Bit allocation and encoding for vector sources

摘要: This paper considers the problem of efficient transmission vector sources over a digital noiseless channel. It treats optimal allocation total number available bits to components memoryless stationary source with independent components. is applied various encoding schemes, such as minimum mean-square error, sample-by-sample quantization, or entropy quantization. We also give optimally decorrelating scheme for whose are dependent and treat problems selecting optimum characteristic that overall mean-squared error minimized. Several examples including ideal encoder achieves rated istortion bound, related practical discussed.