[XviD-devel] dark blocks in b-frames

Christoph Lampert chl at math.uni-bonn.de
Fri Feb 14 19:04:45 CET 2003


On Sat, 15 Feb 2003, Radek Czyz wrote:
> First, I just used DC value after fdct. When I saw what's happening I
> also checked the values before fdct, to see if it's not fdct problem.
> 
> Take a look at this dark block:
> 
> block number 2 before fdct:
> 
> -2 -4 -1 -1 0 -4 -3 -3
> -2 -2 -1 1 0 -2 -3 -2 
> 1 -1 -2 -2 1 -3 -2 -2 
> -2 -1 -3 1 1 -2 -4 -2 
> -2 -1 -3 -3 -1 -3 -2 -4 
> -3 -3 -3 -3 -2 -2 -3 -3 
> -1 0 -1 -3 -1 0 1 -2 
> -3 -1 -1 -3 -3 0 0 -3 
> 
> DC value is -14, quantized to -1. AC values are 0. Mode direct,
> compensated with "MODE_DIRCT" as this is quarterpel and _NO4V is not
> available.

And there are no such blocks with potive values? DC==+14 quantized to +1 ?


> > If I'm not mistaken, the "too bright" effect is there also
> > without "too dark" blocks, but not always visible due to
> > quantization.
> 
> No, the 8-bit compensated image is never too bright in the decoder.

I didn't mean that. I meant do _all_ blocks have this slightly negative
values (-3, -3, 0, -2, 1), but most of the time this is quantized to
0. Only in rare cases it's quantized to 1 and the block become "dark
flickering". 
Or are the blocks with negative value exceptions? 

gruel



More information about the XviD-devel mailing list