1. Introduction

FFmpeg is a very fast video and audio converter. It can also grab from a live audio/video source. The command line interface is designed to be intuitive, in the sense that ffmpeg tries to figure out all the parameters, when possible. You have usually to give only the target bitrate you want.

FFmpeg can also convert from any sample rate to any other, and resize video on the fly with a high quality polyphase filter.

2. Quick Start

2.1 Video and Audio grabbing

FFmpeg can use a video4linux compatible video source and any Open Sound System audio source:

ffmpeg /tmp/out.mpg 

Note that you must activate the right video source and channel before launching ffmpeg. You can use any TV viewer such as xawtv (http://bytesex.org/xawtv/) by Gerd Knorr which I find very good. You must also set correctly the audio recording levels with a standard mixer.

2.2 Video and Audio file format conversion

* ffmpeg can use any supported file format and protocol as input:

Examples:

* You can input from YUV files:

ffmpeg -i /tmp/test%d.Y /tmp/out.mpg 

It will use the files:

/tmp/test0.Y, /tmp/test0.U, /tmp/test0.V,
/tmp/test1.Y, /tmp/test1.U, /tmp/test1.V, etc...

The Y files use twice the resolution of the U and V files. They are raw files, without header. They can be generated by all decent video decoders. You must specify the size of the image with the `-s' option if ffmpeg cannot guess it.

* You can input from a RAW YUV420P file:

ffmpeg -i /tmp/test.yuv /tmp/out.avi

The RAW YUV420P is a file containing RAW YUV planar, for each frame first come the Y plane followed by U and V planes, which are half vertical and horizontal resolution.

* You can output to a RAW YUV420P file:

ffmpeg -i mydivx.avi -o hugefile.yuv

* You can set several input files and output files:

ffmpeg -i /tmp/a.wav -s 640x480 -i /tmp/a.yuv /tmp/a.mpg

Convert the audio file a.wav and the raw yuv video file a.yuv to mpeg file a.mpg

* You can also do audio and video conversions at the same time:

ffmpeg -i /tmp/a.wav -ar 22050 /tmp/a.mp2

Convert the sample rate of a.wav to 22050 Hz and encode it to MPEG audio.

* You can encode to several formats at the same time and define a mapping from input stream to output streams:

ffmpeg -i /tmp/a.wav -ab 64 /tmp/a.mp2 -ab 128 /tmp/b.mp2 -map 0:0 -map 0:0

Convert a.wav to a.mp2 at 64 kbits and b.mp2 at 128 kbits. '-map file:index' specify which input stream is used for each output stream, in the order of the definition of output streams.

* You can transcode decrypted VOBs

ffmpeg -i snatch_1.vob -f avi -vcodec mpeg4 -b 800 -g 300 -bf 2 -acodec mp3 -ab 128 snatch.avi

This is a typical DVD ripper example, input from a VOB file, output to an AVI file with MPEG-4 video and MP3 audio, note that in this command we use B frames so the MPEG-4 stream is DivX5 compatible, GOP size is 300 that means an INTRA frame every 10 seconds for 29.97 fps input video. Also the audio stream is MP3 encoded so you need LAME support which is enabled using --enable-mp3lame when configuring. The mapping is particularly useful for DVD transcoding to get the desired audio language.

NOTE: to see the supported input formats, use ffmpeg -formats.

3. Invocation

3.1 Syntax

The generic syntax is:

ffmpeg [[infile options][`-i' infile]]... {[outfile options] outfile}...

As a general rule, options are applied to the next specified file. For example, if you give the `-b 64' option, it sets the video bitrate of the next file. Format option may be needed for raw input files.

By default, ffmpeg tries to convert as losslessly as possible: it uses the same audio and video parameter for the outputs as the one specified for the inputs.

3.2 Main options

`-L'

show license

`-h'

show help

`-formats'

show available formats, codecs, protocols, ...

`-f fmt'

force format

`-i filename'

input file name

`-y'

overwrite output files

`-t duration'

set the recording time in seconds. hh:mm:ss[.xxx] syntax is also supported.

`-ss position'

seek to given time position. hh:mm:ss[.xxx] syntax is also supported.

`-title string'

set the title

`-author string'

set the author

`-copyright string'

set the copyright

`-comment string'

set the comment

`-target type'

specify target file type ("vcd", "svcd", "dvd", "pal-vcd", "ntsc-svcd", ... ). All the format options (bitrate, codecs, buffer sizes) are automatically set by this option. You can just type:

ffmpeg -i myfile.avi -target vcd /tmp/vcd.mpg

Nevertheless you can specify additional options as long as you know they do not compromise the standard, as in:

ffmpeg -i myfile.avi -target vcd -bf 2 /tmp/vcd.mpg

`-hq'

activate high quality settings

`-itsoffset offset'

set the input time offset in seconds. [-]hh:mm:ss[.xxx] syntax is also supported. This option affects all the input files that follow it. The offset is added to the input files' timestamps; specifying a positive offset means that the corresponding streams are delayed by 'offset' seconds.

