The overall strategy for structuring the problem is as follows. The respiratory system will be considered to be out of scope; the lung pressures applied to drive the simulations will simply be posited, using realistic temporal contours. The nasal tract is practically immobile and will be included only as a static, acoustically equivalent tube, to which acoustic access is controlled by movements of a simulated velum (soft palate). The bulk of the project concerns the larynx and the vocal tract. Each of these is assigned two work packages. The first is for numerical simulations (WP2 and WP5). The second is for experimental validation using mechanical replicas (WP3 and WP6). Such validation is essential: it allows for precise quantitative control of the model parameters (e.g., vocal fold elasticity, constriction geometry, opening/closing gestures) which are also needed for the numerical simulations. Mechanical replicas also confirm the validity of approximations that must be made in the simulations.
A central issue is the integration of the fluid-structure simulations with the resulting acoustic pressure field, involving a reconciliation of a non-compressible and a compressible domain of simulation. This task is addressed by a dedicated WP4, which aims at the breakthrough of achieving an entirely unified computational domain, including the acoustics.
The simulation of dynamic changes in the larynx and the vocal tract will need a rich set of time-varying motor control parameters, which ultimately are to be derived from muscle activation models and phonetic representations. Designing the control architecture and its interface to external users is the task of WP7, which will interact with WP’s 2, 3, 5 and 6.
Finally, the management of the project as a whole is in WP1, while dissemination activities are organised and implemented in WP8.