Opens in a new tab

AES E-Library

← Back to search

Conference Paper

Drum Pattern Based Genre Classification of Popular Music

Authors: Dittmar, Christian

AES Conference: 25th International Conference: Metadata for Audio · Paper 6-1 · June 2004

Abstract

This publication addresses the identification of drum patterns and the classification of their musical genres. Drum patterns are estimated from audio data automatically. This process involves the transcription of percussive un-pitched instruments with a method based on Independent Subspace Analysis and a robust estimation of tatum grid and musical meter. Subsequently, patterns are identified from histograms, describing the frequency of occurrence of the percussive events on distinct metric positions. Various high-level features are derived from the drum patterns by identifying typical playing styles. The classification procedure evaluates these features in combination with information on musical meter and occurring instruments.

Details

Published in
AES Conference: 25th International Conference: Metadata for Audio
Paper number
6-1
Publication date
June 6, 2004
Session subject
Metadata for Audio
Affiliation
Fraunhofer Institute for Digital Media Technology IDMT, Ilmenau, Germany (See document for exact affiliation information.)
Type
Conference Paper