On the Efficient Interpolation of Stochastic Band-Limited Signals

Interpolation of stochastic band-limited signals is of relevance both from the theoretical point of view and application purposes. It is well-known that the classical interpolation method for band-limited signals, which follows from the Whittaker-Kotel'nikov- Shannon sampling theorem, provides the appropriate answer to exactly reconstruct them. Unfortunately, whenever an open domain is in question, only truncated versions of cardinal series expansions can be employed and this generates artifacts on the interpolated data. In particular, the quality of the interpolation scheme affects the data statistics and is a relevant issue which must be carefully examined. An interpolation scheme, which minimizes statistical artifacts, is therefore of great interest. In this paper we propose the use of two efficient interpolation schemes and demonstrate that they provide much better results than the truncated cardinal series expansion.