3.3 Video Options

`-b bitrate'

set the video bitrate in kbit/s (default = 200 kb/s)

`-r fps'

set frame rate (default = 25)

`-s size'

set frame size. The format is `WxH' (default 160x128). The following abbreviations are recognized:

`sqcif'

128x96

`qcif'

176x144

`cif'

352x288

`4cif'

704x576

`-aspect aspect'

set aspect ratio (4:3, 16:9 or 1.3333, 1.7777)

`-croptop size'

set top crop band size (in pixels)

`-cropbottom size'

set bottom crop band size (in pixels)

`-cropleft size'

set left crop band size (in pixels)

`-cropright size'

set right crop band size (in pixels)

`-padtop size'

set top pad band size (in pixels)

`-padbottom size'

set bottom pad band size (in pixels)

`-padleft size'

set left pad band size (in pixels)

`-padright size'

set right pad band size (in pixels)

`-padcolor (hex color)'

set color of padded bands. The value for pad color is expressed as a six digit hexidecimal number where the first two digits represent red, middle two digits green and last two digits blue. Defaults to 000000 (black)

`-vn'

disable video recording

`-bt tolerance'

set video bitrate tolerance (in kbit/s)

`-maxrate bitrate'

set max video bitrate tolerance (in kbit/s)

`-minrate bitrate'

set min video bitrate tolerance (in kbit/s)

`-bufsize size'

set ratecontrol buffere size (in kbit)

`-vcodec codec'

force video codec to codec. Use the copy special value to tell that the raw codec data must be copied as is.

`-sameq'

use same video quality as source (implies VBR)

`-pass n'

select the pass number (1 or 2). It is useful to do two pass encoding. The statistics of the video are recorded in the first pass and the video at the exact requested bit rate is generated in the second pass.

`-passlogfile file'

select two pass log file name to file.

3.4 Advanced Video Options

`-g gop_size'

set the group of picture size

`-intra'

use only intra frames

`-qscale q'

use fixed video quantiser scale (VBR)

`-qmin q'

min video quantiser scale (VBR)

`-qmax q'

max video quantiser scale (VBR)

`-qdiff q'

max difference between the quantiser scale (VBR)

`-qblur blur'

video quantiser scale blur (VBR)

`-qcomp compression'

video quantiser scale compression (VBR)

`-rc_init_cplx complexity'

initial complexity for 1-pass encoding

`-b_qfactor factor'

qp factor between p and b frames

`-i_qfactor factor'

qp factor between p and i frames

`-b_qoffset offset'

qp offset between p and b frames

`-i_qoffset offset'

qp offset between p and i frames

`-rc_eq equation'

set rate control equation (see section 3.8 FFmpeg formula evaluator). Default is tex^qComp.

`-rc_override override'

rate control override for specific intervals

`-me method'

set motion estimation method to method. Available methods are (from lower to best quality):

`zero'

Try just the (0, 0) vector.

`phods'

`log'

`x1'

`epzs'

(default method)

`full'

exhaustive search (slow and marginally better than epzs)

`-dct_algo algo'

set dct algorithm to algo. Available values are:

`0'

FF_DCT_AUTO (default)

`1'

FF_DCT_FASTINT

`2'

FF_DCT_INT

`3'

FF_DCT_MMX

`4'

FF_DCT_MLIB

`5'

FF_DCT_ALTIVEC

`-idct_algo algo'

set idct algorithm to algo. Available values are:

`0'

FF_IDCT_AUTO (default)

`1'

FF_IDCT_INT

`2'

FF_IDCT_SIMPLE

`3'

FF_IDCT_SIMPLEMMX

`4'

FF_IDCT_LIBMPEG2MMX

`5'

FF_IDCT_PS2

`6'

FF_IDCT_MLIB

`7'

FF_IDCT_ARM

`8'

FF_IDCT_ALTIVEC

`9'

FF_IDCT_SH4

`10'

FF_IDCT_SIMPLEARM

`-er n'

set error resilience to n.

`1'

FF_ER_CAREFULL (default)

`2'

FF_ER_COMPLIANT

`3'

FF_ER_AGGRESSIVE

`4'

FF_ER_VERY_AGGRESSIVE

`-ec bit_mask'

set error concealment to bit_mask. bit_mask is a bit mask of the following values:

`1'

FF_EC_GUESS_MVS (default=enabled)

`2'

FF_EC_DEBLOCK (default=enabled)

`-bf frames'

use 'frames' B frames (supported for MPEG-1, MPEG-2 and MPEG-4)

`-mbd mode'

macroblock decision

`0'

FF_MB_DECISION_SIMPLE: use mb_cmp (cannot change it yet in ffmpeg)

`1'

FF_MB_DECISION_BITS: chooses the one which needs the fewest bits

`2'

FF_MB_DECISION_RD: rate distoration

`-4mv'

use four motion vector by macroblock (only MPEG-4)

`-part'

use data partitioning (only MPEG-4)

`-bug param'

workaround not auto detected encoder bugs

`-strict strictness'

how strictly to follow the standarts

`-aic'

enable Advanced intra coding (h263+)

`-umv'

enable Unlimited Motion Vector (h263+)

`-deinterlace'

deinterlace pictures

`-interlace'

force interlacing support in encoder (only MPEG-2 and MPEG-4). Use this option if your input file is interlaced and if you want to keep the interlaced format for minimum losses. The alternative is to deinterlace the input stream with `-deinterlace', but deinterlacing introduces more losses.

