RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-07-26 00:27:43 +00:00
Code Issues Activity Actions

LibVideo: Read Matroska lazily so that large files can start quickly

Browse source 
The Demuxer class was changed to return errors for more functions so
that all of the underlying reading can be done lazily. Other than that,
the demuxer interface is unchanged, and only the underlying reader was
modified.

The MatroskaDocument class is no more, and MatroskaReader's getter
functions replace it. Every MatroskaReader getter beyond the Segment
element's position is parsed lazily from the file as needed. This means
that all getter functions can return DecoderErrors which must be
handled by callers.
This commit is contained in:
Zaggy1024 2022-11-11 17:14:27 -06:00 • committed by Andreas Kling
parent f4c476b26f
commit 393cfdd5c5
11 changed files with 576 additions and 310 deletions
Download patch file Download diff file Expand all files Collapse all files

76
Userland/Libraries/LibVideo/Containers/Matroska/MatroskaDemuxer.cpp
View file

@ -10,15 +10,15 @@ namespace Video::Matroska {
DecoderErrorOr<NonnullOwnPtr<MatroskaDemuxer>> MatroskaDemuxer::from_file(StringView filename)
{
return make<MatroskaDemuxer>(TRY(Reader::parse_matroska_from_file(filename)));
return make<MatroskaDemuxer>(TRY(Reader::from_file(filename)));
}
DecoderErrorOr<NonnullOwnPtr<MatroskaDemuxer>> MatroskaDemuxer::from_data(ReadonlyBytes data)
{
return make<MatroskaDemuxer>(TRY(Reader::parse_matroska_from_data(data)));
return make<MatroskaDemuxer>(TRY(Reader::from_data(data)));
}
Vector<Track> MatroskaDemuxer::get_tracks_for_type(TrackType type)
DecoderErrorOr<Vector<Track>> MatroskaDemuxer::get_tracks_for_type(TrackType type)
{
TrackEntry::TrackType matroska_track_type;
@ -35,21 +35,20 @@ Vector<Track> MatroskaDemuxer::get_tracks_for_type(TrackType type)
}
Vector<Track> tracks;
for (auto const& track_table_entry : m_document->tracks()) {
auto const& track_entry = track_table_entry.value;
if (matroska_track_type == track_entry->track_type())
tracks.append(Track(type, track_entry->track_number()));
}
// FIXME: Sort the vector, presumably the hashtable will not have a consistent order.
TRY(m_reader.for_each_track_of_type(matroska_track_type, [&](TrackEntry const& track_entry) -> DecoderErrorOr<IterationDecision> {
VERIFY(track_entry.track_type() == matroska_track_type);
DECODER_TRY_ALLOC(tracks.try_append(Track(type, track_entry.track_number())));
return IterationDecision::Continue;
}));
return tracks;
}
ErrorOr<MatroskaDemuxer::TrackStatus*> MatroskaDemuxer::get_track_status(Track track)
DecoderErrorOr<MatroskaDemuxer::TrackStatus*> MatroskaDemuxer::get_track_status(Track track)
{
if (!m_track_statuses.contains(track))
TRY(m_track_statuses.try_set(track, TrackStatus()));
if (!m_track_statuses.contains(track)) {
auto iterator = TRY(m_reader.create_sample_iterator(track.identifier()));
DECODER_TRY_ALLOC(m_track_statuses.try_set(track, { iterator }));
}
return &m_track_statuses.get(track).release_value();
}
@ -57,10 +56,8 @@ ErrorOr<MatroskaDemuxer::TrackStatus*> MatroskaDemuxer::get_track_status(Track t
DecoderErrorOr<void> MatroskaDemuxer::seek_to_most_recent_keyframe(Track track, size_t timestamp)
{
if (timestamp == 0) {
auto track_status = DECODER_TRY_ALLOC(get_track_status(track));
track_status->m_cluster_index = 0;
track_status->m_block_index = 0;
track_status->m_frame_index = 0;
// Removing the track status will cause us to start from the beginning.
m_track_statuses.remove(track);
return {};
}
@ -69,41 +66,26 @@ DecoderErrorOr<void> MatroskaDemuxer::seek_to_most_recent_keyframe(Track track,
DecoderErrorOr<NonnullOwnPtr<Sample>> MatroskaDemuxer::get_next_sample_for_track(Track track)
{
auto track_status = DECODER_TRY_ALLOC(get_track_status(track));
// FIXME: This makes a copy of the sample, which shouldn't be necessary.
// Matroska should make a RefPtr<ByteBuffer>, probably.
auto& status = *TRY(get_track_status(track));
for (; track_status->m_cluster_index < m_document->clusters().size(); track_status->m_cluster_index++) {
auto const& cluster = m_document->clusters()[track_status->m_cluster_index];
for (; track_status->m_block_index < cluster.blocks().size(); track_status->m_block_index++) {
auto const& block = cluster.blocks()[track_status->m_block_index];
if (block.track_number() != track.identifier())
continue;
if (track_status->m_frame_index < block.frame_count()) {
switch (track.type()) {
case TrackType::Video: {
// FIXME: This makes a copy of the sample, which shouldn't be necessary.
// Matroska should make a RefPtr<ByteBuffer>, probably.
auto cicp = m_document->track_for_track_number(track.identifier())->video_track()->color_format.to_cicp();
Time timestamp = Time::from_nanoseconds((cluster.timestamp() + block.timestamp()) * m_document->segment_information()->timestamp_scale());
return make<VideoSample>(block.frame(track_status->m_frame_index++), cicp, timestamp);
}
default:
return DecoderError::not_implemented();
}
}
track_status->m_frame_index = 0;
}
track_status->m_block_index = 0;
if (!status.block.has_value() || status.frame_index >= status.block->frame_count()) {
status.block = TRY(status.iterator.next_block());
status.frame_index = 0;
}
return DecoderError::with_description(DecoderErrorCategory::EndOfStream, "End of stream reached."sv);
auto const& cluster = status.iterator.current_cluster();
Time timestamp = Time::from_nanoseconds((cluster.timestamp() + status.block->timestamp()) * TRY(m_reader.segment_information()).timestamp_scale());
auto cicp = TRY(m_reader.track_for_track_number(track.identifier())).video_track()->color_format.to_cicp();
return make<VideoSample>(status.block->frame(status.frame_index++), cicp, timestamp);
}
Time MatroskaDemuxer::duration()
DecoderErrorOr<Time> MatroskaDemuxer::duration()
{
if (!m_document->segment_information().has_value())
auto segment_information = TRY(m_reader.segment_information());
if (!segment_information.duration().has_value())
return Time::zero();
if (!m_document->segment_information()->duration().has_value())
return Time::zero();
return Time::from_nanoseconds(m_document->segment_information()->duration().value() * m_document->segment_information()->timestamp_scale());
return Time::from_nanoseconds(static_cast<i64>(segment_information.duration().value() * segment_information.timestamp_scale()));
}
}