Abstract
Digital interpolation, or the process of increasing the sampling rate of a sampled-data signal is the main concern of this paper. After briefly reviewing the subject and emphasising the linear filtering approach to interpolation, it investigates the possibility of improving the performance of interpolators for band-limited signals regarding which no additional information is available. The approach suggested is to estimate the MEM spectrum and design optimum interpolators based on this estimate. Using this approach, extensive studies on a representative set of signals were carried out. A comparative evaluation of the performance of the various interpolation schemes is made on the basis of mean-square error criterion. The degree of improvement resulting from the suggested method is shown to depend on the nature of spectrum. Some useful computer programmes used in the study are also included.