The prerequisite for the robot being able to learn how to produce timbres is that it has to have a striking mechanism that actually can produce a suitable variety of timbers. So you know you will go down some unexpected paths in your PhD, and for me it was the manufacture of synthetic hands, which turns out to be more difficult than you might think to get one that sounds good. This one is actually a fully lifelike reconstruction of my own hand, made by alginate casting silicone. It sounded okay, but this one is quite a bit too heavy to actuate by practical means. So I have had a bunch of different hands on there over time.