Trends in onset detection

  • Authors:
  • Carlos Rosão;Ricardo Ribeiro

  • Affiliations:
  • L2F -- INESC-ID/ISCTE-IUL, Rua Alves Redol, Lisboa, Portugal;L2F -- INESC-ID/ISCTE-IUL, Rua Alves Redol, Lisboa, Portugal

  • Venue:
  • Proceedings of the 2011 Workshop on Open Source and Design of Communication
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Onset detection is the basis for several high-level musical processing tasks such as Beat Tracking, Chord Estimation or Music Genre Classification. In this paper, the complete process that leads from the "raw signal" to onset detection is described. The most important onset detection methods are reviewed, explained and organized in categories, according to the properties of their reduction functions: time domain reduction functions, spectral domain reduction functions, probabilistic reduction functions, pitch-based onset detection techniques and data-driven reduction functions. We also discuss the applications of onset detection to Human-Computer Interaction, more properly on the change of paradigm provided by the new Music Visualization and Browsing methods.