1 @chapter Filtering Introduction
2 @c man begin FILTERING INTRODUCTION
4 Filtering in FFmpeg is enabled through the libavfilter library.
6 Libavfilter is the filtering API of FFmpeg. It is the substitute of
7 the now deprecated 'vhooks' and started as a Google Summer of Code
10 Audio filtering integration into the main FFmpeg repository is a work in
11 progress, so audio API and ABI should not be considered stable yet.
13 In libavfilter, it is possible for filters to have multiple inputs and
15 To illustrate the sorts of things that are possible, we can
16 use a complex filter graph. For example, the following one:
19 input --> split --> fifo -----------------------> overlay --> output
22 +------> fifo --> crop --> vflip --------+
25 splits the stream in two streams, sends one stream through the crop filter
26 and the vflip filter before merging it back with the other stream by
27 overlaying it on top. You can use the following command to achieve this:
30 ffmpeg -i input -vf "[in] split [T1], fifo, [T2] overlay=0:H/2 [out]; [T1] fifo, crop=iw:ih/2:0:ih/2, vflip [T2]" output
33 The result will be that in output the top half of the video is mirrored
36 Filters are loaded using the @var{-vf} or @var{-af} option passed to
37 @command{ffmpeg} or to @command{ffplay}. Filters in the same linear
38 chain are separated by commas. In our example, @var{split, fifo,
39 overlay} are in one linear chain, and @var{fifo, crop, vflip} are in
40 another. The points where the linear chains join are labeled by names
41 enclosed in square brackets. In our example, that is @var{[T1]} and
42 @var{[T2]}. The special labels @var{[in]} and @var{[out]} are the points
43 where video is input and output.
45 Some filters take in input a list of parameters: they are specified
46 after the filter name and an equal sign, and are separated from each other
49 There exist so-called @var{source filters} that do not have an
50 audio/video input, and @var{sink filters} that will not have audio/video
53 @c man end FILTERING INTRODUCTION
56 @c man begin GRAPH2DOT
58 The @file{graph2dot} program included in the FFmpeg @file{tools}
59 directory can be used to parse a filter graph description and issue a
60 corresponding textual representation in the dot language.
67 to see how to use @file{graph2dot}.
69 You can then pass the dot description to the @file{dot} program (from
70 the graphviz suite of programs) and obtain a graphical representation
73 For example the sequence of commands:
75 echo @var{GRAPH_DESCRIPTION} | \
76 tools/graph2dot -o graph.tmp && \
77 dot -Tpng graph.tmp -o graph.png && \
81 can be used to create and display an image representing the graph
82 described by the @var{GRAPH_DESCRIPTION} string.
86 @chapter Filtergraph description
87 @c man begin FILTERGRAPH DESCRIPTION
89 A filtergraph is a directed graph of connected filters. It can contain
90 cycles, and there can be multiple links between a pair of
91 filters. Each link has one input pad on one side connecting it to one
92 filter from which it takes its input, and one output pad on the other
93 side connecting it to the one filter accepting its output.
95 Each filter in a filtergraph is an instance of a filter class
96 registered in the application, which defines the features and the
97 number of input and output pads of the filter.
99 A filter with no input pads is called a "source", a filter with no
100 output pads is called a "sink".
102 @anchor{Filtergraph syntax}
103 @section Filtergraph syntax
105 A filtergraph can be represented using a textual representation, which is
106 recognized by the @option{-filter}/@option{-vf} and @option{-filter_complex}
107 options in @command{ffmpeg} and @option{-vf} in @command{ffplay}, and by the
108 @code{avfilter_graph_parse()}/@code{avfilter_graph_parse2()} function defined in
109 @file{libavfilter/avfiltergraph.h}.
111 A filterchain consists of a sequence of connected filters, each one
112 connected to the previous one in the sequence. A filterchain is
113 represented by a list of ","-separated filter descriptions.
115 A filtergraph consists of a sequence of filterchains. A sequence of
116 filterchains is represented by a list of ";"-separated filterchain
119 A filter is represented by a string of the form:
120 [@var{in_link_1}]...[@var{in_link_N}]@var{filter_name}=@var{arguments}[@var{out_link_1}]...[@var{out_link_M}]
122 @var{filter_name} is the name of the filter class of which the
123 described filter is an instance of, and has to be the name of one of
124 the filter classes registered in the program.
125 The name of the filter class is optionally followed by a string
128 @var{arguments} is a string which contains the parameters used to
129 initialize the filter instance, and are described in the filter
132 The list of arguments can be quoted using the character "'" as initial
133 and ending mark, and the character '\' for escaping the characters
134 within the quoted text; otherwise the argument string is considered
135 terminated when the next special character (belonging to the set
136 "[]=;,") is encountered.
138 The name and arguments of the filter are optionally preceded and
139 followed by a list of link labels.
140 A link label allows to name a link and associate it to a filter output
141 or input pad. The preceding labels @var{in_link_1}
142 ... @var{in_link_N}, are associated to the filter input pads,
143 the following labels @var{out_link_1} ... @var{out_link_M}, are
144 associated to the output pads.
146 When two link labels with the same name are found in the
147 filtergraph, a link between the corresponding input and output pad is
150 If an output pad is not labelled, it is linked by default to the first
151 unlabelled input pad of the next filter in the filterchain.
152 For example in the filterchain:
154 nullsrc, split[L1], [L2]overlay, nullsink
156 the split filter instance has two output pads, and the overlay filter
157 instance two input pads. The first output pad of split is labelled
158 "L1", the first input pad of overlay is labelled "L2", and the second
159 output pad of split is linked to the second input pad of overlay,
160 which are both unlabelled.
162 In a complete filterchain all the unlabelled filter input and output
163 pads must be connected. A filtergraph is considered valid if all the
164 filter input and output pads of all the filterchains are connected.
166 Libavfilter will automatically insert scale filters where format
167 conversion is required. It is possible to specify swscale flags
168 for those automatically inserted scalers by prepending
169 @code{sws_flags=@var{flags};}
170 to the filtergraph description.
172 Follows a BNF description for the filtergraph syntax:
174 @var{NAME} ::= sequence of alphanumeric characters and '_'
175 @var{LINKLABEL} ::= "[" @var{NAME} "]"
176 @var{LINKLABELS} ::= @var{LINKLABEL} [@var{LINKLABELS}]
177 @var{FILTER_ARGUMENTS} ::= sequence of chars (eventually quoted)
178 @var{FILTER} ::= [@var{LINKNAMES}] @var{NAME} ["=" @var{ARGUMENTS}] [@var{LINKNAMES}]
179 @var{FILTERCHAIN} ::= @var{FILTER} [,@var{FILTERCHAIN}]
180 @var{FILTERGRAPH} ::= [sws_flags=@var{flags};] @var{FILTERCHAIN} [;@var{FILTERGRAPH}]
183 @c man end FILTERGRAPH DESCRIPTION
185 @chapter Audio Filters
186 @c man begin AUDIO FILTERS
188 When you configure your FFmpeg build, you can disable any of the
189 existing filters using @code{--disable-filters}.
190 The configure output will show the audio filters included in your
193 Below is a description of the currently available audio filters.
197 Convert the input audio format to the specified formats.
199 The filter accepts a string of the form:
200 "@var{sample_format}:@var{channel_layout}".
202 @var{sample_format} specifies the sample format, and can be a string or the
203 corresponding numeric value defined in @file{libavutil/samplefmt.h}. Use 'p'
204 suffix for a planar sample format.
206 @var{channel_layout} specifies the channel layout, and can be a string
207 or the corresponding number value defined in @file{libavutil/audioconvert.h}.
209 The special parameter "auto", signifies that the filter will
210 automatically select the output format depending on the output filter.
212 Some examples follow.
216 Convert input to float, planar, stereo:
222 Convert input to unsigned 8-bit, automatically select out channel layout:
230 Convert the input audio to one of the specified formats. The framework will
231 negotiate the most appropriate format to minimize conversions.
233 The filter accepts the following named parameters:
237 A comma-separated list of requested sample formats.
240 A comma-separated list of requested sample rates.
242 @item channel_layouts
243 A comma-separated list of requested channel layouts.
247 If a parameter is omitted, all values are allowed.
249 For example to force the output to either unsigned 8-bit or signed 16-bit stereo:
251 aformat=sample_fmts\=u8\,s16:channel_layouts\=stereo
256 Merge two or more audio streams into a single multi-channel stream.
258 The filter accepts the following named options:
263 Set the number of inputs. Default is 2.
267 If the channel layouts of the inputs are disjoint, and therefore compatible,
268 the channel layout of the output will be set accordingly and the channels
269 will be reordered as necessary. If the channel layouts of the inputs are not
270 disjoint, the output will have all the channels of the first input then all
271 the channels of the second input, in that order, and the channel layout of
272 the output will be the default value corresponding to the total number of
275 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
276 is FC+BL+BR, then the output will be in 5.1, with the channels in the
277 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
278 first input, b1 is the first channel of the second input).
280 On the other hand, if both input are in stereo, the output channels will be
281 in the default order: a1, a2, b1, b2, and the channel layout will be
282 arbitrarily set to 4.0, which may or may not be the expected value.
284 All inputs must have the same sample rate, and format.
286 If inputs do not have the same duration, the output will stop with the
289 Example: merge two mono files into a stereo stream:
291 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
294 Example: multiple merges:
297 amovie=input.mkv:si=0 [a0];
298 amovie=input.mkv:si=1 [a1];
299 amovie=input.mkv:si=2 [a2];
300 amovie=input.mkv:si=3 [a3];
301 amovie=input.mkv:si=4 [a4];
302 amovie=input.mkv:si=5 [a5];
303 [a0][a1][a2][a3][a4][a5] amerge=inputs=6" -c:a pcm_s16le output.mkv
308 Mixes multiple audio inputs into a single output.
312 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
314 will mix 3 input audio streams to a single output with the same duration as the
315 first input and a dropout transition time of 3 seconds.
317 The filter accepts the following named parameters:
321 Number of inputs. If unspecified, it defaults to 2.
324 How to determine the end-of-stream.
328 Duration of longest input. (default)
331 Duration of shortest input.
334 Duration of first input.
338 @item dropout_transition
339 Transition time, in seconds, for volume renormalization when an input
340 stream ends. The default value is 2 seconds.
346 Pass the audio source unchanged to the output.
350 Resample the input audio to the specified sample rate.
352 The filter accepts exactly one parameter, the output sample rate. If not
353 specified then the filter will automatically convert between its input
354 and output sample rates.
356 For example, to resample the input audio to 44100Hz:
361 @section asetnsamples
363 Set the number of samples per each output audio frame.
365 The last output packet may contain a different number of samples, as
366 the filter will flush all the remaining samples when the input audio
369 The filter accepts parameters as a list of @var{key}=@var{value} pairs,
374 @item nb_out_samples, n
375 Set the number of frames per each output audio frame. The number is
376 intended as the number of samples @emph{per each channel}.
377 Default value is 1024.
380 If set to 1, the filter will pad the last audio frame with zeroes, so
381 that the last frame will contain the same number of samples as the
382 previous ones. Default value is 1.
385 For example, to set the number of per-frame samples to 1234 and
386 disable padding for the last frame, use:
388 asetnsamples=n=1234:p=0
393 Show a line containing various information for each input audio frame.
394 The input audio is not modified.
396 The shown line contains a sequence of key/value pairs of the form
397 @var{key}:@var{value}.
399 A description of each shown parameter follows:
403 sequential number of the input frame, starting from 0
406 presentation TimeStamp of the input frame, expressed as a number of
407 time base units. The time base unit depends on the filter input pad, and
408 is usually 1/@var{sample_rate}.
411 presentation TimeStamp of the input frame, expressed as a number of
415 position of the frame in the input stream, -1 if this information in
416 unavailable and/or meaningless (for example in case of synthetic audio)
422 channel layout description
425 number of samples (per each channel) contained in the filtered frame
428 sample rate for the audio frame
431 Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame
434 Adler-32 checksum (printed in hexadecimal) for each input frame plane,
435 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3} @var{c4} @var{c5}
441 Split input audio into several identical outputs.
443 The filter accepts a single parameter which specifies the number of outputs. If
444 unspecified, it defaults to 2.
448 [in] asplit [out0][out1]
451 will create two separate outputs from the same input.
453 To create 3 or more outputs, you need to specify the number of
456 [in] asplit=3 [out0][out1][out2]
460 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
462 will create 5 copies of the input audio.
467 Forward two audio streams and control the order the buffers are forwarded.
469 The argument to the filter is an expression deciding which stream should be
470 forwarded next: if the result is negative, the first stream is forwarded; if
471 the result is positive or zero, the second stream is forwarded. It can use
472 the following variables:
476 number of buffers forwarded so far on each stream
478 number of samples forwarded so far on each stream
480 current timestamp of each stream
483 The default value is @code{t1-t2}, which means to always forward the stream
484 that has a smaller timestamp.
486 Example: stress-test @code{amerge} by randomly sending buffers on the wrong
487 input, while avoiding too much of a desynchronization:
489 amovie=file.ogg [a] ; amovie=file.mp3 [b] ;
490 [a] [b] astreamsync=(2*random(1))-1+tanh(5*(t1-t2)) [a2] [b2] ;
498 The filter accepts exactly one parameter, the audio tempo. If not
499 specified then the filter will assume nominal 1.0 tempo. Tempo must
500 be in the [0.5, 2.0] range.
