You are here: MIMS > EPrints
MIMS EPrints

2012.48: A fast algorithm for spectral interpolation of sampled data

2012.48: D.F. Tang (2012) A fast algorithm for spectral interpolation of sampled data.

Full text available as:

PDF - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
163 Kb

Abstract

This paper describes a fast algorithm to interpolate between samples of a bandwidth-limited signal that has been sampled at regular intervals. The algorithm is most suited to situations when a small amount of error in the interpolated values is acceptable, the larger the acceptable error, the more efficient is the algorithm. This can be useful in singnal processing, image processing and as an alternative to finite difference calculations. The same algorithm can be used to calculate the value of an $n^{th}$ degree polynomial at the $2n^{th}$ degree Chebyshev points, given its values at the $n^{th}$ degree Chebyshev points, thus providing a fast algorithm for the approximate multiplicaltion of polynomials. This could be used in, for example, computer algebra systems. We consider signals of finite duration, but the same technique can be used to deal with infinite signals so could be used to achieve fast and power-efficient interpolation of streamed data. On a prarallel architecture, the algorithm also requires less inter-process communication than interpolation using a Fast Fourier Transform method.

Item Type:MIMS Preprint
Subjects:MSC 2000 > 65 Numerical analysis
MSC 2000 > 68 Computer science
MIMS number:2012.48
Deposited By:Mr D.F. Tang
Deposited On:30 April 2012

Download Statistics: last 4 weeks
Repository Staff Only: edit this item