[Bug 671909] Port base video encoding and decoding classes from gst-plugins-bad

[GStreamer (bugzilla.gnome.org)]

https://bugzilla.gnome.org/show_bug.cgi?id=671909
  GStreamer | gst-plugins-base | git

Mark Nauwelaerts <mnauw> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |mnauw at users.sourceforge.net

--- Comment #14 from Mark Nauwelaerts <mnauw at users.sourceforge.net> 2012-04-02 11:15:27 UTC ---
Additional comments:

[video encoder]
* almost unused variable stuff (e.g. priv->clipping always TRUE)
* at_eos set to TRUE upon EOS, set to FALSE when new segment (where afaik only
FLUSH can clear eos state?)
* draining (e.g. call to gst_video_encoder_drain) happens only once where this
happens more in e.g. audio encoder case, and some (or all) might also make
sense here.  Also semantics/documentation are unclear there, since the subclass
reset method is actually being used to perform draining, not a reset (as in
post-seek, as suggested in .h documentation)
* lots of frame number stuff seems to involve almost unused fields, or badly
used ones (e.g. system_frame_number is never really properly initialized,
frame_decode_number is used to determine dts, but this may have no relation at
all to the pts since it does not consider segment info at all).
Might also consider (maybe depending on a subclass settable setting) to obtain
a frame's dts as the oldest queued pts.

[video decoder]
* see previous comment w.r.t. parsing
* roughly similar comments as for the encoder apply
(a.o. various frame number and ts issues, e.g. presentation_timestamp might get
determined based on (a.o.) a never set base_picture_number, as noted in a
FIXME).

-- 
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.