502 For example, to slow down audio to 80% tempo:
507 For example, to speed up audio to 125% tempo:
514 Make audio easier to listen to on headphones.
516 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
517 so that when listened to on headphones the stereo image is moved from
518 inside your head (standard for headphones) to outside and in front of
519 the listener (standard for speakers).
525 Mix channels with specific gain levels. The filter accepts the output
526 channel layout followed by a set of channels definitions.
528 This filter is also designed to remap efficiently the channels of an audio
531 The filter accepts parameters of the form:
532 "@var{l}:@var{outdef}:@var{outdef}:..."
536 output channel layout or number of channels
539 output channel specification, of the form:
540 "@var{out_name}=[@var{gain}*]@var{in_name}[+[@var{gain}*]@var{in_name}...]"
543 output channel to define, either a channel name (FL, FR, etc.) or a channel
544 number (c0, c1, etc.)
547 multiplicative coefficient for the channel, 1 leaving the volume unchanged
550 input channel to use, see out_name for details; it is not possible to mix
551 named and numbered input channels
554 If the `=' in a channel specification is replaced by `<', then the gains for
555 that specification will be renormalized so that the total is 1, thus
556 avoiding clipping noise.
558 @subsection Mixing examples
560 For example, if you want to down-mix from stereo to mono, but with a bigger
561 factor for the left channel:
563 pan=1:c0=0.9*c0+0.1*c1
566 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
569 pan=stereo: FL < FL + 0.5*FC + 0.6*BL + 0.6*SL : FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
572 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
573 that should be preferred (see "-ac" option) unless you have very specific
576 @subsection Remapping examples
578 The channel remapping will be effective if, and only if:
581 @item gain coefficients are zeroes or ones,
582 @item only one input per channel output,
585 If all these conditions are satisfied, the filter will notify the user ("Pure
586 channel mapping detected"), and use an optimized and lossless method to do the
589 For example, if you have a 5.1 source and want a stereo audio stream by
590 dropping the extra channels:
592 pan="stereo: c0=FL : c1=FR"
595 Given the same source, you can also switch front left and front right channels
596 and keep the input channel layout:
598 pan="5.1: c0=c1 : c1=c0 : c2=c2 : c3=c3 : c4=c4 : c5=c5"
601 If the input is a stereo audio stream, you can mute the front left channel (and
602 still keep the stereo channel layout) with:
607 Still with a stereo audio stream input, you can copy the right channel in both
608 front left and right:
610 pan="stereo: c0=FR : c1=FR"
613 @section silencedetect
615 Detect silence in an audio stream.
617 This filter logs a message when it detects that the input audio volume is less
618 or equal to a noise tolerance value for a duration greater or equal to the
619 minimum detected noise duration.
621 The printed times and duration are expressed in seconds.
625 Set silence duration until notification (default is 2 seconds).
628 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
629 specified value) or amplitude ratio. Default is -60dB, or 0.001.
632 Detect 5 seconds of silence with -50dB noise tolerance:
634 silencedetect=n=-50dB:d=5
637 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
638 tolerance in @file{silence.mp3}:
640 ffmpeg -f lavfi -i amovie=silence.mp3,silencedetect=noise=0.0001 -f null -
645 Adjust the input audio volume.
647 The filter accepts exactly one parameter @var{vol}, which expresses
648 how the audio volume will be increased or decreased.
650 Output values are clipped to the maximum value.
652 If @var{vol} is expressed as a decimal number, the output audio
653 volume is given by the relation:
655 @var{output_volume} = @var{vol} * @var{input_volume}
658 If @var{vol} is expressed as a decimal number followed by the string
659 "dB", the value represents the requested change in decibels of the
660 input audio power, and the output audio volume is given by the
663 @var{output_volume} = 10^(@var{vol}/20) * @var{input_volume}
666 Otherwise @var{vol} is considered an expression and its evaluated
667 value is used for computing the output audio volume according to the
670 Default value for @var{vol} is 1.0.
676 Half the input audio volume:
681 The above example is equivalent to:
687 Decrease input audio power by 12 decibels:
693 @section volumedetect
695 Detect the volume of the input video.
697 The filter has no parameters. The input is not modified. Statistics about
698 the volume will be printed in the log when the input stream end is reached.
700 In particular it will show the mean volume (root mean square), maximum
701 volume (on a per-sample basis), and the beginning of an histogram of the
702 registered volume values (from the maximum value to a cumulated 1/1000 of
705 All volumes are in decibels relative to the maximum PCM value.
707 Here is an excerpt of the output:
709 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
710 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
711 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
712 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
713 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
714 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
715 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
716 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
717 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
723 The mean square energy is approximately -27 dB, or 10^-2.7.
725 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
727 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
730 In other words, raising the volume by +4 dB does not cause any clipping,
731 raising it by +5 dB causes clipping for 6 samples, etc.
734 Synchronize audio data with timestamps by squeezing/stretching it and/or
735 dropping samples/adding silence when needed.
737 The filter accepts the following named parameters:
741 Enable stretching/squeezing the data to make it match the timestamps. Disabled
742 by default. When disabled, time gaps are covered with silence.
745 Minimum difference between timestamps and audio data (in seconds) to trigger
746 adding/dropping samples. Default value is 0.1. If you get non-perfect sync with
747 this filter, try setting this parameter to 0.
750 Maximum compensation in samples per second. Relevant only with compensate=1.
754 Assume the first pts should be this value.
755 This allows for padding/trimming at the start of stream. By default, no
756 assumption is made about the first frame's expected pts, so no padding or
757 trimming is done. For example, this could be set to 0 to pad the beginning with
758 silence if an audio stream starts after the video stream.
762 @section channelsplit
763 Split each channel in input audio stream into a separate output stream.
765 This filter accepts the following named parameters:
768 Channel layout of the input stream. Default is "stereo".
771 For example, assuming a stereo input MP3 file
773 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
775 will create an output Matroska file with two audio streams, one containing only
776 the left channel and the other the right channel.
778 To split a 5.1 WAV file into per-channel files
780 ffmpeg -i in.wav -filter_complex
781 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
782 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
783 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
788 Remap input channels to new locations.
790 This filter accepts the following named parameters:
793 Channel layout of the output stream.
796 Map channels from input to output. The argument is a comma-separated list of
797 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
798 @var{in_channel} form. @var{in_channel} can be either the name of the input
799 channel (e.g. FL for front left) or its index in the input channel layout.
800 @var{out_channel} is the name of the output channel or its index in the output
801 channel layout. If @var{out_channel} is not given then it is implicitly an
802 index, starting with zero and increasing by one for each mapping.
805 If no mapping is present, the filter will implicitly map input channels to
806 output channels preserving index.
808 For example, assuming a 5.1+downmix input MOV file
810 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL\,DR-FR' out.wav
812 will create an output WAV file tagged as stereo from the downmix channels of
815 To fix a 5.1 WAV improperly encoded in AAC's native channel order
817 ffmpeg -i in.wav -filter 'channelmap=1\,2\,0\,5\,3\,4:channel_layout=5.1' out.wav
821 Join multiple input streams into one multi-channel stream.
823 The filter accepts the following named parameters:
827 Number of input streams. Defaults to 2.
830 Desired output channel layout. Defaults to stereo.
833 Map channels from inputs to output. The argument is a comma-separated list of
834 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
835 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
836 can be either the name of the input channel (e.g. FL for front left) or its
837 index in the specified input stream. @var{out_channel} is the name of the output
841 The filter will attempt to guess the mappings when those are not specified
842 explicitly. It does so by first trying to find an unused matching input channel
843 and if that fails it picks the first unused input channel.
845 E.g. to join 3 inputs (with properly set channel layouts)
847 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
850 To build a 5.1 output from 6 single-channel streams:
852 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
853 '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'
858 Convert the audio sample format, sample rate and channel layout. This filter is
859 not meant to be used directly.
861 @c man end AUDIO FILTERS
863 @chapter Audio Sources
864 @c man begin AUDIO SOURCES
866 Below is a description of the currently available audio sources.
870 Buffer audio frames, and make them available to the filter chain.
872 This source is mainly intended for a programmatic use, in particular
873 through the interface defined in @file{libavfilter/asrc_abuffer.h}.
875 It accepts the following mandatory parameters:
876 @var{sample_rate}:@var{sample_fmt}:@var{channel_layout}
881 The sample rate of the incoming audio buffers.
884 The sample format of the incoming audio buffers.
885 Either a sample format name or its corresponging integer representation from
886 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
889 The channel layout of the incoming audio buffers.
890 Either a channel layout name from channel_layout_map in
891 @file{libavutil/audioconvert.c} or its corresponding integer representation
892 from the AV_CH_LAYOUT_* macros in @file{libavutil/audioconvert.h}
898 abuffer=44100:s16p:stereo
901 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
902 Since the sample format with name "s16p" corresponds to the number
903 6 and the "stereo" channel layout corresponds to the value 0x3, this is
911 Generate an audio signal specified by an expression.
913 This source accepts in input one or more expressions (one for each
914 channel), which are evaluated and used to generate a corresponding
917 It accepts the syntax: @var{exprs}[::@var{options}].
918 @var{exprs} is a list of expressions separated by ":", one for each
919 separate channel. In case the @var{channel_layout} is not
920 specified, the selected channel layout depends on the number of
921 provided expressions.
923 @var{options} is an optional sequence of @var{key}=@var{value} pairs,
926 The description of the accepted options follows.
930 @item channel_layout, c
931 Set the channel layout. The number of channels in the specified layout
932 must be equal to the number of specified expressions.
935 Set the minimum duration of the sourced audio. See the function
936 @code{av_parse_time()} for the accepted format.
937 Note that the resulting duration may be greater than the specified
938 duration, as the generated audio is always cut at the end of a
941 If not specified, or the expressed duration is negative, the audio is
942 supposed to be generated forever.
945 Set the number of samples per channel per each output frame,
949 Specify the sample rate, default to 44100.
952 Each expression in @var{exprs} can contain the following constants:
956 number of the evaluated sample, starting from 0
959 time of the evaluated sample expressed in seconds, starting from 0
978 Generate a sin signal with frequency of 440 Hz, set sample rate to
981 aevalsrc="sin(440*2*PI*t)::s=8000"
985 Generate a two channels signal, specify the channel layout (Front
986 Center + Back Center) explicitly:
988 aevalsrc="sin(420*2*PI*t):cos(430*2*PI*t)::c=FC|BC"
992 Generate white noise:
994 aevalsrc="-2+random(0)"
998 Generate an amplitude modulated signal:
1000 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
1004 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
1006 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) : 0.1*sin(2*PI*(360+2.5/2)*t)"
1013 Null audio source, return unprocessed audio frames. It is mainly useful
1014 as a template and to be employed in analysis / debugging tools, or as
1015 the source for filters which ignore the input data (for example the sox
1018 It accepts an optional sequence of @var{key}=@var{value} pairs,
1021 The description of the accepted options follows.
1025 @item sample_rate, s
1026 Specify the sample rate, and defaults to 44100.
1028 @item channel_layout, cl
1030 Specify the channel layout, and can be either an integer or a string
1031 representing a channel layout. The default value of @var{channel_layout}
1034 Check the channel_layout_map definition in
1035 @file{libavcodec/audioconvert.c} for the mapping between strings and
1036 channel layout values.
1039 Set the number of samples per requested frames.
1043 Follow some examples:
1045 # set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
1046 anullsrc=r=48000:cl=4
1049 anullsrc=r=48000:cl=mono
1053 Buffer audio frames, and make them available to the filter chain.
1055 This source is not intended to be part of user-supplied graph descriptions but
1056 for insertion by calling programs through the interface defined in
1057 @file{libavfilter/buffersrc.h}.
1059 It accepts the following named parameters:
1063 Timebase which will be used for timestamps of submitted frames. It must be
1064 either a floating-point number or in @var{numerator}/@var{denominator} form.
1070 Name of the sample format, as returned by @code{av_get_sample_fmt_name()}.
1072 @item channel_layout
1073 Channel layout of the audio data, in the form that can be accepted by
1074 @code{av_get_channel_layout()}.
1077 All the parameters need to be explicitly defined.
1081 Synthesize a voice utterance using the libflite library.
1083 To enable compilation of this filter you need to configure FFmpeg with
1084 @code{--enable-libflite}.
1086 Note that the flite library is not thread-safe.
1088 The source accepts parameters as a list of @var{key}=@var{value} pairs,
1091 The description of the accepted parameters follows.
1096 If set to 1, list the names of the available voices and exit
1097 immediately. Default value is 0.
1100 Set the maximum number of samples per frame. Default value is 512.
1103 Set the filename containing the text to speak.
1106 Set the text to speak.
1109 Set the voice to use for the speech synthesis. Default value is
1110 @code{kal}. See also the @var{list_voices} option.
1113 @subsection Examples
1117 Read from file @file{speech.txt}, and synthetize the text using the
1118 standard flite voice:
1120 flite=textfile=speech.txt
1124 Read the specified text selecting the @code{slt} voice:
1126 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
1130 Make @file{ffplay} speech the specified text, using @code{flite} and
1131 the @code{lavfi} device:
1133 ffplay -f lavfi flite='No more be grieved for which that thou hast done.'
