Vector quantization for speech coding and recognition
Robert M. Gray · The Journal of the Acoustical Society of America · 1986
Vector quantization (VQ) is the coding or mapping of continuous or high-resolution vectors into a finite rate digital representation. The general goal is either analog-to-digital conversion or data compression with a minimum loss of fidelity. In this paper, the basic vector quantizer structures and code design techniques for fixed rate vector quantizers are surveyed and applications to speech coding and recognition are described.