Design Method of Kernel for Sampling Rate Converter Using Block Structure

Abstract

We present a kernel for sampling rate conversion based on Nth polynomials and ratio of each block. The proposed kernel is approximated a main robe of Sinc function by the Nth polynomial and side robes by the different Nth polynomial function and/or its amplitude ratio. The optimization of kernel is carried out by semidefinite programming (SDP). The proposed kernel has the advantage of not redesigning one whenever a sampling rate is changed. Moreover we present the kernel realized in the Farrow structure form. The usefulness of the proposed kernel is verified through examples.