RE: FlyLoops and electronic instruments

[Mech <mech@m3ch.net>]

At 11:50 AM -0800 11/4/07, Aaron Leese wrote:
>
>I've implemented the frequency shift before ...

Aaron, I think you've made reference to this twice now, so I'd like 
to get some clarification if you don't mind.

IIRC, frequency shifting usually refers to transposing a signal 
without shifting the frequencies proportionally to each other.  The 
term pitch shifting is used for the more musical practice of 
transposing those frequencies in a proportional manner.

For instance, let's say that you have a sound composed of three sine 
frequencies at 1200, 1300, and 1400 hertz.  This audio doubled in 
frequency using pitch shifting would result in a sound composed of 
2400, 2600, and 2800 Hz -- musically, a one octave transposition 
upward.  However, using frequency shifting, it would merely add the 
baseline 1200 Hz to each of the values and you'd end up with a sound 
composed of 2400, 2500, and 2600 Hz.  While the lowest (let's assume 
fundamental) pitch is transposed up an octave perfectly, the other 
sine frequencies are actually transposed to inharmonic intervals and 
appear more as aliasing artifacts than true pitch elements.  In the 
real world, merely multiply this example by ~20,000 frequencies.

Pitch shifting is usually the more musically useful of the two. 
Because of that, frequency shifting is much less common (I can only 
think of a few Alesis products and some modular synth modules, right 
off the top of my head).  Many of the results of a frequency shifter 
sound somewhat like a bizarre form of ring modulation [sic].

So, I'm just curious: which algorithm are you talking about possibly 
implementing?  If you were implementing true *frequency* shifting 
into FlyLoops, then I'd imagine there are quite a few noise junkies 
out there (myself included) who'd be keenly interested, although it 
might alienate a few of the more traditional crowd too.  :)

        --m.
-- 
_____
"the wind in my heart; the dust in my head...."