1@chapter Muxers 2@c man begin MUXERS 3 4Muxers are configured elements in FFmpeg which allow writing 5multimedia streams to a particular type of file. 6 7When you configure your FFmpeg build, all the supported muxers 8are enabled by default. You can list all available muxers using the 9configure option @code{--list-muxers}. 10 11You can disable all the muxers with the configure option 12@code{--disable-muxers} and selectively enable / disable single muxers 13with the options @code{--enable-muxer=@var{MUXER}} / 14@code{--disable-muxer=@var{MUXER}}. 15 16The option @code{-formats} of the ff* tools will display the list of 17enabled muxers. 18 19A description of some of the currently available muxers follows. 20 21@anchor{aiff} 22@section aiff 23 24Audio Interchange File Format muxer. 25 26@subsection Options 27 28It accepts the following options: 29 30@table @option 31@item write_id3v2 32Enable ID3v2 tags writing when set to 1. Default is 0 (disabled). 33 34@item id3v2_version 35Select ID3v2 version to write. Currently only version 3 and 4 (aka. 36ID3v2.3 and ID3v2.4) are supported. The default is version 4. 37 38@end table 39 40@anchor{crc} 41@section crc 42 43CRC (Cyclic Redundancy Check) testing format. 44 45This muxer computes and prints the Adler-32 CRC of all the input audio 46and video frames. By default audio frames are converted to signed 4716-bit raw audio and video frames to raw video before computing the 48CRC. 49 50The output of the muxer consists of a single line of the form: 51CRC=0x@var{CRC}, where @var{CRC} is a hexadecimal number 0-padded to 528 digits containing the CRC for all the decoded input frames. 53 54See also the @ref{framecrc} muxer. 55 56@subsection Examples 57 58For example to compute the CRC of the input, and store it in the file 59@file{out.crc}: 60@example 61ffmpeg -i INPUT -f crc out.crc 62@end example 63 64You can print the CRC to stdout with the command: 65@example 66ffmpeg -i INPUT -f crc - 67@end example 68 69You can select the output format of each frame with @command{ffmpeg} by 70specifying the audio and video codec and format. For example to 71compute the CRC of the input audio converted to PCM unsigned 8-bit 72and the input video converted to MPEG-2 video, use the command: 73@example 74ffmpeg -i INPUT -c:a pcm_u8 -c:v mpeg2video -f crc - 75@end example 76 77@anchor{framecrc} 78@section framecrc 79 80Per-packet CRC (Cyclic Redundancy Check) testing format. 81 82This muxer computes and prints the Adler-32 CRC for each audio 83and video packet. By default audio frames are converted to signed 8416-bit raw audio and video frames to raw video before computing the 85CRC. 86 87The output of the muxer consists of a line for each audio and video 88packet of the form: 89@example 90@var{stream_index}, @var{packet_dts}, @var{packet_pts}, @var{packet_duration}, @var{packet_size}, 0x@var{CRC} 91@end example 92 93@var{CRC} is a hexadecimal number 0-padded to 8 digits containing the 94CRC of the packet. 95 96@subsection Examples 97 98For example to compute the CRC of the audio and video frames in 99@file{INPUT}, converted to raw audio and video packets, and store it 100in the file @file{out.crc}: 101@example 102ffmpeg -i INPUT -f framecrc out.crc 103@end example 104 105To print the information to stdout, use the command: 106@example 107ffmpeg -i INPUT -f framecrc - 108@end example 109 110With @command{ffmpeg}, you can select the output format to which the 111audio and video frames are encoded before computing the CRC for each 112packet by specifying the audio and video codec. For example, to 113compute the CRC of each decoded input audio frame converted to PCM 114unsigned 8-bit and of each decoded input video frame converted to 115MPEG-2 video, use the command: 116@example 117ffmpeg -i INPUT -c:a pcm_u8 -c:v mpeg2video -f framecrc - 118@end example 119 120See also the @ref{crc} muxer. 121 122@anchor{framemd5} 123@section framemd5 124 125Per-packet MD5 testing format. 126 127This muxer computes and prints the MD5 hash for each audio 128and video packet. By default audio frames are converted to signed 12916-bit raw audio and video frames to raw video before computing the 130hash. 131 132The output of the muxer consists of a line for each audio and video 133packet of the form: 134@example 135@var{stream_index}, @var{packet_dts}, @var{packet_pts}, @var{packet_duration}, @var{packet_size}, @var{MD5} 136@end example 137 138@var{MD5} is a hexadecimal number representing the computed MD5 hash 139for the packet. 140 141@subsection Examples 142 143For example to compute the MD5 of the audio and video frames in 144@file{INPUT}, converted to raw audio and video packets, and store it 145in the file @file{out.md5}: 146@example 147ffmpeg -i INPUT -f framemd5 out.md5 148@end example 149 150To print the information to stdout, use the command: 151@example 152ffmpeg -i INPUT -f framemd5 - 153@end example 154 155See also the @ref{md5} muxer. 156 157@anchor{gif} 158@section gif 159 160Animated GIF muxer. 161 162It accepts the following options: 163 164@table @option 165@item loop 166Set the number of times to loop the output. Use @code{-1} for no loop, @code{0} 167for looping indefinitely (default). 168 169@item final_delay 170Force the delay (expressed in centiseconds) after the last frame. Each frame 171ends with a delay until the next frame. The default is @code{-1}, which is a 172special value to tell the muxer to re-use the previous delay. In case of a 173loop, you might want to customize this value to mark a pause for instance. 