Commit graph

191 commits

Author SHA1 Message Date
Thiago Santos
33ebda8ecf qtdemux: handle fragmented files with mdat before moofs
Assume a file with atoms in the following order: moov, mdat, moof,
mdat, moof ...

The first moov usually doesn't contain any sample entries atoms (or
they are all set to 0 length), because the real samples are signaled
at the moofs. In push mode, qtdemux parses the moov and then finds the mdat,
but then it has 0 entries and assumes it is EOS.

This patch makes it continue parsing in case it is a fragmented file so that
it might find the moofs and play the media.

https://bugzilla.gnome.org/show_bug.cgi?id=710623
2013-11-07 11:22:04 -03:00
Thiago Santos
0e78ffc9d6 qtdemux: When using a buffered mdat, store all received data for later use
In push mode, when qtdemux can't use a seek to skip the mdat buffer it has
to buffer it for later use.

The issue is that after parsing the next moov/moof, there might be some
trailing bytes from the next atom in the file. This data was being discarded
along with the already parsed moov/moof and playback would fail to continue
after the contents of this moov/moof are played.

This is particularly bad on fragmented files that have the mdat before the
corresponding moof. So you'd get:

mdat|moof|mdat|moof ...

When a moof was received, it usually came with some extra bytes that would
belong to the next mdat (because upstream doesn't care about atoms alignment).
So those bytes were being discarded and playback would fail.

This patch makes qtdemux store those extra bytes to reuse them later after the
mdat is emptied.

https://bugzilla.gnome.org/show_bug.cgi?id=710623
2013-11-07 11:22:03 -03:00
Thiago Santos
2eec7909aa qtdemux: check if the end_time is defined before using it
Avoids sending EOS too soon because of overflow. Can happen on
fragmented mp4 playback.
2013-10-25 11:30:36 -03:00
Thiago Santos
673301ef48 qtdemux: use correct unref function
Events aren't GstObjects, but GstMiniObjects
2013-10-23 13:38:56 -03:00
Stefan Sauer
ae1150e85c qtdemux: rename chunks_are_chunks to chunks_are_samples and flip the logic
As the variable name suggests, sometimes chunks are chunks. Rename the variable
to tell what they are when they are not chunks.
2013-10-15 09:53:30 +02:00
Stefan Sauer
6789ba1ece qtdemux: fix typos and add more logging for unhandled parts 2013-10-15 09:53:30 +02:00
Sebastian Dröge
21947f9d13 qtdemux: Add support for the mp2v fourcc for MPEG-2 video
https://bugzilla.gnome.org/show_bug.cgi?id=709270
2013-10-03 11:59:25 +02:00
Sreerenj Balachandran
e779b6587b qtdemux: Add HEVC support
https://bugzilla.gnome.org/show_bug.cgi?id=709093
2013-10-02 11:54:24 +02:00
Thiago Santos
dc02d91c14 qtdemux: add code to parse creation time earlier than 1970
Use g_date_time seconds manipulation to allow to cover the quicktime
spec for creation_time. It uses seconds since 1904.

Both paths could be done using the generic approach of seconds since
1904 with GDateTime handling, but the first path using seconds from
1970 should be more commonly found and avoids a few objects creation and
ref/unref, so keep it there for performance.

Additionally, the code for handling seconds since 1970 changed from >
to >= because having 0 seconds since 1970 is also a valid case for that
path to handle.

https://bugzilla.gnome.org/show_bug.cgi?id=707975
2013-09-24 15:16:54 -07:00
Thiago Santos
566b0dce40 qtdemux: only update stop position if seek requests it
Check for GST_SEEK_TYPE_NONE for stop poistion and only update
the stop time if it is requested. Otherwise just maintain whatever
was stored at the segment

https://bugzilla.gnome.org/show_bug.cgi?id=707530
2013-09-13 09:21:12 -03:00
Thiago Santos
be0eeae491 qtdemux: track streams that are EOS on push mode to finish earlier
When the segment has a defined stop position, qtdemux should check
when streams reach this position and mark those as EOS. When all
streams are EOS it will return GST_FLOW_EOS to upstream to allow
the pipeline to finish instead of continuously consume buffers
from upstream that are not useful for the segment.

https://bugzilla.gnome.org/show_bug.cgi?id=707530
2013-09-10 16:43:17 -03:00
Thiago Santos
33cf8b679d qtdemux: preserve stop of segment when doing seeks in push mode
When handling seeks in push mode, qtdemux converts the seek to bytes
and pushes upstream. It needs to keep track of the seek and the
subsequent segment to be able to map them back to the requested
seek time and properly preserve the segment stop of the seek.

