Quasi-optimum filter selection for the separation of a signal spectrum from a noise spectrum



Journal Title

Journal ISSN

Volume Title



In this thesis the mean-square error introduced when lowpass filters separate adjacent signal and noise spectra is analyzed, calculated, and tabulated. The error consists of a filter error and a crosstalk error from the noise spectrum. The variables involved are the filter type and order, the separation between the spectra, and the spectrum model. Tables were compiled using a digital computer program. The purpose of the tables is to enable a designer to select a filter that yields the desired signal quality. A graphical summary conveniently indicates the best filter for any condition considered in the tables. The first eight orders of the Bessel, Butterworth, and Chebychev filters are covered. The digital program included in Appendix B complements the tables.