Treffer: Optimum discrete filtering of band limited signals

The problem of designing a finite impulse response (FIR) digital filter to approximate a desired spectral response is presented. The FIR filter is used as a structure to provide a minimal maximal (MINIMAX) error estimate of an energy bounded band limited design signal. Algorithms and the computer programs are included for designing low pass filters. A comparison of this filter to filters designed from an estimator structure which provides a minimum mean-squared error (MMSE) estimate of stochastic classes of signals is shown. When the MINIMAX designed filter is used at its optimum conditions, filter performance exceeds that of the MMSE designed filters. However, when design conditions for transition bands are wider than optimum conditions for the MINIMAX design, it cannot be used whereas the MMSE filter exhibited improved performance. Both techniques used the Toeplitz matrix inversion which provides for minimal computation time. (See more in text.) ; Includes bibliographical references (page 41) ; California State University, Northridge. Department of Engineering.