[XviD-devel] Naive/newbie questions

[Greg Handi]

> > 2) also per the interpolation routines, is there a way to get the
> > "src" parameter to always be a four byte aligned memory address?
> > (I realize this may not be possible/practicle given how it is
> >  derived but I thought I should ask)
> 
>   Usually, the "src" reference samples can be anywhere in the
picture,
>   and in particular are not aligned at all. Now, you can do some
>   aligned reads of more samples and shift the whole in the right
place
>   ("funnel shift"), but ... it's a patented trick:
>
http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO1&Sect2=HITOFF&d=PALL&p=1&u=%2Fnetahtml%2FPTO%2Fsrchnum.htm&r=1&f=G&l=50&s1=6,449,706.PN.&OS=PN/6,449,706&RS=PN/6,449,706
> 
>    too bad hey? :)

Interesting patent; there are funnel shift op-codes on the chip I'm
using so that does work but introduces 3 more operations per 4 bytes
(extra read + extra average + funnel shift) but at least big-endian vs.
little-endian doesn't require a byte swap. ;)  I think it is still 1/3
as many instructions in the inner loop but unfortunately introduces
extra branching...

Thanks for answering, hopefully I can be of some use to the project
some day. :)


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com