Like Auret, I immediately thought to add offsets to the V/Oct signal.
In Purf’s original post he stated he had an arpeggio, and I assumed there is a polyphonic chord that is arpeggiated. My understanding was that first 5 semitones are to be added to each step at random times, followed by either 7 additional semitones to make an octave, or else an additional octave. I chose to add an additional 7.
I used a ramp down to trigger a dual sample and hold that generates two sets of thresholds for the 4 note arpeggio. The first ranges between 0 and 5 for the 5 semitones. The second is offset to range from 5 to 10 for the 7 semitones. A ramp up is compared against each of the thresholds, and the A>B 10V gate is scaled to add the appropriate number of semitones. Unlike the bernoulli gate solution, once an offset is added to a step, it remains until the ramp finishes its cycle. New thresholds are generated for each ramp cycle.
The ramp frequency is set to 1/64 the master arpeggio clock, so the offsets are added within 16 arpeggio cycles, 5 semitones during the first 8 cycles, and 7 during the second 8 cycles.
Arp Random Offset Injection.vcv (4.1 KB)
Update
I find I get a bit more satisfying results if I scale the first threshold set from 0.75 - 4.75, and the second set from 5.25 - 9.25. That way the beginning and end of each ramp cycle are more consistent. But I won’t bother posting another video.