Numerical computation of the human voice source

The voice is the carrier signal of speech. The process of voice production, also called phonation, can be described by the interaction between the tracheal airflow and the two elastic vocal folds in the larynx which are excited to periodical oscillations. Thus, the two oscillating vocal folds (normally between 100 Hz and 300 Hz) periodically interrupt the expiration air stream forming the primary acoustic voice signal. Although we use our voice continuously and take it for granted, the exact causalities between airflow, vocal fold dynamics, and resulting acoustic voice signal, especially for disturbed or dysphonic voice, are still not fully understood Our central objective is to develop an aeroacoustic computational model simVoice for clinical applicability in future. The simVoice model will be a hybrid 3D-FVM (computational fluid with driven structural dynamics) and 3D-FEM (aeroacoustics) model, being optimized in computing time due to reduced complexity but still able to resolve the phonatory components to the needed degree. Innovative scientific aspects of this project include the knowledge to which amount turbulent scales have to be resolved for sustaining critical acoustic characteristics, revealing the cause and effect chain of dynamics-airflow-acoustics for the phonation process and the first detailed numerical study on the dependencies of vocal fold dynamics towards the acoustic quality. The expected clinical valuable outcomes of simVoice are to (1) help understanding pathological and physiological voice production processes, (2) identify new treatment approaches and to (3) simulate conservative and surgical treatment outcome.

Today, communication disorders are of high social and economic relevance. Depending on the study, for teachers between 11% and 63% were reported to have voice problems, compared to around 6% within the normal population. The central objective of this research project was to develop an aeroacoustic computational model simVoice for clinical applicability in future. The incompressible CFD using a LES (Large Eddy Simulation) turbulence model is based on prescribed vocal fold oscillations identified first from synthetic and then from in-vivo and ex-vivo high-speed imaging. A pressure-driven airflow is used for the model. In this way the fluid-solid interaction problem, whose accuracy critically depends on reliable geometrical and material parameters of all layers of the vocal folds, is circumvented. According to a perturbation ansatz, the acoustic model is based on the perturbed convective wave equation (PCWE) with the substantial derivative of the incompressible pressure as a source term. In doing so, we can state that the computational model simVoice is capable to compute acoustic signals comparable to measured microphone signals of humans and enables to suggest successful conservative and surgical strategies: simVoice reveals laryngeal interrelations between airflow, vocal fold dynamics and resulting acoustics. Based on that knowledge, specific treatment strategies can be suggested. The knowledge to which amount turbulent scales have to be resolved could be determined and demonstrated. The detailed numerical studies on the dependencies of vocal fold dynamics towards the acoustic quality revealed the following: A high level of glottal insufficiency worsens the acoustic signal quality more than an asymmetric or aperiodic oscillation, but all symptoms combined further reduce the quality of the sound signal. Therefore, our simulation model simVoice is on the way for teaching young medical doctors and scientists.