174@end table 175 176For example, to encode a gif looping 10 times, with a 5 seconds delay between 177the loops: 178@example 179ffmpeg -i INPUT -loop 10 -final_delay 500 out.gif 180@end example 181 182Note 1: if you wish to extract the frames in separate GIF files, you need to 183force the @ref{image2} muxer: 184@example 185ffmpeg -i INPUT -c:v gif -f image2 "out%d.gif" 186@end example 187 188Note 2: the GIF format has a very small time base: the delay between two frames 189can not be smaller than one centi second. 190 191@anchor{hls} 192@section hls 193 194Apple HTTP Live Streaming muxer that segments MPEG-TS according to 195the HTTP Live Streaming (HLS) specification. 196 197It creates a playlist file and numbered segment files. The output 198filename specifies the playlist filename; the segment filenames 199receive the same basename as the playlist, a sequential number and 200a .ts extension. 201 202For example, to convert an input file with @command{ffmpeg}: 203@example 204ffmpeg -i in.nut out.m3u8 205@end example 206 207See also the @ref{segment} muxer, which provides a more generic and 208flexible implementation of a segmenter, and can be used to perform HLS 209segmentation. 210 211@subsection Options 212 213This muxer supports the following options: 214 215@table @option 216@item hls_time @var{seconds} 217Set the segment length in seconds. Default value is 2. 218 219@item hls_list_size @var{size} 220Set the maximum number of playlist entries. If set to 0 the list file 221will contain all the segments. Default value is 5. 222 223@item hls_wrap @var{wrap} 224Set the number after which the segment filename number (the number 225specified in each segment file) wraps. If set to 0 the number will be 226never wrapped. Default value is 0. 227 228This option is useful to avoid to fill the disk with many segment 229files, and limits the maximum number of segment files written to disk 230to @var{wrap}. 231 232@item start_number @var{number} 233Start the playlist sequence number from @var{number}. Default value is 2340. 235 236@item hls_base_url @var{baseurl} 237Append @var{baseurl} to every entry in the playlist. 238Useful to generate playlists with absolute paths. 239 240Note that the playlist sequence number must be unique for each segment 241and it is not to be confused with the segment filename sequence number 242which can be cyclic, for example if the @option{wrap} option is 243specified. 244@end table 245 246@anchor{ico} 247@section ico 248 249ICO file muxer. 250 251Microsoft's icon file format (ICO) has some strict limitations that should be noted: 252 253@itemize 254@item 255Size cannot exceed 256 pixels in any dimension 256 257@item 258Only BMP and PNG images can be stored 259 260@item 261If a BMP image is used, it must be one of the following pixel formats: 262@example 263BMP Bit Depth FFmpeg Pixel Format 2641bit pal8 2654bit pal8 2668bit pal8 26716bit rgb555le 26824bit bgr24 26932bit bgra 270@end example 271 272@item 273If a BMP image is used, it must use the BITMAPINFOHEADER DIB header 274 275@item 276If a PNG image is used, it must use the rgba pixel format 277@end itemize 278 279@anchor{image2} 280@section image2 281 282Image file muxer. 283 284The image file muxer writes video frames to image files. 285 286The output filenames are specified by a pattern, which can be used to 287produce sequentially numbered series of files. 288The pattern may contain the string "%d" or "%0@var{N}d", this string 289specifies the position of the characters representing a numbering in 290the filenames. If the form "%0@var{N}d" is used, the string 291representing the number in each filename is 0-padded to @var{N} 292digits. The literal character '%' can be specified in the pattern with 293the string "%%". 294 295If the pattern contains "%d" or "%0@var{N}d", the first filename of 296the file list specified will contain the number 1, all the following 297numbers will be sequential. 298 299The pattern may contain a suffix which is used to automatically 300determine the format of the image files to write. 301 302For example the pattern "img-%03d.bmp" will specify a sequence of 303filenames of the form @file{img-001.bmp}, @file{img-002.bmp}, ..., 304@file{img-010.bmp}, etc. 305The pattern "img%%-%d.jpg" will specify a sequence of filenames of the 306form @file{img%-1.jpg}, @file{img%-2.jpg}, ..., @file{img%-10.jpg}, 307etc. 308 309@subsection Examples 310 311The following example shows how to use @command{ffmpeg} for creating a 312sequence of files @file{img-001.jpeg}, @file{img-002.jpeg}, ..., 313taking one image every second from the input video: 314@example 315ffmpeg -i in.avi -vsync 1 -r 1 -f image2 'img-%03d.jpeg' 316@end example 317 318Note that with @command{ffmpeg}, if the format is not specified with the 319@code{-f} option and the output filename specifies an image file 320format, the image2 muxer is automatically selected, so the previous 321command can be written as: 322@example 323ffmpeg -i in.avi -vsync 1 -r 1 'img-%03d.jpeg' 324@end example 325 326Note also that the pattern must not necessarily contain "%d" or 327"%0@var{N}d", for example to create a single image file 328@file{img.jpeg} from the input video you can employ the command: 329@example 330ffmpeg -i in.avi -f image2 -frames:v 1 img.jpeg 331@end example 332 333The @option{strftime} option allows you to expand the filename with 334date and time information. Check the documentation of 335the @code{strftime()} function for the syntax. 