Second European Conference on Speech Communication and Technology

Genova, Italy
September 24-26, 1991


Noise Reduction for Speech Enhancement in Cars: Non-Linear Spectral Subtraction / Kalman Filtering

P. Lockwood (1), C. Baillargeat (1), J. M. Gillot (1), J. Boudy (1), G. Faucon (2)

(1) MATRA-Communication, Bois d'Arcy, France
(2) University of Rennes I / IRIS A, Campus de Beaulieu, Rennes, France

Achieving a good quality of speech transmission and reliable performance in speech recognition in the car is an important challenge especially in the context of mobile telephony applications where the user can access the telephone functions by voice. The break through such a technology is appealing, since the driver can concentrate completely and safely on his task while composing and conversing in a full handfree mode. This paper compares approaches enhancing noisy speech with applications both for speech transmission and speech recognition, two approaches have been retained: a temporal approach based on Kalman filtering and a frequential approach based on spectral subtraction. These methods have been extended further yielding new algorithms for speech processing in noise. We report comparative experiments made with various implementations of the speech enhancers. These include a non-linear spectral subtracter [12] and several implementations of a Kalman filter, including new noise whitening schemes.

Full Paper

Bibliographic reference.  Lockwood, P. / Baillargeat, C. / Gillot, J. M. / Boudy, J. / Faucon, G. (1991): "Noise reduction for speech enhancement in cars: non-linear spectral subtraction / kalman filtering", In EUROSPEECH-1991, 83-86.