Self-Supervised Learning of Context-Aware Pitch Prosody Representations
This work addresses the need for better feature spaces in Music Information Retrieval, though it appears incremental as it builds on existing self-supervised learning methods.
The paper tackled the problem of inferring meaning from short and long contexts in music and speech by learning contextual representations of sung vocal lines from fundamental frequency, showing that these representations can enhance downstream classification tasks by up to 15% compared to traditional features.
In music and speech, meaning is derived at multiple levels of context. Affect, for example, can be inferred both by a short sound token and by sonic patterns over a longer temporal window such as an entire recording. In this letter, we focus on inferring meaning from this dichotomy of contexts. We show how contextual representations of short sung vocal lines can be implicitly learned from fundamental frequency ($F_0$) and thus be used as a meaningful feature space for downstream Music Information Retrieval (MIR) tasks. We propose three self-supervised deep learning paradigms which leverage pseudotask learning of these two levels of context to produce latent representation spaces. We evaluate the usefulness of these representations by embedding unseen pitch contours into each space and conducting downstream classification tasks. Our results show that contextual representation can enhance downstream classification by as much as 15\% as compared to using traditional statistical contour features.