I am working on a game in which I want to use my cello as an input device. I play cello to generate and control the game world, and the audience uses Xbox controllers to play with/against me.

Now it is easy enough to read microphone input from my cello to my game, but what I really want to know is which notes I play and when notes start and end.

So here's the question: are there any libraries that can do real-time waveform analysis and tell me what notes are being played? Either in the form of a VST (FMOD can use VSTs, as far as I know), or as a separate library that I can call from C++. So far I have not found something suitable, while I know this should technically be possible: Rocksmith and guitar synths must use the same kind of technology.

Bonus features (not necessary, but would be nice to have):
-volume envelope within the note
-how much vibrato there is
-able to recognise very fast notes
-can recognise two-note chords (cello cannot do more than two notes at a time)

So, does anyone have any suggestions for libraries that can do this, or an approach for this?

Thanks in advance!