ASA 125th Meeting Ottawa 1993 May

2aSP7. Proposal of a wavelet model for vocalic speech coarticulation.

Robert C. Lange

Grad. Prog. in Acoust., Penn State Univ., 117 Appl. Sci. Bldg., University Park, PA 16802

A recent development in the theory of wavelets is wavelet system characterization [R. K. Young, Wavelet Theory and its Applications (Kluwer, Boston, 1993)]. In wavelet system theory, the wavelet transform can be used to describe the time-frequency behavior of a transmission channel, by virtue of its ability to describe the time-frequency content of the system's input and output signals. The present research proposes a wavelet-system model of speech production; wherein, the system is the process of transformation from one speech state to another. The transmission channel generates an effected utterance from a control utterance. A speech production effect (such as voice quality, or the influence of the speaker) is thereby characterized in a time-frequency manner. The effect of CVC coarticulation is investigated in particular, using samples of real speech. In this case, an isolated vowel utterance (the control) is contrasted with an utterance of the same vowel imbedded within a /C--C/ context (the effected). The channel thus describes (in terms of wavelets) the CVC coarticulation. [Work supported by Appl. Res. Lab.]