Klaus Frieler and I have been working a bit on singing intonation, and it's super interesting.
Even solo singing is massively expressive, singers can use pitch (detune, slur, vibrato, ...), intensity (tremolo, crescendo/decrescendo, ...), and timbre (husky vs. clear vs. breathy) to make a melody their own.
I'm currently working on a system that lets musicologists annotate these things more easily than has been possible, but I'm using very simple data structures.
Wouldn't it be great to have an ontology to describe that? If you're into ontologies, I'm certain that you'd say yes! So give me a shout if you'd like to collaborate!