posted 09 January 2004 10:59            

y = ax**2 + bx + c

In other words, it is any expression that has an x "squared" in it. The shape is like a cup (or inverted cup) with a single minimum (or maximum) value at the center.

Here's a site for experimenting with parabolic functions http://www-gap.dcs.st-and.ac.uk/~history/Java/Parabola.html

A sinusoidal function is different in that it is periodic (repeating) and the slope of the curve is completely different. It has an infinite number of minima and maxima (because it repeats)

y = sin x
http://jwilson.coe.uga.edu/EMT668/EMAT6680.2001/Mealor/writeup1/assignment%201.html

posted 15 January 2004 14:03            

Does anybody have an idea what might be interesting to use a parabolic function instead of a sine as waveform for an oscillator?
More dirt?
More energy?

posted 15 January 2004 14:14            

At a discontinuity, you generate an infinite number of harmonics (hence, the "click" you hear when you make a bad splice). If you repeat this discontinuity in an oscillator, you will hear aliasing and "buzziness". In the digital world, all frequencies are assumed to lie between 0 and half the sample rate. So the frequencies above half the sample rate are re-interpreted as

n*SR - freq

Just out of curiosity, why do you ask? Is someone recommending this technique? They might be using it to get a kind of harmonic distortion sound (albeit with some nonharmonics resulting from the aliased components).

posted 16 January 2004 02:33            

I stumbled across parabolic functions first in context to Jomox products: http://www.jomox.de/manual/aireman_e.html (make a search on parabolic).
Then I discovered that there is a synced parabolic function prototype in Reaktor which is used in some of their drum programs (as bass drum).
Following the idea of the Jomox Bass drum Box MBase01 (at www.jomox.de) I started to prgram my "ultimate" Bass drum Synthesizer inside Kyma. It will have a lot more to offer than this Jomox thing but I´m still experimenting with waveforms. This MBase01 is supposed to use "modified" triangle, for example.

What I also suspect that the plane of this parabolic function is bigger than of a sine, is this correct? That would result in more energy and a more compressed sound, I assume.
(By the way everyting in Reaktor seems to be somehow compressed. I do have the suspicion that there is some kind of waveshaper built in in every module. But maybe I´m turning to conspiracy theories now...)