[XviD-devel] Rounding

[Christoph Lampert]

On Wed, 30 Oct 2002, Michael Militzer wrote:
> the sum of the K luminance vectors, that corresponds to K 8x8 blocks that
> *do not lie outside* the VOP shape [...]"
> 
> so what? are luminance vectors which are pointing outside the vop not used
> for the calculation of the chrominance vector??

I think the emphasis lies on _shape_, and I will forget about it until we
implemented arbitrary shaped VOPs.

gruel