A modification of the Whittaker-Kotelnikov-Shannon sampling series (Q1069074)

The Whittaker-Kotel'nikov-Shannon sampling theorem states that any bandlimited function f with Fourier transform f{\^{\ }} having finite support (-\(\pi\) W,\(\pi\) W) can be reconstructed from its samples taken at instances k/W by its sampling series. In case the support of f{\^{\ }} is unbounded, the sampling series approximate f under additional (continuity) conditions for \(W\to \infty\), the order of approximation being known. Better orders can be achieved by using generalized sampling series constructed by discretizing bandlimited convolution integrals, see e.g. \textit{W. Splettstößer} [Arch. Elek. Übertr. 32, 267-275 (1978)] for examples. In the present paper the concept of discrete convolution operators for generalizing the sampling series is dealt with under a strong mathematical viewpoint. The authors' main theorem on the approximation of (Lipschitz-) continuous functions by a broad class of generalized sampling series includes direct and inverse results, results on simultaneous approximation of derivatives, and a saturation theorem. The proofs of their fundamental assertions are based on comparisons with the element of best approximation in an appropriate Paley-Wiener space \(\beta^{\infty}_{\sigma}\).