1 paper · 1 filter
Chuyang Chen, Bea Steers, Brian McFee +1
We propose a benchmark for evaluating compositionality in audio representations. Audio compositionality refers to representing sound scenes in terms of constituent sources and attr…