School of Physics, Engineering and Technology
- School of Physics, Engineering and Technology
- Research and innovation
- Communication technologies
- Projects
- Dynamic 3D digital waveguide mesh voice synthesis
Dynamic 3D digital waveguide mesh voice synthesis
This project uses MRI data to inform a 3D digital waveguide mesh (DWM) model of the vocal tract. A novel impedance-mapping technique is used which allows the effective shape of the mesh to change during synthesis, resulting in highly natural sounding dynamic vowel sounds and eventually the full range of speech sounds. This method has the potential to significantly increase the naturalness of synthetic speech, enhance our understanding of the human vocal system and improve assistive technologies.
Members
- Amelia Gully
- Damian Murphy
- Helena Daffern
Funding
- EPSRC
Research