Ben Jefferys wrote:

> Charles Baker wrote:
>
> > Why the heck isn't there *more* emphasis on floating point in musical
> > signal processing ? Makes all the difference
>
> Of course it doesn't necessarily have to be floating point. I'd guess
> a large part of the reason why floating point stuff is better is
> because it uses more bits - typically 32/64/80 (in think they're
> the standards). If you used 32 bit ints the quality would improve
> by a similar amount to your standard C floats, although there would
> be subtleties lost in quieter audio compared to using a float due to
> the "logarithmic" nature of floating point.

Exactly: only floating point gives you full usage of word, no matterwhat
the absolute value of the  sample:within limits, of course, but
*practically*
it gives better results ! Even if we're talking short floats, no longer
than the
int formats, the results are often quite noticably better: if only because
darn few
of us make sure that all samples input and output are scaled to use all of
the
word. And this brings up another advantage of F.P.:  when int sounds are
used,
one has to be extremely careful in combining them, scaling the input
samples
to assure that you don't overflow. This is not a worry with float samples:

they can be combined freely & without concern over the absolute value of
the result,
and then scaled/converted into ints to play. CLM and jpff's WinCsound (for
example) both
provide this as a way to synthesize your sound, a way many experienced
composers
prefer. Try it , you'll like it!
As for the processor speed issue:I don't care if floating point is a bit
slower: "real-time" is not
particularly a big issue with me ( I am rather tired of simple synth
sounds, such as are used to give
polyphony in real-time on PCs). If making your pc act like a very poor
realtime synthesizer is your
raison d'etre, go to it. I'll go buy a much cheaper GM MIDI module for
that. I use csound for the
sounds you *don't* find on the GM modules, thank you.



> But the 16-24-32 bit
> transition within the integer realm is very significant, and on
> some processors integer performance is superior.

 Well, if one is careful with normalization of sounds, and if you
carefully attend to scaling issues in
mixing,  I agree larger int word sizes will produce better results. Too
many ifs there for me, and I
don't  have the money right now to go buy a true 32-bit machine (hmmm...
what's the word on
DEC Alpha Csound? And is csound built in irix32? Anybody got ~$6k lying
around? )

>   Bye!
>   Ben
>


Fare thee well, Ben. Where are you going? I'm right here.....
Char lieB

--
*********************************************
Charlie Baker              baker@charlieb.com
"Das Ewig-Weibliche Zieht uns hinan." -Goethe
*********************************************