Design and Analysis of a Successively Refinable Lattice Quantizer for i.i.d. Gaussian Sources

摘要

A successively refinable quantizer is proposed for independent identically distributed (i.i.d.) Gaussian sources. Our quantizer is a successively refinable extension of the Gaussian source coding algorithm based on wrapped spherical codes designed by Hamkins and Zeger in "Gaussian Source Coding With Spherical Codes," IEEE TRANSACTIONS ON INFORMATION THEORY, vol. 4, no. 11, pp. 2980-2989, November 2002. The proposed successively refinable quantizer is analyzed for the case of high resolution and optimum bit allocation is computed. Our simulation results show that our proposed quantizer outperforms other successively refinable quantizers designed for Gaussian sources.