Synthesis of Kannada Isolated Consonant-Vowel Coarticulations using Formant Synthesizer Alfred Vivek D’Souza 1 and D.J Ravi 2 1 Assistant Professor, Vidyavardhaka College of Engineering, Mysuru, Visvesvaraya Technological University, Belagavi Email: alfreddsouza@vvce.ac.in 2 Professor, Vidyavardhaka College of Engineering, Mysuru, Visvesvarya Technological University, Belagavi Email: ravidj@vvce.ac.in Abstract—Formant analysis of various sound units play an important role in developing a functional Formant Synthesizer. This paper presents the result of formant analysis made on 15 different consonants occuring in context of 10 vowels and 2 diphthongs of Kannada language. The results were used to generate isolated Consonant-Vowel Coarticulations using Klatt like Formant Synthesizer. Index TermsCoarticulations, Klatt Formant Synthesizer and Kannda CV synthesis. I. INTRODUCTION Considerable efforts are being made since the early days of speech processing in understanding the generation of speech. Speech Synthesis is the branch of speech processing which deals with generation of human speech artificially using machines, known as Speech Synthesizers. Human vocal system generates a series of sounds in specific order to convey information. A Speech Synthesizer which aims to generate intelligible speech must have the capability of producing different types of sounds the human vocal system can produce. The number and types of sounds that are produced when speaking, differs from one language to another. Sound units such as vowels and diphthongs can be pronounced independently where as consonats are often pronounced in conjunction with a vowel or diphthong, forming coarticulations. Coarticulation refers to the process of transition from one utterence to another. Coarticulations cause the vocal tract to vary its parameter in a complex way as opposed to the same sound units successively uttered in isolation. The goal of this study is to analyze how the vocal tract parameters change its configuation during coarticulations of Kannada language and synthesize such coarticulation using Formant Synthesizer. Kannada is one of the widely spoken languages of southern India having over 50 million speakers and has a rich set of sounds. A. Speech Synthesizers and Current Trends in Speech Synthesis Speech Synthesizers form the core of Text-To-Speech(TTS) systems and have variety of applications such as screen readers, that reads out the text on computer screens for visually impaired, talking help for those who Grenze ID: 02.ICCTEST.2017.1.4 © Grenze Scientific Society, 2017 Proc. of Int. Conf. on Current Trends in Eng., Science and Technology, ICCTEST