[XviD-devel] SAD of last visited neighbour positions

[Marco Al]

Michael Militzer wrote:

> The whole thing might become even more interesting for qpel: Currently we're
> using image based interpolation for the halfpel positions using a 6tap
> low-pass filter (as proposed by the JVT specs). In my tests this seemed to
> be a pretty good approximation (at least for ME), however since I made
> several changes to the interpolation (MC) code  since then, I feel I should
> redo these tests. Maybe the correct way (block-based interpolation using the
> 8-tap low-pass filter for calcualting the halfpel positions during the ME
> step) will give better results now. If so, a fast block-based refinement
> will give a substantial speed improvement here, since the needed
> interpolated pixels are much harder to calculate in qpel than in halfpel
> mode...

Shouldnt that be the other way around? The more computations needed for
interpolation the less a block based approach will get you right? (Since you are
basically trading extra computations for less memory accesses, so the more
computations the worse the deal.)

Marco