mpv/DOCS/edl-mpv.rst

257 lines
9.2 KiB
ReStructuredText
Raw Normal View History

EDL files
=========
EDL files basically concatenate ranges of video/audio from multiple source
files into a single continuous virtual file. Each such range is called a
segment, and consists of source file, source offset, and segment length.
For example::
# mpv EDL v0
f1.mkv,10,20
f2.mkv
f1.mkv,40,10
This would skip the first 10 seconds of the file f1.mkv, then play the next
20 seconds, then switch to the file f2.mkv and play all of it, then switch
back to f1.mkv, skip to the 40 second mark, and play 10 seconds, and then
stop playback. The difference to specifying the files directly on command
line (and using ``--{ --start=10 --length=20 f1.mkv --}`` etc.) is that the
2013-12-05 19:16:15 +00:00
virtual EDL file appears as a virtual timeline (like a single file), instead
as a playlist.
2013-12-05 19:16:15 +00:00
The general simplified syntax is::
# mpv EDL v0
<filename>
<filename>,<start in seconds>,<length in seconds>
If the start time is omitted, 0 is used. If the length is omitted, the
2014-10-11 23:32:42 +00:00
estimated remaining duration of the source file is used.
Note::
Usage of relative or absolute paths as well as any protocol prefixes may be
prevented for security reasons.
Syntax of mpv EDL files
=======================
Generally, the format is relatively strict. No superfluous whitespace (except
empty lines and commented lines) are allowed. You must use UNIX line breaks.
The first line in the file must be ``# mpv EDL v0``. This designates that the
file uses format version 0, which is not frozen yet and may change any time.
(If you need a stable EDL file format, make a feature request. Likewise, if
you have suggestions for improvements, it's not too late yet.)
The rest of the lines belong to one of these classes:
1) An empty or commented line. A comment starts with ``#``, which must be the
first character in the line. The rest of the line (up until the next line
break) is ignored. An empty line has 0 bytes between two line feed bytes.
2) A header entry if the line starts with ``!``.
3) A segment entry in all other cases.
Each segment entry consists of a list of named or unnamed parameters.
Parameters are separated with ``,``. Named parameters consist of a name,
followed by ``=``, followed by the value. Unnamed parameters have only a
value, and the name is implicit from the parameter position.
Syntax::
segment_entry ::= <param> ( <param> ',' )*
param ::= [ <name> '=' ] ( <value> | '%' <number> '%' <valuebytes> )
The ``name`` string can consist of any characters, except ``=%,;\n!``. The
``value`` string can consist of any characters except of ``,;\n!``.
The construct starting with ``%`` allows defining any value with arbitrary
contents inline, where ``number`` is an integer giving the number of bytes in
``valuebytes``. If a parameter value contains disallowed characters, it has to
be guarded by a length specifier using this syntax.
The parameter name defines the meaning of the parameter:
1) ``file``, the source file to use for this segment.
2) ``start``, a time value that specifies the start offset into the source file.
3) ``length``, a time value that specifies the length of the segment.
See the section below for the format of timestamps.
Unnamed parameters carry implicit names. The parameter position determines
which of the parameters listed above is set. For example, the second parameter
implicitly uses the name ``start``.
Example::
# mpv EDL v0
%18%filename,with,.mkv,10,length=20,param3=%13%value,escaped,param4=value2
this sets ``file`` to ``filename,with,.mkv``, ``start`` to ``10``, ``length``
to ``20``, ``param3`` to ``value,escaped``, ``param4`` to ``value2``.
Instead of line breaks, the character ``;`` can be used. Line feed bytes and
``;`` are treated equally.
Header entries start with ``!`` as first character after a line break. Header
entries affect all other file entries in the EDL file. Their format is highly
implementation specific. They should generally follow the file header, and come
before any file entries.
Disabling chapter generation and copying
========================================
By default, chapters from the source ranges are copied to the virtual file's
chapters. Also, a chapter is inserted after each range. This can be disabled
with the ``no_chapters`` header.
Example::
!no_chapters
MP4 DASH
========
This is a header that helps implementing DASH, although it only provides a low
level mechanism.