1137 For more information about libflite, check:
1138 @url{http://www.speech.cs.cmu.edu/flite/}
1140 @c man end AUDIO SOURCES
1142 @chapter Audio Sinks
1143 @c man begin AUDIO SINKS
1145 Below is a description of the currently available audio sinks.
1147 @section abuffersink
1149 Buffer audio frames, and make them available to the end of filter chain.
1151 This sink is mainly intended for programmatic use, in particular
1152 through the interface defined in @file{libavfilter/buffersink.h}.
1154 It requires a pointer to an AVABufferSinkContext structure, which
1155 defines the incoming buffers' formats, to be passed as the opaque
1156 parameter to @code{avfilter_init_filter} for initialization.
1160 Null audio sink, do absolutely nothing with the input audio. It is
1161 mainly useful as a template and to be employed in analysis / debugging
1164 @section abuffersink
1165 This sink is intended for programmatic use. Frames that arrive on this sink can
1166 be retrieved by the calling program using the interface defined in
1167 @file{libavfilter/buffersink.h}.
1169 This filter accepts no parameters.
1171 @c man end AUDIO SINKS
1173 @chapter Video Filters
1174 @c man begin VIDEO FILTERS
1176 When you configure your FFmpeg build, you can disable any of the
1177 existing filters using @code{--disable-filters}.
1178 The configure output will show the video filters included in your
1181 Below is a description of the currently available video filters.
1183 @section alphaextract
1185 Extract the alpha component from the input as a grayscale video. This
1186 is especially useful with the @var{alphamerge} filter.
1190 Add or replace the alpha component of the primary input with the
1191 grayscale value of a second input. This is intended for use with
1192 @var{alphaextract} to allow the transmission or storage of frame
1193 sequences that have alpha in a format that doesn't support an alpha
1196 For example, to reconstruct full frames from a normal YUV-encoded video
1197 and a separate video created with @var{alphaextract}, you might use:
1199 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
1202 Since this filter is designed for reconstruction, it operates on frame
1203 sequences without considering timestamps, and terminates when either
1204 input reaches end of stream. This will cause problems if your encoding
1205 pipeline drops frames. If you're trying to apply an image as an
1206 overlay to a video stream, consider the @var{overlay} filter instead.
1210 Draw ASS (Advanced Substation Alpha) subtitles on top of input video
1211 using the libass library.
1213 To enable compilation of this filter you need to configure FFmpeg with
1214 @code{--enable-libass}.
1216 This filter accepts the syntax: @var{ass_filename}[:@var{options}],
1217 where @var{ass_filename} is the filename of the ASS file to read, and
1218 @var{options} is an optional sequence of @var{key}=@var{value} pairs,
1221 A description of the accepted options follows.
1225 Specifies the size of the original video, the video for which the ASS file
1226 was composed. Due to a misdesign in ASS aspect ratio arithmetic, this is
1227 necessary to correctly scale the fonts if the aspect ratio has been changed.
1230 For example, to render the file @file{sub.ass} on top of the input
1231 video, use the command:
1238 Compute the bounding box for the non-black pixels in the input frame
1241 This filter computes the bounding box containing all the pixels with a
1242 luminance value greater than the minimum allowed value.
1243 The parameters describing the bounding box are printed on the filter
1246 @section blackdetect
1248 Detect video intervals that are (almost) completely black. Can be
1249 useful to detect chapter transitions, commercials, or invalid
1250 recordings. Output lines contains the time for the start, end and
1251 duration of the detected black interval expressed in seconds.
1253 In order to display the output lines, you need to set the loglevel at
1254 least to the AV_LOG_INFO value.
1256 This filter accepts a list of options in the form of
1257 @var{key}=@var{value} pairs separated by ":". A description of the
1258 accepted options follows.
1261 @item black_min_duration, d
1262 Set the minimum detected black duration expressed in seconds. It must
1263 be a non-negative floating point number.
1265 Default value is 2.0.
1267 @item picture_black_ratio_th, pic_th
1268 Set the threshold for considering a picture "black".
1269 Express the minimum value for the ratio:
1271 @var{nb_black_pixels} / @var{nb_pixels}
1274 for which a picture is considered black.
1275 Default value is 0.98.
1277 @item pixel_black_th, pix_th
1278 Set the threshold for considering a pixel "black".
1280 The threshold expresses the maximum pixel luminance value for which a
1281 pixel is considered "black". The provided value is scaled according to
1282 the following equation:
1284 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
1287 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
1288 the input video format, the range is [0-255] for YUV full-range
1289 formats and [16-235] for YUV non full-range formats.
1291 Default value is 0.10.
1294 The following example sets the maximum pixel threshold to the minimum
1295 value, and detects only black intervals of 2 or more seconds:
1297 blackdetect=d=2:pix_th=0.00
1302 Detect frames that are (almost) completely black. Can be useful to
1303 detect chapter transitions or commercials. Output lines consist of
1304 the frame number of the detected frame, the percentage of blackness,
1305 the position in the file if known or -1 and the timestamp in seconds.
1307 In order to display the output lines, you need to set the loglevel at
1308 least to the AV_LOG_INFO value.
1310 The filter accepts the syntax:
1312 blackframe[=@var{amount}:[@var{threshold}]]
1315 @var{amount} is the percentage of the pixels that have to be below the
1316 threshold, and defaults to 98.
1318 @var{threshold} is the threshold below which a pixel value is
1319 considered black, and defaults to 32.
1323 Apply boxblur algorithm to the input video.
1325 This filter accepts the parameters:
1326 @var{luma_radius}:@var{luma_power}:@var{chroma_radius}:@var{chroma_power}:@var{alpha_radius}:@var{alpha_power}
1328 Chroma and alpha parameters are optional, if not specified they default
1329 to the corresponding values set for @var{luma_radius} and
1332 @var{luma_radius}, @var{chroma_radius}, and @var{alpha_radius} represent
1333 the radius in pixels of the box used for blurring the corresponding
1334 input plane. They are expressions, and can contain the following
1338 the input width and height in pixels
1341 the input chroma image width and height in pixels
1344 horizontal and vertical chroma subsample values. For example for the
1345 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
1348 The radius must be a non-negative number, and must not be greater than
1349 the value of the expression @code{min(w,h)/2} for the luma and alpha planes,
1350 and of @code{min(cw,ch)/2} for the chroma planes.
1352 @var{luma_power}, @var{chroma_power}, and @var{alpha_power} represent
1353 how many times the boxblur filter is applied to the corresponding
1356 Some examples follow:
1361 Apply a boxblur filter with luma, chroma, and alpha radius
1368 Set luma radius to 2, alpha and chroma radius to 0
1374 Set luma and chroma radius to a fraction of the video dimension
1376 boxblur=min(h\,w)/10:1:min(cw\,ch)/10:1
1381 @section colormatrix
1383 The colormatrix filter allows conversion between any of the following color
1384 space: BT.709 (@var{bt709}), BT.601 (@var{bt601}), SMPTE-240M (@var{smpte240m})
1385 and FCC (@var{fcc}).
1387 The syntax of the parameters is @var{source}:@var{destination}:
1390 colormatrix=bt601:smpte240m
1395 Copy the input source unchanged to the output. Mainly useful for
1400 Crop the input video to @var{out_w}:@var{out_h}:@var{x}:@var{y}:@var{keep_aspect}
1402 The @var{keep_aspect} parameter is optional, if specified and set to a
1403 non-zero value will force the output display aspect ratio to be the
1404 same of the input, by changing the output sample aspect ratio.
1406 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
1407 expressions containing the following constants:
1411 the computed values for @var{x} and @var{y}. They are evaluated for
1415 the input width and height
1418 same as @var{in_w} and @var{in_h}
1421 the output (cropped) width and height
1424 same as @var{out_w} and @var{out_h}
1427 same as @var{iw} / @var{ih}
1430 input sample aspect ratio
1433 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
1436 horizontal and vertical chroma subsample values. For example for the
1437 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
1440 the number of input frame, starting from 0
1443 the position in the file of the input frame, NAN if unknown
1446 timestamp expressed in seconds, NAN if the input timestamp is unknown
1450 The @var{out_w} and @var{out_h} parameters specify the expressions for
1451 the width and height of the output (cropped) video. They are
1452 evaluated just at the configuration of the filter.
1454 The default value of @var{out_w} is "in_w", and the default value of
1455 @var{out_h} is "in_h".
1457 The expression for @var{out_w} may depend on the value of @var{out_h},
1458 and the expression for @var{out_h} may depend on @var{out_w}, but they
1459 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
1460 evaluated after @var{out_w} and @var{out_h}.
1462 The @var{x} and @var{y} parameters specify the expressions for the
1463 position of the top-left corner of the output (non-cropped) area. They
1464 are evaluated for each frame. If the evaluated value is not valid, it
1465 is approximated to the nearest valid value.
1467 The default value of @var{x} is "(in_w-out_w)/2", and the default
1468 value for @var{y} is "(in_h-out_h)/2", which set the cropped area at
1469 the center of the input image.
1471 The expression for @var{x} may depend on @var{y}, and the expression
1472 for @var{y} may depend on @var{x}.
1474 Follow some examples:
1476 # crop the central input area with size 100x100
1479 # crop the central input area with size 2/3 of the input video
1480 "crop=2/3*in_w:2/3*in_h"
1482 # crop the input video central square
1485 # delimit the rectangle with the top-left corner placed at position
1486 # 100:100 and the right-bottom corner corresponding to the right-bottom
1487 # corner of the input image.
1488 crop=in_w-100:in_h-100:100:100
1490 # crop 10 pixels from the left and right borders, and 20 pixels from
1491 # the top and bottom borders
1492 "crop=in_w-2*10:in_h-2*20"
1494 # keep only the bottom right quarter of the input image
1495 "crop=in_w/2:in_h/2:in_w/2:in_h/2"
1497 # crop height for getting Greek harmony
1498 "crop=in_w:1/PHI*in_w"
1501 "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)"
1503 # erratic camera effect depending on timestamp
1504 "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)"
1506 # set x depending on the value of y
1507 "crop=in_w/2:in_h/2:y:10+10*sin(n/10)"
1512 Auto-detect crop size.
1514 Calculate necessary cropping parameters and prints the recommended
1515 parameters through the logging system. The detected dimensions
1516 correspond to the non-black area of the input video.
1518 It accepts the syntax:
1520 cropdetect[=@var{limit}[:@var{round}[:@var{reset}]]]
1526 Threshold, which can be optionally specified from nothing (0) to
1527 everything (255), defaults to 24.
1530 Value which the width/height should be divisible by, defaults to
1531 16. The offset is automatically adjusted to center the video. Use 2 to
1532 get only even dimensions (needed for 4:2:2 video). 16 is best when
1533 encoding to most video codecs.
1536 Counter that determines after how many frames cropdetect will reset
1537 the previously detected largest video area and start over to detect
1538 the current optimal crop area. Defaults to 0.
1540 This can be useful when channel logos distort the video area. 0
1541 indicates never reset and return the largest area encountered during
1547 This filter drops frames that do not differ greatly from the previous
1548 frame in order to reduce framerate. The main use of this filter is
1549 for very-low-bitrate encoding (e.g. streaming over dialup modem), but
1550 it could in theory be used for fixing movies that were
1551 inverse-telecined incorrectly.
1553 It accepts the following parameters:
1554 @var{max}:@var{hi}:@var{lo}:@var{frac}.
1559 Set the maximum number of consecutive frames which can be dropped (if
1560 positive), or the minimum interval between dropped frames (if
1561 negative). If the value is 0, the frame is dropped unregarding the
1562 number of previous sequentially dropped frames.
1567 Set the dropping threshold values.
1569 Values for @var{hi} and @var{lo} are for 8x8 pixel blocks and
1570 represent actual pixel value differences, so a threshold of 64
1571 corresponds to 1 unit of difference for each pixel, or the same spread
1572 out differently over the block.
1574 A frame is a candidate for dropping if no 8x8 blocks differ by more
1575 than a threshold of @var{hi}, and if no more than @var{frac} blocks (1
1576 meaning the whole image) differ by more than a threshold of @var{lo}.
1578 Default value for @var{hi} is 64*12, default value for @var{lo} is
1579 64*5, and default value for @var{frac} is 0.33.
1584 Suppress a TV station logo by a simple interpolation of the surrounding
1585 pixels. Just set a rectangle covering the logo and watch it disappear
1586 (and sometimes something even uglier appear - your mileage may vary).
1588 The filter accepts parameters as a string of the form
1589 "@var{x}:@var{y}:@var{w}:@var{h}:@var{band}", or as a list of
1590 @var{key}=@var{value} pairs, separated by ":".
1592 The description of the accepted parameters follows.
1597 Specify the top left corner coordinates of the logo. They must be
1601 Specify the width and height of the logo to clear. They must be
1605 Specify the thickness of the fuzzy edge of the rectangle (added to
1606 @var{w} and @var{h}). The default value is 4.
1609 When set to 1, a green rectangle is drawn on the screen to simplify
1610 finding the right @var{x}, @var{y}, @var{w}, @var{h} parameters, and
1611 @var{band} is set to 4. The default value is 0.
1615 Some examples follow.
1620 Set a rectangle covering the area with top left corner coordinates 0,0
1621 and size 100x77, setting a band of size 10:
1623 delogo=0:0:100:77:10
1627 As the previous example, but use named options:
1629 delogo=x=0:y=0:w=100:h=77:band=10
1636 Attempt to fix small changes in horizontal and/or vertical shift. This
1637 filter helps remove camera shake from hand-holding a camera, bumping a
1638 tripod, moving on a vehicle, etc.
