pseudo-random .. tremelo...

[vdm]

hey there fellas
i was checking out RG's work on pseudo-random LFO circuits and the such, and have always wanted a really crazy random effect. what came to mind was a box that is just a simple tremelo circuit with a switch for square or triangle(sine) envelope, but without a rate knob.

i'm a little confused as to how to incorporate RG's work, and was wondering if anyone knew if it was possible to create a random pot, ie. a board with three 'lugs' and certain changable components to give different pot values.

any ideas/help are most welcome and i'd be very interested in figuring this little idea out.

trent

[bdevlin]

Wow, this is an old post but I too would like to have an option of a Tremelo with a random signal level output.  A post from another thread indicated hooking up Schmitt triggers to the depth pot.  In series with one of the lugs??? Parallel??  Has anyone done this?

[R.G.]

Pseudorandom is a delicate thing in terms of human perception.

The Maximal-Length-Sequence (LMS) pseudorandom generators all repeat at some point, the repetition frequency being 2^N-1 clocks where N is the shift length. If that happens to be less than about two seconds, the human ear will pick out that it's repeating, even if you're listening to it as hiss. That's bad if you're wanting to listen to the hiss as hiss.

However, if what you want to do is listen to something semi-repetitive but more random than an LFO, it's great. There seems to be a human liking for a repetitiveness about like the ocean surf - randomness imposed on a repetitive  pattern, sets of events happening  in slightly non repeating ways. We seem to like that. This amount of random can be done with non-locked oscilators.

There are a couple of ways to go about this. One is to just integrate the output of an MLS or run it through a low pass filter and use that instead of an LFO. IF you generate your LFO with the Schmitt trigger gates, you could use an opamp mixer to mix the capacitor voltages, then mildly lowpass and amplify them.

Something like the EA trem could be fixed up like that easily. Just pipe your new random LFO into the place where the LFO used to go.

[Mark Hammer]

Delicate it is.  I suspect that when people talk about "random", what they really mean is aperiodic.  That is, something that does not just go up-down-up-down-up-down in a fully predictable way.

Why is periodic bad or less desirable?  Because it commands our attention too much, I suspect.  The high predictability of the up-down-up-down-up-down leads us to be unable to pay attention to anything else, sometimes.  Does that make "random" a better choice?  Not necessarily, because large divergences in whatever parameter is undergoing "random" shifts can be just as jarring and distracting, and also command too much of our attention.  Imagine a tremolo had sudden bursts where it went from 1hz to 10hz for a moment, then down to 3hz, down to .3hz and up to 12hz.  Just a little distracting.

My sense is that the perfect balance is often something like a pair of unsynchronized (and smoothed) LFOs summed together.  This achieves something that is certainly less periodic than one LFO, but not too wild.  Experimenting with a pair of nonsynchronized choruses or phasers in series, set to slightly different speeds, yields a really pleasing noncyclical quality, without being too outlandish.  You get to appreciate the thickening, without being distracted by the sweep.  The trick is to have a decent combination of speeds, and balance of each source.

I might point out that even when synths use "random" voltage sources, these are often highly smoothed (or at least smoothable) to reduce those jarring contrasts.

[Paul Perry (Frostwave)]

Are people asking for a sine (or square) LFO that has
1. a randomly varying frequency or
2. a randomly varying maximum peak voltage or
3. both?
And do they want to be able to set limits for the min and max period of the LFO,
and for the peak amplitude?
And do they want an uncorrelated random variable, or something with a bit of structure to it?
Or maybe 1/F noise?

[bdevlin]

varying frequency would be fun but I thought that migh be too much for me to tackle. I want a varying output level.  In other words the pulses of the tremelo would be more like peaks, valleys, and everything in between. 

[Paul Perry (Frostwave)]

You mean like a Zvex tremorama but with randomly selecting the steps?
Because it would be pretty easy to do that based on the random step thing at geofex..