Abstract:
In this paper, a transmission-line model of the dark and light allophones of the lateral approximant /l/ is presented. Modeling was based on vocal-tract area functions derived from MRI data of four speakers of American English (two males and two females). The transmission line consisted of 50--60 sections; each T-section is represented by four components (R, L, C, and G). Different topologies in which the vocal tract was modeled with and without side branches were investigated. Simulations were performed using the analog-circuit simulator HSPICE [Rael et al., J. Acoust. Soc. Am. 97, 3245(A) (1995)]. Small-signal analysis was used to determine the formant frequencies from the frequency response of the transmission-line model. The articulatory-to-acoustic mappings in /l/ for all four speakers will be discussed in detail. [Work supported by NSF Grant No. IRI-95-03089.]