Second European Conference on Speech Communication and Technology

Genova, Italy
September 24-26, 1991


Synthesizing Intonation for Speech in Hindi

A. S. Madhukumar, S. Rajendran, C. Chandra Sekhar, B. Yegnanarayana

Department of Computer Science and Engineering, Indian Institute of Technology, Madras, India

The present work lays emphasis on acquisition and incorporation of intonational rules in a text-to-speech system for Hindi, an Indian language. The phonetic characteristics of Indian languages are exploited in building the basic system. Intonation refers to the stream of fundamental frequency values which models the timing of glottal pulse source of voiced speech sounds. Properties of intonational patterns such as declination tendency, local fall-rise pattern, resetting of pitch contour and inherent fundamental frequency are discussed. The rules governing these properties are coded into production system format. The synthesized speech becomes more intelligible and natural when these are incorporated into the system. Keywords: text-to-speech system, fundamental frequency (F0), declination tendency, fall-rise pattern, word order, resetting, production system

Full Paper

Bibliographic reference.  Madhukumar, A. S. / Rajendran, S. / Sekhar, C. Chandra / Yegnanarayana, B. (1991): "Synthesizing intonation for speech in hindi", In EUROSPEECH-1991, 1153-1156.