Building an FX Device

[James Coker <jcoker@interaccess.com>]

Jon Durant wrote:
> 
> 
> The big issue is: is there a market out there? So far the returns on 
>Jam/Plex
> etc have indicated not. If someone asked me if I'd be involved in a 
>company that
> made *only* looping devices, I'd say no. Too risky. But a new 
>effect/processing
> company with a little bit of creative vision *is* a real option. 
>Consider how
> many people hate Digitech/Alesis/etc but still buy the stuff anyway. 
>Well, what
> if someone new came in and answered the issues, made the right products 
>at the
> right prices. Thought about the user for a change. Hmmm....


I have to agree with Mr. Durant here.  I think the market for a
dedicated 
looping device is rather small, but a truly interesting and flexible
effects device could be a different story entirely.

Off and on for the past several weeks I've been looking at products such
as SuperCollider and Kyma for creative sound maipulation.  Both of these
products allow one virtually unlimited creativity in sound production.
How about a Vortex Fractal B patch with four dual delay paths, each
delay 
adjustable from a few millisecs to several seconds, with various lfo's 
to control panning, and perhaps some FM thrown in for good measure?

SuperCollider is a relatively inexpensive ($300) program for power
macintoshes
that uses the CPU for signal processing. It can input straight from the 
mac's sound in ports, and send the sound right back out.  The bad part
is
that thanks to various output buffers, there is a minimum delay from
signal
input to signal output, which in the few tests I've been able to do with
the demo program, is a minimum of around 0.4 seconds.  That is workable
for
doing loops, but very frustrating given the tonal modifications it can
do.
Another issue with SuperCollider is that it's patches are written in a 
relatively terse lisp dialect.  This is great if you're a programmer,
anything
can be modifed sample by sample, but would suck if all you want to do is 
modify a few obvious parameters. SuperCollider does have the ability to
build
simple dialogs to control parameters, though.

The Kyma system is just about opposite in all parameters.  It is *VERY*
expensive,
at $4500, does not have a large minimum delay (reportedly 10ms for most 
operations), and does have a graphic patch editor.  Kyma comes with a
DSP
mainframe that holds 8 cards, each with a 66MHz 56K DSP chip and 3MB ram
for samples.  The base unit (included in the price) comes with two
cards.
Add a couple more cards ($600 each) and you can do some truly outrageous
stuff,
such as a detailed FFT analysis of the signal, modify the analysis data
in the frequency domain, then resynthesize it, all in real time (but
with
a 0.5 sec or so delay).  I doubt that even Eventide can do pitch
shifting
that accurately.

So, Kyma is very nice, but proprietary hardware makes me nervous, what
with
dual CPU 250 MHz Macs coming on the scene, and, well, gosh, that price
is 
a bit steep.  Though, giving them credit, it's not that bad compared to
say an Eventide 4000, and there really isn't any competition for it.  

The approach I'd like to take is to build a simple PCI card to provide
high 
quality analog ins and outs with very little delay, modify SuperCollider
to 
use that card for i/o, and build a gui patch editor for it.  The selling
price 
could easily be kept under $1000 (or less?). That doesn't include the
computer, 
but then many folks have them anyway, and the software/expansion card
solution 
avoids investing in proprietary hardware. In return, you get a sound 
creation/manipulation platform that is utterly flexible.  Someone could 
program it to be the ultimate analog synth, or try physical modeling,
or build a guitar synth that merges the two, have bizzare arpeggiators
track audio signals without midi conversion delays.  And then add a
second
i/o card for quadraphonic sound...with quadraphonic delays, of course!


Jim