[FFmpeg-devel] [PATCH v2 0/3] Properly decode ogg metadata in ogg/flac chained bitstreams

[Romain Beauxis]

Le mar. 4 févr. 2025 à 06:15, Romain Beauxis
<romain.beauxis at gmail.com> a écrit :
>
> This is a series of 3 patches to allow proper decoding of ogg metadata
> in chained ogg/flac streams.
>
> ogg/flac streams are pretty important because there are perhaps the only
> combination of lossless audio codec and open-source container that
> allows for proper transmittion of lossless audio data accross systems
> such as Icecast, browser media tags and more.
>
> In the context of long-running audio streams, the ogg bitstream specs[1]
> have historically been very badly implemented. For each new track and
> each new metadata, the specs require the logical bitstream to come to a
> full EOF and then start with a full new logical stream.
>
> These specs have often been confused with a gobal EOF by most
> implementations.
>
> Furtunately, FFmpeg is a little better at that in that it is capable to
> parsing chained logical ogg bitstreams and properly output either
> encoded ogg packets or decoded audio.
>
> Chained bitstreams with more than one underlying type of content
> (audio+video, etc) is not yet supported though this is a much less needed
> feature.
>
> The purpose of these changes is to also allow proper decoding of
> metadata associated with subsequent streams in chained ogg/flac
> bitstream.
>
> This is done by simply intercepting ogg packets with comments in the
> flac decoded, parsing the comment block and retaining it to be attached
> with the next decoded audio frame.
>
> Along with the changes is a new FATE test validating the implementation.
>
> This solution keeps a proper separation of concerns: ogg packets are
> sill output by the demuxer (as shown in the test) but consumer of decoded
> data see decoded metadata in the decoded frames.
>
> Only drawback is that this adds a dependency on libavformat to
> libavcodec.
>
> I have looked at moving the vorbis metadata parsing to libavutil but a
> lot of definitions and utilities related to metadata are in fact located
> in libavformat so perhaps this makes sense.

Of course, it was only after posting all this that I figured out how
to properly do this.

About to send an updated patch set sorry for the noise.

-- Romain

> Follow-up work not addressed in this series of patch:
> * Ensure valid PTS in decoded frames of subsequent streams?
> * Generalize this approach to other chained ogg codec.
>
> Thanks,
> -- Romain
>
> Romain Beauxis (3):
>   libavformat/oggdec: Allow first parameter in ff_vorbis_comment to be a
>     generic AVClass struct
>   libavcodec/flacdec: parse vorbis metadata from ogg packets, add them
>     to the next decoded frame.
>   Add stream dump test with test for ogg/flac.
>
>  libavcodec/flacdec.c                  |  12 +-
>  libavformat/oggdec.h                  |   5 +-
>  libavformat/oggparsevorbis.c          |   4 +-
>  tests/Makefile                        |   2 +
>  tests/api/Makefile                    |   2 +-
>  tests/api/api-dump-stream-meta-test.c | 169 ++++++++++++++++++++++++++
>  tests/fate/api.mak                    |   5 +
>  tests/fate/ogg-flac.mak               |  11 ++
>  8 files changed, 205 insertions(+), 5 deletions(-)
>  create mode 100644 tests/api/api-dump-stream-meta-test.c
>  create mode 100644 tests/fate/ogg-flac.mak
>
> --
> 2.39.5 (Apple Git-154)
>