Accurate glottal model parametrization by integrating audio and high-speed endoscopic video data

Abstract

The aim of this paper is to evaluate the effectiveness of using video data for voice source parametrization in the representation of voice production through physical modeling. Laryngeal imaging techniques can be effectively used to obtain vocal fold video sequences and to derive time patterns of relevant glottal cues, such as folds edge position or glottal area. In many physically based numerical models of the vocal folds, these parameters are estimated from the inverse filtered glottal flow waveform, obtained from audio recordings of the sound pressure at lips. However, this model inversion process is often problematic and affected by accuracy and robustness issues. It is here discussed how video analysis of the fold vibration might be effectively coupled to the parametric estimation algorithms based on voice recordings, to improve accuracy and robustness of model inversion.