1640 The filter accepts parameters as a string of the form
1641 "@var{x}:@var{y}:@var{w}:@var{h}:@var{rx}:@var{ry}:@var{edge}:@var{blocksize}:@var{contrast}:@var{search}:@var{filename}"
1643 A description of the accepted parameters follows.
1648 Specify a rectangular area where to limit the search for motion
1650 If desired the search for motion vectors can be limited to a
1651 rectangular area of the frame defined by its top left corner, width
1652 and height. These parameters have the same meaning as the drawbox
1653 filter which can be used to visualise the position of the bounding
1656 This is useful when simultaneous movement of subjects within the frame
1657 might be confused for camera motion by the motion vector search.
1659 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
1660 then the full frame is used. This allows later options to be set
1661 without specifying the bounding box for the motion vector search.
1663 Default - search the whole frame.
1666 Specify the maximum extent of movement in x and y directions in the
1667 range 0-64 pixels. Default 16.
1670 Specify how to generate pixels to fill blanks at the edge of the
1671 frame. An integer from 0 to 3 as follows:
1674 Fill zeroes at blank locations
1676 Original image at blank locations
1678 Extruded edge value at blank locations
1680 Mirrored edge at blank locations
1683 The default setting is mirror edge at blank locations.
1686 Specify the blocksize to use for motion search. Range 4-128 pixels,
1690 Specify the contrast threshold for blocks. Only blocks with more than
1691 the specified contrast (difference between darkest and lightest
1692 pixels) will be considered. Range 1-255, default 125.
1695 Specify the search strategy 0 = exhaustive search, 1 = less exhaustive
1696 search. Default - exhaustive search.
1699 If set then a detailed log of the motion search is written to the
1706 Draw a colored box on the input image.
1708 It accepts the syntax:
1710 drawbox=@var{x}:@var{y}:@var{width}:@var{height}:@var{color}
1716 Specify the top left corner coordinates of the box. Default to 0.
1719 Specify the width and height of the box, if 0 they are interpreted as
1720 the input width and height. Default to 0.
1723 Specify the color of the box to write, it can be the name of a color
1724 (case insensitive match) or a 0xRRGGBB[AA] sequence.
1727 Follow some examples:
1729 # draw a black box around the edge of the input image
1732 # draw a box with color red and an opacity of 50%
1733 drawbox=10:20:200:60:red@@0.5"
1738 Draw text string or text from specified file on top of video using the
1739 libfreetype library.
1741 To enable compilation of this filter you need to configure FFmpeg with
1742 @code{--enable-libfreetype}.
1744 The filter also recognizes strftime() sequences in the provided text
1745 and expands them accordingly. Check the documentation of strftime().
1747 The filter accepts parameters as a list of @var{key}=@var{value} pairs,
1750 The description of the accepted parameters follows.
1755 Used to draw a box around text using background color.
1756 Value should be either 1 (enable) or 0 (disable).
1757 The default value of @var{box} is 0.
1760 The color to be used for drawing box around text.
1761 Either a string (e.g. "yellow") or in 0xRRGGBB[AA] format
1762 (e.g. "0xff00ff"), possibly followed by an alpha specifier.
1763 The default value of @var{boxcolor} is "white".
1766 Set an expression which specifies if the text should be drawn. If the
1767 expression evaluates to 0, the text is not drawn. This is useful for
1768 specifying that the text should be drawn only when specific conditions
1771 Default value is "1".
1773 See below for the list of accepted constants and functions.
1776 If true, check and fix text coords to avoid clipping.
1779 The color to be used for drawing fonts.
1780 Either a string (e.g. "red") or in 0xRRGGBB[AA] format
1781 (e.g. "0xff000033"), possibly followed by an alpha specifier.
1782 The default value of @var{fontcolor} is "black".
1785 The font file to be used for drawing text. Path must be included.
1786 This parameter is mandatory.
1789 The font size to be used for drawing text.
1790 The default value of @var{fontsize} is 16.
1793 Flags to be used for loading the fonts.
1795 The flags map the corresponding flags supported by libfreetype, and are
1796 a combination of the following values:
1803 @item vertical_layout
1804 @item force_autohint
1807 @item ignore_global_advance_width
1809 @item ignore_transform
1816 Default value is "render".
1818 For more information consult the documentation for the FT_LOAD_*
1822 The color to be used for drawing a shadow behind the drawn text. It
1823 can be a color name (e.g. "yellow") or a string in the 0xRRGGBB[AA]
1824 form (e.g. "0xff00ff"), possibly followed by an alpha specifier.
1825 The default value of @var{shadowcolor} is "black".
1827 @item shadowx, shadowy
1828 The x and y offsets for the text shadow position with respect to the
1829 position of the text. They can be either positive or negative
1830 values. Default value for both is "0".
1833 The size in number of spaces to use for rendering the tab.
1837 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
1838 format. It can be used with or without text parameter. @var{timecode_rate}
1839 option must be specified.
1841 @item timecode_rate, rate, r
1842 Set the timecode frame rate (timecode only).
1845 The text string to be drawn. The text must be a sequence of UTF-8
1847 This parameter is mandatory if no file is specified with the parameter
1851 A text file containing text to be drawn. The text must be a sequence
1852 of UTF-8 encoded characters.
1854 This parameter is mandatory if no text string is specified with the
1855 parameter @var{text}.
1857 If both @var{text} and @var{textfile} are specified, an error is thrown.
1860 The expressions which specify the offsets where text will be drawn
1861 within the video frame. They are relative to the top/left border of the
1864 The default value of @var{x} and @var{y} is "0".
1866 See below for the list of accepted constants and functions.
1869 The parameters for @var{x} and @var{y} are expressions containing the
1870 following constants and functions:
1874 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
1877 horizontal and vertical chroma subsample values. For example for the
1878 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
1881 the height of each text line
1889 @item max_glyph_a, ascent
1890 the maximum distance from the baseline to the highest/upper grid
1891 coordinate used to place a glyph outline point, for all the rendered
1893 It is a positive value, due to the grid's orientation with the Y axis
1896 @item max_glyph_d, descent
1897 the maximum distance from the baseline to the lowest grid coordinate
1898 used to place a glyph outline point, for all the rendered glyphs.
1899 This is a negative value, due to the grid's orientation, with the Y axis
1903 maximum glyph height, that is the maximum height for all the glyphs
1904 contained in the rendered text, it is equivalent to @var{ascent} -
1908 maximum glyph width, that is the maximum width for all the glyphs
1909 contained in the rendered text
1912 the number of input frame, starting from 0
1914 @item rand(min, max)
1915 return a random number included between @var{min} and @var{max}
1918 input sample aspect ratio
1921 timestamp expressed in seconds, NAN if the input timestamp is unknown
1924 the height of the rendered text
1927 the width of the rendered text
1930 the x and y offset coordinates where the text is drawn.
1932 These parameters allow the @var{x} and @var{y} expressions to refer
1933 each other, so you can for example specify @code{y=x/dar}.
1936 If libavfilter was built with @code{--enable-fontconfig}, then
1937 @option{fontfile} can be a fontconfig pattern or omitted.
1939 Some examples follow.
1944 Draw "Test Text" with font FreeSerif, using the default values for the
1945 optional parameters.
1948 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
1952 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
1953 and y=50 (counting from the top-left corner of the screen), text is
1954 yellow with a red box around it. Both the text and the box have an
1958 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
1959 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
1962 Note that the double quotes are not necessary if spaces are not used
1963 within the parameter list.
1966 Show the text at the center of the video frame:
1968 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h-line_h)/2"
1972 Show a text line sliding from right to left in the last row of the video
1973 frame. The file @file{LONG_LINE} is assumed to contain a single line
1976 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
1980 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
1982 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
1986 Draw a single green letter "g", at the center of the input video.
1987 The glyph baseline is placed at half screen height.
1989 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
1993 Show text for 1 second every 3 seconds:
1995 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:draw=lt(mod(t\\,3)\\,1):text='blink'"
1999 Use fontconfig to set the font. Note that the colons need to be escaped.
2001 drawtext='fontfile=Linux Libertine O-40\\:style=Semibold:text=FFmpeg'
2006 For more information about libfreetype, check:
2007 @url{http://www.freetype.org/}.
2009 For more information about fontconfig, check:
2010 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
2014 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
2016 This filter accepts the following optional named parameters:
2020 Set low and high threshold values used by the Canny thresholding
2023 The high threshold selects the "strong" edge pixels, which are then
2024 connected through 8-connectivity with the "weak" edge pixels selected
2025 by the low threshold.
2027 @var{low} and @var{high} threshold values must be choosen in the range
2028 [0,1], and @var{low} should be lesser or equal to @var{high}.
2030 Default value for @var{low} is @code{20/255}, and default value for @var{high}
2036 edgedetect=low=0.1:high=0.4
2041 Apply fade-in/out effect to input video.
2043 It accepts the parameters:
2044 @var{type}:@var{start_frame}:@var{nb_frames}[:@var{options}]
2046 @var{type} specifies if the effect type, can be either "in" for
2047 fade-in, or "out" for a fade-out effect.
2049 @var{start_frame} specifies the number of the start frame for starting
2050 to apply the fade effect.
2052 @var{nb_frames} specifies the number of frames for which the fade
2053 effect has to last. At the end of the fade-in effect the output video
2054 will have the same intensity as the input video, at the end of the
2055 fade-out transition the output video will be completely black.
2057 @var{options} is an optional sequence of @var{key}=@var{value} pairs,
2058 separated by ":". The description of the accepted options follows.
2065 @item start_frame, s
2066 See @var{start_frame}.
2069 See @var{nb_frames}.
2072 If set to 1, fade only alpha channel, if one exists on the input.
2076 A few usage examples follow, usable too as test scenarios.
2078 # fade in first 30 frames of video
2081 # fade out last 45 frames of a 200-frame video
2084 # fade in first 25 frames and fade out last 25 frames of a 1000-frame video
2085 fade=in:0:25, fade=out:975:25
2087 # make first 5 frames black, then fade in from frame 5-24
2090 # fade in alpha over first 25 frames of video
2091 fade=in:0:25:alpha=1
2096 Transform the field order of the input video.
2098 It accepts one parameter which specifies the required field order that
2099 the input interlaced video will be transformed to. The parameter can
2100 assume one of the following values:
2104 output bottom field first
2106 output top field first
2109 Default value is "tff".
2111 Transformation is achieved by shifting the picture content up or down
2112 by one line, and filling the remaining line with appropriate picture content.
2113 This method is consistent with most broadcast field order converters.
2115 If the input video is not flagged as being interlaced, or it is already
2116 flagged as being of the required output field order then this filter does
2117 not alter the incoming video.
2119 This filter is very useful when converting to or from PAL DV material,
2120 which is bottom field first.
2124 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
2129 Buffer input images and send them when they are requested.
2131 This filter is mainly useful when auto-inserted by the libavfilter
2134 The filter does not take parameters.
2138 Convert the input video to one of the specified pixel formats.
2139 Libavfilter will try to pick one that is supported for the input to
2142 The filter accepts a list of pixel format names, separated by ":",
2143 for example "yuv420p:monow:rgb24".
2145 Some examples follow:
2147 # convert the input video to the format "yuv420p"
2150 # convert the input video to any of the formats in the list
2151 format=yuv420p:yuv444p:yuv410p
2156 Convert the video to specified constant framerate by duplicating or dropping
2157 frames as necessary.
2159 This filter accepts the following named parameters:
2163 Desired output framerate.
2169 Select one frame every N.
2171 This filter accepts in input a string representing a positive
2172 integer. Default argument is @code{1}.
2177 Apply a frei0r effect to the input video.
2179 To enable compilation of this filter you need to install the frei0r
2180 header and configure FFmpeg with @code{--enable-frei0r}.
2182 The filter supports the syntax:
2184 @var{filter_name}[@{:|=@}@var{param1}:@var{param2}:...:@var{paramN}]
2187 @var{filter_name} is the name to the frei0r effect to load. If the
2188 environment variable @env{FREI0R_PATH} is defined, the frei0r effect
2189 is searched in each one of the directories specified by the colon
2190 separated list in @env{FREIOR_PATH}, otherwise in the standard frei0r
2191 paths, which are in this order: @file{HOME/.frei0r-1/lib/},
2192 @file{/usr/local/lib/frei0r-1/}, @file{/usr/lib/frei0r-1/}.
2194 @var{param1}, @var{param2}, ... , @var{paramN} specify the parameters
2195 for the frei0r effect.
2197 A frei0r effect parameter can be a boolean (whose values are specified
2198 with "y" and "n"), a double, a color (specified by the syntax
2199 @var{R}/@var{G}/@var{B}, @var{R}, @var{G}, and @var{B} being float
2200 numbers from 0.0 to 1.0) or by an @code{av_parse_color()} color
2201 description), a position (specified by the syntax @var{X}/@var{Y},
2202 @var{X} and @var{Y} being float numbers) and a string.
2204 The number and kind of parameters depend on the loaded effect. If an
2205 effect parameter is not specified the default value is set.
