Context-based entropy coding of VQ indexes
Xiaolin Wu, Jiang Wen, Wai-Yin Ng · 2002
Limited by the inherent computational intractability of vector quantization in dimensions, practical VQ coders all employ signal blocks of rather modest sizes, failing to capture all higher-order statistical dependency. Therefore, context-based, adaptive entropy coding of VQ indexes can significantly reduce the bit rate of VQ coders for a given distortion.