I know the Access Virus has a wavefolder but frankly it sounds crappy to me when compared with the likes of VCV rack, Reaktor and Eurorack modules, my question is,

are there any keyboard modules with a wavefolder in them that you can pass an external source through, preferably analogue and not eurorack?
- or even Guitar Pedals?

In case it's called something else on whatever model, this is the effect in general I'm talking about:

The Arturia Minibrutes have a folder .. but only for the oscillator triangle wave. They call it the metallizer. Lots of softsythns with wavefolding. Pretty common in eurorack. The new Korg Opsix has an oscillator wavefolder apparently.

An old thread: Why no wavefolder based synths for the masses?

MiniBrute 2 will let you pass anything through its "metalizer"/wavefolder, however it's not a module.

I'd generally look into other West Coast leaning semi-modular synths like the 0-Coast or microvolt 3900. I'm not familiar enough with them to say if they'll let you do what you want or not, but it's likely they can.

Wavefolding sounds good (but limited) for triangle waves. Not as good for other waveforms.

This makes sense because wavefolding is waveshaping with a triangle-wave function. If the triangle-wave function is antisymmetric, you make odd 'harmonics' at critical points (triangle waves at odd multiples of the fundamental). If symmetric, even 'harmonics'.

If the waveshaping is done with a sine function, and you have sine input, this is just FM.

So triangle functions match well with triangles (and mixtures), sine functions match well with sines. And this suggests, there may be something related in the 'sound' of wavefolding and FM.

Wavefolding sounds especially good for mixtures of triangle waves. But the oscillators have to be synchronized, usually by soft-sync, which is rare in analog designs.

Timbral transitions can also be rough, if there's a change in the average amplitude of the input wave, or the average (DC) level. This roughness, which sounds like a 'scrambling' of the timbre, is unfortunately all over youtube examples, so it may put people off.

So in mixing different triangle waves, if you want to vary the mixture and get a smooth result, you need to crossfade them.

So you need a weird architecture to take advantage of wavefolding, which almost has to be digital to be practical, especially in a commercial polysynth context: 4-6 triangle waves, synchronized to harmonic or subharmonic relationships, several wavefolders, and also plenty of crossfaders for timbral control.

But then digital with fixed clock rate introduces aliasing, especially under these kinds of nonlinear transformations. So you either need one of these 10 Mhz-FPGA platforms, or you need to apply some techniques discovered only recently, which may actually work quite well at lower sampling rates: https://research.aalto.fi/en/publica...nonlinearities

Assuming that's solved, the complex architecture is still a lot for the user to manage.

So a commercial design will probably give a choice of fixed 'algorithms' (i.e. routings of the oscillators, wavefolders, and crossfaders), and give the user control only over the frequencies and some assignable envelopes for timbral transition.

Sound familiar?

Buchla Easel
Minibrute 1 and 2
MicroBrute
MatrixBrute
PolyBrute
Minilogue

There are modules like the 0 Coast as well

Having options like that are definitely how you can maximize the usefulness of your wavefolder. But I don't think you absolutely need several oscillators and crossfaders.

The main thing for me is modulation. Rather than taking a complex shape and filtering it like in most subtractive models, you can take a simple sine or triangle and add the harmonics via modulation. Multiple oscillators can make that richer but so will audio rate modulation or a complex EG.

Unfortunately I think a lot of people turn up the Metallizer on their Brute or watch those YT videos and just decide they don't like wavefolders at all without really exploring their capabilities.

That’s correct only if the carrier is anything but 0hz.

1Hz, 2hz, 3, fine, FM. But at 0hz, we specifically call that wavefolding, or alternatively, waveshaping: a “sine shaper”

I have equations if anyone wants to get that nerdy

Yeah, please share, I was always wondering what do they mean by that (0 Hz FM = waveshaping) but wasn't able to google up any reasonable info on the web. Thanks.

I see this, but the point is one can get at least the range of tones from triangle-based wavefolding, than one can get from FM, as long as there's a mix of more than one triangle wave at the input.

The point is they have similar potential as synthesis techniques.

Serge and Buchla examples generally come closer to fulfilling this potential, because there are all these other supporting modules, and odd/even wavefolders are supported.

But to do any kind of modular thing polyphonically, especially with multiple oscillators, mixers, crossfaders etc. per voice, gets expensive and unwieldy very quickly, which is why a better realization of the method probably needs to be non-modular and digital.

The full potential of wavefolding just hasn't arrived yet.

I’m in a rush as it’s lunch time but here’s a paragraph from one of the books I often reference.

It generally does one of two things, either people get a total maths-boner, or it makes their head hurt

Great, thanks. Now it makes sense, I didn't connected it with the fact that the SY line of synths that introduced 0 Hz operators also had a phase setting.

You could even share the name of the book(s)/article(s) you are using when you will have time.

I'll add the Yamaha Montage (since OS 3.0) to the list, with a dedicated wavefolder FX module.

(I assume the whole MODX line probably have it too)

https://yamahasynth.com/learn/montag...older-in-os-v3


I managed to get a pretty close approximation of a very simple Buchla 259e self patch with it (by using a "complex osc" architecture algorythm and linking several parameters to replicate the "timbre" knob.

...before it devolved into a cyber howling banshee's scream when I changed the filter resonance by accident (probably one of my only real "happy" accident with the montage sound design, when I think about it )

Qun synth has a wavefolder.

Seems I still need to clarify the difference between waveSHAPING and waveFOLDING for myself... (Seems wavefolder is a waveshaper where the shaping function just "bounces back" the signal that's over a threshold, whereas waveshaper is the more general case with arbitrary shaping function.)

