Currently, AVStream contains an embedded AVCodecContext instance, which
is used by demuxers to export stream parameters to the caller and by
muxers to receive stream parameters from the caller. It is also used
internally as the codec context that is passed to parsers.
In addition, it is also widely used by the callers as the decoding (when
demuxer) or encoding (when muxing) context, though this has been
officially discouraged since Libav 11.
There are multiple important problems with this approach:
- the fields in AVCodecContext are in general one of
* stream parameters
* codec options
* codec state
However, it's not clear which ones are which. It is consequently
unclear which fields are a demuxer allowed to set or a muxer allowed to
read. This leads to erratic behaviour depending on whether decoding or
encoding is being performed or not (and whether it uses the AVStream
embedded codec context).
- various synchronization issues arising from the fact that the same
context is used by several different APIs (muxers/demuxers,
parsers, bitstream filters and encoders/decoders) simultaneously, with
there being no clear rules for who can modify what and the different
processes being typically delayed with respect to each other.
- avformat_find_stream_info() making it necessary to support opening
and closing a single codec context multiple times, thus
complicating the semantics of freeing various allocated objects in the
codec context.
Those problems are resolved by replacing the AVStream embedded codec
context with a newly added AVCodecParameters instance, which stores only
the stream parameters exported by the demuxers or read by the muxers.
Some (de)muxers open additional files beyond the main IO context.
Currently, they call avio_open() directly, which prevents the caller
from using custom IO for such streams.
This commit adds callbacks to AVFormatContext that default to
avio_open2()/avio_close(), but can be overridden by the caller. All
muxers and demuxers using AVIO are switched to using those callbacks
instead of calling avio_open()/avio_close() directly.
(de)muxers that use the URLProtocol layer directly instead of AVIO
remain unconverted for now. This should be fixed in later commits.
All encoders set pts and dts properly now (and have been doing that for
a while), so there is no good reason to do any timestamp guessing in the
muxer.
The newly added AVStreamInternal will be later used for storing all the
private fields currently living in AVStream.
Note that convergence_duration had another meaning, one which was in
practice never used. The only real use for it was a 64 bit replacement
for the duration field. It's better just to make duration 64 bits, and
to get rid of it.
Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
In these cases, only drop dts. Because if we drop both we have no
timestamps at all for some files.
This improves playback of HLS streams from GoPro cameras.
Signed-off-by: Martin Storsjö <martin@martin.st>
When AVFMT_FLAG_NOBUFFER is set, the packets are not added to the
AVFormatContext packet list, so they need to be freed when they are
no longer needed.
Signed-off-by: Anton Khirnov <anton@khirnov.net>
It is basically a wrapper around av_get_audio_frame_duration(), with a
fallback to AVCodecContext.frame_size. However, that field is set only
when the stream codec context is actually used for encoding or decoding,
which is discouraged.
For muxing, it is generally the responsibility of the caller to set the
packet duration.
For demuxing, if the duration is not stored at the container level, it
should be set by the parser.
Therefore, removing the frame_size fallback should not break any
important case.
If any option named "metadata" is set inside the context, it is pulled up to
the context and then the option is cleared.
Signed-off-by: Anton Khirnov <anton@khirnov.net>
Previously, AVStream.codec.time_base was used for that purpose, which
was quite confusing for the callers. This change also opens the path for
removing AVStream.codec.
The change in the lavf-mkv test is due to the native timebase (1/1000)
being used instead of the default one (1/90000), so the packets are now
sent to the crc muxer in the same order in which they are demuxed
(previously some of them got reordered because of inexact timestamp
conversion).
