Polynomial implementation structure for lagrange-type variable fractional delay filters

By investigating the coefficient matrix of Farrow structure based Lagrange-type variable fractional delay (VFD) FIR filters, a new implementation strategy is proposed for this type of VFD. Due to the dependence of coefficient values within each subfilter and between adjacent subfilters, the proposed technique reduces the number of independent coefficient values to be synthesized to 25% of that of the original one. Furthermore, all subfilter coefficients could be scaled to integers by a common factor. Thus, multiplierless filtering techniques could be used to synthesize the coefficient values without introducing any quantization errors and therefore reduce the implementation complexity further.