[MPlayer-cvslog] CVS: main/DOCS/tech snow.txt,1.4,1.5

[Loren Merritt]

On Fri, 22 Apr 2005, Guillaume Poirier wrote:
> On 4/22/05, Oded Shimon <ods15 at ods15.dyndns.org> wrote:
>>
>> Doesn't this NOTE seem out of place?.. I don't see why it's said
>> twice, and here..
>
> You're probably right. It's just that if you have a look a the full paragraph:
>
>  * cmp, subcmp, mbcmp
>   Set the comparison function, default: 0 (SAD).
>   useful values = 0 (SAD), 1 (SSD), 11 (5/3 wavelet), 12 (9/7 wavelet).
>   Experience shows that SSD is the best most of the time, while SAD is
>   slightly better the remainder of the time.
>   You can add 256 to any of the options to enable chroma motion
>   estimation for that comparison (e.g. mbcmp=257 for SSD with chroma),
>   but it doesn't seem to help much for the moment.
>   NOTE: 9/7 wavelet doesn't work with lossless mode.
>
> This paragraphs does talk about 9/7 wavelet.
> Now, maybe I got confused by it, and it doesn't need to be there, or
> maybe I'm right! ;-)

*cmp=12 does work in lossless, even if 9/7 wavelet doesn't, because 
cmp is only used to choose MVs, not part of the residual coding. It's just 
that even if wavelet cmp helped, you wouldn't use cmp=12 with pred=1.

--Loren Merritt