[Theora-dev] Theora camera/fast computers

[Andrey Filippov]

> Exactly, I think it is frame 256, though I can not see the "turbulence"
> on my screen, now looking at the frames over vnc from a remote host.

I've seen some (non-accumulating) faint trail after the moving objects
(especially over the uniform background) and tried to explain it in one of
the previous posts - I believe it is because quantization tables have
larger theshold for the INETR frames and the difference in this trail was
visible, higher then for the INTRA, but still too low to be coded in
INTER. But I can not be 100% sure that it is not a bug still hiding.