Yesterday I found a surprising way to generate noise: Feed an HPF or BPF's input with the absolute value of its output at a high gain. Depending on the filter and parameters used, it will generate white noise, brown noise, narrowband noise, regular or random bursts of noise, crackly sounds, etc.

I did this in Softube Modular (it works with the Korgasmatron II and the Doepfer VCF-8), and was wondering if someone could try it IRL and let me know if it works. I don't own anything in hardware that has an absolute value function. The two above-mentioned filters have input gain knobs that go to clipping, which helps a lot.

Is this already a known trick?


Addendum: After a couple searches online, I can't find any hardware Eurorack modules that have an absolute value function. That seems weird―software modulars always have it.

As a substitute, you can square the feedback signal with a four-quadrant multiplier, but this is less effective. The square function further attenuates quiet signals (making the chaotic behaviour harder to initiate), and further amplifies loud ones (which tends to peg the filter to +10 volts) so the range of filter parameters that will produce noise becomes very narrow.


@re-touch, this is interesting. Some brief replies:

-- IRL, a full rectifier should give the absolute value function you're speaking of (unless I'm misunderstanding)

-- do you use VCV rack? (https://vcvrack.com/Host) That would give you another (and deeper than Softube) software platform to play in and test the above hypothesis, though it is still software

-- while I'm still pretty new to hardware modular, I find my "more unique" filters REALLY surprise me if I go beyond "routine" settings. So the response per the settings you mention above MAY vary meaningfully depending on the filter unit.

That said, a more experienced modular user or electrical engineer could maybe give you a definitive answer. Maybe baseline unit noise, with feedback, gained into overdrive (with gain varying the spectrum or "color" or results) would produce similar results across filter types.

Cheers!

NG


Makes sense...and the reason this won't work in analog is because that feedback loop is generating digital fullcode. Basically, all of that gain + feedback loop = massive digital overload, which is what that "noise" is. So it's a valid method, but only in an emulator such as VCV because you literally CANNOT make that happen in analog.