Error spectrum shaping in 3-D separable-denominator digital filters based on reduced-dimensional decomposition

摘要

This paper is concerned with the error spectrum shaping of tree-dimensional (3-D) digital filters with separable denominator. The 3-D separable-denominator digital filters are described here by using minimal decomposition. A technique is developed for designing the error feedback coefficients that minimize the noise variance at the filter output due to quantization errors of signals after multiplications and summations. The optimal coefficients can be obtained in closed forms. Finally, a numerical example is given to illustrate the utility of the proposed technique.