INFORMATION RATES OF PRE POST-FILTERED DITHERED QUANTIZERS/

We consider encoding of a source with pre-specified second-order stati stics, but otherwise arbitrary, by Entropy-Coded Dithered (lattice) Qu antization (ECDQ) incorporating linear pre- and post-filters, In the d esign and analysis of this scheme we utilize the equivalent additive-n oise channel model of the ECDQ. For Gaussian sources and square error distortion measure, the coding performance of the pre/post filtered EC DQ approaches the rate-distortion function, as the dimension of the (o ptimal) lattice quantizer becomes large; actually, in this case the pr oposed coding scheme simulates the optimal forward channel realization of the rate-distortion function. For non-Gaussian sources and finite- dimensional lattice quantizers, the coding rate exceeds the rate-disto rtion function by at most the sum of two terms: the ''information dive rgence of the source from Gaussianity'' and the ''information divergen ce of the quantization noise from Gaussianity.'' Additional hounds on the excess rate of the scheme from the rate distortion function are al so provided.