Abstract: The paper presents an overview of the automatic speech recognition (ASR) systems with the goal of bringing forth how the advancements in field of ASR have been applied in attempt to develop a natural language ASR. Initial sections of this paper will introduce the basic building blocks of an ASR and interactions among them to implement a speech recognition system. The techniques used in each of these building blocks are briefly introduced. Later sections of this tutorial will focus on efforts at building these essential building blocks for implementing a Kannada ASR and techniques employed for this purpose.

Keywords: ASR, Acoustic, Vector Quantization, DTW.