Aliasing and Poisson summation in the sampling theory of Paley-Wiener spaces (Q1907298)

The sampling theorem interfaces between analog and digital signal analysis and presents a leitmotiv for signal reconstruction from samples [\textit{J. P. Mugler III} and \textit{J. R. Brookeman}, The optimum data sampling period for maximum signal-to-noise ratio in MR imaging, Rev. Magn. Res. Med. 3, 1--51 (1988)]. It is based on the Poisson summation formula which interfaces between the Fourier series of a periodized function and its Fourier transform. In fact, the Whittaker-Kotel'nikov-Shannon sampling theorem for bandlimited signals is mathematically equivalent to the Poisson summation formula. If the signal of interest has a frequency beyond the band which is limited by the Nyquist frequencies, a higher frequency appears to be at a lower frequency, a falsification of frequency phenomenon commonly referred to as aliasing. In most practical applications it is not possible to guarantee that the signal wavelet being processed will be strictly band-limited, and hence before every analog-to-digital converter there is always a low pass filter, known as anti-aliasing filter, which, ideally, removes components above the baseband [\textit{R. A. Haddad} and \textit{T. W. Parsons}, Digital signal processing: theory, applications, and hardware. Computer Science Press, New York, Oxford (1991)]. In the field of magnetic resonance imaging, the sampling phenomenon of Brillouin zoning due to undersampling is also known as wraparound or foldover artifact. Aliasing in the phase encoding direction is more difficult to correct in practice than alias distortions occurring in the readout direction [\textit{R. M. Henkelman} and \textit{M. J. Bronskill}, Artifacts in magnetic resonance imaging, Rev. Magn. Res. Med. 2, 1--126 (1987)]. The paper under review deals with bounds for the pointwise aliasing error of one- and two-channel sampling series in terms of Paley-Wiener type spaces. Unfortunately, the paper does not emphasize the symplectic affine structure of the Poisson summation formula which is basic for practical sampling procedures of signals in quadrature format by using the phase routing or swapping technique [the reviewer, Wavelets in high resolution radar imaging and clinical magnetic resonance imaging. Cybern. Syst. 28, No. 1, 1--24 (1997; Zbl 0864.94011)], nor does it provide a detailed analysis of the error due to aliased components present in subband filter banks which interfere with efficient coding [\textit{N. J. Fliege}, Multirate systems, filter banks, wavelets (Wiley, 1995)]. Because the sampling theorem should be regarded as a guide rather than an explicit signal reconstruction formula, the results derived in the paper have limited applicability for hardware design [\textit{P. Horowitz} and \textit{W. Hill}, The art of electronics. 2nd ed., Cambridge University Press, Cambridge, New York, Port Chester 1990].