[FFmpeg-devel] [PATCH] FFV1 2pass

[Michael Niedermayer]

On Sun, Oct 24, 2010 at 12:16:50AM -0700, Jason Garrett-Glaser wrote:
> On Sat, Oct 23, 2010 at 11:24 AM, Michael Niedermayer <michaelni at gmx.at> wrote:
> > Hi
> >
> > The patches below add a 2pass mode to ffv1 that optimally sorts the range
> > coder states. The compression gain is for my test quite small but more things
> > can be added in the future to be finetuned in 2 passes ...
> > Also ive not been able to meassure speedloss here for the added code.
> >
> > Ill apply this patchset soon as it seems quite trivial.
> 
> Couldn't this be done per gop in some fashion?  Does it make sense to
> do it globally for the whole file?

i dont know, the gains from this patch are small (biggest ive seen is 0.15%)
and in ver 1.2 that table is not stored by gop anymore, we can of course
undo that and store by gop again but there are 2 issues with that
1. the table needs space too so this would negatively impact 1pass (and 2pass
   does not seem really worth it for end users with this patch yet)
2. i wanted to add tables storing initial states in extradata and if we do and
   if this does lead to some gains we have the problem that changing the state
   transition table per gop will change the inital state tables.

that said i agree its worth a try to see how much can be gained by per gop
sorting

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Why not whip the teacher when the pupil misbehaves? -- Diogenes of Sinope
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20101024/711cbe1c/attachment.pgp>