @code{1} overlap is set to recommended size for specific
25844 window function currently used.
25847 Set scale gain for calculating intensity color values.
25848 Default value is @code{1}.
25851 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
25854 Set color rotation, must be in [-1.0, 1.0] range.
25855 Default value is @code{0}.
25858 Set start frequency from which to display spectrogram. Default is @code{0}.
25861 Set stop frequency to which to display spectrogram. Default is @code{0}.
25864 Set upper frame rate limit. Default is @code{auto}, unlimited.
25867 Draw time and frequency axes and legends. Default is disabled.
25870 The usage is very similar to the showwaves filter; see the examples in that
25873 @subsection Examples
25877 Large window with logarithmic color scaling:
25879 showspectrum=s=1280x480:scale=log
25883 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
25885 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
25886 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
25890 @section showspectrumpic
25892 Convert input audio to a single video frame, representing the audio frequency
25895 The filter accepts the following options:
25899 Specify the video size for the output. For the syntax of this option, check the
25900 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25901 Default value is @code{4096x2048}.
25904 Specify display mode.
25906 It accepts the following values:
25909 all channels are displayed in the same row
25911 all channels are displayed in separate rows
25913 Default value is @samp{combined}.
25916 Specify display color mode.
25918 It accepts the following values:
25921 each channel is displayed in a separate color
25923 each channel is displayed using the same color scheme
25925 each channel is displayed using the rainbow color scheme
25927 each channel is displayed using the moreland color scheme
25929 each channel is displayed using the nebulae color scheme
25931 each channel is displayed using the fire color scheme
25933 each channel is displayed using the fiery color scheme
25935 each channel is displayed using the fruit color scheme
25937 each channel is displayed using the cool color scheme
25939 each channel is displayed using the magma color scheme
25941 each channel is displayed using the green color scheme
25943 each channel is displayed using the viridis color scheme
25945 each channel is displayed using the plasma color scheme
25947 each channel is displayed using the cividis color scheme
25949 each channel is displayed using the terrain color scheme
25951 Default value is @samp{intensity}.
25954 Specify scale used for calculating intensity color values.
25956 It accepts the following values:
25961 square root, default
25971 Default value is @samp{log}.
25974 Specify frequency scale.
25976 It accepts the following values:
25984 Default value is @samp{lin}.
25987 Set saturation modifier for displayed colors. Negative values provide
25988 alternative color scheme. @code{0} is no saturation at all.
25989 Saturation must be in [-10.0, 10.0] range.
25990 Default value is @code{1}.
25993 Set window function.
25995 It accepts the following values:
26019 Default value is @code{hann}.
26022 Set orientation of time vs frequency axis. Can be @code{vertical} or
26023 @code{horizontal}. Default is @code{vertical}.
26026 Set scale gain for calculating intensity color values.
26027 Default value is @code{1}.
26030 Draw time and frequency axes and legends. Default is enabled.
26033 Set color rotation, must be in [-1.0, 1.0] range.
26034 Default value is @code{0}.
26037 Set start frequency from which to display spectrogram. Default is @code{0}.
26040 Set stop frequency to which to display spectrogram. Default is @code{0}.
26043 @subsection Examples
26047 Extract an audio spectrogram of a whole audio track
26048 in a 1024x1024 picture using @command{ffmpeg}:
26050 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
26054 @section showvolume
26056 Convert input audio volume to a video output.
26058 The filter accepts the following options:
26065 Set border width, allowed range is [0, 5]. Default is 1.
26068 Set channel width, allowed range is [80, 8192]. Default is 400.
26071 Set channel height, allowed range is [1, 900]. Default is 20.
26074 Set fade, allowed range is [0, 1]. Default is 0.95.
26077 Set volume color expression.
26079 The expression can use the following variables:
26083 Current max volume of channel in dB.
26089 Current channel number, starting from 0.
26093 If set, displays channel names. Default is enabled.
26096 If set, displays volume values. Default is enabled.
26099 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
26100 default is @code{h}.
