Re: [motm] Partial solution, or up your ante!
2000-03-24 by jwbarlow@aol.com
"What do you mean up my auntie? Up your auntie Barlow!!!" I go to work one day, with a couple of ideas about generating non-harmonic overtones, and come back to a bunch of mail about guitar synths! One thought I have about them involves my experiences using an old 360 Degree Systems Slavedriver about 20 years ago: a good device......if you need a paper weight. OK, I was thinking of Ken's ideas (below), and would like to add a couple of my own. But first, I think you've got an interesting idea about single side band shifting, have you tried taking a RM out and running it into about three or more 420s (or other VCFs)? I'd be interested in what you find. That steep filter slope is what you'd need, but everyone hates it. First, I'd imagined ganging two VCOs together such that they shared the same 1V/oct inputs (maybe the same expo converter -- like the old Moog Oscillator Driver module), but had separate linear FM inputs which I would hope might minimize the error encountered during FM. I imagine this is not easily accomplished (maybe not even possible) but it might be worth looking at if it is since many people use simultaneous VCOs in order to thicken the sound. The other idea I had (and have had for a long time) is quite digital, but has a certain similarity to Ken's second, additive, idea. We could use all those stale old computers that are accumulating in everyone's garage as VC digital oscillators. I imagine using the monitor to display harmonic amplitude of say the first 16 partials. One could specify the relationship between the harmonics (yeah Ken, you'd have to use a keyboard/mouse). I think such a system would be quite useful for some of the things that Ken mentions, and I think the display would (as Ken intimated earlier) be much more intuitive than the typical display/draw the waveform type. JB In a message dated 3/22/2000 2:17:42 PM, ken.tkacs@... writes:
Show quoted textHide quoted text
>Okay, here are some thoughts on modules that change harmonic spacing. >For a modifier, what about something that basically does what a ring >modulator does (but not through simple multiplication obviously)---change >the spacing between harmonics plus or minus---but along a controllable >curve. Is that clear? I'm fumbling for language here. A ring modulator >ends >up expanding or contracting the harmonic spacing in a linear way, and >because pitch has an exponential relationship to frequency, we get the >weird >metallic non-harmonic sound. What if you could do something similar, but >with "curves" that could be voltage-controllable? As well as the deviation >amount? >Hope that description makes some kind of sense. Due to the nature of a >ring >modulator, the higher harmonics in the upper sideband get affected more >than >the "closer" ones, and the reverse for the lower sideband. If you could >control the effect along a curve, you could really control the effect and >make it very subtle. At close to exponential I think it would approximate >a >pitch shifter. At linear, it would create the upper sideband of a ring >mod, >and inverse linear, the lower. >For an additive source module, I guess some kind of high-frequency VCO >that >feeds, say, eight "divide-by-N" chains that can divide by large numbers. >Each of these eight chains goes to a VCA and a square-to-sine waveshaper >and >are then mixed. These modules could be ganged to add more harmonics off >the >same source VCO (or maybe the HF VCO is one module, like a 'driver,' and >there's a second module for the divider banks that can all be fed from >one >driver). I realize that this gets to be a lot of circuitry. Probably more >expensive than eight VCOs. But with these things feeding off the same clock >they are locked tight in "tune." Ideally you'd want to be able to specify >the phase of each harmonic, too, but that's getting to be crazy. >Anyway, that's first pass at daydreaming. I wonder what that first module >would sound like. I bet it would be a very weird choruser if used subtly >and >mixed back in with the original signal.