Tuned Noise Gen (.. but not as we know it Jim)

[HughBanton]

I thought of this when I woke up this morning - but I admit I didn't expect it to be so effective. It sounds exactly like random noise with a precision-tuned filter, but there's no noise gen or filter in sight.

This method must have been done before, no? I don't recall having seen it. If it's worth pursuing it could obviously be massively optimised into one piece of code. Anything to be gained .. dunno.

Well ... I'm sure I can find a use for it anyway

H

osc_ring.fsm

(15.96 KiB) Downloaded 1112 times

[Spogg]

Very clever!

This doesn't even register CPU on my PC, so it's less than 0.1% I guess. This is great because there's no filter to run to get the pitching. Organ pipe noise??

Cheers

Spogg

[tulamide]

This method must have been done before, no?

No idea, if exactly that way, but basically yes. For example, the Tyrell synth uses its oscillators (who run seemlessly between sine and square) to create noise. But probably this is different.
https://u-he.com/products/tyrelln6/

[HughBanton]

And also, I have realised, not unrelated to how a DX7 creates noise, if you employ an FM algorithm with feedback loops. Which is in effect what I've set up here.

I've also discovered that if you reduce the input values drastically you can get the thing to run in slow motion, demonstrates how it works. It's kind-of randomly sweeping up & down between 0Hz and the input note-frequency'. Done at full speed that's going to thow up hf sidebands, which I guess sound like noise, mostly below the note-frequency. Summat like that.

Yes, it's going to be very useful in HB organ world

All I need now is for MV to dive in with the triple-oscillator code, hoho. I dare say I can succesfully string together the code for three of them, but you can bet there'll be a way better method. (Would help if I understood how they work, hmm ...!)

H

[k brown]

Could be used for an interesting alternative to something like Majken's Chimera.

[martinvicanek]

I think FS' stock noise generator uses a sine oscillator with feedback. I always thought that was a very expensive method. The interesting feature of Hugh's three stages design is the sharp tunable cutoff.

[k brown]

Indeed the old noise gen is good old Yammy-style FM (which was actually phase modulation). Exactly DX7 Operator Feedback.

[adamszabo]

I like it more when there is a single normal whitenoise osc, that goes through a band pass filter with high resonance, and thats keytracked. Those make nice musical noise sounds

[k brown]

That's is what Chimera is.

[HughBanton]

My accidental tuned noise generator seems to be a bit of a 'one trick pony'; I've tried several variations but yesterday's original setup still seems to be optimal. But I do rather like the sharp rolloff - can anyone suggest a method of measuring it?

Anyway .. here's my take on merging all the code into one, so that its cpu-sage is minimised and can be compared against the traditional method of a regular noise gen + filtering. I dare say it can be shrunk further .. there's usually something!

osc_ring_assem.fsm

(17.34 KiB) Downloaded 1113 times

Hope others find it useful.

H