336 337For example to generate image files from the @code{strftime()} 338"%Y-%m-%d_%H-%M-%S" pattern, the following @command{ffmpeg} command 339can be used: 340@example 341ffmpeg -f v4l2 -r 1 -i /dev/video0 -f image2 -strftime 1 "%Y-%m-%d_%H-%M-%S.jpg" 342@end example 343 344@subsection Options 345 346@table @option 347@item start_number 348Start the sequence from the specified number. Default value is 1. Must 349be a non-negative number. 350 351@item update 352If set to 1, the filename will always be interpreted as just a 353filename, not a pattern, and the corresponding file will be continuously 354overwritten with new images. Default value is 0. 355 356@item strftime 357If set to 1, expand the filename with date and time information from 358@code{strftime()}. Default value is 0. 359@end table 360 361The image muxer supports the .Y.U.V image file format. This format is 362special in that that each image frame consists of three files, for 363each of the YUV420P components. To read or write this image file format, 364specify the name of the '.Y' file. The muxer will automatically open the 365'.U' and '.V' files as required. 366 367@section matroska 368 369Matroska container muxer. 370 371This muxer implements the matroska and webm container specs. 372 373@subsection Metadata 374 375The recognized metadata settings in this muxer are: 376 377@table @option 378@item title 379Set title name provided to a single track. 380 381@item language 382Specify the language of the track in the Matroska languages form. 383 384The language can be either the 3 letters bibliographic ISO-639-2 (ISO 385639-2/B) form (like "fre" for French), or a language code mixed with a 386country code for specialities in languages (like "fre-ca" for Canadian 387French). 388 389@item stereo_mode 390Set stereo 3D video layout of two views in a single video track. 391 392The following values are recognized: 393@table @samp 394@item mono 395video is not stereo 396@item left_right 397Both views are arranged side by side, Left-eye view is on the left 398@item bottom_top 399Both views are arranged in top-bottom orientation, Left-eye view is at bottom 400@item top_bottom 401Both views are arranged in top-bottom orientation, Left-eye view is on top 402@item checkerboard_rl 403Each view is arranged in a checkerboard interleaved pattern, Left-eye view being first 404@item checkerboard_lr 405Each view is arranged in a checkerboard interleaved pattern, Right-eye view being first 406@item row_interleaved_rl 407Each view is constituted by a row based interleaving, Right-eye view is first row 408@item row_interleaved_lr 409Each view is constituted by a row based interleaving, Left-eye view is first row 410@item col_interleaved_rl 411Both views are arranged in a column based interleaving manner, Right-eye view is first column 412@item col_interleaved_lr 413Both views are arranged in a column based interleaving manner, Left-eye view is first column 414@item anaglyph_cyan_red 415All frames are in anaglyph format viewable through red-cyan filters 416@item right_left 417Both views are arranged side by side, Right-eye view is on the left 418@item anaglyph_green_magenta 419All frames are in anaglyph format viewable through green-magenta filters 420@item block_lr 421Both eyes laced in one Block, Left-eye view is first 422@item block_rl 423Both eyes laced in one Block, Right-eye view is first 424@end table 425@end table 426 427For example a 3D WebM clip can be created using the following command line: 428@example 429ffmpeg -i sample_left_right_clip.mpg -an -c:v libvpx -metadata stereo_mode=left_right -y stereo_clip.webm 430@end example 431 432@subsection Options 433 434This muxer supports the following options: 435 436@table @option 437@item reserve_index_space 438By default, this muxer writes the index for seeking (called cues in Matroska 439terms) at the end of the file, because it cannot know in advance how much space 440to leave for the index at the beginning of the file. However for some use cases 441-- e.g. streaming where seeking is possible but slow -- it is useful to put the 442index at the beginning of the file. 443 444If this option is set to a non-zero value, the muxer will reserve a given amount 445of space in the file header and then try to write the cues there when the muxing 446finishes. If the available space does not suffice, muxing will fail. A safe size 447for most use cases should be about 50kB per hour of video. 448 449Note that cues are only written if the output is seekable and this option will 450have no effect if it is not. 451@end table 452 453@anchor{md5} 454@section md5 455 456MD5 testing format. 457 458This muxer computes and prints the MD5 hash of all the input audio 459and video frames. By default audio frames are converted to signed 46016-bit raw audio and video frames to raw video before computing the 461hash. 462 463The output of the muxer consists of a single line of the form: 464MD5=@var{MD5}, where @var{MD5} is a hexadecimal number representing 465the computed MD5 hash. 466 467For example to compute the MD5 hash of the input converted to raw 468audio and video, and store it in the file @file{out.md5}: 469@example 470ffmpeg -i INPUT -f md5 out.md5 471@end example 472 473You can print the MD5 to stdout with the command: 474@example 475ffmpeg -i INPUT -f md5 - 476@end example 477 478See also the @ref{framemd5} muxer. 