On Wednesday 05 December 2007, Rob wrote:
Sure, but those things were probably designed and tuned for such CPUs,
whereas modern softsynths in general, are designed and tuned for
P-III and better.
It's possible to get pretty decent sound out of "simple" integer DSP
code, but it's hard work and doesn't scale well. You often have to
decide whether you want to resolution or speed - and that's before
even considering algorithm variants, such as different interpolation
methods, oversampling etc. In short, more work for much less - so if
you can get away with floating point (which you can on anything like
a P-II or better), integer doesn't make sense.
> The Via and Geode chips are pretty weak, and I'm sure Jack and
It shouldn't have to be, theoretically, but if it relies on the FPU,
it could still be a lot slower than an integer based synth on such
> And for what it's worth, you can get micro-ATX boards that take
Yeah, and Mini-ITX boards for multicore CPUs are showing up now.
We're using Mini-ITX boards with 2+ GHz "P4" Celeron CPUs in lab
instruments. Compared to the usual industrial SBCs, they're lower
cost, many times faster and include more "might come in handy"
integrated hardware. More power than we really need, theoretically -
but it's easier to code in EEL (real time scripting language) than C
or C++, so you won't hear me complaining! ;-)
//David Olofson - Programmer, Composer, Open Source Advocate
.------- http://olofson.net - Games, SDL examples -------.
| http://zeespace.net - 2.5D rendering engine |
| http://audiality.org - Music/audio engine |
| http://eel.olofson.net - Real time scripting |
'-- http://www.reologica.se - Rheology instrumentation --'
Linux-audio-user mailing list