A computer assisted learning system for pronunciation of Japanese word accent on the Internet

摘要

This paper describes a system of computer assisted language learning on the Internet with Java servlet to evaluate Japanese word accent. It is known to be difficult to master for non-native speakers. A learner for Japanese sends his/her own speech to the web server. Next, the sound file is automatically judged by a Java servlet program on the learner's word accent type. The program uses two features: relative position of accent and inclination of logarithmic fundamental frequency series. Finally, the learner's accent type judged by Java servlet and an image of learner's fundamental frequency pattern are displayed on the learner's screen.