479 480@section mov, mp4, ismv 481 482MOV/MP4/ISMV (Smooth Streaming) muxer. 483 484The mov/mp4/ismv muxer supports fragmentation. Normally, a MOV/MP4 485file has all the metadata about all packets stored in one location 486(written at the end of the file, it can be moved to the start for 487better playback by adding @var{faststart} to the @var{movflags}, or 488using the @command{qt-faststart} tool). A fragmented 489file consists of a number of fragments, where packets and metadata 490about these packets are stored together. Writing a fragmented 491file has the advantage that the file is decodable even if the 492writing is interrupted (while a normal MOV/MP4 is undecodable if 493it is not properly finished), and it requires less memory when writing 494very long files (since writing normal MOV/MP4 files stores info about 495every single packet in memory until the file is closed). The downside 496is that it is less compatible with other applications. 497 498@subsection Options 499 500Fragmentation is enabled by setting one of the AVOptions that define 501how to cut the file into fragments: 502 503@table @option 504@item -moov_size @var{bytes} 505Reserves space for the moov atom at the beginning of the file instead of placing the 506moov atom at the end. If the space reserved is insufficient, muxing will fail. 507@item -movflags frag_keyframe 508Start a new fragment at each video keyframe. 509@item -frag_duration @var{duration} 510Create fragments that are @var{duration} microseconds long. 511@item -frag_size @var{size} 512Create fragments that contain up to @var{size} bytes of payload data. 513@item -movflags frag_custom 514Allow the caller to manually choose when to cut fragments, by 515calling @code{av_write_frame(ctx, NULL)} to write a fragment with 516the packets written so far. (This is only useful with other 517applications integrating libavformat, not from @command{ffmpeg}.) 518@item -min_frag_duration @var{duration} 519Don't create fragments that are shorter than @var{duration} microseconds long. 520@end table 521 522If more than one condition is specified, fragments are cut when 523one of the specified conditions is fulfilled. The exception to this is 524@code{-min_frag_duration}, which has to be fulfilled for any of the other 525conditions to apply. 526 527Additionally, the way the output file is written can be adjusted 528through a few other options: 529 530@table @option 531@item -movflags empty_moov 532Write an initial moov atom directly at the start of the file, without 533describing any samples in it. Generally, an mdat/moov pair is written 534at the start of the file, as a normal MOV/MP4 file, containing only 535a short portion of the file. With this option set, there is no initial 536mdat atom, and the moov atom only describes the tracks but has 537a zero duration. 538 539Files written with this option set do not work in QuickTime. 540This option is implicitly set when writing ismv (Smooth Streaming) files. 541@item -movflags separate_moof 542Write a separate moof (movie fragment) atom for each track. Normally, 543packets for all tracks are written in a moof atom (which is slightly 544more efficient), but with this option set, the muxer writes one moof/mdat 545pair for each track, making it easier to separate tracks. 546 547This option is implicitly set when writing ismv (Smooth Streaming) files. 548@item -movflags faststart 549Run a second pass moving the index (moov atom) to the beginning of the file. 550This operation can take a while, and will not work in various situations such 551as fragmented output, thus it is not enabled by default. 552@item -movflags rtphint 553Add RTP hinting tracks to the output file. 554@end table 555 556@subsection Example 557 558Smooth Streaming content can be pushed in real time to a publishing 559point on IIS with this muxer. Example: 560@example 561ffmpeg -re @var{<normal input/transcoding options>} -movflags isml+frag_keyframe -f ismv http://server/publishingpoint.isml/Streams(Encoder1) 562@end example 563 564@section mp3 565 566The MP3 muxer writes a raw MP3 stream with an ID3v2 header at the beginning and 567optionally an ID3v1 tag at the end. ID3v2.3 and ID3v2.4 are supported, the 568@code{id3v2_version} option controls which one is used. Setting 569@code{id3v2_version} to 0 will disable the ID3v2 header completely. The legacy 570ID3v1 tag is not written by default, but may be enabled with the 571@code{write_id3v1} option. 572 573The muxer may also write a Xing frame at the beginning, which contains the 574number of frames in the file. It is useful for computing duration of VBR files. 575The Xing frame is written if the output stream is seekable and if the 576@code{write_xing} option is set to 1 (the default). 577 578The muxer supports writing ID3v2 attached pictures (APIC frames). The pictures 579are supplied to the muxer in form of a video stream with a single packet. There 580can be any number of those streams, each will correspond to a single APIC frame. 581The stream metadata tags @var{title} and @var{comment} map to APIC 582@var{description} and @var{picture type} respectively. See 583@url{http://id3.org/id3v2.4.0-frames} for allowed picture types. 