Article: Real-time feedback system for English listening comprehension using speech recognition and synthesis Journal: International Journal of Information and Communication Technology (IJICT) 2025 Vol.26 No.33 pp.76 - 90 Abstract: Addressing the lack of instant feedback in English listening practice, this study introduces a real-time feedback system leveraging integrated speech recognition and synthesis. We achieve low-latency recognition (<50 ms) via a lightweight streaming conformer model. An end-to-end feedback pipeline is constructed by innovatively integrating confidence-driven keyword localisation with intelligibility-enhanced FastSpeech2 synthesis. Evaluations on LibriSpeech and a customised dataset (200 non-native speakers) demonstrate a mean system latency of 230 ms. User studies reveal a 28.3% relative improvement in listening comprehension accuracy and a user satisfaction rating of 4.7/5.0. This system provides effective technical support for adaptive language learning frameworks. Inderscience Publishers - linking academia, business and industry through research

Title: Real-time feedback system for English listening comprehension using speech recognition and synthesis

Authors: Ling Zhang

Addresses: School of Foreign Languages, Chifeng University, Chifeng, 024000, China

Abstract: Addressing the lack of instant feedback in English listening practice, this study introduces a real-time feedback system leveraging integrated speech recognition and synthesis. We achieve low-latency recognition (<50 ms) via a lightweight streaming conformer model. An end-to-end feedback pipeline is constructed by innovatively integrating confidence-driven keyword localisation with intelligibility-enhanced FastSpeech2 synthesis. Evaluations on LibriSpeech and a customised dataset (200 non-native speakers) demonstrate a mean system latency of 230 ms. User studies reveal a 28.3% relative improvement in listening comprehension accuracy and a user satisfaction rating of 4.7/5.0. This system provides effective technical support for adaptive language learning frameworks.

Keywords: real-time feedback system; streaming speech recognition; speech synthesis; English listening training; confidence thresholds.

DOI: 10.1504/IJICT.2025.148658

International Journal of Information and Communication Technology, 2025 Vol.26 No.33, pp.76 - 90

Received: 30 Jun 2025
Accepted: 23 Jul 2025
Published online: 17 Sep 2025 *

Title: Real-time feedback system for English listening comprehension using speech recognition and synthesis

Keep up-to-date