Toward Fully Self-Supervised Multi-Pitch Estimation

arXiv:2402.15569v1 Announce Type: cross
Abstract: Multi-pitch estimation is a decades-long research problem involving the detection of pitch activity associated with concurrent musical events within multi-instrument mixtures. Supervised learning techniques have demonstrated solid performance on more narrow characterizations of the task, but suffer from limitations concerning the shortage of large-scale and diverse polyphonic music datasets with multi-pitch annotations. We present a suite of self-supervised learning objectives for multi-pitch estimation, which encourage the concentration of support around harmonics, invariance to timbral transformations, …

abstract arxiv cs.lg cs.sd datasets detection diverse eess.as events limitations music narrow performance pitch research scale shortage solid supervised learning type