Speech-to-Text-Mediated Vowel Pronunciation Training in Korean EFL Learners: An Acoustic Analysis of Pre-Service English Teachers

This study examined whether short-term Speech-to-Text (STT)-based pronunciation training is associated with acoustic changes in English vowel production by Korean learners of English as a foreign language (EFL). Ten pre-service English teachers participated in a two-week training program involving repeated speech production with feedback provided through STT transcription accuracy. Pre- and post-training productions of English vowels were acoustically analyzed, focusing on formant frequencies (F1, F2) and vowel duration using linear mixed-effects models. The results revealed a limited and dimension-specific pattern of change. While no significant effects were observed for vowel height (F1) or duration, a condition-by-backness interaction in F2 suggested localized adjustments, particularly for back vowels. These patterns were observed across both male and female speakers without significant gender-related interaction effects. The findings suggest that STT-based practice may be associated with selective acoustic adjustments rather than global phonetic restructuring. Given the small sample size and short training period, the results should be interpreted as exploratory. Further research with larger samples and longitudinal designs is needed to determine the robustness of these effects.