I breadboarded it tonight, and it works fine, at least it works as I expected it to work after reading the code.
-Tracking is not perfect (having the guitar in tune helps, though
)
-You need to pick very cleanly.
-The unit is only tracking pitch at the attack of the note, so if after picking you modify the pitch (bending, slide, pull up...), it is not tracked.
-Volume of the note attack is translated into midi note velocity. But velocity is constant until note fades.
Not bad, considering that I skipped the part of the circuit that references the ADC max and min, so probably accuracy can be improved in the final build.
I will investigate if adding some filtering by hardware, the tracking can be improved.
Modify it to have bend tracking will be tricky... because of the way the algorithm works. It tracks pitch at the note attack on purpose to avoid harmonics... so if we continuously track pitch, probably will have wrong pitch detected at the note decay. And, if we add a schmidt trigger type of filter to focus on the main harmonic, we will lose the dynamics.
I cannot compare with the Sonnus G2M, because I have never tried it.
Going back to the hex pickup topic, IMHO if you use a PIC with enough ADC inputs (6), it should not be an issue to modify the code to "mix" the midi signals of each string. The timing could be challenging, though (you have to be tracking pitch in all strings at the same time, or "multiplexed"). But taking into account that in this project the PIC works at 4Mhz but it could be configured up to 20Mhz, the PIC should be able to handle the workload (although the ADC could be the bottleneck if it is shared by the inputs).
Potul
