Optimization and efficient implementation of fractional delay FIR filters

An efficient design method is introduced for optimizing polynomial-based fractional delay filters. For these filters, the worst-case amplitude distortion occurs in the given passband when the delay is half the sampling interval. The corresponding discrete-time filter is optimized to minimize the passband amplitude distortion, binding some degrees of freedom. The remaining degrees of freedom are used for minimizing the worst-case deviation of the phase delay response from the desired delay value. Because these filters are polynomial-based, they can be efficiently implemented using the so-called Farrow structure.