Articulatory synthesis of words in six voice qualities using a modified two-mass model of the vocal folds

A modified two-mass model of the vocal folds is introduced and applied to the articulatory synthesis of words in six voice qualities. The modified two-mass model uses mass elements that are inclined, instead of parallel, with respect to the dorsoventral axis as a function of the degree of abduction. This allows to produce the continuum of voice qualities from pressed over modal to breathy voices. Furthermore, the model is extended by a variable posterior chink to represent the space between the arytenoid cartilages, like in whispery phonation. Five words were each synthesized with different glottal settings to simulate modal voice, pressed voice, breathy voice, whispery voice, vocal fry, and falsetto. The stimuli were judged by a group of listeners in a forced-choice experiment with respect to the perceived voice qualities. Apart from whispery voice, which was more often judged as breathy than whispery, all voice types were identified as intended with probabilities between 50% (modal voice) and 94% (falsetto), which are well above the chance level of 16.1%.