26103 Set step size, allowed range is [0, 5]. Default is 0, which means
26107 Set background opacity, allowed range is [0, 1]. Default is 0.
26110 Set metering mode, can be peak: @code{p} or rms: @code{r},
26111 default is @code{p}.
26114 Set display scale, can be linear: @code{lin} or log: @code{log},
26115 default is @code{lin}.
26119 If set to > 0., display a line for the max level
26120 in the previous seconds.
26121 default is disabled: @code{0.}
26124 The color of the max line. Use when @code{dm} option is set to > 0.
26125 default is: @code{orange}
26130 Convert input audio to a video output, representing the samples waves.
26132 The filter accepts the following options:
26136 Specify the video size for the output. For the syntax of this option, check the
26137 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26138 Default value is @code{600x240}.
26143 Available values are:
26146 Draw a point for each sample.
26149 Draw a vertical line for each sample.
26152 Draw a point for each sample and a line between them.
26155 Draw a centered vertical line for each sample.
26158 Default value is @code{point}.
26161 Set the number of samples which are printed on the same column. A
26162 larger value will decrease the frame rate. Must be a positive
26163 integer. This option can be set only if the value for @var{rate}
26164 is not explicitly specified.
26167 Set the (approximate) output frame rate. This is done by setting the
26168 option @var{n}. Default value is "25".
26170 @item split_channels
26171 Set if channels should be drawn separately or overlap. Default value is 0.
26174 Set colors separated by '|' which are going to be used for drawing of each channel.
26177 Set amplitude scale.
26179 Available values are:
26197 Set the draw mode. This is mostly useful to set for high @var{n}.
26199 Available values are:
26202 Scale pixel values for each drawn sample.
26205 Draw every sample directly.
26208 Default value is @code{scale}.
26211 @subsection Examples
26215 Output the input file audio and the corresponding video representation
26218 amovie=a.mp3,asplit[out0],showwaves[out1]
26222 Create a synthetic signal and show it with showwaves, forcing a
26223 frame rate of 30 frames per second:
26225 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
26229 @section showwavespic
26231 Convert input audio to a single video frame, representing the samples waves.
26233 The filter accepts the following options:
26237 Specify the video size for the output. For the syntax of this option, check the
26238 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26239 Default value is @code{600x240}.
26241 @item split_channels
26242 Set if channels should be drawn separately or overlap. Default value is 0.
26245 Set colors separated by '|' which are going to be used for drawing of each channel.
26248 Set amplitude scale.
26250 Available values are:
26270 Available values are:
26273 Scale pixel values for each drawn sample.
26276 Draw every sample directly.
26279 Default value is @code{scale}.
26282 Set the filter mode.
26284 Available values are:
26287 Use average samples values for each drawn sample.
26290 Use peak samples values for each drawn sample.
26293 Default value is @code{average}.
26296 @subsection Examples
26300 Extract a channel split representation of the wave form of a whole audio track
26301 in a 1024x800 picture using @command{ffmpeg}:
26303 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
26307 @section sidedata, asidedata
26309 Delete frame side data, or select frames based on it.
26311 This filter accepts the following options:
26315 Set mode of operation of the filter.
26317 Can be one of the following:
26321 Select every frame with side data of @code{type}.
26324 Delete side data of @code{type}. If @code{type} is not set, delete all side
26330 Set side data type used with all modes. Must be set for @code{select} mode. For
26331 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
26332 in @file{libavutil/frame.h}. For example, to choose
26333 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
26337 @section spectrumsynth
26339 Synthesize audio from 2 input video spectrums, first input stream represents
26340 magnitude across time and second represents phase across time.
26341 The filter will transform from frequency domain as displayed in videos back
26342 to time domain as presented in audio output.
26344 This filter is primarily created for reversing processed @ref{showspectrum}
26345 filter outputs, but can synthesize sound from other spectrograms too.
