ABSTRACT
This chapter discusses the compute parameters of a window that are useful in choosing a suitable window for particular applications, such as power spectral estimation via discrete Fourier transform and the design of FIR digital filters. The reduction of bandwidth is unavoidable when applying data windows. When a data window is applied, it causes, increase in the bandwidth of analysis. The important window parameters in the context of performance, following the definitions of Refs. An approximate linear relationship between the side-lobe levels and the main-lobe widths is obtained for Kaiser's modified zeroth-order Bessel family, is given by where x represents the NHMLW and y gives the MSLL in dB. Thus, the selection of an optimal data window for spectral estimation turns out to be a judicious compromise among the various parameters presented in the Tables. The graphs presented in are useful in selecting a data window that is closest to the near-optimum Kaiser—Bessel window, namely, the modified zeroth-order Bessel family.
