octapus
23rd February 2002, 22:25
I was doing a second pass with stats which was curve-scaled in GuardianKnot and got a 0% weight on credits.
So i did the 2nd pass with external method (modulated) but i put a constraint on Intra frames with values 2-4.
When movie finished main body and got to the end where credits were there were some stange things happenning, maybe because of the constraint because inter values were usually from 21 to 31 and when suddenly (every 200 frames) an intra frame was formed it took a vale greater than 31. I saw values like 35,39,56,69,45 e.t.c
Watching the movie afterwards there were problems on those specific parts.
Now normally there shouldn't be any frames (inter or intra) bigger than 31.
Largest quantizer value given in XVID Coded Dialog box was 31. And even if there were something wrong there, the codec itself puts maximum value to 31 from what i saw in the sources.
This is the first time something like this has happened but also the first time using consteraints on a guradianknot-edited stats.
P.S I was using Koepi's 17 February binaries.
So i did the 2nd pass with external method (modulated) but i put a constraint on Intra frames with values 2-4.
When movie finished main body and got to the end where credits were there were some stange things happenning, maybe because of the constraint because inter values were usually from 21 to 31 and when suddenly (every 200 frames) an intra frame was formed it took a vale greater than 31. I saw values like 35,39,56,69,45 e.t.c
Watching the movie afterwards there were problems on those specific parts.
Now normally there shouldn't be any frames (inter or intra) bigger than 31.
Largest quantizer value given in XVID Coded Dialog box was 31. And even if there were something wrong there, the codec itself puts maximum value to 31 from what i saw in the sources.
This is the first time something like this has happened but also the first time using consteraints on a guradianknot-edited stats.
P.S I was using Koepi's 17 February binaries.