26346 But in such case results are going to be poor if the phase data is not
26347 available, because in such cases phase data need to be recreated, usually
26348 it's just recreated from random noise.
26349 For best results use gray only output (@code{channel} color mode in
26350 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
26351 @code{lin} scale for phase video. To produce phase, for 2nd video, use
26352 @code{data} option. Inputs videos should generally use @code{fullframe}
26353 slide mode as that saves resources needed for decoding video.
26355 The filter accepts the following options:
26359 Specify sample rate of output audio, the sample rate of audio from which
26360 spectrum was generated may differ.
26363 Set number of channels represented in input video spectrums.
26366 Set scale which was used when generating magnitude input spectrum.
26367 Can be @code{lin} or @code{log}. Default is @code{log}.
26370 Set slide which was used when generating inputs spectrums.
26371 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
26372 Default is @code{fullframe}.
26375 Set window function used for resynthesis.
26378 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
26379 which means optimal overlap for selected window function will be picked.
26382 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
26383 Default is @code{vertical}.
26386 @subsection Examples
26390 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
26391 then resynthesize videos back to audio with spectrumsynth:
26393 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
26394 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
26395 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
26399 @section split, asplit
26401 Split input into several identical outputs.
26403 @code{asplit} works with audio input, @code{split} with video.
26405 The filter accepts a single parameter which specifies the number of outputs. If
26406 unspecified, it defaults to 2.
26408 @subsection Examples
26412 Create two separate outputs from the same input:
26414 [in] split [out0][out1]
26418 To create 3 or more outputs, you need to specify the number of
26421 [in] asplit=3 [out0][out1][out2]
26425 Create two separate outputs from the same input, one cropped and
26428 [in] split [splitout1][splitout2];
26429 [splitout1] crop=100:100:0:0 [cropout];
26430 [splitout2] pad=200:200:100:100 [padout];
26434 Create 5 copies of the input audio with @command{ffmpeg}:
26436 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
26442 Receive commands sent through a libzmq client, and forward them to
26443 filters in the filtergraph.
26445 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
26446 must be inserted between two video filters, @code{azmq} between two
26447 audio filters. Both are capable to send messages to any filter type.
26449 To enable these filters you need to install the libzmq library and
26450 headers and configure FFmpeg with @code{--enable-libzmq}.
26452 For more information about libzmq see:
26453 @url{http://www.zeromq.org/}
26455 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
26456 receives messages sent through a network interface defined by the
26457 @option{bind_address} (or the abbreviation "@option{b}") option.
26458 Default value of this option is @file{tcp://localhost:5555}. You may
26459 want to alter this value to your needs, but do not forget to escape any
26460 ':' signs (see @ref{filtergraph escaping}).
26462 The received message must be in the form:
26464 @var{TARGET} @var{COMMAND} [@var{ARG}]
26467 @var{TARGET} specifies the target of the command, usually the name of
26468 the filter class or a specific filter instance name. The default
26469 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
26470 but you can override this by using the @samp{filter_name@@id} syntax
26471 (see @ref{Filtergraph syntax}).
26473 @var{COMMAND} specifies the name of the command for the target filter.
26475 @var{ARG} is optional and specifies the optional argument list for the
26476 given @var{COMMAND}.
26478 Upon reception, the message is processed and the corresponding command
26479 is injected into the filtergraph. Depending on the result, the filter
26480 will send a reply to the client, adopting the format:
26482 @var{ERROR_CODE} @var{ERROR_REASON}
26486 @var{MESSAGE} is optional.
26488 @subsection Examples
26490 Look at @file{tools/zmqsend} for an example of a zmq client which can
26491 be used to send commands processed by these filters.
26493 Consider the following filtergraph generated by @command{ffplay}.
26494 In this example the last overlay filter has an instance name. All other
26495 filters will have default instance names.
