FINITE-WORD-LENGTH DESIGN OF 2-D FIR DIGITAL-FILTERS FOR SAMPLING STRUCTURE CONVERSION

In this paper, an effective approach is proposed for designing discret e coefficient 2-D FIR digital filters for sampling structure conversio n. After obtaining the initial continuous solution, the conventional L agrange multiplier approach associated with an appropriate tree search algorithm is used iteratively to optimize the remaining unquantized c oefficients of the designed filter in the least-squares sense when one or more of the coefficients take on discrete values, till all of the filter coefficients are quantized. The method is simple and the perfor mance is comparable with that of the existing methods.