[Bug 656412] vorbisdec: discard starting samples according to granpos

[GStreamer (bugzilla.gnome.org)]

https://bugzilla.gnome.org/show_bug.cgi?id=656412
  GStreamer | gst-plugins-base | unspecified

--- Comment #18 from Mark Nauwelaerts <mnauw at users.sourceforge.net> 2011-11-28 22:59:10 UTC ---
FWIW, fairly certain about 3rd point, since only out_timestamp is essentially
affected (in addition to some segment stuff), and out_timestamp only affects
the current packet, and since that one's timestamp is forcibly clamped/shifted
to 0, it would create a ts discont with the subsequent packet [*].  segment
data can only "affect by shift" and not modify a discont somewhere in between.

That is, the above is all assuming no trick stuff happens in a subsequent
decoder, and there should be no(t much) there.  At least, performing some ts
shifting is beyond/besides GstAudioDecoder's purview (or any decoder for that
matter) though it seems the semantics here should not require shifting anyway.

So, trying to do this without segment stuff brings us back to Comment #6, with
some ts tricks/hacks, which can probably be thought of and might be
special-case handled in a vorbisdec pre_push method or so.


[*] well, probably not a ts discont considering ts and duration set on buffers
with encoded data, but when decoded into samples the subsequent "real
durations" lead to discont

-- 
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.