[Bug 743900] oggdemux gets first packet timestamp wrong - theora

GStreamer (GNOME Bugzilla) bugzilla at gnome.org
Mon Feb 16 09:22:50 PST 2015


https://bugzilla.gnome.org/show_bug.cgi?id=743900

--- Comment #4 from Vincent Penquerc'h <vincent.penquerch at collabora.co.uk> ---
I think this is actually correct.

On a sample file here, the first data page contains 30 packets, and has a
granpos of 15360. The sum of the durations of the 30 packets is 15360.
The first packet has duration... well, this one's a bit odd. The packet itself
has duration 128, but oggstream says 0, as it's the first packet in the stream
(see ext/ogg/oggstream.c, in packet_duration_vorbis, for the use of
pad->last_size). I assume that's per the Vorbis spec. The 15360 total duration
is counting that first duration as 0.

Now, the first packet (with duration 0) gets a pts of 0, and a duration of 0.
The second packet has a pts of 0 too, an a duration of 576-0, 576, which is
0.13ish at 44100 Hz. This appears to be correct to me.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.


More information about the gstreamer-bugs mailing list