GitLab

This includes a lot of formating, obviously the stream Id changes as
usual, but we now have bitrate with AVI, and chrome-format field
with h264.

PTS/DTS were modified by https://bugzilla.gnome.org/show_bug.cgi?id=785119

And update `op1a-mpeg2-wave_hd.mxf` audio frames now that they are
odered by PTS (not sure why but it shouldn't be wrong).

PTSs have changed since:

830e89b7e9418c7039021ebc4060d5c922e928b0:

mxfdemux: Take temporal reordering from the index table into account
This is needed to know the PTS, without that we only know the DTS and
using that also for the PTS is wrong unless we have an intra-only codec.

If we can't get the temporal reordering from the index table, don't set
any PTS for non-intra-only codecs and let decoders figure out something.

https://bugzilla.gnome.org/show_bug.cgi?id=784027

Edward Hervey authored May 25, 2017 and Edward Hervey committed May 25, 2017

This now contains more than just DASH-IF streams and eventually could
also contain HLS/MSS/... streams

Edward Hervey authored Apr 10, 2017 and Edward Hervey committed Apr 14, 2017

Allows testing various DASH-IF Test Assets

As h264parse now adds the information.

Edward Hervey authored Jan 31, 2017 and Edward Hervey committed Jan 31, 2017

The PTS/DTS no longer reset. Fixed by the following commit:

Commit: 1fc3d42f8ae2e0b5b4241cb42c25806149e9ae45
URL:    http://cgit.freedesktop.org/gstreamer/gst-plugins-good/commit/?id=1fc3d42f8ae2e0b5b4241cb42c25806149e9ae45

Author: Jan Schmidt <jan@centricular.com>
Date:   Sat Jan  7 23:55:42 2017 +1100

qtdemux: Don't reset output timestamps when no tfdt

If a fragmented stream doesn't have a tfdt, don't
reset the output timestamps at each fragment boundary
by erroneously using the default value of 0. Introduced
by commit 69fc48

https://bugzilla.gnome.org/show_bug.cgi?id=754230

Fixing places where updates were needed but I missed them.

Huge changes happened in frame detection as we used to check
unparsed frames (ie buffers outputed by demuxers) but this was
not correct at all, new information should be much more occurate.

This reverts commit 8cd6d99a.

Basically the video buffers offsets used to go back to 0 when DTS
was going back to 0 and now they are always increasing, which is
also fine.

Edward Hervey authored Oct 18, 2016 and Edward Hervey committed Oct 18, 2016

We know have the program-id also in there

The file was generated with a non-seekable sink, so some values in
the headers couldn't be rewritten and that requires some extra
care when playing the file.

Related to bug https://bugzilla.gnome.org/show_bug.cgi?id=765805

Edward Hervey authored Mar 09, 2016 and Edward Hervey committed Mar 09, 2016

We now properly detect profile/level for these streams

They need to have *exactly* the expected stream-id since they come from
a specific URI. This was broken in the last updates of mkv/webm stream-id

Edward Hervey authored Mar 08, 2016 and Edward Hervey committed Mar 08, 2016

The output of h264parse has changed again after this commit:
    videoparsers: h264: Disable passthorugh mode enabling

Edward Hervey authored Mar 08, 2016 and Edward Hervey committed Mar 08, 2016

Wasn't updated in the last round of stream-id updates due to matroska demuxer
changes.

Synchronization is made on running-time, store it in the media-check
files to to be able to use it and compare them with new data extracted
in tests

commit e8908f5aeef952566f6bccde743c7735d3f8c6ef
h264parse: Don't switch to passthrough on set_caps()

This commit changes the offset and DTS of the first frame
of the mkv h264 stream. The changes seem harmless and are not
relevant in the context of this file.

Also changes the number of frames pushed in the mpegts streams.

It now has the stream headers prepended to the stream

commit c3bcbadd5452d5b3450f70e49dad3e64f14de00a
baseparse: Use new gst_adapter_get_buffer() API instead of gst_adapter_map()

This makes the buffers preserve their offsets, needing to
update this media-info file.

commit dd72267a8da0d2884ff9ecff6731ab753562258d
qtdemux: Add cslg support

This commit changed how pts/dts is calculated for the streams
to allow aligning of the PTS and DTS and accomodate possible
negative DTS on some streams. The stream infos
needed to be updated to reflect the changes.

It's now properly set to 0/1, letting GStreamer entirely rely on
timestamps as it should.

Some media info had outdate stream-id, which was preventing
regression testing on the produce caps. Also, few media info where
missing, which endup enabling few more tests.

+ Fix the png media info