The bound for the lossless coding rates as shown below: : R_X\geq H(X|Y), \, : R_Y\geq H(Y|X), \, : R_X+R_Y\geq H(X,Y). \, If both the encoder and the decoder of the two sources are independent, the lowest rate it can achieve for lossless compression is H(X) and H(Y) for X and Y respectively, where H(X) and H(Y) are the entropies of X and Y. However, with joint decoding, if vanishing error probability for long sequences is accepted, the Slepian–Wolf theorem shows that much better compression rate can be achieved. As long as the total rate of X and Y is larger than their joint entropy H(X,Y) and none of the sources is encoded with a rate smaller than its
entropy, distributed coding can achieve arbitrarily small
error probability for long sequences. A special case of distributed coding is compression with decoder side information, where source Y is available at the decoder side but not accessible at the encoder side. This can be treated as the condition that R_Y=H(Y) has already been used to encode Y, while we intend to use H(X|Y) to encode X. In other words, two isolated sources can compress data as efficiently as if they were communicating with each other. The whole system is operating in an asymmetric way (compression rate for the two sources are asymmetric). This bound has been extended to the case of more than two correlated sources by
Thomas M. Cover in 1975, and similar results were obtained in 1976 by
Aaron D. Wyner and
Jacob Ziv with regard to lossy coding of joint Gaussian sources. Later work studied variants in which the source statistics or source model are less restricted, including universal coding and strong-converse results for the Slepian–Wolf data-compression system, correlated general sources, and information-spectrum formulations. ==See also==