Design of sinusoid-based variable fractional delay FIR filter using weighted least squares method

In this paper, the design of sinusoid-based variable fractional delay FIR filter is presented. First, the coefficients of filters are expressed as the linear combination of sinusoidal basis by using the Fourier series expansion. Then, by minimizing the weighted squares errors of frequency response, the optimal linear combination coefficients are obtained by solving a set of linear simultaneous equations. Finally, the design examples are used to demonstrate that the design error of proposed sinusoid-based method is smaller than the one of conventional polynomial-based method for the same truncation order.