A unified method for quality assessment of signal reconstruction

In this paper we consider two classes of signals and treat them separately: continuous-frequency (continuous spectrum) signals and finite discrete-frequency (discrete spectrum) signals. The quality of signal reconstruction is assessed for both of cases from the Digital Alias-free Signal Processing (DASP) point of view. In the case of continuous-frequency signals, we concentrate on the effects of dealing with not sufficient number of samples needed for signal reconstruction. In the case of discrete-frequency signals, such signals are fully characterised by finite number of parameters. When the collected signal samples are subject to errors the reconstructions are not unique. We propose a reliability function to facilitate giving a sensible signal reconstruction.