Kevin Fisher <kgf at golden.net> wrote:
> Well I don't use ESD, I use aRts and artsdsp BUT...
> 
> Using artsdsp with Squeak there is a small delay...this is somewhat
> configurable since aRts lets you specify the priority, right down to
> 'real time' (which uses a certain amount of CPU).  If I crank the 
> responsiveness the delay 'appears' to go away to my ears.
> 

Thanks.  I'm surprised it's this good.  I wonder if esd can match it?

By the way, priority shouldn't matter; the issue is *buffer sizes*.  In
ESD, as far as I can tell, and as the authors told me, there isn't a way
to minimize the buffer to, say, 1 second of audio.  Instead, you have to
wait for the network socket to block, which might hold *several* seconds
of data.  A three-second delay just seems so poor of an implementation,
for Squeak's purposes, that it wouldn't be worth bothering with.

Just imagine a morph bouncing off a wall, and then going halfway back
across the screen before the sound is heard....




-Lex