Structure of a sampling rate converter based on elementary signal processing blocks

Sampling rate conversion is a technique widely used in many areas. For the most general case of an irrational and time varying conversion factor, a discrete time varying filter is needed together with a sample skip and/or sample duplicate operation. The functionality of this skip/duplicate block has not been treated in proper depth in the bibliography and in many cases it has been ambiguously described and located in the structure of a sampling rate converter. In this paper a detailed analysis of a sampling rate converter for an arbitrary and variable ratio is revisited. The operations involved to carry out the conversion and the order in which they must be performed lead to a novel structure where only elementary signal processing blocks (adders, multipliers, delays, and decimators) are employed to describe both the filtering and the skip/duplicate operations. This structure constitutes an improvement in clarity with respect to previous works.