>There are artefacts, little rectangles not update for a few frames - I
>don't know if this is a performance issue. Audio is faster than video,
>they're out of sync. And for some files, it crashes after a while.

Andreas mentioned something about transparency at OOPSLA, not sure 
what should go in the Alpha byte, or if it was a bug? in morphic. If 
Andreas is listening perhaps he could comment on what the issue is.

Right now I might not drop enough frames per second to keep the frame 
rate upto the audio rate, since I do limit the drop rate, I'll look 
at changing that.

>
>And I had to change the color mode from 13 to 1 (MPEG3_RGBAF8888 to
>MPEG3_BGRA8888) in the player. But this should be changed depending on
>"Smalltalk endianness".
>
>Apropos constants: I'd rename MPEG3_RGBA555 to MPEG3_BGR555 and RGBAF8888
>to ABGR8888 - or whatever these modes actually do ;) Or is the F in
>RGBAF8888 for "big endian"? Anyway, I'd need a byteswapped BGR555 mode ...

Correct RGBAF should be ARGB (I think). Since I've not looked at the 
endianness issue I'm not sure of the ordering issue for Squeak Forms, 
does ARGB become BGRA? .

At 1:15 AM +0100 11/5/00, Bert Freudenberg wrote:
>
>I got nasm now, tried it - there's no observable difference, even with
>3Dnow. It seems decoding speed is not the bottleneck ...
>
>-- Bert

Time to profile and look. Certainly would one go to the hassle of 
writing assembler unless they thought they could buy some performance?
-- 
--
===========================================================================
John M. McIntosh <johnmci at smalltalkconsulting.com> 1-800-477-2659
Corporate Smalltalk Consulting Ltd.  http://www.smalltalkconsulting.com
===========================================================================
Custom Macintosh programming & various Smalltalk dialects
PGP Key: DSS/Diff/46FC3BE6
Fingerprint=B22F 7D67 92B7 5D52 72D7  E94A EE69 2D21 46FC 3BE6
===========================================================================