ðòïåëôù 


  áòèé÷ 


Apache-Talk @lexa.ru 

Inet-Admins @info.east.ru 

Filmscanners @halftone.co.uk 

Security-alerts @yandex-team.ru 

nginx-ru @sysoev.ru 

  óôáôøé 


  ðåòóïîáìøîïå 


  ðòïçòáííù 



ðéûéôå
ðéóøíá












     áòèé÷ :: Filmscanners
Filmscanners mailing list archive (filmscanners@halftone.co.uk)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[filmscanners] RE: JPEG2000 > Paul





>-----Original Message-----
>From: filmscanners_owner@halftone.co.uk
>[mailto:filmscanners_owner@halftone.co.uk]On Behalf Of Paul D. DeRocco
>
>My guess is that JPEG algorithms have been boiled down to carefully hand
>coded SIMD routines (i.e., using MMX on a Pentium II or higher), but no
>one's gotten around to doing that yet for JPEG2000. On the other hand, it's
>possible that JPEG2000 doesn't lend itself to that sort of programming, and
>will never be as fast. We'll see.

Jpeg is a fairly simple algorithm and as you said has been optimized a lot.
Nevertheless, kakadu has been optimized a lot, too. It uses quite a bit of
MMX. I have also added SSE for certain parts of the code. But on top of that
kakadu uses some other sophisticated approaches to considerably improve
performance. Despite that, j2k is still much slower then jpeg. The reason
for that is that the algorithm is much more complex then jpeg. On top of
that the most compute intensive part is not very suitable for SIMD.
Nevertheless, keep in mind that when jpeg did become popular it was fairly
slow on the machines available at that time.

Robert

----------------------------------------------------------------------------------------
Unsubscribe by mail to listserver@halftone.co.uk, with 'unsubscribe 
filmscanners'
or 'unsubscribe filmscanners_digest' (as appropriate) in the message title or 
body



 




Copyright © Lexa Software, 1996-2009.