A general method for the efficient analysis of sensitivities in digital filter designed by linear transformations

Abstract

A method is presented for fast and accurate calculation of multiplier coefficient sensitivities in linear transformation digital filters. Since the method is based on dealing with analog filter prototypes, it is suitable for all kinds of analog filters. Making use of a general theorem for signal-flow networks, an efficient approach is presented, which is based mainly on the computation of voltages and currents in the analog prototype and its transportation.<>