[theora] Theora/Thusnelda features questions.

[xiphmont at xiph.org]

> It is said that Theora performs really well with low-resolution
> content, but less with high-def content. How and why is Theora's
> performance related to the size of the video? Is it an aspect of the
> Theora spec (or of the VP3 spec), or is this caused by the
> implementation?

The two biggest reasons Theora is said to be best for low res content
is becasue Theora's maximum motion vector size is relatively small
(+/-31 halfpel) and the minimum quantizer of DCT coefficients
relatively coarse compared to codecs designed for HD.

Monty