If this header is set, the given url designates an mp4 init fragment. It's
downloaded, and every URL in the EDL is prefixed with the init fragment on the
byte stream level. This is mostly for use by mpv's internal ytdl support. The
ytdl script will call youtube-dl, which in turn actually processes DASH
manifests. It may work only for this very specific purpose and fail to be
2017-02-05 15:02:58 +00:00
useful in other scenarios. It can be removed or changed in incompatible ways
at any times.
Example::
!mp4_dash,init=url
The ``url`` is encoded as parameter value as defined in the general EDL syntax.
It's expected to point to an "initialization fragment", which will be prefixed
to every entry in the EDL on the byte stream level.
The current implementation will
- ignore stream start times
- use durations as hint for seeking only
- not adjust source timestamps
- open and close segments (i.e. fragments) as needed
- not add segment boundaries as chapter points
- require full compatibility between all segments (same codec etc.)
Another header part of this mechanism is ``no_clip``. This header is similar
to ``mp4_dash``, but does not include on-demand opening/closing of segments,
and does not support init segments. It also exists solely to support internal
ytdl requirements.
The ``mp4_dash`` and ``no_clip`` headers are not part of the core EDL format.
They may be changed or removed at any time, depending on mpv's internal
requirements.
demux, demux_edl: add extension for tracks sourced from separate streams This commit adds an extension to mpv EDL, which basically allows you to do the same as --audio-file, --external-file, etc. in a single EDL file. This is a relatively quick & dirty implementation. The dirty part lies in the fact that several shortcuts are taken. For example, struct timeline now forms a singly linked list, which is really weird, but also means the other timeline using demuxers (cue, mkv) don't need to be touched. Also, memory management becomes even worse (weird object ownership rules that are just fragile WTFs). There are some other dubious small changes, mostly related to the weird representation of separate streams. demux_timeline.c contains the actual implementation of the separate stream handling. For the most part, most things that used to be on the top level are now in struct virtual_source, of which one for each separate stream exists. This is basically like running multiple demux_edl.c in parallel. Some changes could strictly speaking be split into a separate commit, such as the stream_map type change. Mostly untested. Seems to work for the intended purpose. Potential for regressions for other timeline uses (like ordered chapters) is probably low. One thing which could definitely break and which I didn't test is the pseudo-DASH fragmented EDL code, of which ytdl can trigger various forms in obscure situations. (Uh why don't we have a test suite.) Background: The intention is to use this for the ytdl wrapper. A certain streaming site from a particularly brain damaged and plain evil Silicon Valley company usually provides streams as separate audio and video streams. The ytdl wrapper simply does use audio-add (i.e. adding it as external track, like with --audio-file), which works mostly fine. Unfortunately, mpv manages caching completely separately for external files. This has the following potential problems: 1. Seek ranges are rendered incorrectly. They always use the "main" stream, in this case the video stream. E.g. clicking into a cached range on the OSC could trigger a low level seek if the audio stream is actually not cached at the target position. 2. The stream cache bloats unnecessarily. Each stream may allocate the full configured maximum cache size, which is not what the user intends to do. Cached ranges are not pruned the same way, which creates disjoint cache ranges, which only use memory and won't help with fast seeking or playback. 3. mpv will try to aggressively read from both streams. This is done from different threads, with no regard which stream is more important. So it might happen that one stream starves the other one, especially if they have different bitrates. 4. Every stream will use a separate thread, which is an unnecessary waste of system resources. In theory, the following solutions are available (this commit works towards D): A. Centrally manage reading and caching of all streams. A single thread would do all I/O, and decide from which stream it should read next. As long as the total TCP/socket buffering is not too high, this should be effective to avoid starvation issues. This can also manage the cached ranges better. It would also get rid of the quite useless additional demuxer threads. This solution is conceptually simple, but requires refactoring the entire demuxer middle layer. B. Attempt to coordinate the demuxer threads. This would maintain a shared cache and readahead state to solve the mentioned problems explicitly. While this sounds simple and like an incremental change, it's probably hard to implement, creates more messy special cases, solution A. seems just a better and simpler variant of this. (On the other hand, A. requires refactoring more code.) C. Render an intersection of the seek ranges across all streams. This fixes only problem 1. D. Merge all streams in a dedicated wrapper demuxer. The general demuxer layer remains unchanged, and reading from separate streams is handled as special case. This effectively achieves the same as A. In particular, caching is simply handled by the usual demuxer cache layer, which sees the wrapper demuxer as a single stream of interleaved packets. One implementation variant of this is to reuse the EDL infrastructure, which this commit does. All in all, solution A would be preferable, because it's cleaner and works for all external streams in general. Some previous commit tried to prepare for implementing solution A. This could still happen. But it could take years until this is finally seriously started and finished. In any case, this commit doesn't block or complicate such attempts, which is also why it's the way to go. It's worth mentioning that original mplayer handles external files by creating a wrapper demuxer. This is like a less ideal mixture of A. and D. (The similarity with A. is that extending the mplayer approach to be fully dynamic and without certain disadvantages caused by the wrapper would end up with A. anyway. The similarity with D. is that due to the wrapper, no higher level code needs to be changed.)
