A Robust Syllable Centric Pronunciation Model for Tamil Text To Speech Synthesizer

ABSTRACT

The Human–Computer Interaction era contrived the researchers to work on speech and languages to develop interactive interfaces. A speech synthesizer is one such interface facilitating people to amalgamate with the digital era. The present work is focused on developing a Letter-To-Sound mapping for a Tamil speech synthesizer, which is an intriguing task due to the script to sound mapping irregularities in Tamil. Tamil is a syllable-timed language, hence a new syllable centric rule-based approach is formulated in the present work with a more extended set of rules than the existing rule-bases in the literature. This proposed rule-based system outperforms the existing rule-based systems with a low Character Error Rate and High Mean Similarity Score.

KEYWORDS:

• Grapheme to phoneme
• Letter to sound
• Letter to syllable
• Lexicon
• Pronunciation model
• Seed lexicon
• Syllables
• Tamil
• Text to speech

ACKNOWLEDGMENTS

We express our fervent gratitude to Dr Va.Mu.Se.Muthuramalinga Andavar, associate professor in PG and Research Department of Tamil, Pachaiyappa's College, Chennai, Tamil Nadu, India; Dr S. Ganesh, assistant professor, Department of Tamil, Arul Anandar College, Karumathur, Madurai, Tamil Nadu, India; and Dr R. Vimala Devi, assistant professor, Department of Tamil, Chellammal Women's College, Chennai, Tamil Nadu, India, for their valuable help in building the pronunciation test set. We also thank the Tamil native speakers who actively took part and shared their opinion in the analysis of the pronunciation generation.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the authors.

Additional information

Notes on contributors

Vaibhavi Rajendran

Vaibhavi Rajendran holds a bachelor's degree in information technology and a master's degree in software engineering. She is currently pursuing her PhD degree in the field of computer science and engineering. Her research interests include natural language processing, speech synthesis and artificial intelligence.

Corresponding author. E-mail: [email protected]

G. Bharadwaja Kumar

G. Bharadwaja Kumar holds a PhD degree in computer science and his research interest include machine learning, data analytics, Internet of things, speech and natural language processing. He is very passionate about developing resources and applications for Indian Languages in the areas of Natural Language Processing and Speech.

E-mail: [email protected]