Re[15]: [XviD-devel] New Motion Estimation (from sysKin) committed to branch

Radoslaw 'sysKin' Czyz xvid-devel@xvid.org
Sun, 29 Sep 2002 23:57:51 +0930


> After I commented out the following lines:

//   if (!(current->>global_flags & XVID_LUMIMASKING)) {
//    pMB->dquant = NO_CHANGE;
//    pMB->quant = current->quant; }

> in MotionEstimation() (motion_est.c), the artifacts were gone... This is
> extremely strange and I'm sure that these three lines cannot be the real
> source of the artifact problems, but I didn't have the time yet to further
> look into this...

It sounds like inter4v problem to me - if pMB->dquant != NO_CHANGE,
then inter4v is off. If lumimasking is off, pMB->dquant is
uninitialized, so it's very likely to be != NO_CHANGE.

However, there was a very stupid problem with older motion_est.c,
which in reality limited motion vector range to 0,0 + halfpel IF
inter4v was off.
I have no idea how I made it...... but please check if the beginning
of SearchP in your file is the same as in cvs. In particular, there
should never be any MV check before normal loop. How did I make it..?

Best Regards,
Radek