Simulation of flow and acoustics in the vocal tract

A widely used approach for the simulation of vocal tract acoustics is the transmission line model (TLM) with lumped elements [2, 3]. The TLM is based on onedimensional acoustic propagation in a tube with a piecewise constant cross-sectional area. The governing equations for the TLM were originally derived under the following assumptions: The sound field causes only small perturbations of the thermodynamic equilibrium and the particle velocities are small compared to the speed of sound. However, the assumption of small fluid velocities in the vocal tract is often not justifiable. For instance at the constriction of fricatives or shortly after the release of plosives the particle velocities can exceed 0.1 Mach [1, p. 47-51]. The peak velocities of the airflow through the glottis during normal phonation are of the same order of magnitude, too. The consequences of these high velocities are the Bernoulli effect and, if the occasion arises, energy losses due to flow separation from the vocal tract walls at sudden expansions (shock losses) or due to flow contractions at sudden constrictions (e.g., at the inlet of the glottis).