5th International Conference on Spoken Language Processing

Sydney, Australia
November 30 - December 4, 1998

Enhancing Speech Processing of Japanese Learners of English Utilizing Time-Scale Expansion With Constant Pitch

Kaoru Tomita-Nakayama (1), Kazuo Nakayama (1), Masayuki Misaki (2)

(1) Yamagata University, Japan
(2) Matsushita Electrical Industries Co. Ltd., Japan

This study demonstrated that the time-scale expansion of speech with the constant pitch (henceforth, expanded speech) enhanced the speech recognition of Japanese learners of English in contrast to the previous studies in which the time-scale expanded speeches did not contribute to the speech recognition chiefly because of the severe distortion of original speech and pitch change. Experiments were administered with the stimuli of original normal speech and the corresponding expanded speech. These results showed that the expanded speech stimuli were intelligible to many of the subjects. The hypotheses are that the expanded speech enhances listeners' speech processing and also enables the listeners to call into play virtual memory capacity for an on-line speech processing, which are more apparent in a longer stimulus. The expanded speech worked well with most subjects. Another prescriptions should be prepared for the rest of the subjects for whom the expanded speech was not solely much effective.

Full Paper

Bibliographic reference.  Tomita-Nakayama, Kaoru / Nakayama, Kazuo / Misaki, Masayuki (1998): "Enhancing speech processing of Japanese learners of English utilizing time-scale expansion with constant pitch", In ICSLP-1998, paper 0180.