The following is an abstract for the selected article. A PDF download of the full text of this article is available here. Members may download full texts at no charge. Non-members may be charged a small fee for certain articles.

Estimating Optimum Filter Length in Linear Prediction

Volume 51, Number 5 (May 1997) Page 718-720

Sievanen, O.-P.

In this article a new method to estimate optimum filter length in linear prediction is described. Linear prediction was used to en- hance resolution of a spectrum. In particular, the dependence of prediction error on filter length has been studied. With calculations of simulated spectra it is shown that the prediction error falls rapidly when the filter length attains its optimum value. This effect is quite pronounced when the spectrum has a good signal-to-noise ratio and the modified covariance method is used to calculate prediction filter coefficients. The method is illustrated with applications to real Raman spectra.