Abstract

For many real speech applications such as speech synthesis, speech conversion it is essential to reproduce the glottal flow as close as possible to the real glottal flow in order to synthetize a natural speech. In the literature, glottal flow model such as Liljencrants-Fant (LF) synthetize the glottal flow using mathematical expressions. Generally, the synthetized glottal flow are nicely smoothed and do not follow the dynamic of the real glottal flow. The common point for all synthesis methods is the non-consideration of the glottal flow phase. In this framework, we propose to investigate the importance of the phase on the glottal flow synthesis. The non-natural behavior of the phase of the synthesized glottal flow is discus. Then, a new method to synthesis the glottal flow is presented introducing the phase component. Experimental results show the validity of the proposed idea.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call