Dan Ellis :

[music image] Music Content Analysis by Machine Learning

This page provides the materials for a short course on Music Content Analysis. This course was first presented at the Music Technology Group of the Universitat Pompeu Fabra in Barcelona from March 18-21, 2003.

Much of this material is adapted from my course at Columbia on Speech and Audio Processing and Recognition, but with the addition of many music-specific examples.

Slide Packs

The course is presented in three, 3-hour sessions. Here are the slide packs for each session (in PDF format).

A condensed version of this material was presented as a single 75 minute lecture as part of the Johns Hopkins CLSP Workshop 2003. The slides are here: Pattern Recognition Applied to Music Signals. These slides serve as an introduction for the self-paced practical below.

Matlab resources

The course includes illustrations mainly based in Matlab. Below are some Matlab diaries and other resources related to the material:

For the July 2003 session at Johns Hopkins CLSP, I structured this material into a 3-hour self-paced Matlab practical in which you go through building and evaluating singing detection systems using GMMs and Neural Nets:


This material is based in part upon work supported by the National Science Foundation under Grant No. IIS-0238301. Any opinions, findings and conclusions or recomendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation (NSF).

Last updated: $Date: 2003/07/02 11:10:35 $

Dan Ellis <dpwe@ee.columbia.edu>