Syllable as the basic unit for Kannada speech synthesis

Abstract

Speech is the primary mode of communication. Speech synthesis is a very important application in the linguistic communication process and it is the artificial production of human speech using a computer system. A text to speech synthesis (TTS) system converts text into speech. TTS can be developed using synthesis methods like formant synthesis, articulatory synthesis, and concatenative synthesis. This paper describes concatenative speech synthesis using syllable as a basic unit of concatenation. A syllable is the basic unit of language, which may be spoken independently of adjacent phones. The primary reason for choosing the syllable as a basic unit is that the Indian languages are syllable-centered. The work proposes syllable-based speech units for concatenative speech synthesis for Kannada language. The work takes into account the position of the syllables in a word. We divide the process into multiple steps: the entered text file is analyzed, then syllabication is performed based on the syllabification rules of Kannada language, and later, the syllables are stored separately in a variable. Finally, the corresponding speech file for the syllable is retrieved, and the resulting synthesized speech of good quality is produced.