584 585Note that the APIC frames must be written at the beginning, so the muxer will 586buffer the audio frames until it gets all the pictures. It is therefore advised 587to provide the pictures as soon as possible to avoid excessive buffering. 588 589Examples: 590 591Write an mp3 with an ID3v2.3 header and an ID3v1 footer: 592@example 593ffmpeg -i INPUT -id3v2_version 3 -write_id3v1 1 out.mp3 594@end example 595 596To attach a picture to an mp3 file select both the audio and the picture stream 597with @code{map}: 598@example 599ffmpeg -i input.mp3 -i cover.png -c copy -map 0 -map 1 600-metadata:s:v title="Album cover" -metadata:s:v comment="Cover (Front)" out.mp3 601@end example 602 603Write a "clean" MP3 without any extra features: 604@example 605ffmpeg -i input.wav -write_xing 0 -id3v2_version 0 out.mp3 606@end example 607 608@section mpegts 609 610MPEG transport stream muxer. 611 612This muxer implements ISO 13818-1 and part of ETSI EN 300 468. 613 614The recognized metadata settings in mpegts muxer are @code{service_provider} 615and @code{service_name}. If they are not set the default for 616@code{service_provider} is "FFmpeg" and the default for 617@code{service_name} is "Service01". 618 619@subsection Options 620 621The muxer options are: 622 623@table @option 624@item -mpegts_original_network_id @var{number} 625Set the original_network_id (default 0x0001). This is unique identifier 626of a network in DVB. Its main use is in the unique identification of a 627service through the path Original_Network_ID, Transport_Stream_ID. 628@item -mpegts_transport_stream_id @var{number} 629Set the transport_stream_id (default 0x0001). This identifies a 630transponder in DVB. 631@item -mpegts_service_id @var{number} 632Set the service_id (default 0x0001) also known as program in DVB. 633@item -mpegts_pmt_start_pid @var{number} 634Set the first PID for PMT (default 0x1000, max 0x1f00). 635@item -mpegts_start_pid @var{number} 636Set the first PID for data packets (default 0x0100, max 0x0f00). 637@item -mpegts_m2ts_mode @var{number} 638Enable m2ts mode if set to 1. Default value is -1 which disables m2ts mode. 639@item -muxrate @var{number} 640Set a constant muxrate (default VBR). 641@item -pcr_period @var{numer} 642Override the default PCR retransmission time (default 20ms), ignored 643if variable muxrate is selected. 644@item -pes_payload_size @var{number} 645Set minimum PES packet payload in bytes. 646@item -mpegts_flags @var{flags} 647Set flags (see below). 648@item -mpegts_copyts @var{number} 649Preserve original timestamps, if value is set to 1. Default value is -1, which 650results in shifting timestamps so that they start from 0. 651@item -tables_version @var{number} 652Set PAT, PMT and SDT version (default 0, valid values are from 0 to 31, inclusively). 653This option allows updating stream structure so that standard consumer may 654detect the change. To do so, reopen output AVFormatContext (in case of API 655usage) or restart ffmpeg instance, cyclically changing tables_version value: 656@example 657ffmpeg -i source1.ts -codec copy -f mpegts -tables_version 0 udp://1.1.1.1:1111 658ffmpeg -i source2.ts -codec copy -f mpegts -tables_version 1 udp://1.1.1.1:1111 659... 660ffmpeg -i source3.ts -codec copy -f mpegts -tables_version 31 udp://1.1.1.1:1111 661ffmpeg -i source1.ts -codec copy -f mpegts -tables_version 0 udp://1.1.1.1:1111 662ffmpeg -i source2.ts -codec copy -f mpegts -tables_version 1 udp://1.1.1.1:1111 663... 664@end example 665@end table 666 667Option mpegts_flags may take a set of such flags: 668 669@table @option 670@item resend_headers 671Reemit PAT/PMT before writing the next packet. 672@item latm 673Use LATM packetization for AAC. 674@end table 675 676@subsection Example 677 678@example 679ffmpeg -i file.mpg -c copy \ 680 -mpegts_original_network_id 0x1122 \ 681 -mpegts_transport_stream_id 0x3344 \ 682 -mpegts_service_id 0x5566 \ 683 -mpegts_pmt_start_pid 0x1500 \ 684 -mpegts_start_pid 0x150 \ 685 -metadata service_provider="Some provider" \ 686 -metadata service_name="Some Channel" \ 687 -y out.ts 688@end example 689 690@section null 691 692Null muxer. 693 694This muxer does not generate any output file, it is mainly useful for 695testing or benchmarking purposes. 696 697For example to benchmark decoding with @command{ffmpeg} you can use the 698command: 699@example 700ffmpeg -benchmark -i INPUT -f null out.null 701@end example 702 703Note that the above command does not read or write the @file{out.null} 704file, but specifying the output file is required by the @command{ffmpeg} 705syntax. 706 707Alternatively you can write the command as: 708@example 709ffmpeg -benchmark -i INPUT -f null - 710@end example 711 712@section nut 713 714@table @option 715@item -syncpoints @var{flags} 716Change the syncpoint usage in nut: 717@table @option 718@item @var{default} use the normal low-overhead seeking aids. 719@item @var{none} do not use the syncpoints at all, reducing the overhead but making the stream non-seekable; 720 Use of this option is not recommended, as the resulting files are very damage 721 sensitive and seeking is not possible. Also in general the overhead from 722 syncpoints is negligible. Note, -@code{write_index} 0 can be used to disable 723 all growing data tables, allowing to mux endless streams with limited memory 724 and wihout these disadvantages. 725@item @var{timestamped} extend the syncpoint with a wallclock field. 