Re[4]: [XviD-devel] motion estimation for B/P/I decision

[Marc FD]

>
> > For I/PB frame decision, maybe my histogramm detection could be very
useful.
> > but i need to work a lot on it to get it ready.
>
> You mean I/PB or I/P/B ?

for any combinaison. depends on how it's used.

> Of course your histogram could do better. I just wrote a very simple
> piece of code ;) that just seems to work. It doesn't mean it's the
> best.

i saw it. sort of sad of MB decision ?

>
> > is SKIP mode possible for BVops ?
>
> Yeah, it's simply direct mode with 0,0 vector. I think they could
> spare this coded/not coded bit. Whatever.
>
> However, I meant skip-mode in P frames. If a macroblock is skipped
> it's also (automatically) encoded as forward-mode 0,0 in all B-frames
> before.
> As I do an analisys of all B-frames before P-frame, I could also make
> sure that skip-mode is allowed without any computational cost (I make
> a SAD of 0,0 forward-mode anyway).

i want to say it CLEAR. i've done LOTS of test/coding/compare on
SKIP decision the last past month, and i'm almost sure of a thing now :
you can't avoid smearing with SAD/SSD or anything else
(Hadamrad based,ect..) because it's impossible to know if it's due
to noise or to very slow gradient move.

the only thing who is precise enough to avoid smearing is a precise ME.

just my humble opinion.
MarcFD  marc.fd@libertysurf.fr