Modular synth conversion to guitar use

[Mick Bailey]

I've built quite a few synths over the years but I'm now thinking I'd get more use out of my most recent build by replacing the keyboard controller with a guitar interface. I've been looking at the Electrax Synthax-100 fundamental extractor and how it drives a PLL to give a sustained square wave output, along with a trigger signal. My thought is to use these outputs with my existing EG/LFO/VCF/VCA/Sub-Octave, etc. to generate a held pedal note (rather like an EH 'Freeze', though mono) and be able to play over this. Bit like the opening to 'Shine On You Crazy Diamond'.

Could need a bit of creativity to apply PWM to the PLL output, and it would be nice to track the signal with another VCO, but the basic idea would probably work OK.

So the question is; anyone tried this with the Synthax circuit (or anything else)?

PDF here;
http://www.google.co.uk/url?sa=t&rct...82001339,d.ZWU

[Mark Hammer]

I've slowly (VERY slowly!) been assembling a modular setup for the same purpose, but I'm not there yet. I gather you're aiming for monophonic?

[Mick Bailey]

Yes - mono is fine for what I want to get out of the setup. I thought if I can get a trigger (easy) and a held, glitch free square wave (more difficult) I'd be most of the way there. I prototyped part of the E&MM harmony Generator and it was all over the place, but did hold a note. Usually unrelated to the one being played! The Synthax looks a much better circuit in that it operates more like a traditional synth sample and hold setup. It seems like the only limitation in note duration would be the usual S&H droop caused by cap leakage.

I had a Roland GR-33 guitar synth for repair and whilst it's impressive it doesn't sound anywhere near as good as a mono synth - the rawness isn't there, nor the immediacy of knob twiddling.

Currently my system is;

Ray Wilson Mk2 MFOS single-contact keyboard controller
Thomas Henry VCO
Thomas Henry XR-VCO
2x MFOS LFO
MFOS ring mod
MFOS noise generator
6-octave range parallel sub octave generator
Sample & Hold
2x MFOS Envelope Generator, one with variable delay
Mini Moog VCF
Steiner-Parker VCF
2x Thomas Henry VCA.

Even with a 'straight' square wave from a single oscillator this sounds impressive, especially with the sub-oct to fatten the sound. Being able to control this with a guitar even to just get a pedal or drone note would be really nice.

[Mark Hammer]

That's a nice range of modules you have there, Mick. A tip of the hat from me.

The difficulty with guitars as input source is that the signal is highly unstable. Even at its best, it starts out with more harmonic content, then quickly loses both the harmonic content as well as amplitude, making it necessary to impose a note-on/note-off based on setting thresholds for triggering. But it's even worse, since there are often infrasonic variations in amplitude making it even harder to use those arbitrary thresholds with any reliability.

If you look at Craig Anderton's 35 year-old solution to that challenge, here: http://hammer.ampage.org/files/Device1-1.PDF you'll see that he employs the compressor half of a compander chip to smooth out the signal. If one is simply going to process a guitar signal with filters, and modulation of various forms, rather than extracting a pitch and driving VCOs, Craig's solution is excellent. But if one wants to drive oscillators, then I would think the amplitude and harmonic content needs to be kept as simple and invariant as possible.

I wonder if some sort of gated multi-band overdrive would fit the bill to make a P-to-V converter (like this one: http://hammer.ampage.org/files/GuitarP2V.PDF ) behave better. So, since you can't have an adaptive filter, at least split the input signal into a few different frequency bands, clip the bejeezus out of it, and filter it sharply, such that each band has roughly the same relative amount/absence of harmonic content, and the pooled output is gated, prior to feeding it to the P2V module. The gate would be driven by an envelope follower detecting the raw input.

[g1]

Did I hear someone say "traditional"?

[Mark Hammer]

What in tarnation IS that?

[g1]

It's part of the tone generator for a Cordovox electric accordian setup. As you probably know, unlike vacuum tube "delay pedals", tube synths are viable (ok, maybe not so practical ).

