Uh... wow. That's far more of a challenge than I thought it would be. My thought was much simpler... compare input to a known frequency, and report on the difference. I didn't consider that harmonics would complicate the whole matter. And the FFTs, well, I've heard *of* them.

Instead of measuring the current tuning of the string, we could use the DSPs tone generators to play a note that we could tune against. This would avoid the hassle of coding FFTs, or dealing with guitar signal levels. In fact, the userland coding would be trivial.

This is more or less what I did -- I just dumped a .wav file on the Empeg.