There generally doesn't seem to be much interest in building guitar synths - not just signal processors, but true synths with one or more VCOs, envelope shapers etc., that just accept a regular guitar signal. I suppose an earlier reasonably successful DIY design (if modified slightly) is the Penfold guitar tracker, though this requires a separate mono 1v/oct synth to generate the audio. I haven't seen much else around - even on the synth sites there's almost a complete absence of DIY builds.

The idea of using the guitar as a controller has always appealed to me. I bought a Korg X-911 when they came out, but it was largely disappointing even for the time. In more recent years I've experimented with Roland stuff but this has limited appeal in that I'm never really happy playing through stuff I haven't built myself.

Is the lack of interest down to a gulf between keyboard players and guitarists? There's an assertion in some quarters that if you want synth sounds, play a synth. I see the world of synthesis opening up the range of sounds achievable with the guitar - in particular the ability to extend note duration way beyond the natural sustain of the string and using different waveforms and processing.



 

Isn't the lack of such things partly down to the technical problems? An actual guitar doesn't make a great controller because the individual string signals are all mixed together. Hex pickups help a lot, but hardly anyone makes them. And even if you get the string tracking sorted out, you've then got to build six voices to go with it! That's quite a lot of synth!

In a similar way, you don't see many DIY harmonisers or vocoders. They're difficult and/or need a large amount of circuitry.

I suspect interest in guitar synths was somewhat greater when we were back in the era of monophonic analog keyboard synths, and the goal of adapters for tracking guitar were focused on converting one note from pitch to CV/gate.  Obviously there were polyphonic devices from Roland and others, but once we came to expect two-handed chords and such from keyboards, the hurdles imposed by tracking multiple strings likely became too much of a disincentive for DIY.

I've mentioned before this oddball polyphonic octave divider I have, produced by Guild in the very early '70s.  The proprietary hex pickup it came with was about the same surface area as a P90, meaning that it could not be positioned against the bridge, unless the guitar had no bridge pickup.  That really buggered up the tracking, by inviting bleedthrough from adjacent strings.

The moral of that story is that guitar synths require a very slender divided pickup that CAN be butted up against the bridge, where string-wiggle is minimal, and bleedthrough easily defeated.

Some 22 years ago, a bunch of us had a "synth summit" in my basement, and Harry Bissell came up from the Detroit area, with his DIY guitar synth/processor.  He had an Ibanez solid-body with a hex pickup mounted on it that was (I think) used for the G Vox system.  Each string was fuzzed to provide plenty of harmonic content to carve away with filters, and six independent EG/VCA pairs.  It was a phenomenally responsive system.

My own particular interest is in some of the mono sounds from the 70s - thinking of the opening hook in Sweet's 'Fox on the run' and such like. I'm just finishing off my latest synth build that fits in a 1590B enclosure that gets this type of sound. Digital, but capable of adding further analogue processing. I've been working on note tracking for a couple of years now and at least for mono playing it tracks really well. Just a handful of parts and it incorporates intelligent harmony and arpeggio with a minimum of controls. Just two minor bugs that should be an easy fix, but I'm not seeing the  source of the problem right now.

I'd be interest to see any builds that people have put together. I'd been in touch with a guy in Germany who'd developed the Penfold tracker into a complete working synth, but it sure was complicated.

Some years ago, I put together a switching guitar to iPhone box then used MidiGuitar2 on the iPhone with either internal sounds or those from Soundtank or other software synth.

That gave me a polyphonic system, albeit with some latency, with which to gauge whether I really wanted to go down this route. Effectively used as a stompbox in my pedal chain, it sounded really good through my guitar amps on most setting though I really struggled with note tracking on any kind of piano. It was useful for laying down simple bass or synthy swells for backing, the sax solo on Baker Street and annoying everyone with the bagpipes but ultimately, my inability to play anything other than  guitar parts was such a limiting factor that the setup was consigned to the bottom drawer.

I've seen demos of the same setup where the guitarist really makes every instrument come alive but I think its to do with their ability to play the correct parts for the particular instrument that's been chosen. Me, I'm still too busy struggling with actual guitar parts to wander off down yet another rabbit hole

I found the same thing with the Roland setups I've experimented with - playing the relevant instrument-specific parts is the only way to make the patch sound convincing. A fast minor pentatonic run with bends and hammer-ons isn't something a cellist would do, but when I learnt some actual cello pieces it sounded pretty authentic.

The same even with mono synth parts - they sound best when played in a keyboard style.

There's that parasitstudio guy who makes synthy type of guitar pedals with cmos chips.

I messed around with a cd4046(?). Probably used the circuit in the datasheet because it was barebones. I just wanted to see if I could get sound out of it before investing much time to it.

It tracked surprisingly well with worn out strings. Pitch tracking issues mostly occurred when picking a string so an ebow might be handy. I didn't bother continuing because I don't have the know-how to fix the shortcomings and the sound wasn't all that pleasant anyway.

The answer might be additional electronics along with the hex pickup.  Suppose each fret was divided into six isolated electrical sections and you grounded the string so that as you pressed the string, you would get a digital low signal at the fret you selected on each string.  You would also take the parallel outputs of each pickup section of the hex pickup.  You would then get a signal of the type:

G string 7th fret down, amplitude 75 mV.

