Interspeech'2005 - Eurospeech

Lisbon, Portugal
September 4-8, 2005

An Improved GMM-Based Voice Quality Predictor

Tiago H. Falk (1), Wai-Yip Chan (1), Peter Kabal (2)

(1) Queen's University, Canada; (2) McGill University, Canada

A voice quality prediction method based on Gaussian mixture models (GMMs) is improved by constructing a feature selection algorithm to provide the best GMM-based prediction quality. The proposed sequential selection algorithm performs N-survivor search, allowing for trading between design complexity and performance. Simulation shows that predictors designed using the proposed algorithm outperform two benchmark selection algorithms. Performance improvements over the ITU-T P.862 PESQ standard are also attained.

Full Paper

Bibliographic reference.  Falk, Tiago H. / Chan, Wai-Yip / Kabal, Peter (2005): "An improved GMM-based voice quality predictor", In INTERSPEECH-2005, 2733-2736.