New realizations of second-order all-pass transfer functions

This paper presents an implementation of a systematic method for realization of the second-order all-pass transfer functions using the minimum number of multipliers. The presenteded realizations are analyzed with regard to the error due to roundoff of the products. The realizations are compared in order to determine optimum realization for particular pole position. Using these filter sections for realisation of some recursive transfer function as the parallel connection of two all-pass networks, will lead us to the computationally efficient realization with low pass-band sensitivity.