Abstract :
In this paper, we first derive an explicit formula for expressing the coefficients of an arbitrary-order Lagrange-type variable fractional-delay (VFD) digital filter as polynomials of the VFD parameter , and then develop some useful symmetries for both even- and odd-order Lagrange-type VFD filter coefficients. The coefficient-symmetries facilitate the evaluations of VFD filter coefficients as well as variable frequency responses with reduced computational complexity. More importantly, the coefficient-symmetries can be exploited for efficiently implementing both even- and odd-order Lagrange-type VFD filters as Farrow structure and a more efficient one called even-odd structure such that the subfilters have symmetric or anti-symmetric coefficients, which saves the storage for VFD filter coefficients and reduces the number of multiplications required in VFD filtering process by almost 50%. Therefore, exploiting the coefficient-symmetries not only speeds up the VFD filtering, but also reduces implementation cost.
Keywords :
delay filters; digital filters; filtering theory; frequency response; interpolation; Farrow structure; arbitrary-order Lagrange-type filters; coefficient symmetriy; computational complexity; digital filters; even-odd structure; frequency response; interpolating polynomial; variable fractional-delay; Computational complexity; Costs; Digital filters; Filtering; Finite impulse response filter; Frequency; Interpolation; Lagrangian functions; Polynomials; Signal processing; Coefficient-symmetry; Farrow structure; Lagrange polynomial; Lagrange-type variable fractional-delay (VFD) filter; even–odd structure; interpolating polynomial; variable fractional-delay (VFD) digital filter;