Running audio through osc FX? Technically feasible?

[Sendy]

I was just wondering... I know Osc FX aren't 'audio' effects, they work on the spectral/waveform data of the oscillator, right? Well what if you could convert a stream of digital audio into this data on the fly, and mangle it with the Osc FX? I'd love to run some breakbeats through some of those effects

Not really a request, so much as, wondering if it's technically feasible without a big engine rewrite.

[Urs]

It would be very difficult. The material would have to be devided into proper harmonic single cycle waveforms because that's what the effects are working on (fundamental, phase dist, pretty much all of them... they all take the actual cycle into account). Sampled material however doesn't really consist of homogenuous cycles.

Nevertheless, Tone2's Gladiator synth for instance seems to have sequences of waveforms that indeed resemble sample-like contant that can be recognised as language or loops, baring similarities to vocoded material. So if one had an algorithm that kind of converts the audio into a stream of single cycle waveforms akin to what Gladiator seems to do, then it should be possible to add Zebra's OscFX (I could probably ask Markus about that)

On the other hand side it should also be possible to go for an offline process with proper pitch detection (think Melodyne) or at least a bit more intense analysis like things Camel Audio do.

Somewhere inbetween these might be a solution.

No promises, but I must admit I've been thinking about that a lot.

[Sendy]

It would be very difficult. The material would have to be devided into proper harmonic single cycle waveforms because that's what the effects are working on (fundamental, phase dist, pretty much all of them... they all take the actual cycle into account). Sampled material however doesn't really consist of homogenuous cycles.

Nevertheless, Tone2's Gladiator synth for instance seems to have sequences of waveforms that indeed resemble sample-like contant that can be recognised as language or loops, baring similarities to vocoded material. So if one had an algorithm that kind of converts the audio into a stream of single cycle waveforms akin to what Gladiator seems to do, then it should be possible to add Zebra's OscFX (I could probably ask Markus about that)

On the other hand side it should also be possible to go for an offline process with proper pitch detection (think Melodyne) or at least a bit more intense analysis like things Camel Audio do.

Somewhere inbetween these might be a solution.

No promises, but I must admit I've been thinking about that a lot.

Really? I was going to ask about the harmonic series. I assumed that inharmonic partials were supported, especially when you use some of the more abrasive Osc FX, because things can sound very noisy. That's awesome that you've been thinking about this as well, though.

I have Gladiator, it's an excellent synth, but I haven't used it as much as Zebra, because I prefer to work directly with the waveform (even if it is a spectral engine that unltimately renders the waveform). It feels a bit like 'pick a bunch of waves and processes, twiddle and hope for the best' most of the time for me, lol. Still, I'm a fan of tone2 stuff in general.