726@end table 727The @var{none} and @var{timestamped} flags are experimental. 728@item -write_index @var{bool} 729Write index at the end, the default is to write an index. 730@end table 731 732@example 733ffmpeg -i INPUT -f_strict experimental -syncpoints none - | processor 734@end example 735 736@section ogg 737 738Ogg container muxer. 739 740@table @option 741@item -page_duration @var{duration} 742Preferred page duration, in microseconds. The muxer will attempt to create 743pages that are approximately @var{duration} microseconds long. This allows the 744user to compromise between seek granularity and container overhead. The default 745is 1 second. A value of 0 will fill all segments, making pages as large as 746possible. A value of 1 will effectively use 1 packet-per-page in most 747situations, giving a small seek granularity at the cost of additional container 748overhead. 749@end table 750 751@anchor{segment} 752@section segment, stream_segment, ssegment 753 754Basic stream segmenter. 755 756This muxer outputs streams to a number of separate files of nearly 757fixed duration. Output filename pattern can be set in a fashion similar to 758@ref{image2}. 759 760@code{stream_segment} is a variant of the muxer used to write to 761streaming output formats, i.e. which do not require global headers, 762and is recommended for outputting e.g. to MPEG transport stream segments. 763@code{ssegment} is a shorter alias for @code{stream_segment}. 764 765Every segment starts with a keyframe of the selected reference stream, 766which is set through the @option{reference_stream} option. 767 768Note that if you want accurate splitting for a video file, you need to 769make the input key frames correspond to the exact splitting times 770expected by the segmenter, or the segment muxer will start the new 771segment with the key frame found next after the specified start 772time. 773 774The segment muxer works best with a single constant frame rate video. 775 776Optionally it can generate a list of the created segments, by setting 777the option @var{segment_list}. The list type is specified by the 778@var{segment_list_type} option. The entry filenames in the segment 779list are set by default to the basename of the corresponding segment 780files. 781 782See also the @ref{hls} muxer, which provides a more specific 783implementation for HLS segmentation. 784 785@subsection Options 786 787The segment muxer supports the following options: 788 789@table @option 790@item reference_stream @var{specifier} 791Set the reference stream, as specified by the string @var{specifier}. 792If @var{specifier} is set to @code{auto}, the reference is chosen 793automatically. Otherwise it must be a stream specifier (see the ``Stream 794specifiers'' chapter in the ffmpeg manual) which specifies the 795reference stream. The default value is @code{auto}. 796 797@item segment_format @var{format} 798Override the inner container format, by default it is guessed by the filename 799extension. 800 801@item segment_list @var{name} 802Generate also a listfile named @var{name}. If not specified no 803listfile is generated. 804 805@item segment_list_flags @var{flags} 806Set flags affecting the segment list generation. 807 808It currently supports the following flags: 809@table @samp 810@item cache 811Allow caching (only affects M3U8 list files). 812 813@item live 814Allow live-friendly file generation. 815@end table 816 817@item segment_list_type @var{type} 818Select the listing format. 819@table @option 820@item @var{flat} use a simple flat list of entries. 821@item @var{hls} use a m3u8-like structure. 822@end table 823 824@item segment_list_size @var{size} 825Update the list file so that it contains at most the last @var{size} 826segments. If 0 the list file will contain all the segments. Default 827value is 0. 828 829@item segment_list_entry_prefix @var{prefix} 830Prepend @var{prefix} to each entry. Useful to generate absolute paths. 831By default no prefix is applied. 832 833The following values are recognized: 834@table @samp 835@item flat 836Generate a flat list for the created segments, one segment per line. 837 838@item csv, ext 839Generate a list for the created segments, one segment per line, 840each line matching the format (comma-separated values): 841@example 842@var{segment_filename},@var{segment_start_time},@var{segment_end_time} 843@end example 844 845@var{segment_filename} is the name of the output file generated by the 846muxer according to the provided pattern. CSV escaping (according to 847RFC4180) is applied if required. 848 849@var{segment_start_time} and @var{segment_end_time} specify 850the segment start and end time expressed in seconds. 851 852A list file with the suffix @code{".csv"} or @code{".ext"} will 853auto-select this format. 854 855@samp{ext} is deprecated in favor or @samp{csv}. 856 857@item ffconcat 858Generate an ffconcat file for the created segments. The resulting file 859can be read using the FFmpeg @ref{concat} demuxer. 860 861A list file with the suffix @code{".ffcat"} or @code{".ffconcat"} will 862auto-select this format. 863 864@item m3u8 865Generate an extended M3U8 file, version 3, compliant with 866@url{http://tools.ietf.org/id/draft-pantos-http-live-streaming}. 867 868A list file with the suffix @code{".m3u8"} will auto-select this format. 869@end table 870 871If not specified the type is guessed from the list file name suffix. 