The study constructs and evaluates an automatic speech recognition ASR-based system with multiple levels of corrective feedback to support adult learners with opportunities of practicing English speaking with immediate diagnosis. The corrective feedback includes three levels: the first level shows the learner's pronunciation score and audio waveform, the second level provides learners with a comment, a list of words that are pronounced accurately and inaccurately, and an audio toolbar for replaying the learner's utterance, and the third level shows a demonstration of the accurate utterances with both full sentence and single-word form at normal and slow speed. A total of 38 adults from Taiwan participated in this experiment, divided into an experimental and a control group. The control group practiced English speaking using the single-level-feedback system while the experimental group was given the three-level-feedback system. The results of the study indicate that the ASR-based system serves as a helpful tool for Taiwanese learners and the learners were satisfied with the system for self-paced learning. Besides, the learners in the experimental group with three-level feedback made significant progress in English speaking. Some research issues and suggestions are also presented for future reference.
[1]
H. Franco,et al.
Effects of Speech Recognition-based Pronunciation Feedback on Second-Language Pronunciation Ability
,
2007
.
[2]
Howard Hao-Jan Chen.
Developing and evaluating an oral skills training website supported by automatic speech recognition technology*
,
2011,
ReCALL.
[3]
Diego Giuliani,et al.
The effectiveness of computer assisted pronunciation training for foreign language learning by children
,
2008
.
[4]
Hsien-Chin Liou,et al.
A Study of web-based oral activities enhanced by Automatic Speech Recognition for EFL college learning
,
2007
.
[5]
Olle Bälter,et al.
Designing the user interface of the computer-based speech training system ARTUR based on early user tests
,
2006,
Behav. Inf. Technol..