[Fluxus] puzzle

[Kassen]

Dave;


> I'm wondering where this curve comes from - it must be based on some
> frequency response.
>
>
Well, basically we want to go from 1 to 255 in steps that are equal in size
as multiplications. This is because of how our hearing works. To our ear the
distance from 40 to 80Hz is as large as the distance from 10.000 to 20.000
Hz (one octave, in this case. If we'd just cut the spectrum (from 0 to
20KHz) into -say- 10 equal parts on the linear Hz scale then half (5 in this
case) of the parts would be in the top octave (from 10 to 20 KHz). That
clearly isn't desirable as that would mean half of the bands would basically
be tracking snare hits and the like. Hence it makes sense to use a
multiplication per step, rather than a addition, these would basically
equate to musical intervals.

If you want to get clever about it all you could have slightly higher
resolution in the area where our hearing is most sensitive (say 300Hz to
1KHz) but I'm not sure that would be worthwhile as I'd expect Fluxus to be
used a lot with highly synthetic sources.

Classical Vocoders (moog, etc) use a series of bands that take a range of
factors into account, including gut feeling of the creator. These can sound
great but manually tuning a system with that many parameters takes ages so
if variable numbers of bands are a concern I'd go with a exp curve and leave
it at that.

Hope that helps and that it even answers the question. It wasn't quite clear
to me what exactly the question was or what was unclear; appologies if some
of this was obvious. I didn't do the math, I'll trust Matt that it checks
out. I would BTW expect these numbers to depend on the samplerate that Jack
is running at as that will dictate the maximum frequency that's of interest.
Proper behaviour would likely be polling Jack for the sample rate, then
taking the desired number of bands and basing these numbers on that.

Yours,
Kas.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.pawfal.org/pipermail/fluxus-pawfal.org/attachments/20100117/dd21b434/attachment-0001.html>