872 873@item segment_time @var{time} 874Set segment duration to @var{time}, the value must be a duration 875specification. Default value is "2". See also the 876@option{segment_times} option. 877 878Note that splitting may not be accurate, unless you force the 879reference stream key-frames at the given time. See the introductory 880notice and the examples below. 881 882@item segment_atclocktime @var{1|0} 883If set to "1" split at regular clock time intervals starting from 00:00 884o'clock. The @var{time} value specified in @option{segment_time} is 885used for setting the length of the splitting interval. 886 887For example with @option{segment_time} set to "900" this makes it possible 888to create files at 12:00 o'clock, 12:15, 12:30, etc. 889 890Default value is "0". 891 892@item segment_time_delta @var{delta} 893Specify the accuracy time when selecting the start time for a 894segment, expressed as a duration specification. Default value is "0". 895 896When delta is specified a key-frame will start a new segment if its 897PTS satisfies the relation: 898@example 899PTS >= start_time - time_delta 900@end example 901 902This option is useful when splitting video content, which is always 903split at GOP boundaries, in case a key frame is found just before the 904specified split time. 905 906In particular may be used in combination with the @file{ffmpeg} option 907@var{force_key_frames}. The key frame times specified by 908@var{force_key_frames} may not be set accurately because of rounding 909issues, with the consequence that a key frame time may result set just 910before the specified time. For constant frame rate videos a value of 9111/(2*@var{frame_rate}) should address the worst case mismatch between 912the specified time and the time set by @var{force_key_frames}. 913 914@item segment_times @var{times} 915Specify a list of split points. @var{times} contains a list of comma 916separated duration specifications, in increasing order. See also 917the @option{segment_time} option. 918 919@item segment_frames @var{frames} 920Specify a list of split video frame numbers. @var{frames} contains a 921list of comma separated integer numbers, in increasing order. 922 923This option specifies to start a new segment whenever a reference 924stream key frame is found and the sequential number (starting from 0) 925of the frame is greater or equal to the next value in the list. 926 927@item segment_wrap @var{limit} 928Wrap around segment index once it reaches @var{limit}. 929 930@item segment_start_number @var{number} 931Set the sequence number of the first segment. Defaults to @code{0}. 932 933@item reset_timestamps @var{1|0} 934Reset timestamps at the begin of each segment, so that each segment 935will start with near-zero timestamps. It is meant to ease the playback 936of the generated segments. May not work with some combinations of 937muxers/codecs. It is set to @code{0} by default. 938 939@item initial_offset @var{offset} 940Specify timestamp offset to apply to the output packet timestamps. The 941argument must be a time duration specification, and defaults to 0. 942@end table 943 944@subsection Examples 945 946@itemize 947@item 948To remux the content of file @file{in.mkv} to a list of segments 949@file{out-000.nut}, @file{out-001.nut}, etc., and write the list of 950generated segments to @file{out.list}: 951@example 952ffmpeg -i in.mkv -codec copy -map 0 -f segment -segment_list out.list out%03d.nut 953@end example 954 955@item 956As the example above, but segment the input file according to the split 957points specified by the @var{segment_times} option: 958@example 959ffmpeg -i in.mkv -codec copy -map 0 -f segment -segment_list out.csv -segment_times 1,2,3,5,8,13,21 out%03d.nut 960@end example 961 962@item 963As the example above, but use the @command{ffmpeg} @option{force_key_frames} 964option to force key frames in the input at the specified location, together 965with the segment option @option{segment_time_delta} to account for 966possible roundings operated when setting key frame times. 967@example 968ffmpeg -i in.mkv -force_key_frames 1,2,3,5,8,13,21 -codec:v mpeg4 -codec:a pcm_s16le -map 0 \ 969-f segment -segment_list out.csv -segment_times 1,2,3,5,8,13,21 -segment_time_delta 0.05 out%03d.nut 970@end example 971In order to force key frames on the input file, transcoding is 972required. 973 974@item 975Segment the input file by splitting the input file according to the 976frame numbers sequence specified with the @option{segment_frames} option: 977@example 978ffmpeg -i in.mkv -codec copy -map 0 -f segment -segment_list out.csv -segment_frames 100,200,300,500,800 out%03d.nut 979@end example 980 981@item 982To convert the @file{in.mkv} to TS segments using the @code{libx264} 983and @code{libfaac} encoders: 984@example 985ffmpeg -i in.mkv -map 0 -codec:v libx264 -codec:a libfaac -f ssegment -segment_list out.list out%03d.ts 986@end example 987 988@item 989Segment the input file, and create an M3U8 live playlist (can be used 990as live HLS source): 991@example 992ffmpeg -re -i in.mkv -codec copy -map 0 -f segment -segment_list playlist.m3u8 \ 993-segment_list_flags +live -segment_time 10 out%03d.mkv 994@end example 995@end itemize 996 997@section smoothstreaming 998 999Smooth Streaming muxer generates a set of files (Manifest, chunks) suitable for serving with conventional web server. 