The long-term goal of this project is to provide new insight into how the basic structure of the vocal tract and the kinematic perturbations superimposed upon it to produce speech, are acoustically encoded in the time variation of the formant frequencies, and perceptually decoded by the listener into phonetic elements.
The specific aims are: 1) To study the acoustic effects of consonantal constriction variability within adult and child-like vocal tracts;2) To conduct perceptual tests that assess listeners'identification of consonants in simulated VCV utterances;3) To study the temporal variability of vowel and consonant perturbations at the word and phrase levels;and 4) To conduct perceptual tests to assess the intelligibility of simulated and natural speech. Much of our knowledge of speech perception has been derived from experiments in which acoustic parameters such as formant frequencies are manipulated. In the experiments proposed here, it is the parameters of a kinematic model tract model that will be manipulated to produce speech for perceptual tests. Experiments in which we apply various scaling factors to the vocal tract length to simulate children's vocal tracts may provide important insights into speech development in children. Though we propose to study only normal speech in the current research plan, our methods may eventually contribute toward and an understanding of speech movement deficits, associated with neurogenic disorders, on vocal tract output and speech intelligibility. The combined results of these studies will enhance our understanding of 1) how kinematic variations of the vocal tract are related to the time-varying formant frequencies contained in the speech signal, and 2) how those kinematically produced formant frequencies are related to the perception of speech.
Showing the most recent 10 out of 21 publications
