Online Speech-to-Text for Speech Training: The Perspective of Learners of English

Page No.: 
Joao Carlos Koch, Jr., Sapporo Gakuin University

This project analysed the effectiveness of a speech-to-text tool for language learning. Speech-to-text has existed for a while, and its main purpose is to allow users to dictate text in their native language with accuracy. With most tools, accuracy is enhanced through trainability; that is, the tool identifies the user’s pronunciation patterns and adapts itself to them. However, this is not desirable for language-learning situations because it would produce unrealistic, flawed feedback with decreased usefulness for the language learner. This research was conducted by choosing one tool suited to learners’ needs, which participants then used in an e-learning situation, with focus on speech training and phonemic and phonetic awareness-raising processes, guided by the tutor. Participants reflected on and evaluated the experience and the chosen tool. Results indicate that the experience is valuable and promising, although the tutor role is paramount in both providing feedback as well as promoting learner autonomy.