This paper describes the effectiveness of CALL (Computer-Assisted Language Learning) systems with regard to speech learning methods in second language acquisition. The discussion includes a report on the construction of CALL classrooms which are available for multimedia devices, such as CCD cameras, audiovisual equipment, speech analysis programs and so forth. We introduce our original multimedia CALL teaching materials, which are particularly useful for aural training. We also introduce our automatic speech evaluation CALL systems for second language acquisition. Finally, we describe an analysis of the actual usage of our CALL system in second language classrooms.
This paper describes a summary of acoustic analysis systems for speech CALL systems. First, speech signal recording and processing systems using a personal computer are described. In particular, a USB sound unit is strongly recommended for high quality recording. Furthermore, the concept of acoustic analysis methods and several software programs for acoustic analysis and speech recognition are introduced. Finally, several speech CALL systems using the programs are roughly explained. Using a speech recognition and acoustic analysis program, it is very easy to point out errors in pronunciation, intonation and rhythm for learners.
In this paper, we survey automatic evaluation methods for pronunciation proficiency of English and Japanese. We also introduce an English speech database read by Japanese students and a Japanese speech database read by foreign students, which were used in developing the evaluation systems and testing the systems. Finally, we describe our automatic evaluation methods for pronunciation of English and Japanese. Our methods for the evaluation of pronunciation proficiency approximate evaluation by teachers.
Pronunciation training for non-native speakers is viewed focusing on prosodic features. With a brief explanation on prosodic features of speech, necessary functions for pronunciation training system are first pointed out. Then several research works mainly conducted in the author's group are introduced; Japanese accent type training, intonation training, and English stress pattern training. In the work on Japanese accent type training, the fundamental frequency in mora unit was defined which well corresponded to the perceived pitch value. Based on the definition, a method was developed for the accurate recognition of Japanese lexical accents, and was applied to a system for teaching non-Japanese learners pronunciation of Japanese accents. Pronunciation training on intonation and rhythm in sentence level is introduced as an important issue for the future works.
The purpose of this study is to investigate the effect of self-monitoring activity on the pronunciation learning of Japanese initial voiced plosives by Korean L2 learners. Subjects were divided into two groups. The experimental group performed a self-monitoring task whereas the control group performed a traditional pronunciation learning task. Data from an identification test using words and from a perception test using stimuli were collected. Each subject's pronunciation was evaluated by native speakers and by acoustic analysis. The results showed that the experimental group outperformed the control group in terms of both perception and production of the target sounds.
This study examines the effect of the duration of vowels that precede or follow a geminate stop on the identification of the geminate consonant. Stimuli were made by lengthening or shortening the first and second /a/ in /uta(Q)tane/ while the length of the geminate stop was systematically modified. Twenty-four Japanese speakers were presented with these stimuli and were asked to judge whether or not each stimulus contains /Q/. The results revealed the significant effects of the duration of both the preceding and following vowels. It is also suggested that the VC unit may play a role in the perception of a geminate stop in Japanese.