`-psnr'

calculate PSNR of compressed frames

`-vstats'

dump video coding statistics to `vstats_HHMMSS.log'.

`-vhook module'

insert video processing module. module contains the module name and its parameters separated by spaces.

3.5 Audio Options

`-ar freq'

set the audio sampling freq (default = 44100 Hz)

`-ab bitrate'

set the audio bitrate in kbit/s (default = 64)

`-ac channels'

set the number of audio channels (default = 1)

`-an'

disable audio recording

`-acodec codec'

force audio codec to codec. Use the copy special value to tell that the raw codec data must be copied as is.

3.6 Audio/Video grab options

`-vd device'

set video grab device (e.g. `/dev/video0')

`-vc channel'

set video grab channel (DV1394 only)

`-tvstd standard'

set television standard (NTSC, PAL (SECAM))

`-dv1394'

set DV1394 grab

`-ad device'

set audio device (e.g. `/dev/dsp')

3.7 Advanced options

`-map file:stream'

set input stream mapping

`-debug'

print specific debug info

`-benchmark'

add timings for benchmarking

`-hex'

dump each input packet

`-bitexact'

only use bit exact algorithms (for codec testing)

`-ps size'

set packet size in bits

`-re'

read input at native frame rate. Mainly used to simulate a grab device.

`-loop'

loop over the input stream. Currently it works only for image streams. This option is used for ffserver automatic testing.

3.8 FFmpeg formula evaluator

When evaluating a rate control string, FFmpeg uses an internal formula evaluator.

The following binary operators are available: +, -, *, /, ^.

The following unary operators are available: +, -, (...).

The following functions are available:

sinh(x)

cosh(x)

tanh(x)

sin(x)

cos(x)

tan(x)

exp(x)

log(x)

squish(x)

gauss(x)

abs(x)

max(x, y)

min(x, y)

gt(x, y)

lt(x, y)

eq(x, y)

bits2qp(bits)

qp2bits(qp)

The following constants are available:

PI

E

iTex

pTex

tex

mv

fCode

iCount

mcVar

var

isI

isP

isB

avgQP

qComp

avgIITex

avgPITex

avgPPTex

avgBPTex

avgTex

3.9 Protocols

The filename can be `-' to read from the standard input or to write to the standard output.

ffmpeg handles also many protocols specified with the URL syntax.

Use 'ffmpeg -formats' to have a list of the supported protocols.

The protocol http: is currently used only to communicate with ffserver (see the ffserver documentation). When ffmpeg will be a video player it will also be used for streaming :-)

4. Tips

• For streaming at very low bit rate application, use a low frame rate and a small gop size. This is especially true for real video where the Linux player does not seem to be very fast, so it can miss frames. An example is:

  ffmpeg -g 3 -r 3 -t 10 -b 50 -s qcif -f rv10 /tmp/b.rm

• The parameter 'q' which is displayed while encoding is the current quantizer. The value of 1 indicates that a very good quality could be achieved. The value of 31 indicates the worst quality. If q=31 too often, it means that the encoder cannot compress enough to meet your bit rate. You must either increase the bit rate, decrease the frame rate or decrease the frame size.

• If your computer is not fast enough, you can speed up the compression at the expense of the compression ratio. You can use '-me zero' to speed up motion estimation, and '-intra' to disable completely motion estimation (you have only I frames, which means it is about as good as JPEG compression).

• To have very low bitrates in audio, reduce the sampling frequency (down to 22050 kHz for mpeg audio, 22050 or 11025 for ac3).

• To have a constant quality (but a variable bitrate), use the option '-qscale n' when 'n' is between 1 (excellent quality) and 31 (worst quality).

• When converting video files, you can use the '-sameq' option which uses in the encoder the same quality factor than in the decoder. It allows to be almost lossless in encoding.

5. Supported File Formats and Codecs

You can use the -formats option to have an exhaustive list.

5.1 File Formats

FFmpeg supports the following file formats through the libavformat library:

./configure --enable-mingw32 --cross-prefix=i386-mingw32msvc-

/**
 * @file mpeg.c
 * mpeg codec.
 * @author ...
 */

/**
 * Summary sentance.
 * more text ...
 * ...
 */
typedef struct Foobar{
    int var1; /**< var1 description */
    int var2; ///< var2 description
    /** var3 description */
    int var3;
} Foobar;

/**
 * Summary sentance.
 * more text ...
 * ...
 * @param my_parameter description of my_parameter
 * @return return value description
 */
int myfunc(int my_parameter)
...