This is done by using the start offset in bytes of the seek,
that should be the same of the segment from upstream. And this
is also backwards compatible with what qtdemux already was using.

https://bugzilla.gnome.org/show_bug.cgi?id=707530
2013-09-10 16:42:36 -03:00
Alex Ashley
a965185dee qtdemux: Add support for the avc3 sample entry format of the AVC file format
Amendment 2 of ISO/IEC 14496-15 (AVC file format) is defining a new
structure for fragmented MP4 called "avc3". The principal difference
between AVC1 and AVC3 is the location of the codec initialisation
data (e.g. SPS, PPS). In AVC1 this data is placed in the initial
MOOV box (moov.trak.mdia.minf.stbl.stsd.avc1) but in AVC3 this data
goes in the first sample of every fragment (i.e. the first sample in
each mdat box).  The principal reason for avc3 is to make it easier
for client implementations, because it removes the requirement to
insert the SPS+PPS in to the decoder pipeline every time there is a
representation change.

This commit adds support for the "avc3" atom, which is almost identical
to the "avc1" atom, except it does not contain any SPS or PPS data.

https://bugzilla.gnome.org/show_bug.cgi?id=702004
2013-09-04 13:33:22 +02:00
Thiago Santos
9549289a18 qtdemux: push buffers after segment stop until reaching a keyframe
This should make decoders able to precisely push buffers until the stop
time in case they need the next keyframe to do it.

Also, according to gst_segment_clip, it should only push a buffer that
the starting ts is strictly smaller than the segment stop, so we change
the min < comparison for <=
2013-08-28 12:58:56 -03:00
Chris Bass
3e9dea3f8c qtdemux: check denominator isn't zero before scaling duration.
When gst_qtdemux_configure_stream sets fps_d, check that n_samples is
non-zero before using it as a denominator to scale the stream duration.

https://bugzilla.gnome.org/show_bug.cgi?id=706076
2013-08-16 10:14:30 +02:00
Matej Knopp
2269ac8f28 qtdemux: elst should offset samples instead of buffers
The current approach where buffers are offset is not ideal, as during seek
and loop current time is compared to sample times.

https://bugzilla.gnome.org/show_bug.cgi?id=700264
2013-08-12 13:48:04 +02:00
Thibault Saunier
38946bd9f4 qtdemux: Fix compilation 2013-08-06 15:31:38 +02:00
Thibault Saunier
593a31f2b4 qtdemux: Raw buffer DTS should always be CLOCK_TIME_NONE 2013-08-06 15:17:44 +02:00
Sebastian Dröge
2e35b36aab gst: Don't swap start/stop for negative rates in the SEGMENT query 2013-07-29 12:12:41 +02:00
Matej Knopp
47ed79fb1c qtdemux: Check for data size when parsing h264 codec data from strf atom 2013-07-29 11:53:07 +02:00
Matej Knopp
2dcdfe07f7 qtdemux: Support H264 fourcc
https://bugzilla.gnome.org/show_bug.cgi?id=704996
2013-07-29 09:11:39 +02:00
Thiago Santos
e49b6e7c35 qtdemux: correctly handle seqnum for seeks and segments
Use the same seqnum on messages and events for derived events.
Fixed for flushes / stream-start / segment after a seek, and segment
after a segment.

Fixes #676242
2013-07-25 15:24:31 -03:00
Edward Hervey
3d48d25756 qtdemux: Add all the mpeg XDCAM variants
This should cover all known XDCAM variants (which are all mpeg2 video)

Fixes #672227
2013-07-23 15:03:31 +02:00
Sebastian Dröge
1a0278ed64 qtdemux: Add support for group-id in the stream-start event 2013-07-22 15:30:13 +02:00
Olivier Crête
96a8fb92e2 qtdemux: Add WRLE support 2013-07-19 14:58:30 -04:00
Tim-Philipp Müller
aa7d597120 qtdemux: make files from Vivotek camera play
Skip tracks of 'vivo' subtype with empty stsd instead of
erroring out saying that the file is broken.

https://bugzilla.gnome.org/show_bug.cgi?id=699791
2013-07-19 19:38:30 +01:00
Wim Taymans
bdd3c31902 qtdemux: simplify some helpers
Some helper functions are not needed anymore or can be simplified.
2013-07-19 17:26:54 +02:00
Wim Taymans
61a8937ced qtdemux: for non-raw video, move palette in caps
We only need to append the palette to raw video buffers, non-raw video has the
palette in the caps still.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=704292
2013-07-19 17:14:46 +02:00
Arnaud Vrac
40ab78825c qtdemux: nitpicking in esds parsing 2013-07-19 14:26:18 +02:00
Arnaud Vrac
d0d25a5e1f qtdemux: set proper caps for mpeg-1 audio
Remove AAC specific fields from mpeg-1 audio caps, remove assumption
that the mpeg1 audio layer is 3, and set `parsed' field.

