[gstreamer-bugs] [Bug 614460] matroskademux: wrong H.264 demux in some file

GStreamer (bugzilla.gnome.org) bugzilla at gnome.org
Wed Mar 31 09:03:26 PDT 2010


https://bugzilla.gnome.org/show_bug.cgi?id=614460
  GStreamer | gst-plugins-good | 0.10.21

--- Comment #1 from Mark Nauwelaerts <mnauw at users.sourceforge.net> 2010-03-31 16:03:23 UTC ---
Created an attachment (id=157606)
 View: https://bugzilla.gnome.org/attachment.cgi?id=157606
 Review: https://bugzilla.gnome.org/review?bug=614460&attachment=157606

Disable syncing for non-subtitle streams

It looks like this sample is not really muxed appropriately interleaved, that
is, the audio data comes in rather big chunks, e.g.
----
| + SimpleBlock (key, track number 2, 8 frame(s), timecode 4.864s =
00:00:04.864000000)
|   + Frame with size 7423
|   + Frame with size 7488
|   + Frame with size 7785
|   + Frame with size 7716
|   + Frame with size 8319
|   + Frame with size 8827
|   + Frame with size 8475
|   + Frame with size 7676
<....>
| + SimpleBlock (key, track number 2, 8 frame(s), timecode 5.547s =
00:00:05.547000000)
|   + Frame with size 7684
|   + Frame with size 8208
|   + Frame with size 7731
|   + Frame with size 8097
|   + Frame with size 7886
|   + Frame with size 7983
|   + Frame with size 8402
|   + Frame with size 8995
----

This leads to lots of newsegment events from matroskademux resync mechanics.

I am not sure what can be done with such irregular interleaving.  Pending other
suggestions or inspiration for some creative heuristic, attached patch naively
restricts the current sync done to subtitle tracks only (which is/was likely
the intention in the first place).

It turns out that increasing the lag to 1 sec (in stead of 0.5 sec) also works
in this case, if there were arguments to go this way instead ...

-- 
Configure bugmail: https://bugzilla.gnome.org/userprefs.cgi?tab=email
------- 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