Comparing onset detection methods based on spectral features

  • Authors:
  • Carlos Rosão;Ricardo Ribeiro;David Martins de Matos

  • Affiliations:
  • ISCTE-IUL L2F/INESC-ID, Lisboa;ISCTE-IUL L2F/INESC-ID, Lisboa;IST/UTL L2F/INESC-ID, Lisboa

  • Venue:
  • Proceedings of the Workshop on Open Source and Design of Communication
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

Onset Detection, the quest for finding transient regions in the audio signal, has been an active research subject as note onset detection is commonly used as a first step in high-level music processing techniques. In this paper, we review several Onset Detection methods using spectral features such as magnitude, phase and complex domain representation, and compare their performances when considering several types of musical signals belonging to a publicly available dataset. Our results show that the accuracy of onset detection varies clearly between onset types and between detection functions used, as well as between performance techniques.