[FFmpeg-devel] [PATCH 1/2] avcodec: add decode_error field to AVFrame
michaelni at gmx.at
Wed Jul 11 01:13:50 CEST 2012
On Tue, Jul 10, 2012 at 11:57:59PM +0200, Marton Balint wrote:
> On Tue, 10 Jul 2012, Reimar Döffinger wrote:
> >On Sun, Jul 08, 2012 at 06:12:08PM +0200, Marton Balint wrote:
> >>+ /**
> >>+ * decode error flag of the frame, set to a nonzero codec-specific error
> >>+ * code if the decoder produced a frame, but there were errors during
> >>+ * the decoding.
> >I don't like the codec specific error. I'd very much prefer this to be a
> >bit field, with at least some things pre-defined.
> >I don't mind a part being reserved for codec-specific stuff though.
> I gave this a thought before sending the patch, but I could not
> think of too many generic decode errors, do you have some ideas?
> And what would be your goal with the generic errors? Indicate the
> reason of the decode error as precisely as possible? In that case
> probably the generic errors won't be specific enough... Or to
> indicate the "quality" of the decoded frame (e.g. it is probably
> complete garbage, or it is roughly the same as it would be without
> the decode error, etc). That is probably not easy (sometimes not
> even possible) to determine...
when error concealment is in use we know how much of the frame was
decoded and how much was produced by the concealment code. The
concealment code also calculates the similarity to the previous
frame so theres also a rough guess at how good the concealment areas
would be. dunno how this cpuld be presented to the user though
and its just h26* mpeg*, the other decoders dont have such information
Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB
The misfortune of the wise is better than the prosperity of the fool.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 198 bytes
Desc: Digital signature
More information about the ffmpeg-devel