Extracting signals of interest from a set of mixed signals is called signal separation. For instance, given a recording of music corrupted with background noise, we might wish to recover the underlying melody. Signal separation is generally an underdetermined problem, with more unknowns than constraints. However, conventional Western music is often highly structured. The central thesis of this work is that exploiting the underlying structure characteristic of music is key for advancing signal separation methods. Not only do we hope that these techniques will enable us to remove ambient noise that lacks musical structure, but we also hope to extract signals that possess distinct musical structure.

I’m looking forward to tackling an open-ended signal processing problem that I find fascinating. I hope to get more experience not only with conducting independent research, but also with communicating technical concepts clearly and concisely.