https://bugzilla.gnome.org/show_bug.cgi?id=704548
2013-07-19 14:26:08 +02:00
Arnaud Vrac
5def061d20 qtdemux: remove chapter stream
Remove all streams that are actually table of contents, since we will
never need the data after parsing them.
2013-07-18 11:48:12 +02:00
Arnaud Vrac
ae67c13416 qtdemux: send gap event for sparse streams in push mode
This allows to pre-roll at least if the next subtitle buffer
is far away.
2013-07-18 11:48:11 +02:00
Arnaud Vrac
1237898351 qtdemux: do not use indexes from sparse stream when seeking in push mode
This makes seeking more accurate in push mode, since the previous
keyframe on a sparse stream might be far away.
2013-07-18 11:48:11 +02:00
Arnaud Vrac
e561d12655 qtdemux: advertise subtitle streams as sparse 2013-07-18 11:48:11 +02:00
Wim Taymans
4c97701650 qtdemux: extract the palette from stsd
Sometimes a palette is inside the stsd, extract it instead of always using
the default one
2013-07-17 15:17:19 +02:00
Wim Taymans
6b82c89562 qtdemux: add support for WRAW
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=704292
2013-07-17 09:57:17 +02:00
Wim Taymans
f698483bb3 qtdemux: palette is appended to buffers, not in caps
Fix the palette handling, in 1.0 we append the palette to the buffer instead of
placing it on the caps.

See also https://bugzilla.gnome.org/show_bug.cgi?id=704292
2013-07-17 09:57:16 +02:00
Arnaud Vrac
54bba4f60c qtdemux: reset segment on flush stop
cca2f555d1 introduces a regression, where the demux segment is not
reset on flush stop, so the next upstream segment event will calculate
an invalid base time on the new segment to be sent downstream.

https://bugzilla.gnome.org/show_bug.cgi?id=704255
2013-07-16 10:47:20 +02:00
Matej Knopp
ca32442f86 qtdemux: offset samples according to edit list
https://bugzilla.gnome.org/show_bug.cgi?id=700264
2013-07-15 09:59:23 +02:00
Matej Knopp
3111161e8a qtdemux: unselect instead of ignoring disabled track, detect chapter track
https://bugzilla.gnome.org/show_bug.cgi?id=704007
2013-07-12 11:45:33 +02:00
Matej Knopp
7b69f427f1 qtdemux: correct argument order in gst_util_uint64_scale_int_round
https://bugzilla.gnome.org/show_bug.cgi?id=703350
2013-07-10 09:20:17 +02:00
Matej Knopp
4053e1d6ac qtdemux: compute framerate from average sample duration
https://bugzilla.gnome.org/show_bug.cgi?id=703350
2013-07-01 12:53:17 +02:00
Wim Taymans
8428423c04 qtdemux: handle SEGMENT query 2013-06-20 11:31:22 +02:00
Sebastian Dröge
b001da2926 qtdemux: Disable usage of allocation queries
This can only reliably work if demuxers have a
separate streaming thread per srcpad. This should be
done in a demuxer base class, which integrates parts
of multiqueue

https://bugzilla.gnome.org/show_bug.cgi?id=701856
2013-06-19 11:07:48 +02:00
Alex Ashley
46a137c810 Avoid skipping moov atoms for fragmented MP4 files.
bug #700505

Following a representation change that causes a resolution change,
the video decoder fails to decode correctly. Dashdemux detects the
representation change and pushes a new caps event and an
initialization segment (a new moov atom) to the downstream qtdemux,
but it doesn't handle this new moov yet, it will only parse the
first one it receives.

This commit changes qtdemux to accept a new moov in a dash bitstream
switching scenario.
2013-06-19 01:44:22 -03:00
Thiago Santos
384e8f6c34 qtdemux: send stream-start only once for each stream
Do not send stream start again when reconfiguring a pad for new caps.
That is common for adaptive streams
2013-06-19 00:55:30 -03:00
Sebastian Dröge
adc9f0bd10 qtdemux: Do allocation query after exposing all pads and no-more-pads
Also configure video streams as early as possible.

Related https://bugzilla.gnome.org/show_bug.cgi?id=701856
but not fixing that.
2013-06-11 12:27:19 +02:00
Thiago Santos
7c12435f9b qtdemux: make sure taglist is writable before adding tags
Avoids assertions
2013-06-02 15:37:06 -03:00
Thiago Santos
78dfdee2aa qtdemux: effectively skip tracks that weren't listed on the 1st moov
Without this, stream is NULL and the code will try to access it, leading
to segfaults.
2013-06-02 13:06:15 -03:00
Thiago Santos
70fca21c28 qtdemux: skip redundant check
!got_moov is already checked the line above
2013-06-02 13:06:15 -03:00