Re: [microsound] (science related crapola) (a note on arithmetical precision)

[Alex Young <alex@xxxxxxx>]

On 14 Jan 2004, at 21:40, Arie van Schutterhoef wrote:
> > Pieter is intested in the atomic data of primary elements, 
> > particularly
> > metals, and has evolved an incredibly complex microtonal and granular
> > synthesis system around this data.
> -He needs 64 bit precision for these kind of things,
>  only SGI machines can do this. Although the new
>  PPC970 and the AMD Opteron are capable of this too.

The precision isn't necessarily bound by your hardware.  Obviously if 
you
want to process things in real time using the CPU's maths functions are
the most efficient, but you can work in any precision you want in 
software.

The GNU Multiple Precision library allows you to work with any precision
you like.  I evaluated several libraries like this as part of my 
dissertation
when I was at university, and found GMP to be quite easy to work with.

http://www.swox.com/gmp/

It would be possible to adapt cSound to use GMP for its internal data,
thus allowing you to work with any level of precision.  I haven't 
looked if
such a thing exists, but a distributed implementation of cSound using 
GMP
would be rather useful.

--
http://alex.bash.sh


---------------------------------------------------------------------
To unsubscribe, e-mail: microsound-unsubscribe@xxxxxxxxxxxxx
For additional commands, e-mail: microsound-help@xxxxxxxxxxxxx
website: http://www.microsound.org