This paper presents a study of an articulatory-based speech synthesis based on a 2D-Digital Waveguide Mesh (2D-DWM) to model acoustic wave propagation in the oral tract. It is employed to study the effects of changing oral tract area, and in particular, of moving the articulators during the production of diphthongs. The operation of the synthesizer including details of how diphthongs are produced are discussed. The results support earlier findings that the wall reflection coefficient is inversely proportional to the formant bandwidth.

Download