Abstract—Nowadays, computer-assisted language learning (CALL) systems are widely used for language education. Since the pronunciation of Chinese is difficult, it is important to build a system to evaluate a learner’s pronunciation in a real-time fashion so as to maintain the motivation of learning. This study tried to develop such a system that not only judges pronunciations from the viewpoint of acoustic phonetics, but also provides a learner an advice on improving his/her pronunciations.
As the first step, we built a system for Chinese monophthong vowels, and analyzed the acoustic features of the pronunciations between Chinese and Japanese. The results show that it is possible to distinguish the characteristics of pronunciation using the formant frequency, which is one of the acoustic features; and it is also possible to distinguish round and unround lips, which has been difficult, by using three kinds of formant frequencies from the first formant to third.
Index Terms—CALL (computer-assisted language learning) system, e-learning, formant frequency, Chinese vowel.
The authors are with Tokyo University of Technology, Tokyo, Japan (e-mail: g2117013f5@edu.teu.ac.jp, {chin, she, kameda, ohno}@stf.teu.ac.jp).
Cite: Junya Shinzawa, Shumei Chen, Jinhua She, Hiroyuki Kameda, and Sumio Ohno, "The First Step towards Automatic Quality Evaluation of Chinese Vowel Pronunciations for Foreign Learners for Self-training," International Journal of Information and Education Technology vol. 9, no. 1, pp. 70-73, 2019.