Okay. So I went ahead and built this small setup.
The audio from Axoloti comes in with the orange cable on the left. My soundcard/headphones are connected to the black cable on the right. Power comes from a eurorack case via the ribbon cable. There are some ground hooks for probes and two 3.5mm jacks for connecting a scope (I had the 3.5mm=>BNC cables already and found this less fiddly than standard oscilloscope probes. For audio signal it doesnt make much of a difference anyway)
It is the circuit posted above. Only difference: There is no high pass filter cap and the diode is pointing from ground upwards to the opamp output (due to the inverted signal, this gives me clipping on the positive side of the waveform).
I used this Axoloti patch to send various waveforms to the circuit:
testsignals.axp (28.7 KB)
You can select between Sin/Tri/Saw/Sqr oscillators (both antialiased and not), repeating clicks/pulses, a drum pattern, a synth line and white noise. I'm not sure yet which of these will prove helpful, but it's a start.
I then observed the waveforms on the scope. Here is a zip file with all my measurements: Measurements.zip.axp (1.5 MB - remove the *.axp file extension. I had to rename it to upload)
The measurements I did (+ a few example pics, the rest is in the zip file):
Spectrum of the sine wave I feed into the circuit as well as the spectrum of the distorted waveform with the master volume int he patch set to 32, 45, 55. Here's the distorted version with the level set to 55:
Sin and Saw waveforms with the pitch control set to -38 and +16. Recorded at various levels of dirstorion (32, 40, 45, 50, 55, 64). Here's a saw wave, pitch=-38, level=50:
And the same waveform recorded with the pitch set to +16:
Also I recorded some shots of the Drumbeat. Here's a closeup of the noise portion of a snare hit:
Seems to me like this could be nicely recreated with just waveshaping. I was expecting to see some diode recovery effects on higher frequency signals, but apparently we're not yet high enough to see anything.
Here's a recording of the drumloop. First the source signal, then the distorted signal at levels of 32, 45 and 55 (all normalized so we can hear the effect of the distortion, not just the difference in volume):Drums_src_32_45_55.flac.axp (1.7 MB - again, remove the *.axp file extension)
So, that's all I can do for today, but I'll start creating a simple model later.