[XviD-devel] Reduced-Resolution P-frame

[skal]

	Hi all,

	does someone feel like implementing the reduced-resolution
	stuff in xvid? If so, I can supply half of the hard work:
	C/x86/MMX/SSE functions for the 1-3-3-1 filtering and edge
	deblocking. Preliminary tests I've conducted seems to show
	that: it's no very worth for I-VOP; it seems to be of some
	help for fast-moving scenes to alternate P-VOP/reduced-P-VOP.
	Nothing definitive, just a "feeling"...
	that's it,
	bye
		Skal


[Some pitfalls to be expected:
a) the norm is not clear about the exact way to re-use the 'scaled
down' motion vectors.
b) it's not clear either about field-predicted macroblock post-filtering
c) deblocking vertical post-filter implies an additional delay
of 1 macroblock row for the per-slice image output.]