2207 Some examples follow:
2211 Apply the distort0r effect, set the first two double parameters:
2213 frei0r=distort0r:0.5:0.01
2217 Apply the colordistance effect, takes a color as first parameter:
2219 frei0r=colordistance:0.2/0.3/0.4
2220 frei0r=colordistance:violet
2221 frei0r=colordistance:0x112233
2225 Apply the perspective effect, specify the top left and top right image
2228 frei0r=perspective:0.2/0.2:0.8/0.2
2232 For more information see:
2233 @url{http://frei0r.dyne.org}
2237 Fix the banding artifacts that are sometimes introduced into nearly flat
2238 regions by truncation to 8bit color depth.
2239 Interpolate the gradients that should go where the bands are, and
2242 This filter is designed for playback only. Do not use it prior to
2243 lossy compression, because compression tends to lose the dither and
2244 bring back the bands.
2246 The filter takes two optional parameters, separated by ':':
2247 @var{strength}:@var{radius}
2249 @var{strength} is the maximum amount by which the filter will change
2250 any one pixel. Also the threshold for detecting nearly flat
2251 regions. Acceptable values range from .51 to 255, default value is
2252 1.2, out-of-range values will be clipped to the valid range.
2254 @var{radius} is the neighborhood to fit the gradient to. A larger
2255 radius makes for smoother gradients, but also prevents the filter from
2256 modifying the pixels near detailed regions. Acceptable values are
2257 8-32, default value is 16, out-of-range values will be clipped to the
2261 # default parameters
2270 Flip the input video horizontally.
2272 For example to horizontally flip the input video with @command{ffmpeg}:
2274 ffmpeg -i in.avi -vf "hflip" out.avi
2279 High precision/quality 3d denoise filter. This filter aims to reduce
2280 image noise producing smooth images and making still images really
2281 still. It should enhance compressibility.
2283 It accepts the following optional parameters:
2284 @var{luma_spatial}:@var{chroma_spatial}:@var{luma_tmp}:@var{chroma_tmp}
2288 a non-negative float number which specifies spatial luma strength,
2291 @item chroma_spatial
2292 a non-negative float number which specifies spatial chroma strength,
2293 defaults to 3.0*@var{luma_spatial}/4.0
2296 a float number which specifies luma temporal strength, defaults to
2297 6.0*@var{luma_spatial}/4.0
2300 a float number which specifies chroma temporal strength, defaults to
2301 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}
2306 Modify the hue and/or the saturation of the input.
2308 This filter accepts the following optional named options:
2312 Specify the hue angle as a number of degrees. It accepts a float
2313 number or an expression, and defaults to 0.0.
2316 Specify the hue angle as a number of degrees. It accepts a float
2317 number or an expression, and defaults to 0.0.
2320 Specify the saturation in the [-10,10] range. It accepts a float number and
2324 The @var{h}, @var{H} and @var{s} parameters are expressions containing the
2325 following constants:
2329 frame count of the input frame starting from 0
2332 presentation timestamp of the input frame expressed in time base units
2335 frame rate of the input video, NAN if the input frame rate is unknown
2338 timestamp expressed in seconds, NAN if the input timestamp is unknown
2341 time base of the input video
2344 The options can also be set using the syntax: @var{hue}:@var{saturation}
2346 In this case @var{hue} is expressed in degrees.
2348 Some examples follow:
2351 Set the hue to 90 degrees and the saturation to 1.0:
2357 Same command but expressing the hue in radians:
2363 Same command without named options, hue must be expressed in degrees:
2369 Note that "h:s" syntax does not support expressions for the values of
2370 h and s, so the following example will issue an error:
2376 Rotate hue and make the saturation swing between 0
2377 and 2 over a period of 1 second:
2379 hue="H=2*PI*t: s=sin(2*PI*t)+1"
2383 Apply a 3 seconds saturation fade-in effect starting at 0:
2388 The general fade-in expression can be written as:
2390 hue="s=min(0\, max((t-START)/DURATION\, 1))"
2394 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
2396 hue="s=max(0\, min(1\, (8-t)/3))"
2399 The general fade-out expression can be written as:
2401 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
2406 @subsection Commands
2408 This filter supports the following command:
2411 Modify the hue and/or the saturation of the input video.
2412 The command accepts the same named options and syntax than when calling the
2413 filter from the command-line.
2415 If a parameter is omitted, it is kept at its current value.
2420 Interlaceing detect filter. This filter tries to detect if the input is
2421 interlaced or progressive. Top or bottom field first.
2423 @section lut, lutrgb, lutyuv
2425 Compute a look-up table for binding each pixel component input value
2426 to an output value, and apply it to input video.
2428 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
2429 to an RGB input video.
2431 These filters accept in input a ":"-separated list of options, which
2432 specify the expressions used for computing the lookup table for the
2433 corresponding pixel component values.
2435 The @var{lut} filter requires either YUV or RGB pixel formats in
2436 input, and accepts the options:
2439 first pixel component
2441 second pixel component
2443 third pixel component
2445 fourth pixel component, corresponds to the alpha component
2448 The exact component associated to each option depends on the format in
2451 The @var{lutrgb} filter requires RGB pixel formats in input, and
2452 accepts the options:
2464 The @var{lutyuv} filter requires YUV pixel formats in input, and
2465 accepts the options:
2468 Y/luminance component
2477 The expressions can contain the following constants and functions:
2481 the input width and height
2484 input value for the pixel component
2487 the input value clipped in the @var{minval}-@var{maxval} range
2490 maximum value for the pixel component
2493 minimum value for the pixel component
2496 the negated value for the pixel component value clipped in the
2497 @var{minval}-@var{maxval} range , it corresponds to the expression
2498 "maxval-clipval+minval"
2501 the computed value in @var{val} clipped in the
2502 @var{minval}-@var{maxval} range
2504 @item gammaval(gamma)
2505 the computed gamma correction value of the pixel component value
2506 clipped in the @var{minval}-@var{maxval} range, corresponds to the
2508 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
2512 All expressions default to "val".
2514 Some examples follow:
2516 # negate input video
2517 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
2518 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
2520 # the above is the same as
2521 lutrgb="r=negval:g=negval:b=negval"
2522 lutyuv="y=negval:u=negval:v=negval"
2527 # remove chroma components, turns the video into a graytone image
2528 lutyuv="u=128:v=128"
2530 # apply a luma burning effect
2533 # remove green and blue components
2536 # set a constant alpha channel value on input
2537 format=rgba,lutrgb=a="maxval-minval/2"
2539 # correct luminance gamma by a 0.5 factor
2540 lutyuv=y=gammaval(0.5)
2545 Apply an MPlayer filter to the input video.
2547 This filter provides a wrapper around most of the filters of
2550 This wrapper is considered experimental. Some of the wrapped filters
2551 may not work properly and we may drop support for them, as they will
2552 be implemented natively into FFmpeg. Thus you should avoid
2553 depending on them when writing portable scripts.
2555 The filters accepts the parameters:
2556 @var{filter_name}[:=]@var{filter_params}
2558 @var{filter_name} is the name of a supported MPlayer filter,
2559 @var{filter_params} is a string containing the parameters accepted by
2562 The list of the currently supported filters follows:
2606 The parameter syntax and behavior for the listed filters are the same
2607 of the corresponding MPlayer filters. For detailed instructions check
2608 the "VIDEO FILTERS" section in the MPlayer manual.
2610 Some examples follow:
2613 Adjust gamma, brightness, contrast:
2619 Add temporal noise to input video:
2625 See also mplayer(1), @url{http://www.mplayerhq.hu/}.
2631 This filter accepts an integer in input, if non-zero it negates the
2632 alpha component (if available). The default value in input is 0.
2636 Force libavfilter not to use any of the specified pixel formats for the
2637 input to the next filter.
2639 The filter accepts a list of pixel format names, separated by ":",
2640 for example "yuv420p:monow:rgb24".
2642 Some examples follow:
2644 # force libavfilter to use a format different from "yuv420p" for the
2645 # input to the vflip filter
2646 noformat=yuv420p,vflip
2648 # convert the input video to any of the formats not contained in the list
2649 noformat=yuv420p:yuv444p:yuv410p
2654 Pass the video source unchanged to the output.
2658 Apply video transform using libopencv.
2660 To enable this filter install libopencv library and headers and
2661 configure FFmpeg with @code{--enable-libopencv}.
2663 The filter takes the parameters: @var{filter_name}@{:=@}@var{filter_params}.
2665 @var{filter_name} is the name of the libopencv filter to apply.
2667 @var{filter_params} specifies the parameters to pass to the libopencv
2668 filter. If not specified the default values are assumed.
2670 Refer to the official libopencv documentation for more precise
2672 @url{http://opencv.willowgarage.com/documentation/c/image_filtering.html}
2674 Follows the list of supported libopencv filters.
2679 Dilate an image by using a specific structuring element.
2680 This filter corresponds to the libopencv function @code{cvDilate}.
2682 It accepts the parameters: @var{struct_el}:@var{nb_iterations}.
2684 @var{struct_el} represents a structuring element, and has the syntax:
2685 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
2687 @var{cols} and @var{rows} represent the number of columns and rows of
2688 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
2689 point, and @var{shape} the shape for the structuring element, and
2690 can be one of the values "rect", "cross", "ellipse", "custom".
2692 If the value for @var{shape} is "custom", it must be followed by a
2693 string of the form "=@var{filename}". The file with name
2694 @var{filename} is assumed to represent a binary image, with each
2695 printable character corresponding to a bright pixel. When a custom
2696 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
2697 or columns and rows of the read file are assumed instead.
2699 The default value for @var{struct_el} is "3x3+0x0/rect".
2701 @var{nb_iterations} specifies the number of times the transform is
2702 applied to the image, and defaults to 1.
2704 Follow some example:
2706 # use the default values
2709 # dilate using a structuring element with a 5x5 cross, iterate two times
2710 ocv=dilate=5x5+2x2/cross:2
2712 # read the shape from the file diamond.shape, iterate two times
2713 # the file diamond.shape may contain a pattern of characters like this:
2719 # the specified cols and rows are ignored (but not the anchor point coordinates)
2720 ocv=0x0+2x2/custom=diamond.shape:2
2725 Erode an image by using a specific structuring element.
2726 This filter corresponds to the libopencv function @code{cvErode}.
2728 The filter accepts the parameters: @var{struct_el}:@var{nb_iterations},
2729 with the same syntax and semantics as the @ref{dilate} filter.
2733 Smooth the input video.
2735 The filter takes the following parameters:
2736 @var{type}:@var{param1}:@var{param2}:@var{param3}:@var{param4}.
2738 @var{type} is the type of smooth filter to apply, and can be one of
2739 the following values: "blur", "blur_no_scale", "median", "gaussian",
2740 "bilateral". The default value is "gaussian".
2742 @var{param1}, @var{param2}, @var{param3}, and @var{param4} are
2743 parameters whose meanings depend on smooth type. @var{param1} and
2744 @var{param2} accept integer positive values or 0, @var{param3} and
2745 @var{param4} accept float values.
2747 The default value for @var{param1} is 3, the default value for the
2748 other parameters is 0.
2750 These parameters correspond to the parameters assigned to the
2751 libopencv function @code{cvSmooth}.
2756 Overlay one video on top of another.
2758 It takes two inputs and one output, the first input is the "main"
2759 video on which the second input is overlayed.
2761 It accepts the parameters: @var{x}:@var{y}[:@var{options}].
2763 @var{x} is the x coordinate of the overlayed video on the main video,
2764 @var{y} is the y coordinate. @var{x} and @var{y} are expressions containing
2765 the following parameters:
2768 @item main_w, main_h
2769 main input width and height
2772 same as @var{main_w} and @var{main_h}
2774 @item overlay_w, overlay_h
2775 overlay input width and height
2778 same as @var{overlay_w} and @var{overlay_h}
2781 @var{options} is an optional list of @var{key}=@var{value} pairs,
2784 The description of the accepted options follows.
2788 If set to 1, force the filter to accept inputs in the RGB
2789 color space. Default value is 0.
2792 Be aware that frames are taken from each input video in timestamp
2793 order, hence, if their initial timestamps differ, it is a a good idea
2794 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
2795 have them begin in the same zero timestamp, as it does the example for
2796 the @var{movie} filter.
2798 Follow some examples:
2800 # draw the overlay at 10 pixels from the bottom right
2801 # corner of the main video.
2802 overlay=main_w-overlay_w-10:main_h-overlay_h-10
2804 # insert a transparent PNG logo in the bottom left corner of the input
2805 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
2807 # insert 2 different transparent PNG logos (second logo on bottom
2809 ffmpeg -i input -i logo1 -i logo2 -filter_complex
2810 'overlay=10:H-h-10,overlay=W-w-10:H-h-10' output
2812 # add a transparent color layer on top of the main video,
2813 # WxH specifies the size of the main input to the overlay filter
2814 color=red@.3:WxH [over]; [in][over] overlay [out]
2816 # play an original video and a filtered version (here with the deshake filter)
2818 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
2820 # the previous example is the same as:
2821 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
2824 You can chain together more overlays but the efficiency of such
2825 approach is yet to be tested.
2829 Add paddings to the input image, and places the original input at the
2830 given coordinates @var{x}, @var{y}.
2832 It accepts the following parameters:
2833 @var{width}:@var{height}:@var{x}:@var{y}:@var{color}.
