OK, out of time. Good thing.
Currently my experiments are yielding multiple spurious note on commands with each button press. That didn't happen before, or Iwas blind to that flaw (wasn't). So I officially give up.
In at least one code version, however, the test for "button pressed" was incorrect. This resulted in every button looking pressed at start up, and then forever upside down during execution.
This I believe accounts for the odd dance you had to do to come all the way 'round to a functioning program.
a7