Practical issues in the development of TTS and SR for the Sinhala language

Rip Wickramasinghe,Kh Kumara,Ngj Dias

doi:10.4038/josuk.v3i0.2739

Abstract

In this study, discussed some issues that are arising in the development of Text-to-Speech (TTS) and Speech Recognition (SR) systems for the Sinhala language are discussed. There are numerous benefits of TTS and SR to Sinhala community. In the context of Sinhala language, the Natural Language Processing (NLP) technologies are in the primary stage in comparison with other languages such as English, French, German, Japanese etc. Since current TTS and SR development environments are far from being language independent, the existing systems cannot be adopted directly to the Sinhala language. This contribution outlines problematic areas that come across in the development of TTS and SR systems for Sinhala language. While discussing those problems, some of the possible solution strategies for both TTS and SR development have been formulated. Keywords: Text-to-Speech; Speech Recognition; Sinhala Language; Natural Language Processing. DOI: 10.4038/josuk.v3i0.2739J. Sci. Univ. Kelaniya 3 (2007) : 63-72

Highlights

TTS can be considered as the automatic production of speech, through a grapheme-to-phoneme transcription of the sentences to utter (Dutoit, 1993)
ISSUES Some basic issues that arise at the development of TTS and Speech Recognition (SR) systems for the Sinhala language are discussed here
The reason is Sinhala corpus does not contain such a word. This type of code-switching is another issue that we need to handle in a SR system of Sinhala language

Summary

Introduction

TTS can be considered as the automatic production of speech, through a grapheme-to-phoneme transcription of the sentences to utter (Dutoit, 1993). ISSUES Some basic issues that arise at the development of TTS and SR systems for the Sinhala language are discussed here. Among the major issues that we are going to discuss are speech and text corpora, code switching of text and utterance, lexicalized foreign origin names or terms, prosody, dealing with dates and numbers.

Results

Conclusion