2835 The parameters @var{width}, @var{height}, @var{x}, and @var{y} are
2836 expressions containing the following constants:
2840 the input video width and height
2843 same as @var{in_w} and @var{in_h}
2846 the output width and height, that is the size of the padded area as
2847 specified by the @var{width} and @var{height} expressions
2850 same as @var{out_w} and @var{out_h}
2853 x and y offsets as specified by the @var{x} and @var{y}
2854 expressions, or NAN if not yet specified
2857 same as @var{iw} / @var{ih}
2860 input sample aspect ratio
2863 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
2866 horizontal and vertical chroma subsample values. For example for the
2867 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
2870 Follows the description of the accepted parameters.
2875 Specify the size of the output image with the paddings added. If the
2876 value for @var{width} or @var{height} is 0, the corresponding input size
2877 is used for the output.
2879 The @var{width} expression can reference the value set by the
2880 @var{height} expression, and vice versa.
2882 The default value of @var{width} and @var{height} is 0.
2886 Specify the offsets where to place the input image in the padded area
2887 with respect to the top/left border of the output image.
2889 The @var{x} expression can reference the value set by the @var{y}
2890 expression, and vice versa.
2892 The default value of @var{x} and @var{y} is 0.
2896 Specify the color of the padded area, it can be the name of a color
2897 (case insensitive match) or a 0xRRGGBB[AA] sequence.
2899 The default value of @var{color} is "black".
2907 Add paddings with color "violet" to the input video. Output video
2908 size is 640x480, the top-left corner of the input video is placed at
2911 pad=640:480:0:40:violet
2915 Pad the input to get an output with dimensions increased by 3/2,
2916 and put the input video at the center of the padded area:
2918 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
2922 Pad the input to get a squared output with size equal to the maximum
2923 value between the input width and height, and put the input video at
2924 the center of the padded area:
2926 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
2930 Pad the input to get a final w/h ratio of 16:9:
2932 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
2936 In case of anamorphic video, in order to set the output display aspect
2937 correctly, it is necessary to use @var{sar} in the expression,
2938 according to the relation:
2940 (ih * X / ih) * sar = output_dar
2941 X = output_dar / sar
2944 Thus the previous example needs to be modified to:
2946 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
2950 Double output size and put the input video in the bottom-right
2951 corner of the output padded area:
2953 pad="2*iw:2*ih:ow-iw:oh-ih"
2957 @section pixdesctest
2959 Pixel format descriptor test filter, mainly useful for internal
2960 testing. The output video should be equal to the input video.
2964 format=monow, pixdesctest
2967 can be used to test the monowhite pixel format descriptor definition.
2971 Suppress a TV station logo, using an image file to determine which
2972 pixels comprise the logo. It works by filling in the pixels that
2973 comprise the logo with neighboring pixels.
2975 This filter requires one argument which specifies the filter bitmap
2976 file, which can be any image format supported by libavformat. The
2977 width and height of the image file must match those of the video
2978 stream being processed.
2980 Pixels in the provided bitmap image with a value of zero are not
2981 considered part of the logo, non-zero pixels are considered part of
2982 the logo. If you use white (255) for the logo and black (0) for the
2983 rest, you will be safe. For making the filter bitmap, it is
2984 recommended to take a screen capture of a black frame with the logo
2985 visible, and then using a threshold filter followed by the erode
2986 filter once or twice.
2988 If needed, little splotches can be fixed manually. Remember that if
2989 logo pixels are not covered, the filter quality will be much
2990 reduced. Marking too many pixels as part of the logo does not hurt as
2991 much, but it will increase the amount of blurring needed to cover over
2992 the image and will destroy more information than necessary, and extra
2993 pixels will slow things down on a large logo.
2997 Scale the input video to @var{width}:@var{height}[:@var{interl}=@{1|-1@}] and/or convert the image format.
2999 The scale filter forces the output display aspect ratio to be the same
3000 of the input, by changing the output sample aspect ratio.
3002 The parameters @var{width} and @var{height} are expressions containing
3003 the following constants:
3007 the input width and height
3010 same as @var{in_w} and @var{in_h}
3013 the output (cropped) width and height
3016 same as @var{out_w} and @var{out_h}
3019 same as @var{iw} / @var{ih}
3022 input sample aspect ratio
3025 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
3028 horizontal and vertical chroma subsample values. For example for the
3029 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
3032 If the input image format is different from the format requested by
3033 the next filter, the scale filter will convert the input to the
3036 If the value for @var{width} or @var{height} is 0, the respective input
3037 size is used for the output.
3039 If the value for @var{width} or @var{height} is -1, the scale filter will
3040 use, for the respective output size, a value that maintains the aspect
3041 ratio of the input image.
3043 The default value of @var{width} and @var{height} is 0.
3045 Valid values for the optional parameter @var{interl} are:
3049 force interlaced aware scaling
3052 select interlaced aware scaling depending on whether the source frames
3053 are flagged as interlaced or not
3056 Unless @var{interl} is set to one of the above options, interlaced scaling will not be used.
3058 Some examples follow:
3060 # scale the input video to a size of 200x100.
3063 # scale the input to 2x
3065 # the above is the same as
3068 # scale the input to 2x with forced interlaced scaling
3069 scale=2*iw:2*ih:interl=1
3071 # scale the input to half size
3074 # increase the width, and set the height to the same size
3077 # seek for Greek harmony
3081 # increase the height, and set the width to 3/2 of the height
3084 # increase the size, but make the size a multiple of the chroma
3085 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
3087 # increase the width to a maximum of 500 pixels, keep the same input aspect ratio
3088 scale='min(500\, iw*3/2):-1'
3092 Select frames to pass in output.
3094 It accepts in input an expression, which is evaluated for each input
3095 frame. If the expression is evaluated to a non-zero value, the frame
3096 is selected and passed to the output, otherwise it is discarded.
3098 The expression can contain the following constants:
3102 the sequential number of the filtered frame, starting from 0
3105 the sequential number of the selected frame, starting from 0
3107 @item prev_selected_n
3108 the sequential number of the last selected frame, NAN if undefined
3111 timebase of the input timestamps
3114 the PTS (Presentation TimeStamp) of the filtered video frame,
3115 expressed in @var{TB} units, NAN if undefined
3118 the PTS (Presentation TimeStamp) of the filtered video frame,
3119 expressed in seconds, NAN if undefined
3122 the PTS of the previously filtered video frame, NAN if undefined
3124 @item prev_selected_pts
3125 the PTS of the last previously filtered video frame, NAN if undefined
3127 @item prev_selected_t
3128 the PTS of the last previously selected video frame, NAN if undefined
3131 the PTS of the first video frame in the video, NAN if undefined
3134 the time of the first video frame in the video, NAN if undefined
3137 the type of the filtered frame, can assume one of the following
3149 @item interlace_type
3150 the frame interlace type, can assume one of the following values:
3153 the frame is progressive (not interlaced)
3155 the frame is top-field-first
3157 the frame is bottom-field-first
3161 1 if the filtered frame is a key-frame, 0 otherwise
3164 the position in the file of the filtered frame, -1 if the information
3165 is not available (e.g. for synthetic video)
3168 value between 0 and 1 to indicate a new scene; a low value reflects a low
3169 probability for the current frame to introduce a new scene, while a higher
3170 value means the current frame is more likely to be one (see the example below)
3174 The default value of the select expression is "1".
3176 Some examples follow:
3179 # select all frames in input
3182 # the above is the same as:
3188 # select only I-frames
3189 select='eq(pict_type\,I)'
3191 # select one frame every 100
3192 select='not(mod(n\,100))'
3194 # select only frames contained in the 10-20 time interval
3195 select='gte(t\,10)*lte(t\,20)'
3197 # select only I frames contained in the 10-20 time interval
3198 select='gte(t\,10)*lte(t\,20)*eq(pict_type\,I)'
3200 # select frames with a minimum distance of 10 seconds
3201 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
3204 Complete example to create a mosaic of the first scenes:
3207 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
3210 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
3213 @section setdar, setsar
3215 The @code{setdar} filter sets the Display Aspect Ratio for the filter
3218 This is done by changing the specified Sample (aka Pixel) Aspect
3219 Ratio, according to the following equation:
3221 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
3224 Keep in mind that the @code{setdar} filter does not modify the pixel
3225 dimensions of the video frame. Also the display aspect ratio set by
3226 this filter may be changed by later filters in the filterchain,
3227 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
3230 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
3231 the filter output video.
3233 Note that as a consequence of the application of this filter, the
3234 output display aspect ratio will change according to the equation
3237 Keep in mind that the sample aspect ratio set by the @code{setsar}
3238 filter may be changed by later filters in the filterchain, e.g. if
3239 another "setsar" or a "setdar" filter is applied.
3241 The @code{setdar} and @code{setsar} filters accept a parameter string
3242 which represents the wanted aspect ratio. The parameter can
3243 be a floating point number string, an expression, or a string of the form
3244 @var{num}:@var{den}, where @var{num} and @var{den} are the numerator
3245 and denominator of the aspect ratio. If the parameter is not
3246 specified, it is assumed the value "0:1".
3248 For example to change the display aspect ratio to 16:9, specify:
3253 The example above is equivalent to:
3258 To change the sample aspect ratio to 10:11, specify:
3265 Force field for the output video frame.
3267 The @code{setfield} filter marks the interlace type field for the
3268 output frames. It does not change the input frame, but only sets the
3269 corresponding property, which affects how the frame is treated by
3270 following filters (e.g. @code{fieldorder} or @code{yadif}).
3272 It accepts a string parameter, which can assume the following values:
3275 Keep the same field property.
3278 Mark the frame as bottom-field-first.
3281 Mark the frame as top-field-first.
3284 Mark the frame as progressive.
3289 Show a line containing various information for each input video frame.
3290 The input video is not modified.
3292 The shown line contains a sequence of key/value pairs of the form
3293 @var{key}:@var{value}.
3295 A description of each shown parameter follows:
3299 sequential number of the input frame, starting from 0
3302 Presentation TimeStamp of the input frame, expressed as a number of
3303 time base units. The time base unit depends on the filter input pad.
3306 Presentation TimeStamp of the input frame, expressed as a number of
3310 position of the frame in the input stream, -1 if this information in
3311 unavailable and/or meaningless (for example in case of synthetic video)
3317 sample aspect ratio of the input frame, expressed in the form
3321 size of the input frame, expressed in the form
3322 @var{width}x@var{height}
3325 interlaced mode ("P" for "progressive", "T" for top field first, "B"
3326 for bottom field first)
3329 1 if the frame is a key frame, 0 otherwise
3332 picture type of the input frame ("I" for an I-frame, "P" for a
3333 P-frame, "B" for a B-frame, "?" for unknown type).
3334 Check also the documentation of the @code{AVPictureType} enum and of
3335 the @code{av_get_picture_type_char} function defined in
3336 @file{libavutil/avutil.h}.
3339 Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame
3341 @item plane_checksum
3342 Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
3343 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]"
3348 Pass the images of input video on to next video filter as multiple
3352 ffmpeg -i in.avi -vf "slicify=32" out.avi
3355 The filter accepts the slice height as parameter. If the parameter is
3356 not specified it will use the default value of 16.
3358 Adding this in the beginning of filter chains should make filtering
3359 faster due to better use of the memory cache.
3363 Blur the input video without impacting the outlines.
3365 The filter accepts the following parameters:
3366 @var{luma_radius}:@var{luma_strength}:@var{luma_threshold}[:@var{chroma_radius}:@var{chroma_strength}:@var{chroma_threshold}]
3368 Parameters prefixed by @var{luma} indicate that they work on the
3369 luminance of the pixels whereas parameters prefixed by @var{chroma}
3370 refer to the chrominance of the pixels.
3372 If the chroma parameters are not set, the luma parameters are used for
3373 either the luminance and the chrominance of the pixels.
3375 @var{luma_radius} or @var{chroma_radius} must be a float number in the
3376 range [0.1,5.0] that specifies the variance of the gaussian filter
3377 used to blur the image (slower if larger).
3379 @var{luma_strength} or @var{chroma_strength} must be a float number in
3380 the range [-1.0,1.0] that configures the blurring. A value included in
3381 [0.0,1.0] will blur the image whereas a value included in [-1.0,0.0]
3382 will sharpen the image.
3384 @var{luma_threshold} or @var{chroma_threshold} must be an integer in
3385 the range [-30,30] that is used as a coefficient to determine whether
3386 a pixel should be blurred or not. A value of 0 will filter all the
3387 image, a value included in [0,30] will filter flat areas and a value
3388 included in [-30,0] will filter edges.
3392 Split input video into several identical outputs.
3394 The filter accepts a single parameter which specifies the number of outputs. If
3395 unspecified, it defaults to 2.
3399 ffmpeg -i INPUT -filter_complex split=5 OUTPUT
3401 will create 5 copies of the input video.
3405 [in] split [splitout1][splitout2];
3406 [splitout1] crop=100:100:0:0 [cropout];
3407 [splitout2] pad=200:200:100:100 [padout];
3410 will create two separate outputs from the same input, one cropped and
3415 Scale the input by 2x and smooth using the Super2xSaI (Scale and
3416 Interpolate) pixel art scaling algorithm.
3418 Useful for enlarging pixel art images without reducing sharpness.
3424 Select the most representative frame in a given sequence of consecutive frames.
3426 It accepts as argument the frames batch size to analyze (default @var{N}=100);
3427 in a set of @var{N} frames, the filter will pick one of them, and then handle
3428 the next batch of @var{N} frames until the end.
