Influence de la modélisation spectrale sur les performances d'un système de conversion de voix
Résumé
Voice conversion is a technique which aims to modify the speech signal uttered by a so called source speaker so that it is perceived as if it was uttered by the desired speaker. In this paper, we study the influence of the spectral modelling on the quality of timbre conversion. In the framework of GMM-based conversion, we compare discrete cepstrum and LSF parameter modelling. Objective tests show that the use of LSF parameters leads to better conversion results.