A new methodology for the automatic recognition of musical recordings is presented. A system has been developed that performs recognition among a set of specific musical recordings. The system claims a high rate of success (greater than 86%), even when the unknown compositions have suffered from up to a medium degree of distortion. It comprises a database of musical characteristics that correspond to a set of model musical recordings. These characteristics are derived by applying novel feature extraction algorithms to every model musical recording selected. In order to determine whether an unknown musical recording corresponds to a piece represented in the database, the same feature extraction algorithm is applied to it, and the characteristics thus derived are compared to the database contents by means of a set of criteria. The system can operate in parallel, essentially in real time, even for a considerable number of model musical recordings, as long as the hardware necessary is available.
Authors:
Papaodysseus, Constantin; Roussopoulos, George; Fragoulis, Dimitrios; Panagopoulos, Athanasios; Alexiou, Constantin
Affiliation:
National Technical University of Athens, Department of Electrical and Computer Engineering, Division of Computer Science, Athens, Greece
JAES Volume 49 Issue 1/2 pp. 23-35; February 2001
Publication Date:
February 1, 2001
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.