Hi Bernardo,

On Tue, 15 Nov 2011, Bernardo Barros wrote:

> The whole idea of writing dsp algorithms directly in haskell is very
> interesting. I think it would be much more intuitive to extend synthesis
> systems like this.

That's what I hope for. Currently there still some low-level clutter 
around.

> But there are so many good options like supercollider (and hsc3 for
> instance), that it is very discouraging using a less efficient system
> for real-time.
>
> You told you could not do much complex synthesis with your system,

Did I tell this? What I wanted to express was, that latency and complexity 
of sounds are two separate issues. I can't reduce latency by reducing 
complexity. But I can handle complex sounds at the same latency as simple 
sounds.

> is this because you don't have a bigger block size?

I can freely choose the block size.

> Or Haskell garbage collector? What is your strategy in this respect?

When I had a problem with the garbage collector I had actually always a 
problem with a space leak. I am now on a good way to eliminate them using 
arrows instead of lazy lists. However I find programming this way more 
complicated.

> There are other systems trying to make decent DSP performance with
> languages other then C. The second version of JSyn is all written in
> pure Java, not c like the first version. That is good! The same language
> for high and low level work,

Yes, that's I achieve with LLVM and Haskell, too.

_______________________________________________
haskell-art mailing list
[email protected]
http://lists.lurk.org/mailman/listinfo/haskell-art

Reply via email to