Interspeech'2005 - Eurospeech

Lisbon, Portugal
September 4-8, 2005

A Spectral Conversion Approach to Feature Denoising and Speech Enhancement

A. Mouchtaris (1), J. Van der Spiegel (2), P. Mueller (3), P. Tsakalides (1)

(1) Foundation for Research and Technology, Greece; (2) University of Pennsylvania, USA; (3) Corticon Inc., USA

In this paper we demonstrate that spectral conversion can be successfully applied to the speech enhancement problem as a feature denoising method. The enhanced spectral features can be used in the context of the Kalman filter for estimating the clean speech signal. In essence, instead of estimating the clean speech features and the clean speech signal using the iterative Kalman filter, we show that is more efficient to initially estimate the clean speech features from the noisy speech features using spectral conversion (using a training speech corpus) and then apply the standard Kalman filter. Our results show an average improvement compared to the iterative Kalman filter that can reach 6 dB in the average segmental output Signal-to-Noise Ratio (SNR), in low input SNR's.

Full Paper

Bibliographic reference.  Mouchtaris, A. / Spiegel, J. Van der / Mueller, P. / Tsakalides, P. (2005): "A spectral conversion approach to feature denoising and speech enhancement", In INTERSPEECH-2005, 2057-2060.