Articulatory synthesis is a technique consisting in numerically simulating the physical phenomena involved in speech production. The
aim is to numerically reproduce a speech signal that contains the observed acoustic features with regards to the actual articulatory and
phonatory gestures of the speaker. The talk will show that this tool may be useful for researchers to solve problems related to phonetics,
or more generally speaking, related to speech production. Our application will be the fricatives.

Recent numerical studies have evidenced three distinct regimes of production of fricatives that are controled by the glottal abduction.
These regimes are characterized by the balance between the contributions of both the voiced and the frication noise sources. Interestingly, the regime of voiced fricatives, i.e. when both sources have similar energy, is a very unstable regime, namely small perturbations of the glottal abduction degree leads to large variations of the acoustic characteristics of the produced fricative. Experiments on real subjects have also evidenced these numerical observations, which could explain the different articulatory strategies used by the speakers to ensure the voiced/voiceless contrast during the production of fricatives. These observations open new ways of investigation about the role of these regimes in the percpetion of the voice quality, as well as its analysis and synthesis.

Benjamin Elie: Synthesis of running speech for studying the mechanisms of speech production : the case of fricatives

Benjamin Elie du Laboratoire Signaux et Systèmes (Supélec/CNRS/UPS) présente : "Synthesis of running speech for studying the mechanisms of speech production : the case of fricatives" Abstract: Articulatory synthesis is a technique consisting in numerically simulating the physical phenomena involved in speech production. The aim is to numerically reproduce a speech signal that contains the observed acoustic features with regards to the actual articulatory and phonatory gestures of the speaker. The talk will show that this tool may be useful for researchers to solve problems related to phonetics, or more generally speaking, related to speech production. Our application will be the fricatives. Recent numerical studies have evidenced three distinct regimes of production of fricatives that are controled by the glottal abduction. These regimes are characterized by the balance between the contributions of both the voiced and the frication noise sources. Interestingly, the regime of voiced fricatives, i.e. when both sources have similar energy, is a very unstable regime, namely small perturbations of the glottal abduction degree leads to large variations of the acoustic characteristics of the produced fricative. Experiments on real subjects have also evidenced these numerical observations, which could explain the different articulatory strategies used by the speakers to ensure the voiced/voiceless contrast during the production of fricatives. These observations open new ways of investigation about the role of these regimes in the percpetion of the voice quality, as well as its analysis and synthesis.