Sanderson, Conrad2006-03-102006-03-102006-03-102002https://infoscience.epfl.ch/handle/20.500.14299/228183In this communication we first review the human speech production process and feature extraction approaches commonly used in a speaker verification system. Experiments on the telephone speech {NTIMIT} database suggest that the performance degradation of a speaker verification system used in noisy conditions can be reduced by extending traditional feature vectors with recently proposed {MACV} features.learningtext::report