Aleksandr Diment


Audio Research Group
Department of Signal Processing
Tampere University of Technology

Google Scholar profile

Archetypal analysis for audio dictionary learning

The following audio examples demonstrate the signals produced by the algorithm proposed in the following paper:

The original signals are from the GRID corpus. The dictionary sizes are 1000 in all the cases, except for NMF (dictionary size 100). The sparsity parameter λ was set to 1, producing better results with the presented test cases. Here, the SDR values are computed across the three three demo mixtures. See the paper for the full evaluation results.


Separated SDR, db
AA, KL 9.54
AA, SPAMS 8.28
VQ 8.14
Exempl 8.33
NMF 3.96

IEEE-Copyrighted Material: Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works, must be obtained from the IEEE. Contact: Manager, Copyrights and Permissions / IEEE Service Center / 445 Hoes Lane / P.O. Box 1331 / Piscataway, NJ 08855-1331, USA. Telephone: +Intl. 908-562-3966.