3.1 INTRODUCTION

The implementation of sophisticated speech-coding algorithms is increasingly costeffective in numerous applications as a result of recent advances in large-scale integration. The purpose of these algorithms is to transmit, store, or synthesize speech at a given quality using fewer bits. This reduction is achieved by removing the redundancy of the speech signal.

It is possible to find common traits in the many coding algorithms that have been developed over the past 20 years. It is the purpose of this chapter first to pinpoint these common traits (fundamental coding functions and goals) and then to illustrate their operation in basic coding schemes. In subsequent sections we deal with the concept and applications of the recently developed technique of vector quantization. Table 3.1 gives a list of the fundamental coding functions, which are discussed together with the algorithms considered. Several good tutorial [Crochiere, 3.1] review papers [Jayant, 3.2; Flanagan 3.3; Gibson, 3.4] and books [Rabiner, 3.5; Jayant, 3.72] published in recent years provide complementary insights and overview of this vast field. Preceding the discussion of the various coding functions we outline the overall characteristics of coding algorithms and briefly state the basic properties of the speech signal.