Deriving Uniformly Spaced Samples from a Nonuniformly Sampled Function via Time Warping, with Applications to Fast Transforms

摘要

This paper presents an algorithm for deriving uniformly spaced samples from nonuniformly spaced samples of a signal (and visa versa). The nonuniformity must be a function of linear sampling i.e. to derive uniformly spaced samples from an exponentially sampled signal. One place this situation occurs is in the calculation of some integral transforms like the Hankel transform. The algorithm runs quickly, in linear time and can thus be used with the FFT to create fast transform algorithms and provide a superior alternative to linear interpolation.