[Open-graphics] RGB/YUV converstion... importance?

[Attila Kinali]

On Wed, 13 Sep 2006 22:51:16 +0100
josephhenryblack <josephhenryblack at yahoo.co.nz> wrote:

> I have added croma subsampling and YUV to the glossary, but you may wish 
> to review it http://wiki.duskglow.com/tiki-index.php?page=Glossary

Some comments:

4:2:2: Upsampling should be always done by interpolation
and not how it is mentioned by nearest neighbor (aka
copying the chroma value from the left).
The YUYV format outlined there is just a packed representation
that somehow fits, because U and V together are as large
as Y.

4:2:1 does that even exist? I've never seen it nor
heard of it

4:2:0, the most important subsampling format is missing

4:1:1 and 4:1:0 are similar to 4:2:2 resp 4:2:0 but use
a smaller sample size in horizontal direction.
But those are only relevant for a handfull of specific
codecs. Also it's easily possible to upsample these
in software to 4:2:2 and 4:2:0 respectively
(though i doubt that the hardware required for this would
need much area, but if we need a few gatters more for something
else, it would be the first function i'd drop)

Important to note here is that the chroma samples are often
half a pixel off the luma pixel positions. Thus every sample
needs to be interpolated and noone can be just copied.
But there are different standards even for that, depending
on the codec. 


Maybe a link to 
http://vektor.theorem.ca/graphics/ycbcr/Chroma_subsampling_notation.pdf
Would be something. It explains a few of the things we need
(though not all of it).

				Attila Kinali

-- 
egp ist vergleichbar mit einem ikea bausatz fuer flugzeugtraeger
			-- reeler in +kaosu