Supercollider dev question

[Martin Naef]

On 26.10.2011 08:25, Marc Nostromo [M-.-n] wrote:
> Can anyone tell me what's the dynamic range of the audio representation
> in SuperCollider ? My current framework is currently using a fixed point
> reflection of 16bit range (i.e. +/-32767) but I was wondering if +1/-1
> wouldn't be better. The main point would be that implementing signal
> multiplication (for modulation,ring mod, and so on) wouldn't need some
> kind of additional processing to normalise the result.

There's a lot of good info from THB, so just a few short comments.

Whatever range you choose, it doesn't really matter because it's purely 
a question of convention. For the way my brain works, I'd be most 
comfortable with something that's based around [-1:1] representing the 
maximum output of the digital output, plus a few bits extra to avoid 
overflow and to enable multiplication with gain > 1. Such a definition 
also allows easy conversion to floating point.

You might want to take a look at the fixed point format in the cheap 
Alesis DSPs they used to build the Ion. I think they had a +/- 8.0000 
range expressed with 24bits.

Of course, there's also the question why you're going for fixed point 
(unless you're targeting an embedded system that doesn't offer FP 
support). While it's true that 32bit fixed point can be more precise 
than 32bit float, you also have to be extremely careful to actually 
achieve that extra precision (e.g. your gain structure must really fit 
the chosen range).

Greetings from the US...
Martin