26498 ffplay -dumpgraph 1 -f lavfi "
26499 color=s=100x100:c=red [l];
26500 color=s=100x100:c=blue [r];
26501 nullsrc=s=200x100, zmq [bg];
26502 [bg][l] overlay [bg+l];
26503 [bg+l][r] overlay@@my=x=100 "
26506 To change the color of the left side of the video, the following
26507 command can be used:
26509 echo Parsed_color_0 c yellow | tools/zmqsend
26512 To change the right side:
26514 echo Parsed_color_1 c pink | tools/zmqsend
26517 To change the position of the right side:
26519 echo overlay@@my x 150 | tools/zmqsend
26523 @c man end MULTIMEDIA FILTERS
26525 @chapter Multimedia Sources
26526 @c man begin MULTIMEDIA SOURCES
26528 Below is a description of the currently available multimedia sources.
26532 This is the same as @ref{movie} source, except it selects an audio
26538 Read audio and/or video stream(s) from a movie container.
26540 It accepts the following parameters:
26544 The name of the resource to read (not necessarily a file; it can also be a
26545 device or a stream accessed through some protocol).
26547 @item format_name, f
26548 Specifies the format assumed for the movie to read, and can be either
26549 the name of a container or an input device. If not specified, the
26550 format is guessed from @var{movie_name} or by probing.
26552 @item seek_point, sp
26553 Specifies the seek point in seconds. The frames will be output
26554 starting from this seek point. The parameter is evaluated with
26555 @code{av_strtod}, so the numerical value may be suffixed by an IS
26556 postfix. The default value is "0".
26559 Specifies the streams to read. Several streams can be specified,
26560 separated by "+". The source will then have as many outputs, in the
26561 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
26562 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
26563 respectively the default (best suited) video and audio stream. Default
26564 is "dv", or "da" if the filter is called as "amovie".
26566 @item stream_index, si
26567 Specifies the index of the video stream to read. If the value is -1,
26568 the most suitable video stream will be automatically selected. The default
26569 value is "-1". Deprecated. If the filter is called "amovie", it will select
26570 audio instead of video.
26573 Specifies how many times to read the stream in sequence.
26574 If the value is 0, the stream will be looped infinitely.
26575 Default value is "1".
26577 Note that when the movie is looped the source timestamps are not
26578 changed, so it will generate non monotonically increasing timestamps.
26580 @item discontinuity
26581 Specifies the time difference between frames above which the point is
26582 considered a timestamp discontinuity which is removed by adjusting the later
26586 It allows overlaying a second video on top of the main input of
26587 a filtergraph, as shown in this graph:
26589 input -----------> deltapts0 --> overlay --> output
26592 movie --> scale--> deltapts1 -------+
26594 @subsection Examples
26598 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
26599 on top of the input labelled "in":
26601 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
26602 [in] setpts=PTS-STARTPTS [main];
26603 [main][over] overlay=16:16 [out]
26607 Read from a video4linux2 device, and overlay it on top of the input
26610 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
26611 [in] setpts=PTS-STARTPTS [main];
26612 [main][over] overlay=16:16 [out]
26616 Read the first video stream and the audio stream with id 0x81 from
26617 dvd.vob; the video is connected to the pad named "video" and the audio is
26618 connected to the pad named "audio":
26620 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
26624 @subsection Commands
26626 Both movie and amovie support the following commands:
26629 Perform seek using "av_seek_frame".
26630 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
26633 @var{stream_index}: If stream_index is -1, a default
26634 stream is selected, and @var{timestamp} is automatically converted
26635 from AV_TIME_BASE units to the stream specific time_base.
26637 @var{timestamp}: Timestamp in AVStream.time_base units
26638 or, if no stream is specified, in AV_TIME_BASE units.
26640 @var{flags}: Flags which select direction and seeking mode.
26644 Get movie duration in AV_TIME_BASE units.
26648 @c man end MULTIMEDIA SOURCES