[XviD-devel] SAD vs MSE for ME (warning! acronym overflow!)

[skal]

	Hi everybody,

	just wanted to know: is there any reason to prefer
	SAD as criterion for ME, instead of MSE, e.g.?
	I'd tend to think SAD was faster to compute than
	MSE with mults, but is there any other reasons?

	bye,

		Skal


PS: there are bugs in the interpolate_XXX_xmm() code
I sent lately. Some loops for the _HH_ serie (both
copy and add, 16x8 and 8x8) are unrolled twice too
much! Should be:

  ADD_HH_RND0
  lea ecx,[ecx+edx]
  ADD_HH_RND0
  lea ecx,[ecx+edx]
  ADD_HH_RND0
  lea ecx,[ecx+edx]
  ADD_HH_RND0

only, e.g.
I'm terribly sorry 'bout that :(

PS2: For the CodeVector() I sent recently, I'll try to extract
the old code that I used to generate the big pre-calc-ed table so
it can be initialized at run time...