Stephen G. Lambacher
Ctr. for Lang. Res., Univ. of Aizu, Aizu Wakamatsu, Fukushima Prefecture, 965 Japan
Speech samples are collected from five Japanese university students and five native speakers of English (all Americans) who produce the fricatives /|oC/ and /s/ in sentence frames within the three-vowel environments: i--i, I--I, (ae ligature)--(ae ligature). Critical-band analysis is performed on a 50-ms segment of the fricative consonants and the resulting spectral characteristics of the Japanese and native production of fricatives is compared. Through analysis of native speaker production of the target sounds, the paper attempts to establish what the true parameters are by which to define the acoustical properties of the voiceless fricatives /|oC/ and /s/. The second part of this paper demonstrates how cross-linguistic data can be presented to Japanese learners to effectively enhance their speech production. Analyzed speech samples previously uttered by native and Japanese speakers are electronically transferred from the teacher's computer workstation to the workstations of each of the five Japanese students (each workstation is equipped with a sound analyzer that records speech and measures waveform, intensity, pitch, and formant frequencies). Speech samples from all five students are analyzed. Results show improvement in production of /|oC/ and /s/ after students are instructed and exposed to cross-linguistic data of the target sounds.