3430 Since the filter keeps track of the whole frames sequence, a bigger @var{N}
3431 value will result in a higher memory usage, so a high value is not recommended.
3433 The following example extract one picture each 50 frames:
3438 Complete example of a thumbnail creation with @command{ffmpeg}:
3440 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
3445 Tile several successive frames together.
3447 It accepts as argument the tile size (i.e. the number of lines and columns)
3448 in the form "@var{w}x@var{h}".
3450 For example, produce 8×8 PNG tiles of all keyframes (@option{-skip_frame
3453 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
3455 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
3456 duplicating each output frame to accomodate the originally detected frame
3461 Perform various types of temporal field interlacing.
3463 Frames are counted starting from 1, so the first input frame is
3466 This filter accepts a single parameter specifying the mode. Available
3471 Move odd frames into the upper field, even into the lower field,
3472 generating a double height frame at half framerate.
3475 Only output even frames, odd frames are dropped, generating a frame with
3476 unchanged height at half framerate.
3479 Only output odd frames, even frames are dropped, generating a frame with
3480 unchanged height at half framerate.
3483 Expand each frame to full height, but pad alternate lines with black,
3484 generating a frame with double height at the same input framerate.
3486 @item interleave_top, 4
3487 Interleave the upper field from odd frames with the lower field from
3488 even frames, generating a frame with unchanged height at half framerate.
3490 @item interleave_bottom, 5
3491 Interleave the lower field from odd frames with the upper field from
3492 even frames, generating a frame with unchanged height at half framerate.
3494 @item interlacex2, 6
3495 Double frame rate with unchanged height. Frames are inserted each
3496 containing the second temporal field from the previous input frame and
3497 the first temporal field from the next input frame. This mode relies on
3498 the top_field_first flag. Useful for interlaced video displays with no
3499 field synchronisation.
3502 Numeric values are deprecated but are accepted for backward
3503 compatibility reasons.
3505 Default mode is @code{merge}.
3509 Transpose rows with columns in the input video and optionally flip it.
3511 It accepts a parameter representing an integer, which can assume the
3516 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
3524 Rotate by 90 degrees clockwise, that is:
3532 Rotate by 90 degrees counterclockwise, that is:
3540 Rotate by 90 degrees clockwise and vertically flip, that is:
3548 For values between 4-7 transposition is only done if the input video
3549 geometry is portrait and not landscape.
3553 Sharpen or blur the input video.
3555 It accepts the following parameters:
3556 @var{luma_msize_x}:@var{luma_msize_y}:@var{luma_amount}:@var{chroma_msize_x}:@var{chroma_msize_y}:@var{chroma_amount}
3558 Negative values for the amount will blur the input video, while positive
3559 values will sharpen. All parameters are optional and default to the
3560 equivalent of the string '5:5:1.0:5:5:0.0'.
3565 Set the luma matrix horizontal size. It can be an integer between 3
3566 and 13, default value is 5.
3569 Set the luma matrix vertical size. It can be an integer between 3
3570 and 13, default value is 5.
3573 Set the luma effect strength. It can be a float number between -2.0
3574 and 5.0, default value is 1.0.
3576 @item chroma_msize_x
3577 Set the chroma matrix horizontal size. It can be an integer between 3
3578 and 13, default value is 5.
3580 @item chroma_msize_y
3581 Set the chroma matrix vertical size. It can be an integer between 3
3582 and 13, default value is 5.
3585 Set the chroma effect strength. It can be a float number between -2.0
3586 and 5.0, default value is 0.0.
3591 # Strong luma sharpen effect parameters
3594 # Strong blur of both luma and chroma parameters
3595 unsharp=7:7:-2:7:7:-2
3597 # Use the default values with @command{ffmpeg}
3598 ffmpeg -i in.avi -vf "unsharp" out.mp4
3603 Flip the input video vertically.
3606 ffmpeg -i in.avi -vf "vflip" out.avi
3611 Deinterlace the input video ("yadif" means "yet another deinterlacing
3614 It accepts the optional parameters: @var{mode}:@var{parity}:@var{auto}.
3616 @var{mode} specifies the interlacing mode to adopt, accepts one of the
3621 output 1 frame for each frame
3623 output 1 frame for each field
3625 like 0 but skips spatial interlacing check
3627 like 1 but skips spatial interlacing check
3632 @var{parity} specifies the picture field parity assumed for the input
3633 interlaced video, accepts one of the following values:
3637 assume top field first
3639 assume bottom field first
3641 enable automatic detection
3644 Default value is -1.
3645 If interlacing is unknown or decoder does not export this information,
3646 top field first will be assumed.
3648 @var{auto} specifies if deinterlacer should trust the interlaced flag
3649 and only deinterlace frames marked as interlaced
3653 deinterlace all frames
3655 only deinterlace frames marked as interlaced
3660 @c man end VIDEO FILTERS
3662 @chapter Video Sources
3663 @c man begin VIDEO SOURCES
3665 Below is a description of the currently available video sources.
3669 Buffer video frames, and make them available to the filter chain.
3671 This source is mainly intended for a programmatic use, in particular
3672 through the interface defined in @file{libavfilter/vsrc_buffer.h}.
3674 It accepts a list of options in the form of @var{key}=@var{value} pairs
3675 separated by ":". A descroption of the accepted options follows.
3680 Specify the size (width and height) of the buffered video frames.
3683 A string representing the pixel format of the buffered video frames.
3684 It may be a number corresponding to a pixel format, or a pixel format
3688 Specify the timebase assumed by the timestamps of the buffered frames.
3691 Specify the frame rate expected for the video stream.
3694 Specify the sample aspect ratio assumed by the video frames.
3697 Specify the optional parameters to be used for the scale filter which
3698 is automatically inserted when an input change is detected in the
3699 input size or format.
3704 buffer=size=320x240:pix_fmt=yuv410p:time_base=1/24:pixel_aspect=1/1
3707 will instruct the source to accept video frames with size 320x240 and
3708 with format "yuv410p", assuming 1/24 as the timestamps timebase and
3709 square pixels (1:1 sample aspect ratio).
3710 Since the pixel format with name "yuv410p" corresponds to the number 6
3711 (check the enum PixelFormat definition in @file{libavutil/pixfmt.h}),
3712 this example corresponds to:
3714 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
3717 Alternatively, the options can be specified as a flat string, but this
3718 syntax is deprecated:
3720 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}[:@var{sws_param}]
3724 Create a pattern generated by an elementary cellular automaton.
3726 The initial state of the cellular automaton can be defined through the
3727 @option{filename}, and @option{pattern} options. If such options are
3728 not specified an initial state is created randomly.
3730 At each new frame a new row in the video is filled with the result of
3731 the cellular automaton next generation. The behavior when the whole
3732 frame is filled is defined by the @option{scroll} option.
3734 This source accepts a list of options in the form of
3735 @var{key}=@var{value} pairs separated by ":". A description of the
3736 accepted options follows.
3740 Read the initial cellular automaton state, i.e. the starting row, from
3742 In the file, each non-whitespace character is considered an alive
3743 cell, a newline will terminate the row, and further characters in the
3744 file will be ignored.
3747 Read the initial cellular automaton state, i.e. the starting row, from
3748 the specified string.
3750 Each non-whitespace character in the string is considered an alive
3751 cell, a newline will terminate the row, and further characters in the
3752 string will be ignored.
3755 Set the video rate, that is the number of frames generated per second.
3758 @item random_fill_ratio, ratio
3759 Set the random fill ratio for the initial cellular automaton row. It
3760 is a floating point number value ranging from 0 to 1, defaults to
3763 This option is ignored when a file or a pattern is specified.
3765 @item random_seed, seed
3766 Set the seed for filling randomly the initial row, must be an integer
3767 included between 0 and UINT32_MAX. If not specified, or if explicitly
3768 set to -1, the filter will try to use a good random seed on a best
3772 Set the cellular automaton rule, it is a number ranging from 0 to 255.
3773 Default value is 110.
3776 Set the size of the output video.
3778 If @option{filename} or @option{pattern} is specified, the size is set
3779 by default to the width of the specified initial state row, and the
3780 height is set to @var{width} * PHI.
3782 If @option{size} is set, it must contain the width of the specified
3783 pattern string, and the specified pattern will be centered in the
3786 If a filename or a pattern string is not specified, the size value
3787 defaults to "320x518" (used for a randomly generated initial state).
3790 If set to 1, scroll the output upward when all the rows in the output
3791 have been already filled. If set to 0, the new generated row will be
3792 written over the top row just after the bottom row is filled.
3795 @item start_full, full
3796 If set to 1, completely fill the output with generated rows before
3797 outputting the first frame.
3798 This is the default behavior, for disabling set the value to 0.
3801 If set to 1, stitch the left and right row edges together.
3802 This is the default behavior, for disabling set the value to 0.
3805 @subsection Examples
3809 Read the initial state from @file{pattern}, and specify an output of
3812 cellauto=f=pattern:s=200x400
3816 Generate a random initial row with a width of 200 cells, with a fill
3819 cellauto=ratio=2/3:s=200x200
3823 Create a pattern generated by rule 18 starting by a single alive cell
3824 centered on an initial row with width 100:
3826 cellauto=p=@@:s=100x400:full=0:rule=18
3830 Specify a more elaborated initial pattern:
3832 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
3839 Generate a Mandelbrot set fractal, and progressively zoom towards the
3840 point specified with @var{start_x} and @var{start_y}.
3842 This source accepts a list of options in the form of
3843 @var{key}=@var{value} pairs separated by ":". A description of the
3844 accepted options follows.
3849 Set the terminal pts value. Default value is 400.
3852 Set the terminal scale value.
3853 Must be a floating point value. Default value is 0.3.
3856 Set the inner coloring mode, that is the algorithm used to draw the
3857 Mandelbrot fractal internal region.
3859 It shall assume one of the following values:
3864 Show time until convergence.
3866 Set color based on point closest to the origin of the iterations.
3871 Default value is @var{mincol}.
3874 Set the bailout value. Default value is 10.0.
3877 Set the maximum of iterations performed by the rendering
3878 algorithm. Default value is 7189.
3881 Set outer coloring mode.
3882 It shall assume one of following values:
3884 @item iteration_count
3885 Set iteration cound mode.
3886 @item normalized_iteration_count
3887 set normalized iteration count mode.
3889 Default value is @var{normalized_iteration_count}.
3892 Set frame rate, expressed as number of frames per second. Default
3896 Set frame size. Default value is "640x480".
3899 Set the initial scale value. Default value is 3.0.
3902 Set the initial x position. Must be a floating point value between
3903 -100 and 100. Default value is -0.743643887037158704752191506114774.
3906 Set the initial y position. Must be a floating point value between
3907 -100 and 100. Default value is -0.131825904205311970493132056385139.
3912 Generate various test patterns, as generated by the MPlayer test filter.
3914 The size of the generated video is fixed, and is 256x256.
3915 This source is useful in particular for testing encoding features.
3917 This source accepts an optional sequence of @var{key}=@var{value} pairs,
3918 separated by ":". The description of the accepted options follows.
3923 Specify the frame rate of the sourced video, as the number of frames
3924 generated per second. It has to be a string in the format
3925 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a float
3926 number or a valid video frame rate abbreviation. The default value is
3930 Set the video duration of the sourced video. The accepted syntax is:
3935 See also the function @code{av_parse_time()}.
3937 If not specified, or the expressed duration is negative, the video is
3938 supposed to be generated forever.
3942 Set the number or the name of the test to perform. Supported tests are:
3957 Default value is "all", which will cycle through the list of all tests.
3960 For example the following:
3965 will generate a "dc_luma" test pattern.
3969 Provide a frei0r source.
3971 To enable compilation of this filter you need to install the frei0r
3972 header and configure FFmpeg with @code{--enable-frei0r}.
3974 The source supports the syntax:
3976 @var{size}:@var{rate}:@var{src_name}[@{=|:@}@var{param1}:@var{param2}:...:@var{paramN}]
3979 @var{size} is the size of the video to generate, may be a string of the
3980 form @var{width}x@var{height} or a frame size abbreviation.
3981 @var{rate} is the rate of the video to generate, may be a string of
3982 the form @var{num}/@var{den} or a frame rate abbreviation.
3983 @var{src_name} is the name to the frei0r source to load. For more
3984 information regarding frei0r and how to set the parameters read the
3985 section @ref{frei0r} in the description of the video filters.
3987 For example, to generate a frei0r partik0l source with size 200x200
3988 and frame rate 10 which is overlayed on the overlay filter main input:
3990 frei0r_src=200x200:10:partik0l=1234 [overlay]; [in][overlay] overlay
3995 Generate a life pattern.
3997 This source is based on a generalization of John Conway's life game.
3999 The sourced input represents a life grid, each pixel represents a cell
4000 which can be in one of two possible states, alive or dead. Every cell
4001 interacts with its eight neighbours, which are the cells that are
4002 horizontally, vertically, or diagonally adjacent.
4004 At each interaction the grid evolves according to the adopted rule,
4005 which specifies the number of neighbor alive cells which will make a
4006 cell stay alive or born. The @option{rule} option allows to specify
4009 This source accepts a list of options in the form of
4010 @var{key}=@var{value} pairs separated by ":". A description of the
4011 accepted options follows.
4015 Set the file from which to read the initial grid state. In the file,
4016 each non-whitespace character is considered an alive cell, and newline
4017 is used to delimit the end of each row.