Let's say you have 22 frets and 6 strings for a total of 132 possible fret outputs.  You multiplex that and take the 6 hex amplitudes out to a controller that can identify what fret has been used, what string has been picked and what output amplitude you have on that string.

You might be able to use synthesizer electronics using this as a gesture capture system by converting to a MIDI input.

In this day and age it makes more sense to do it in DSP.

If you want a specific synth-ish sound then a simplified pedal would make sense to use analog but once you start needing "all the functional blocks" of a synth an analog circuit grows out of hand.  It starts to look like a 70's synth circuit, which goes back to the case for DSP.

Take a very simple and specific case of the Boss OC-2 octaver.  They didn't bother making a better analog wheel because you can do a better wheel in DSP with a much simpler circuit, and far less development effort, hence the OC-3.  Sure some of the spin-offs give you more synth sounding options but at the end of the day they are small differences circuit wise - and no significant improvements in tracking.

I moved away from analogue setups because of the complexity in packing the electronics into a foot pedal enclosure. I did develop the OC-2 circuit to generate a  digital signal so that this could be further processed using an Arduino Nano, but it was still tricky to get the tracking accurate - the minor errors that you don't hear with the OC-2 are significant when processed. I coded all of this up to give better control under firmware and at the same time eliminate most of the components needed.

The component count is now down to a Seeed Studio RP2040, a dual opamp, five pots and three pushbuttons plus a small number of discreet components. This gives 5 oscillators with 7 waveform types and an envelope shaper and VCA with 4 modes. Plus transpose, arpeggiator (three modes each with selectable 3 to 8 steps), harmonizer, glide, sample and hold and clipping indicator, parallel FX loop, variable input gain and output mixer. No filter due to ADC restrictions, but could be adapted to use a rotary encoder and OLED screen.

To do this an other way rather than digital would require a lot of parts and a big enclosure, plus the cost would be significant

Which brings me to the EH guitar synth - the Unrelated Activities youtube channel has some interesting stuff on this https://www.youtube.com/results?search_query=EH8000

Some of the circuitry is redrawn very clearly indeed.

As some will know, back in the 1980s I built much of this synth on veroboard, as Mark said "when we were back in the era of monophonic analog keyboard synths". The tracking was very good indeed, and I would recommend this as a build as a next step up from the ADA Flanger (!!!!!!!!!).

Quote from: Mick Bailey on May 10, 2025, 09:06:04 AMI'd be interest to see any builds that people have put together. I'd been in touch with a guy in Germany who'd developed the Penfold tracker into a complete working synth, but it sure was complicated.

Watching some of the videos from the recent Superbooth 25 in Berlin, "complicated" seems to be all the rage.

I see what you mean - lots of equipment with even more knobs than before.....

I have some good news and some bad news.  The good news is that we've provided many new ways for the user to shape and direct the sound.  The bad news is that we've provided many new ways for the user to shape and direct the sound.      

I like real synths!! just got a behringer ub-xa and love it .. the real ones are around 5k i think ..and I just got some korg volcas .. i love these little machines .. one is a full analog synth and the other is a yamaha d7 is basically a guitar pedal size package ..

To be honest Mick, I'm not that interested in guitar synths because I'm not really a guitar player at all (although I did have a bit of a try during covid) but I *am* a programmer, so I'd be interested to hear what you learned about frequency tracking and how you did that digitally.

I know several techniques in theory, but I never tried any practically and my experience of DSP is that you can usually get away with a hell of a lot more than the theory says you should be able to: E.g. They say "don't do this unless X" and you do it anyway and ignore X and actually, mostly it's fine, or fine within certain limitations which often you don't care about. So I'd be interested to hear you experience of implementing an algorithm and how you tweaked it.

Quote from: Mark Hammer on May 12, 2025, 08:24:45 AMI have some good news and some bad news.  The good news is that we've provided many new ways for the user to shape and direct the sound.  The bad news is that we've provided many new ways for the user to shape and direct the sound.      

Yep. Or as we used to put it back at work - the bad news is that the good news IS the bad news.
 

Quote from: StephenGiles on May 11, 2025, 04:51:08 AMWhich brings me to the EH guitar synth - the Unrelated Activities youtube channel has some interesting stuff on this https://www.youtube.com/results?search_query=EH8000

Some of the circuitry is redrawn very clearly indeed.

As some will know, back in the 1980s I built much of this synth on veroboard, as Mark said "when we were back in the era of monophonic analog keyboard synths". The tracking was very good indeed, and I would recommend this as a build as a next step up from the ADA Flanger (!!!!!!!!!).

The video on the adaptive filter is very good.  That guy knows the real problems.  I can see why it tracks well. The detector used on Boss OC-2 only partially addresses the issue and that's why it bobbles.   The stabilizer switch on the Mutron Octave Divider doesn't help those problems much either.   I think the Roland unit had some sort of adaptive filter, even though it uses separate coils (IIRC).

Quote from: ElectricDruid on May 12, 2025, 08:21:38 PMI'd be interested to hear what you learned about frequency tracking and how you did that digitally.

I should have the schematic and final code tweaks finished today and I'll post this as a project along with some video. I'm a guitarist but not a programmer and went down the digital route more out of necessity, initially combining digital and analogue but more recently trying to do everything through code. Rather than use my synths exclusively, I incorporate an FX loop to mix in the analogue guitar signal + distortion etc., and play interactively.

Project file posted under DSP section