Found this article on Fast Company and thought it was interesting:
Zenph sound has been working on something that may either offend or amaze musical purists. They’re using artificial intelligence to analyze old recordings that may not be of the best quality, and then build up a model of the exact nuances of the musician’s performance. The model then allows the company to actually recreate the performances as if they were played today, and recorded using today’s high-definition technology.
Essentially, the algorithm captures the individuality in the touch, tempo, and emphasis of the performer, and it’s then delivered to a specially designed robot piano as a high-definition MIDI file. The piano then physically drives the keys in accordance to the MIDI file, creating music almost as if the original artist was at the keyboard. Zenph will be taking the robot pianos on tour which, slightly creepily, will allow audiences to listen to live performances of long-dead performers–Rachmaninov, say, or Thelonious Monk.
Okay. We can easily apply this to composers and musicians (often the same person) and spin off into some hotly-debated concept of, say, creating new works by Bartok. But what this brings to mind for me is the difference between classical music and other genres. To cover my bases, I will say this is not always true, but I think you can leave your pitchforks in the barn if I declare that for the most part classical music is not tied to one performer. It may be commissioned by, premiered by, or even the signature piece of an individual musician or ensemble, but if you want to play some fiendishly hard piano piece by Liszt that he wrote so that he himself could whip it out it to show off, you can do that — and no one’s going to call it a “cover.”
Therefore — if we recreate Candide exactly how Bernstein would have played it himself, is it, in fact, more authentic than if the BSO goes and plays it tomorrow the way Marin Alsop thinks it ought to be interpreted? Why or why not? Don’t forget to show your work!