2019-01-04 12:09:02 +00:00
Separate files for tracks
=========================
The special ``new_stream`` header lets you specify separate parts and time
offsets for separate tracks. This can for example be used to source audio and
video track from separate files.
Example::
# mpv EDL v0
video.mkv
!new_stream
audio.mkv
This adds all tracks from both files to the virtual track list. Upon playback,
the tracks will be played at the same time, instead of appending them. The files
can contain more than 1 stream; the apparent effect is the same as if the second
part after the ``!new_stream`` part were in a separate ``.edl`` file and added
with ``--external-file``.
Note that all metadata between the stream sets created by ``new_stream`` is
disjoint. Global metadata is taken from the first part only.
In context of mpv, this is redundant to the ``--audio-file`` and
``--external-file`` options, but (as of this writing) has the advantage that
this will use a unified cache for all streams.
The ``new_stream`` header is not part of the core EDL format. It may be changed
or removed at any time, depending on mpv's internal requirements.
Timestamp format
================
Currently, time values are floating point values in seconds.
As an extension, you can set the ``timestamps=chapters`` option. If this option
is set, timestamps have to be integers, and refer to chapter numbers, starting
with 0.
Example::
# mpv EDL v0
file.mkv,2,4,timestamps=chapters
Plays chapter 3 and ends with the start of chapter 7 (4 chapters later).
Implicit chapters
=================
mpv will add one chapter per segment entry to the virtual timeline.
By default, the chapter's titles will match the entries' filenames.
You can override set the ``title`` option to override the chapter title for
that segment.
Example::
# mpv EDL v0
cap.ts,5,240
OP.mkv,0,90,title=Show Opening
The virtual timeline will have two chapters, one called "cap.ts" from 0-240s
and a second one called "Show Opening" from 240-330s.
Entry which defines the track layout
====================================
Normally, you're supposed to put only files with compatible layouts into an EDL
file. However, at least the mpv implementation accepts entries that use
different codecs, or even have a different number of audio/video/subtitle
tracks. In this case, it's not obvious, which virtual tracks the EDL show should
expose when being played.
Currently, mpv will apply an arbitrary heuristic which tracks the EDL file
should expose. (Before mpv 0.30.0, it always used the first source file in the
segment list.)
You can set the ``layout`` option to ``this`` to make a specific entry define
the track layout.
Example::
# mpv EDL v0
file_with_2_streams.ts,5,240
file_with_5_streams.mkv,0,90,layout=this
The way the different virtual EDL tracks are associated with the per-segment
ones is highly implementation-defined, and uses a heuristic. If a segment is
missing a track, there will be a "hole", and bad behavior may result. Improving
this is subject to further development (due to being fringe cases, they don't
have a high priority).
If future versions of mpv change this again, this option may be ignored.
Syntax of EDL URIs
==================
2013-12-05 19:16:15 +00:00
mpv accepts inline EDL data in form of ``edl://`` URIs. Other than the
header, the syntax is exactly the same. It's far more convenient to use ``;``
instead of line breaks, but that is orthogonal.
Example: ``edl://f1.mkv,length=5,start=10;f2.mkv,30,20;f3.mkv``