Modeling Musical Onset Probabilities via Neural Distribution Learning

TL;DR

This paper introduces a neural density prediction model for musical onset detection, estimating time-to-event and time-since-event distributions from mel-spectrograms using CNNs, achieving competitive results on the Bock dataset.

Contribution

It presents a novel sequential density prediction approach for modeling musical onsets with CNNs, advancing the state-of-the-art in onset detection.

Findings

Achieved comparable results to existing deep-learning models.

Successfully modeled TTE and TSE distributions from spectrograms.

Demonstrated effectiveness on the Bock dataset.

Abstract

Musical onset detection can be formulated as a time-to-event (TTE) or time-since-event (TSE) prediction task by defining music as a sequence of onset events. Here we propose a novel method to model the probability of onsets by introducing a sequential density prediction model. The proposed model estimates TTE & TSE distributions from mel-spectrograms using convolutional neural networks (CNNs) as a density predictor. We evaluate our model on the Bock dataset show-ing comparable results to previous deep-learning models.