[CM] sound internal format?

Bill Schottstaedt bil@ccrma.Stanford.EDU
Mon, 31 Jan 2005 04:00:40 -0800

 >  By the way, when a file is loaded into snd is it internally
 >  stored as a floating point format?

The default is to use 32-bit ints with 24-bits of "fraction".
This dates back to the days when floats were too slow to
contemplate -- perhaps it's time to change the default. If
you build Snd with the configure switch --with-float-samples, it
stores the samples internally as floats (I've been using this
flag for a long time -- I sort of forgot about the whole issue).

If you change formats, you'll need to clear out your old
peak-env directory, if any, because it stores its data in
Snd's internal format.

The other original reason to use ints was that back in those
days, all sound data was stored as some sort of integer, and
I felt that an editor ought to have a certain "reverence
for the data" -- bit-wise fidelity to the extent possible.