ASR-based system for promoting pronunciation

Abstract

Abstract In developing English as a Lingua Franca, educators and researchers must employ new methods in language acquisition to make the learners internationally intelligible and comprehensible. This study aimed to determine the implication of infinite access to Automatic Speech Recognition (ASR)-based language-learning program promoting pronunciation skills acquisition of vocational higher education students. In this study, the students learned English as Lingua Franca (ELF) through a collaborative approach. It applied a qualitative approach with 67 first-year university students from three intact classes as participants. One class of 24 students was assigned as the first group utilising the ASR BoldVoice as the additional learning materials in their speaking class. In contrast, the other two classes of 43 students were set as the second and third groups carrying out conventional learning procedures. The development of participants’ utterances was analysed in terms of fluency, completeness, and accuracy at pre- and post-test. In order to support the obtained data, a semi-structured interview was performed right after the post-test. The result showed that the pronunciation skill of the experimental group was significantly improved in particular phones of pronunciation: /ch/, /th/, /sh/, and /j/. The analysis of the interview data confirmed that the students gained substantial improvement with regard to their pronunciation and communicative competence compared to the conventional learning process.