Electronic dance music can be characterised to a large extent by its rhythmic properties. Besides the tempo, the basic rhythmic patterns play a major role. In this work we present a system that uses these features to classify electronic music tracks into subgenres. From each song, a drum pattern of 4 bars length is extracted incorporating source separation techniques, consisting of bass drum and snare drum events quantized to 16th notes. After determining the downbeat, the measure-aligned pattern serves as a feature in a k-nearest-neighbour classification task. The system is evaluated on a dataset containing excerpts from 400 songs from eight electronic subgenres. As a baseline, the classification using solely the tempo as a feature is performed, achieving a classification accuray of 66%. The additional feature of rhythm pattern increases the performance to 71%.
Authors:
Leimeister, Matthias; Gaertner, Daniel; Dittmar, Christian
Affiliation:
Fraunhofer Institute for Digital Media Technology, Ilmenau, Germany
AES Conference:
53rd International Conference: Semantic Audio (January 2014)
Paper Number:
P1-5
Publication Date:
January 27, 2014
Subject:
Content-Based Audio Retrieval
Click to purchase paper as a non-member or you can login as an AES member to see more options.
No AES members have commented on this paper yet.
To be notified of new comments on this paper you can subscribe to this RSS feed. Forum users should login to see additional options.
If you are not yet an AES member and have something important to say about this paper then we urge you to join the AES today and make your voice heard. You can join online today by clicking here.