I think, the problem is that whenever a state-of-the-art synthesiser comes close to an expressive acoustic instrument it lives in the uncanny valley more often than not (see some YouTube videos of the Eigenharp for example). It will need a lot more sensors or better software to overcome that.