Introduction

Abstract

The main aim of speech is for communication. The way of representing the speech is through a signal or waveform, which carries the message content or information. The speech signal is transmitted, stored, and processed in many ways in communication systems. First the message content of the speech signal should be kept safe and the next is representation of the speech signal in a convenient form which is more useful for transformation with out modifying or loss of message content in the speech signal. The speech produced by a human is analog signal which is continuously varying with time and is represented by x(t). Here t represents time variable. For digital speech processing we can convert that analog speech signal into digital. In digital notation it can be represented as a sequence of samples x(n). The more convenient form of representing the speech signal is by sum of sinusoids as it leads to convenient solution to problems like formant estimation, pitch period estimation etc.