A hybrid coding involving ADM and vector quantization for digital video image compression

In the search for reduction of the unacceptably high bit rates of PCM digital images, vector quantization seems to have the edge: its theoretical distortion is lower than that of other block coding methods at comparable bit rates. However, to obtain high quality pictures, when used as a stand-alone process, a vector quantizer requires a codebook of large size and is excessively computational intensive. For the dynamic range of the original images, 0 to 255, it is doubtful that a single optimal and universal codebook of manageable size can be obtained. Hence, to reduce the dynamic range of the signals encoded using VQ, we propose a method that uses Adaptive Delta Modulation with one stage look ahead as a preprocessor in conjunction with vector quantization of the appropriately defined error signals. Using such a scheme, it is shown that good quality composite color video images can be obtained using 1.5 to 2.0 bits per pixel and codebooks which are small and universal.