The Secret Pedal Project (Ongoing Research)
I may be giving away the keys to the citadel here, but... Nah, nobody reads this blog. Seriously, NOBODY reads it, right. Besides, once it's done, my plan is to open source this idea under my usual "modified MIT-for-noncommercial-use-commercial-use-requires-negotiation-of-a-paid-licence" licence, so here goes...
My buddy and former guitarist bandmate, Wayne and I have this idea. Recently, I found a reverse engineered circuit diagram of a similar, all analog(ish), defunct pedal from the 80s, so I've reverse engineered that a little here, but in my own way.But wait, there's more, and those bits, I'm not giving away.
In esscence, we're trying to make a special kind of octave box. This is a proof of the octave methods of that concept. There's nothing new here, nothing to see that hasn't been done before using similar tools, nor is it slavish copy of any of that prior art. But it does open up a number of ways to create a bass signal from a guitar signal. The one potential addition, that I particularly like, is being able to quantize pitches from signal period timing and output bass notes E1 to E3 from guitar input of E2 to D#4, then from guitar E4 to A5, also output bass notes E4 to A3 over a MIDI interface, USB or Serial.
This, of course is not shown here, but what is shown here is a "leaky," "drifty" circuit sim of taking an input signal from a guitar and timing, then filtering, an octave lower output signal using an Atmel Mega328p and a low-pass "switched capacitor" filter that effectively tunes the filter for each note. I'm pretty sure the timing drift is a latency artefact in iCircuit and, that when I get real chips on the breadboard and timing to MIDI quantiseation, timing will be so close to lockstep the circuits will run like a bought one.
To make a proper bass guitar simulator, there are a few things that have to be added to this, and that's the secret source. That's been tried and (kind of) works in prototype... with lots of kludges and ugliness. That's the secret bit. The only work to do on that is refine it, make it easier to use and set up, make it (and the above demo) a tad more stable. Also, make it smaller. The prototype is fing huge! This concept demo is a step closer to fitting it all inside a standard, 60x120x50 squashbox housing. Using a digital frequency meter to quantise notes and "MIDIfy" them is a step towards making this "secret" pedal, a launchable repo and maybe even purchasable kits. I'm probably not going to actually manufacture it myself, so you'll need electronics assebly skills or a guitar tech mate who does for that. Or, the reason I do my modified MIT licence.
This is the code for this demo ->
In the electronics, the input stage is a buffer with a guitar signal input impedance, followed by a schmidtt trigger to square the wave up. This clocks the 328p chip (Arduino) on pin 2, where the sum of the 2 pulseIn() commands is used to calculate the input frequency from the time measured. The 328p generates 2 tones from this, on pin 12 (the octave signal for output) and pin 13, the clock signal for the switched capacitor filter. (I plan to use something like an LTC1064 when I breadboard, and code to quantize my notes. (MIDI and analog, but this was one I mixed up in iCircuit to prve the concept and understand SCFs better.)
Anyway, for the first time in, now, coming close to 6 years of working on this project, I feel like I'm getting somewhere.