[XviD-devel] newpred

[peter ross]

hello,

>From: skal <skal@planet-d.net>
>
>	For the record, there's a reference bitstream with NEWPRED
>	stuff turned on in it:  vcon-a1er1-4.bits
>

yep. in fact, that is the bitstream that got me interested in
newpred.

>	This one's a real pain, for it also contains RESYNC
>	markers with different syntax than the usual (since
>	NEWPRED is turned on). Note also that's there a 'bug'
>	-or let's say a typo- in the norm: the 'marker bit'
>	following 'vop-id-for-prediction' (page 41) is not
>	dependent on the 'vod-id-for-prediction-indication'
>	flag, unlike what the indent might let thinking.

yep; i found that one out the hard way.

>	At least, that's what momusys's pile of code says...
>	[btw, it's 3/4 times in a row we stumble over, hmm,
>	let's say 'discrepancies' in the norm/ref-software/
>	divx/... Maybe a clear policy should be set up
>	("We Abide to The Norm", or "We Abide To Mighty FDAM",
>	"etc...")]

hehe.

>	attached is the log of NEWPRED Id for current and
>	reference segments in vcon-a1er1-4. It seems that
>	references are 2 or 3 P-VOP backward...


-- pete

_________________________________________________________________
Add photos to your e-mail with MSN 8. Get 2 months FREE*. 
http://join.msn.com/?page=features/featuredemail