[Bug 725167] opusdec PLC doesn't seem to work as well as Chrome

[GStreamer (bugzilla.gnome.org)]

https://bugzilla.gnome.org/show_bug.cgi?id=725167
  GStreamer | gst-plugins-bad | 1.2.1

--- Comment #19 from Vincent Penquerc'h <vincent.penquerch at collabora.co.uk> 2015-01-14 10:51:33 UTC ---
Created an attachment (id=294503)
 View: https://bugzilla.gnome.org/attachment.cgi?id=294503
 Review: https://bugzilla.gnome.org/review?bug=725167&attachment=294503

alternative patch for frame size determintion on FEC/PLC cases

Here's another possible patch.
I think (not 100% sure) that the duration should always be taken from the
missing data. This patch also keeps track of any leftover samples when the
number of missing samples is not a multiple of the minimum frame size. This
leftover is carried over non missing frames, which I'm not sure is the best
thing to do (as opposed to re-zeroing it when an actual packet is received).
Doing it this way will desync less over time I think.

Tested with:

gst-launch-1.0 uridecodebin uri=file://path/to/audio.flac  ! audioconvert !
audioresample !  opusenc frame-size=X inband-fec=1 ! identity
drop-probability=0.1 ! opusdec use-inband-fec=1 plc=1 ! pulsesink

X being 2, 5, 10, and using either use-inband-fec=1 and plc=1. This required a
change to identity to send gap events when dropping, which I'll commit shortly.

It sounds better than the original but I don't know how it compares to Chrome.

-- 
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.