Learning Models for Query by Vocal Percussion: A Comparative Study

TL;DR

This paper compares traditional and deep learning methods for retrieving drum sounds from vocal percussion, aiming to improve creative workflows for artists through effective, efficient, and stable query systems.

Contribution

It evaluates various machine learning strategies and data augmentation techniques for vocal percussion-based drum sound retrieval, providing insights into their effectiveness and stability.

Findings

Deep learning models outperform traditional algorithms in accuracy.

Data augmentation improves model generalisation.

Trade-offs exist between speed and accuracy in different models.

Abstract

The imitation of percussive sounds via the human voice is a natural and effective tool for communicating rhythmic ideas on the fly. Thus, the automatic retrieval of drum sounds using vocal percussion can help artists prototype drum patterns in a comfortable and quick way, smoothing the creative workflow as a result. Here we explore different strategies to perform this type of query, making use of both traditional machine learning algorithms and recent deep learning techniques. The main hyperparameters from the models involved are carefully selected by feeding performance metrics to a grid search algorithm. We also look into several audio data augmentation techniques, which can potentially regularise deep learning models and improve generalisation. We compare the final performances in terms of effectiveness (classification accuracy), efficiency (computational speed), stability (performance consistency), and interpretability (decision patterns), and discuss the relevance of these results when it comes to the design of successful query-by-vocal-percussion systems.