Some articles:

Wavefolding
Short article about wavefolding at Keith McMillen
Summary of both and more at Perfect Circuit
(has links to some related modules)

I vaguely remember Korg Prophecy and Z1 also had simple waveshaping that wasn't actually capable of wavefolding (the shaping function could only be very simple, like 1st 1/4 of sine function).

It's at the limits of my maths, but even if I can get basic principles and the subtleties pass me by, it's good to try and understand... otherwise it's "magic", and that's leaving the door open to rubbish... so, I need to get some of this.
A reference to the book would be good @ xanderbeanz ... I might even have it already!

They’re the same thing, it just depends on what the shape of the carrier is. If you use a sine or triangle carrier you’re going to get wavefolding patterns as sh*t bounces around predictably, ‘cause maths, if you use a square wave you’re going to get rectification of the signal, and if you use complex waves (like the multitude in the Kronos/01w) then you will get weird combinations of folds and rectification and end up with something they just lumped under the banner of “waveshaping”.

not exactly the same thing, but a Kurzweil K2500R or K2600R have live mode, which uses the sample hardware to get an external source into the sound engine. Then pull up a WRAP block and you get wave wrapping. (you can also use the SHAPER block if you prefer wave shaping.)

There is a sine filter in the Blofeld.
I believe there is a sine(x) filter in the XT as well.

There is, and it’s glorious!

Isn’t it more that wavefolding is one type of waveshaping? The carrier fed into the circuit will determine the result like you say but the result also depends on the circuit, which can fold, wrap, clip, etc.


This has an overview of the types of waveshapers I mean:
https://www.perfectcircuit.com/signa...is-waveshapers

I understand now. Sorry, I was just focussed on digital low-overhead Non-linear processes as that’s what I have the most experience with.

Yes, of course the choice of circuit will have an effect, I’ve even built a couple of different types myself for pedals (haven’t out anything into an enclosure yet, it was more a soldering exercise) but I guess my point is that you can get very similar (if not exact) results and use low CPU overhead by doing it in the way I said, just using different carriers with differing orientations, polarities etc. That’s why I put it all under one waveshaping umbrella.

Majella Implexus has a complex oscillator with a waveshaper and a wavefolder with some other fun stuff. I kind of wish they had included a LPG, but whatever, it still sounds really good. They only just recently met their kickstarter goal, though, so it'll be a while before they're out in the wild.

Got it, makes sense. Waveshaping pedals seem like an untapped area full of potential now that more guitarists are are open to less traditional, out-there effects like bitcrush, glitch, and so on. There are lots of pedals for clipping various harmonics types but few for waveshaping/folding.

Yes, I have a couple of links if you’re handy with an iron, I’m just learning, and got my first burned finger a few months back

One of the interesting things I forgot to mention is the “scale”, what you might do in a single circuit, you might need to setup up with 5 or 6 FM operators to get a similar effect.

The Soundblox multiwave distortion has a couple of modes on it which sound truly like wavefolding, so that’s a good one to play around with if you can find it second hand.

Oh I’m doing good to even swap pickups or a pot on guitar so I’ve not even thought about building pedals. Luckily I’m also lazy so it works out pretty well!

Thanks for the excellent replies,

I checked out the minibrute2, Microvolt 3900 and 0-Coast they all qualify though it's range seems limited on the Minibrute2, the SynthRotek Fold seems to me like the ultimate version of it with all the extra stages of gain/folding it seems to get you to the realm of noise.

The Blofeld just like the 0-Coast doesn't let me run external stuff through it unfortunately, but good to know it has the function.

The 0-Coast is so tempting the sounds are excellent I must admin, but that will get me into the rabbit hole of eurorack which I'm trying to avoid.

The Microvolt 3900 seems lacking like the Microbrute2 in the depth of the effect.


What I would really love is a pedal version of the SynthRotek Fold, maybe buying a kit, Line in, Line Out and PSU modules + a project box big enough would get me there.

I saw my MODX has the wavefolder effect in it, I'll have a play with it and see how it sounds...

If you can find someone selling a Pgh Modular Voltage Lab, it has probably the best wavefolder you can get.

Technically they are the same. A wavefolder is a waveshaper with an infinite (periodic) characteristic, like triangle or sine wave.

The difference is parameter ranges. With a shaper, it will be applied to an audio signal with only limited internal gain ('drive').

A folder usually operates under very large gains that are first applied to the input. These gains activate the 'periodic' region. For instance, put in a triangle wave with an amplitude of N (odd), the wavefolder outputs a triangle wave of amplitude 1 at N times the frequency. So to get 31st 'harmonic' (a perfectly usable result), you apply a gain of 31 internally.

On the other hand, where you'd want to use a moderate gain is when the input's a sample or complex audio signal, not a pure waveform. Here the 'shaper' can be more useful. The 'folder' will simply produce white noise or some other distorted mess on its way to white noise.

So 'folder' vs. shaper distinction refers more to operating range, and preferred choice of input signals, not the underlying mathematics which is actually the same.

Hi,

I just looked into your equation and have no time to finish it, but at first glance I can see it is not frequency modulation, but phase modulation. When doing phase modulation it should look something like this:

A(t) = Wc[2piFc*IWm(2piFmT)*T + phase]

Do yourself simple mind experiment and substitute Wm with slow sine function and you can see, that in FM it will be swinging from higher pitch to lower, while with phase modulation at sin(pi/2) pitch will be at Fc.

Or am I too tired to understand something here? But I was generally good at math and dealt with a lot more complex functions... :/

Another thing is that some people are pasting pictures of some resonant distorted wave saturation, while wavefolding is mirror image of part of wave that would go above threshold folded below that threshold.

Cheers!