Advances in Voice Modeling for Speech Restoration
Loading...
Author
Schleusing, Olaf
Bertschi, Mattia
Renevey, Philippe
DOI
Abstract
The goal of this project is the restoration of pathological speech for patients with severely disordered voice. A novel approach for glottal inverse filtering has been developed to allow the separation of recorded speech into pathological glottal components to be discarded and articulatory formants to be retained for restoration. This is achieved by joint optimization of glottal and articulatory models, whereby the undesired pathological voice characteristics are captured in appropriate voice model parameters and the distortion of the articulation parameters is reduced.
Publication Reference
CSEM Scientific and Technical Report 2011, p. 58
Year
2011