1000 1001@table @option 1002@item window_size 1003Specify the number of fragments kept in the manifest. Default 0 (keep all). 1004 1005@item extra_window_size 1006Specify the number of fragments kept outside of the manifest before removing from disk. Default 5. 1007 1008@item lookahead_count 1009Specify the number of lookahead fragments. Default 2. 1010 1011@item min_frag_duration 1012Specify the minimum fragment duration (in microseconds). Default 5000000. 1013 1014@item remove_at_exit 1015Specify whether to remove all fragments when finished. Default 0 (do not remove). 1016 1017@end table 1018 1019@section tee 1020 1021The tee muxer can be used to write the same data to several files or any 1022other kind of muxer. It can be used, for example, to both stream a video to 1023the network and save it to disk at the same time. 1024 1025It is different from specifying several outputs to the @command{ffmpeg} 1026command-line tool because the audio and video data will be encoded only once 1027with the tee muxer; encoding can be a very expensive process. It is not 1028useful when using the libavformat API directly because it is then possible 1029to feed the same packets to several muxers directly. 1030 1031The slave outputs are specified in the file name given to the muxer, 1032separated by '|'. If any of the slave name contains the '|' separator, 1033leading or trailing spaces or any special character, it must be 1034escaped (see @ref{quoting_and_escaping,,the "Quoting and escaping" 1035section in the ffmpeg-utils(1) manual,ffmpeg-utils}). 1036 1037Muxer options can be specified for each slave by prepending them as a list of 1038@var{key}=@var{value} pairs separated by ':', between square brackets. If 1039the options values contain a special character or the ':' separator, they 1040must be escaped; note that this is a second level escaping. 1041 1042The following special options are also recognized: 1043@table @option 1044@item f 1045Specify the format name. Useful if it cannot be guessed from the 1046output name suffix. 1047 1048@item bsfs[/@var{spec}] 1049Specify a list of bitstream filters to apply to the specified 1050output. 1051 1052It is possible to specify to which streams a given bitstream filter 1053applies, by appending a stream specifier to the option separated by 1054@code{/}. @var{spec} must be a stream specifier (see @ref{Format 1055stream specifiers}). If the stream specifier is not specified, the 1056bitstream filters will be applied to all streams in the output. 1057 1058Several bitstream filters can be specified, separated by ",". 1059 1060@item select 1061Select the streams that should be mapped to the slave output, 1062specified by a stream specifier. If not specified, this defaults to 1063all the input streams. 1064@end table 1065 1066@subsection Examples 1067 1068@itemize 1069@item 1070Encode something and both archive it in a WebM file and stream it 1071as MPEG-TS over UDP (the streams need to be explicitly mapped): 1072@example 1073ffmpeg -i ... -c:v libx264 -c:a mp2 -f tee -map 0:v -map 0:a 1074 "archive-20121107.mkv|[f=mpegts]udp://10.0.1.255:1234/" 1075@end example 1076 1077@item 1078Use @command{ffmpeg} to encode the input, and send the output 1079to three different destinations. The @code{dump_extra} bitstream 1080filter is used to add extradata information to all the output video 1081keyframes packets, as requested by the MPEG-TS format. The select 1082option is applied to @file{out.aac} in order to make it contain only 1083audio packets. 1084@example 1085ffmpeg -i ... -map 0 -flags +global_header -c:v libx264 -c:a aac -strict experimental 1086 -f tee "[bsfs/v=dump_extra]out.ts|[movflags=+faststart]out.mp4|[select=a]out.aac" 1087@end example 1088 1089@item 1090As below, but select only stream @code{a:1} for the audio output. Note 1091that a second level escaping must be performed, as ":" is a special 1092character used to separate options. 1093@example 1094ffmpeg -i ... -map 0 -flags +global_header -c:v libx264 -c:a aac -strict experimental 1095 -f tee "[bsfs/v=dump_extra]out.ts|[movflags=+faststart]out.mp4|[select=\'a:1\']out.aac" 1096@end example 1097@end itemize 1098 1099Note: some codecs may need different options depending on the output format; 1100the auto-detection of this can not work with the tee muxer. The main example 1101is the @option{global_header} flag. 1102 1103@section webm_dash_manifest 1104 1105WebM DASH Manifest muxer. 1106 1107This muxer implements the WebM DASH Manifest specification to generate the DASH manifest XML. 1108 1109@subsection Options 1110 1111This muxer supports the following options: 1112 1113@table @option 1114@item adaptation_sets 1115This option has the following syntax: "id=x,streams=a,b,c id=y,streams=d,e" where x and y are the 1116unique identifiers of the adaptation sets and a,b,c,d and e are the indices of the corresponding 1117audio and video streams. Any number of adaptation sets can be added using this option. 1118@end table 1119 1120@subsection Example 1121@example 1122ffmpeg -f webm_dash_manifest -i video1.webm \ 1123 -f webm_dash_manifest -i video2.webm \ 1124 -f webm_dash_manifest -i audio1.webm \ 1125 -f webm_dash_manifest -i audio2.webm \ 1126 -map 0 -map 1 -map 2 -map 3 \ 1127 -c copy \ 1128 -f webm_dash_manifest \ 1129 -adaptation_sets "id=0,streams=0,1 id=1,streams=2,3" \ 1130 manifest.xml 1131@end example 1132 1133@c man end MUXERS 1134