iternabe Would it be more accurate to say two sensors are needed: hammer sensor and damper sensor, because both controls how the string behaves?
Correct, there's no other way around it.
iternabe BTW how does NV10S detect note off timing (not velocity) with only the hammer sensor?
I guess it's pretty similar to how I implemented Cybrid: if the hammer is below a certain threshold point (exactly midway), you assume the damper is on the string, hence you can produce a note-off event (provided there was an active note-on event prior to that).
That's a rather silly logic because the hammer is not hard-linked to the keyboard and there are situations in which the hammer would still be near the string (due to e.g. a rebound) while the key has been fully released, hence damper is on the string and a note-off should be produced.
Yeah, it's a limitation with Cybrid too, but then Cybrid was a proof-of-concept, not an expensive commercial product that is marketed as finalized and top-notch 🙂
iternabe Also, after the hammer escapes, strikes, then fall and rest, while the key is held down, will the hammer still move (in small amounts) corresponding to the key stick’s release movement? If so, would a high precision hammer sensor that can sense the full range of hammer position, using a sophisticated algorithm, detect key stick release movement and extrapolate damper position?
Theoretically yes, however if I remember correctly, the NV10 uses some metal notch(es) that will interrupt optical signal and that means it's not a continuous sensor that knows the exact hammer position. It just knows when the hammer has passed 2 or 3 predefined points. That's very limited and can't be used for advanced behavior and prediction, which is why there's no release velocity supported, only a fixed note-off event. I'm not sure if they changed it in the NV12 though, maybe they introduced some continuous sensors?
