Glottal area waveforms study from high speed video-endoscopic recordings and voice production model with aeroacoustic coupling driven by a forced glottal folds model

Source-filter paradigm is one of the most common approaches used by the scientific community to establish models for analysis and synthesis of voice sounds. But, if the source-filter models are easy to compute, they ignore some physical phenomena which are important for the naturalness of the sounds. In this presentation, we investigate how we can modify the usual source-filter model in order to take into account some simplified aeroacoustic coupling between vocal folds and vocal tract while keeping low-cost computation and efficient analysis methods. After presenting a glottal area waveform model, we establish a complete voice model based on the preceding glottal area waveform model and evaluate its relevance by comparing analysis and synthesis methods based on this model and on a usual source-filter model.