4019 If this option is not specified, the initial grid is generated
4023 Set the video rate, that is the number of frames generated per second.
4026 @item random_fill_ratio, ratio
4027 Set the random fill ratio for the initial random grid. It is a
4028 floating point number value ranging from 0 to 1, defaults to 1/PHI.
4029 It is ignored when a file is specified.
4031 @item random_seed, seed
4032 Set the seed for filling the initial random grid, must be an integer
4033 included between 0 and UINT32_MAX. If not specified, or if explicitly
4034 set to -1, the filter will try to use a good random seed on a best
4040 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
4041 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
4042 @var{NS} specifies the number of alive neighbor cells which make a
4043 live cell stay alive, and @var{NB} the number of alive neighbor cells
4044 which make a dead cell to become alive (i.e. to "born").
4045 "s" and "b" can be used in place of "S" and "B", respectively.
4047 Alternatively a rule can be specified by an 18-bits integer. The 9
4048 high order bits are used to encode the next cell state if it is alive
4049 for each number of neighbor alive cells, the low order bits specify
4050 the rule for "borning" new cells. Higher order bits encode for an
4051 higher number of neighbor cells.
4052 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
4053 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
4055 Default value is "S23/B3", which is the original Conway's game of life
4056 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
4057 cells, and will born a new cell if there are three alive cells around
4061 Set the size of the output video.
4063 If @option{filename} is specified, the size is set by default to the
4064 same size of the input file. If @option{size} is set, it must contain
4065 the size specified in the input file, and the initial grid defined in
4066 that file is centered in the larger resulting area.
4068 If a filename is not specified, the size value defaults to "320x240"
4069 (used for a randomly generated initial grid).
4072 If set to 1, stitch the left and right grid edges together, and the
4073 top and bottom edges also. Defaults to 1.
4076 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
4077 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
4078 value from 0 to 255.
4081 Set the color of living (or new born) cells.
4084 Set the color of dead cells. If @option{mold} is set, this is the first color
4085 used to represent a dead cell.
4088 Set mold color, for definitely dead and moldy cells.
4091 @subsection Examples
4095 Read a grid from @file{pattern}, and center it on a grid of size
4098 life=f=pattern:s=300x300
4102 Generate a random grid of size 200x200, with a fill ratio of 2/3:
4104 life=ratio=2/3:s=200x200
4108 Specify a custom rule for evolving a randomly generated grid:
4114 Full example with slow death effect (mold) using @command{ffplay}:
4116 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
4120 @section color, nullsrc, rgbtestsrc, smptebars, testsrc
4122 The @code{color} source provides an uniformly colored input.
4124 The @code{nullsrc} source returns unprocessed video frames. It is
4125 mainly useful to be employed in analysis / debugging tools, or as the
4126 source for filters which ignore the input data.
4128 The @code{rgbtestsrc} source generates an RGB test pattern useful for
4129 detecting RGB vs BGR issues. You should see a red, green and blue
4130 stripe from top to bottom.
4132 The @code{smptebars} source generates a color bars pattern, based on
4133 the SMPTE Engineering Guideline EG 1-1990.
4135 The @code{testsrc} source generates a test video pattern, showing a
4136 color pattern, a scrolling gradient and a timestamp. This is mainly
4137 intended for testing purposes.
4139 These sources accept an optional sequence of @var{key}=@var{value} pairs,
4140 separated by ":". The description of the accepted options follows.
4145 Specify the color of the source, only used in the @code{color}
4146 source. It can be the name of a color (case insensitive match) or a
4147 0xRRGGBB[AA] sequence, possibly followed by an alpha specifier. The
4148 default value is "black".
4151 Specify the size of the sourced video, it may be a string of the form
4152 @var{width}x@var{height}, or the name of a size abbreviation. The
4153 default value is "320x240".
4156 Specify the frame rate of the sourced video, as the number of frames
4157 generated per second. It has to be a string in the format
4158 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a float
4159 number or a valid video frame rate abbreviation. The default value is
4163 Set the sample aspect ratio of the sourced video.
4166 Set the video duration of the sourced video. The accepted syntax is:
4168 [-]HH[:MM[:SS[.m...]]]
4171 See also the function @code{av_parse_time()}.
4173 If not specified, or the expressed duration is negative, the video is
4174 supposed to be generated forever.
4177 Set the number of decimals to show in the timestamp, only used in the
4178 @code{testsrc} source.
4180 The displayed timestamp value will correspond to the original
4181 timestamp value multiplied by the power of 10 of the specified
4182 value. Default value is 0.
4185 For example the following:
4187 testsrc=duration=5.3:size=qcif:rate=10
4190 will generate a video with a duration of 5.3 seconds, with size
4191 176x144 and a frame rate of 10 frames per second.
4193 The following graph description will generate a red source
4194 with an opacity of 0.2, with size "qcif" and a frame rate of 10
4197 color=c=red@@0.2:s=qcif:r=10
4200 If the input content is to be ignored, @code{nullsrc} can be used. The
4201 following command generates noise in the luminance plane by employing
4202 the @code{mp=geq} filter:
4204 nullsrc=s=256x256, mp=geq=random(1)*255:128:128
4207 @c man end VIDEO SOURCES
4209 @chapter Video Sinks
4210 @c man begin VIDEO SINKS
4212 Below is a description of the currently available video sinks.
4216 Buffer video frames, and make them available to the end of the filter
4219 This sink is mainly intended for a programmatic use, in particular
4220 through the interface defined in @file{libavfilter/buffersink.h}.
4222 It does not require a string parameter in input, but you need to
4223 specify a pointer to a list of supported pixel formats terminated by
4224 -1 in the opaque parameter provided to @code{avfilter_init_filter}
4225 when initializing this sink.
4229 Null video sink, do absolutely nothing with the input video. It is
4230 mainly useful as a template and to be employed in analysis / debugging
4233 @c man end VIDEO SINKS
4235 @chapter Multimedia Filters
4236 @c man begin MULTIMEDIA FILTERS
4238 Below is a description of the currently available multimedia filters.
4240 @section asetpts, setpts
4242 Change the PTS (presentation timestamp) of the input frames.
4244 @code{asetpts} works on audio frames, @code{setpts} on video frames.
4246 Accept in input an expression evaluated through the eval API, which
4247 can contain the following constants:
4251 frame rate, only defined for constant frame-rate video
4254 the presentation timestamp in input
4257 the count of the input frame, starting from 0.
4259 @item NB_CONSUMED_SAMPLES
4260 the number of consumed samples, not including the current frame (only
4264 the number of samples in the current frame (only audio)
4270 the PTS of the first frame
4273 the time in seconds of the first frame
4276 tell if the current frame is interlaced
4279 the time in seconds of the current frame
4285 original position in the file of the frame, or undefined if undefined
4286 for the current frame
4292 previous input time in seconds
4298 previous output time in seconds
4301 @subsection Examples
4305 Start counting PTS from zero
4311 Apply fast motion effect:
4317 Apply slow motion effect:
4323 Set fixed rate of 25 frames per second:
4329 Set fixed rate 25 fps with some jitter:
4331 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
4335 Apply an offset of 10 seconds to the input PTS:
4341 @section settb, asettb
4343 Set the timebase to use for the output frames timestamps.
4344 It is mainly useful for testing timebase configuration.
4346 It accepts in input an arithmetic expression representing a rational.
4347 The expression can contain the constants "AVTB" (the
4348 default timebase), "intb" (the input timebase) and "sr" (the sample rate,
4351 The default value for the input is "intb".
4353 @subsection Examples
4357 Set the timebase to 1/25:
4363 Set the timebase to 1/10:
4369 Set the timebase to 1001/1000:
4375 Set the timebase to 2*intb:
4381 Set the default timebase value:
4389 Concatenate audio and video streams, joining them together one after the
4392 The filter works on segments of synchronized video and audio streams. All
4393 segments must have the same number of streams of each type, and that will
4394 also be the number of streams at output.
4396 The filter accepts the following named parameters:
4400 Set the number of segments. Default is 2.
4403 Set the number of output video streams, that is also the number of video
4404 streams in each segment. Default is 1.
4407 Set the number of output audio streams, that is also the number of video
4408 streams in each segment. Default is 0.
4412 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
4413 @var{a} audio outputs.
4415 There are @var{n}×(@var{v}+@var{a}) inputs: first the inputs for the first
4416 segment, in the same order as the outputs, then the inputs for the second
4419 Related streams do not always have exactly the same duration, for various
4420 reasons including codec frame size or sloppy authoring. For that reason,
4421 related synchronized streams (e.g. a video and its audio track) should be
4422 concatenated at once. The concat filter will use the duration of the longest
4423 stream in each segment (except the last one), and if necessary pad shorter
4424 audio streams with silence.
4426 For this filter to work correctly, all segments must start at timestamp 0.
4428 All corresponding streams must have the same parameters in all segments; the
4429 filtering system will automatically select a common pixel format for video
4430 streams, and a common sample format, sample rate and channel layout for
4431 audio streams, but other settings, such as resolution, must be converted
4432 explicitly by the user.
4434 Different frame rates are acceptable but will result in variable frame rate
4435 at output; be sure to configure the output file to handle it.
4440 Concatenate an opening, an episode and an ending, all in bilingual version
4441 (video in stream 0, audio in streams 1 and 2):
4443 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
4444 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
4445 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
4446 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
4450 Concatenate two parts, handling audio and video separately, using the
4451 (a)movie sources, and adjusting the resolution:
4453 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
4454 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
4455 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
4457 Note that a desync will happen at the stitch if the audio and video streams
4458 do not have exactly the same duration in the first file.
4462 @section showspectrum
4464 Convert input audio to a video output, representing the audio frequency
4467 The filter accepts the following named parameters:
4470 Specify the video size for the output. Default value is @code{640x480}.
4473 The usage is very similar to the showwaves filter; see the examples in that
4478 Convert input audio to a video output, representing the samples waves.
4480 The filter accepts the following named parameters:
4484 Set the number of samples which are printed on the same column. A
4485 larger value will decrease the frame rate. Must be a positive
4486 integer. This option can be set only if the value for @var{rate}
4487 is not explicitly specified.
4490 Set the (approximate) output frame rate. This is done by setting the
4491 option @var{n}. Default value is "25".
4494 Specify the video size for the output. Default value is "600x240".
4497 Some examples follow.
4500 Output the input file audio and the corresponding video representation
4503 amovie=a.mp3,asplit[out0],showwaves[out1]
4507 Create a synthetic signal and show it with showwaves, forcing a
4508 framerate of 30 frames per second:
4510 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
4514 @c man end MULTIMEDIA FILTERS
4516 @chapter Multimedia Sources
4517 @c man begin MULTIMEDIA SOURCES
4519 Below is a description of the currently available multimedia sources.
4523 This is the same as @ref{src_movie} source, except it selects an audio
4529 Read audio and/or video stream(s) from a movie container.
4531 It accepts the syntax: @var{movie_name}[:@var{options}] where
4532 @var{movie_name} is the name of the resource to read (not necessarily
4533 a file but also a device or a stream accessed through some protocol),
4534 and @var{options} is an optional sequence of @var{key}=@var{value}
4535 pairs, separated by ":".
4537 The description of the accepted options follows.
4541 @item format_name, f
4542 Specifies the format assumed for the movie to read, and can be either
4543 the name of a container or an input device. If not specified the
4544 format is guessed from @var{movie_name} or by probing.
4546 @item seek_point, sp
4547 Specifies the seek point in seconds, the frames will be output
4548 starting from this seek point, the parameter is evaluated with
4549 @code{av_strtod} so the numerical value may be suffixed by an IS
4550 postfix. Default value is "0".
4553 Specifies the streams to read. Several streams can be specified, separated
4554 by "+". The source will then have as many outputs, in the same order. The
4555 syntax is explained in the @ref{Stream specifiers} chapter. Two special
4556 names, "dv" and "da" specify respectively the default (best suited) video
4557 and audio stream. Default is "dv", or "da" if the filter is called as
4560 @item stream_index, si
4561 Specifies the index of the video stream to read. If the value is -1,
4562 the best suited video stream will be automatically selected. Default
4563 value is "-1". Deprecated. If the filter is called "amovie", it will select
4564 audio instead of video.
4567 Specifies how many times to read the stream in sequence.
4568 If the value is less than 1, the stream will be read again and again.
4569 Default value is "1".
4571 Note that when the movie is looped the source timestamps are not
4572 changed, so it will generate non monotonically increasing timestamps.
4575 This filter allows to overlay a second video on top of main input of
4576 a filtergraph as shown in this graph:
4578 input -----------> deltapts0 --> overlay --> output
4581 movie --> scale--> deltapts1 -------+
4584 Some examples follow.
4588 Skip 3.2 seconds from the start of the avi file in.avi, and overlay it
4589 on top of the input labelled as "in":
4591 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [movie];
4592 [in] setpts=PTS-STARTPTS, [movie] overlay=16:16 [out]
4596 Read from a video4linux2 device, and overlay it on top of the input
4599 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [movie];
4600 [in] setpts=PTS-STARTPTS, [movie] overlay=16:16 [out]
4604 Read the first video stream and the audio stream with id 0x81 from
4605 dvd.vob; the video is connected to the pad named "video" and the audio is
4606 connected to the pad named "audio":
4608 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
4612 @c man end MULTIMEDIA SOURCES