USING OF NATURALREADER: A TEXT-TO-SPEECH TECHNOLOGY USED BY EFL STUDENTS IN LEARNING PRONUNCIATION

Abstract

Pronunciation is a fundamental yet frequently neglected component of English as a Foreign Language (EFL) education, often overshadowed by grammar and vocabulary due to instructional time constraints. This study investigates the effectiveness and user experiences of NaturalReader, an AI-powered Text-to-Speech (TTS) application, as a learning medium to improve pronunciation skills among tenth-grade students at MA Miftahul Jannah Palangka Raya (Academic Year 2025/2026). Employing a mixed-methods approach, quantitative data were gathered through pre-tests and post-tests to measure phonetic accuracy, while qualitative insights were collected via classroom observations and semi-structured interviews to capture student perceptions. The findings reveal a significant improvement in students' command over segmental features (vowels, diphthongs, and consonants) and suprasegmental features (stress and intonation). Furthermore, the application successfully mitigated pronunciation-related anxiety by establishing an autonomous, low-stakes practice environment. This research underscores the strategic value of integrating mobile-friendly TTS technology to bridge the gap between regional phonetic interference and standard English pronunciation.