Research

Paper published in ISMIR 2021

We tested whether Americans, Brazilians and South Koreans perceive the same or different moods in the music that is local or foreign to them. We then compared these ratings with current state-of-the-art machine estimates to examine whether the algorithms are systematically biased to the US (given most training data are English songs).

*nominated for best student paper

video coming soon…

Talk at ICMPC-ESCOM 2021

I presented about the formation of a new database called globalPOP, which I extensively gather representative pop music samples around the world along with various metadata, machine retrieved features, and lyrics. Some preliminary results are also discussed.

Presentation recording

*nominated for best student paper

Talk at CIM19

I introduced the Timbre Perception Test (TPT) a new psychoacoustic test for measuring individual differences in the ability to hear fine-grained variations along three dimensions of timbre space. For more details watch the recorded talk, read the paper, or try out the test yourself.

update: there is now an R package, tptT which is built using the psychTestR and runs locally using R.