[Mick Bailey]

My experiments with P2V haven't produced very good results in the past so I've shifted thoughts towards getting a lock using a PLL and generating a resultant frequency. I've got really good results in fundamental extraction to drive an octave divider with fast, accurate glitch-free tracking and maybe the square-wave output of the divider IC could be used to drive the PLL. If the trigger signal occurred at the correct time during the guitar envelope (before the fundamental decayed and the second harmonic became a problem) the PLL could sample, lock and hold the PLL internal VCO constant. The E&MM harmony generator doesn't do a good job of extracting the fundamental and it seems critical to have a stable frequency for the PLL to operate. Even the slightest variation in input frequency causes problems, but I'm pretty sure this can be resolved.

Looking at the schematic suggest that the Synthax could work OK. Maybe I need to build the input processor and harmony generator to see how well it performs. It does seem to have the basis for a practical, working device to generate an output signal. I think it was a late 70s design so I have to wonder why it subsequently never got anywhere.

[Mark Hammer]

Geez, are you sure that doesn't double as a tanning bed, or at least food warmer?

[Mark Hammer]

If you have the chance, just for the hell of it, try a set of medium-to-heavy gauge flatwounds. I think part of the difficulty we tend to have with devices intended to track the pitch of picked/strummed strings is that: a) most of us tend to like looser strings for the easy bendability they have, and b) we like our strings bright. The unstable pitch of strings that yield to subtle finger vibrato, and the additional harmonic content, confuses pitch-tracking circuits.

Seems to me that a duller-sounding, stiffer, string might provide a more suitable input to something that hopes to translate what the pickups sense into something pitch-related. I won't guarantee that the "picking/fretting experience" will be pleasant, but I suspect the octave-division and pitch-tracking may be more reliable.

[robthequiet]

Not sure of anyone would still be following this, but recently I've been eyeing my Roland GR-300 through several layers of dust and wondering if maybe I should open it up to clean out the spiders and see if it still works. The sucker was very impressive at the time, with Pat Metheny and Robert Fripp actually using it onstage, but to me it was more effect than instrument and it was a bitch to tune, so it never went past a few sessions before just ending up on the shelf. Nevertheless, what it did do fairly well was track pitch, as long as you played by _it's_ rules. So if you were willing to settle for one waveform and the headache of trying to set pitch shift intervals by hand only to have them slip out when a cloud passed over, it did the thing. No MIDI, of course, but poly fuzz worked OK so now i'm wondering if it's worth the effort to a) try to engineer some kind of pitch to MIDI by tapping into the circuitry, or b) try to expand it's capabilities as an analog signal generator, which is why this thread caught my eye. The angle I've bee contemplating is not only the pitch capture from hex pickup but how to possibly implant sensors in the fretboard to get a pitch "map" and trigger the VCOs from a neck "keyboard". I know it's been done somewhere but I'm not sure of the details. So you would have the built in tone generators along with a VCF/envelope follower, chorus, pitch shift and hex fuzz, but also be able to send envelope signals with the fretboard pitch to a modular system. Could be a monster, but the trick to me is where to tap into the existing circuits to get useful signal info. Or build something completely different. So if anyone think's that there might be some potential here to develop the art a bit further, or point me in the direction of advancing the concept, please let me know. Mick, curious to know where your system stands at this point, if you'd like to update. Thanks for the inspiration.

[Mark Hammer]

These days, I don't know that what you have described doing is worth the effort. There are any number of units out there that will permit plugging in a mono guitar signal and turning that into MIDI output. Whether it outputs polyphonic MIDI signals, or simply mono, sort of like the old IVL Pitchrider, is something I am underinformed about. But certainly mono units can be had for under $100. https://www.sweetwater.com/store/detail/G2MV3

[Mick Bailey]

Heck, this seems a good while ago and I missed robthequiet's post altogether. I really wanted there to be a straightforward way to get my synth all working with the guitar, but in the end everything I tried was way to glitchy, though at times the sounds I got were breathtaking. The PLL approach was rock solid when it got a fix on the note, but I couldn't predict whether any particular note would 'lock' when played and I was getting too many false frequencies on anything other than very deliberate, clean picking and slow changes.

A few Roland/Godin setups have come my way in the meantime and they can be glitchy as well - not as good as I was expecting. Fine for sax/piano/cello sounds etc, but the synth sounds aren't anywhere even close to my modular. I did look at a Roland guitar-midi box that uses the GK pickup and feeding this into a midi-CV unit. This would have worked better than how my own experiments turned out but I ended up getting distracted on other FX projects so this one is on the back-burner for now.