Performance and limitation of discrete band-limited signal extrapolation

The discrete band-limited signal extrapolation is a finite-dimensional linear estimation operation. In this paper, we describe the algorithm as a two-step recognition and reconstruction process and analyze its performance and limitation by using the associated singular values as indicators of the algorithm's sensitivity and capability for signal identification. We also indicate that the performance of the algorithm depends on the dimension of the effective vector space, and provide the relationship between the variation of the dimension and control-point spacings.