The SPRECHart (SPEECHart) - project aims at a deeper understanding of the often reported speaker-specific articulation and explains them with respect to individual vocal tract shapes. Thanks to our cooperation partners, we will make use of some already recorded magnetresonance imaging data, computer tomographic and x-ray data of about 50 speakers producing several vowels. On the basis of these anatomical data we will parameterise the individual vocal tract shapes in a first step. In a second step we will implement these vocal tract shapes in a biomechanical tongue model in order to build speaker specific models. These models allow us to examine the link between motor commands, articulation, and acoustics and their respective variability.
Moreover, we will investigate the degree of jaw opening in low vowel production. Jaw opening has often been reported to be speaker-specific. We will discuss jaw opening as a potential compensation mechanism for particular vocal tract shapes.
The project will allow us to understand the role of the vocal tract shape (not only size!) and its constraints on acoustics and articulation.