We determine the rate region of the quadratic Gaussian two-encodersource-coding problem. This rate region is achieved by a simple architecturethat separates the analog and digital aspects of the compression. Furthermore,this architecture requires higher rates to send a Gaussian source than it doesto send any other source with the same covariance. Our techniques can also beused to determine the sum rate of some generalizations of this classicalproblem. Our approach involves coupling the problem to a quadratic Gaussian``CEO problem.''
展开▼