Optimum Smoothing of Infrared Spectroscopic Data with Numerical Filters

Abstract
Numerical smoothing is one of the useful methods to extract the signal components faithfully from noisy spectroscopic data. However, the setting of the filtering parameters has been empirically done to allow the optimum operation because of its complexity. This article discusses the practical methods for finding the relevant parameters of the numerical filter and the data source by using the numerically analyzed characteristics of the least squares convolution filters. The detail procedures are demonstrated in far infrared spectroscopic applications. A result of the quantitative comparison between the numerical filter and the familiar electric RC filter is also given.

This publication has 7 references indexed in Scilit: