JOURNAL OF INFORMATION SYSTEMS & INFORMATION TECHNOLOGY (JISIT)

Call For Papers. Vol.11 No. 1 - 2026

Volume 10. No. 02, 2025

Neural Tamil Text-to-Speech for Children’s Story Narration: Design, Training, and Evaluation
Mohamed Iqbal Fathima Nihla and Koliya Pulasinghe

Tamil is spoken as a first language by more than five million people in Sri Lanka, representing approximately 15% of the population, and serves as the nation's second official language. Despite its importance, many young children experience difficulties in learning and reading Tamil, particularly those enrolled in English-medium preschools. Storytelling is a valuable educational tool that supports language development, knowledge transfer, and the preservation of cultural values. However, the complexity of Tamil grammar and the limited advancement of Tamil Natural Language Processing (NLP) technologies have resulted in a shortage of effective digital learning resources. To address this issue, this research developed a storytelling application that uses Tamil Text-to-Speech (TTS) technology to narrate stories for children. An evaluation comparing TTS-generated narration with human speech showed no significant difference in speech quality, indicating that modern Tamil TTS systems can effectively support children's storytelling and language learning experiences.
Keywords: Text-to-speech, MaryTTS, Tamil Language, Children Story

01-13
PDF