When used for frequency analysis, an FFT analyzer produces a line spectrum on the amplitude‐frequency plane. Each spectrum line represents the output of a filter centered at the frequency associated with the line, the shape of the filter being determined by the effective time window applied to the data being analyzed. While it is generally recognized that the use of an explicit window function is often desirable—e.g., to control spectral “leakage”—many FFT users seem unaware thatuse of a window implies a compensatory calibration adjustment that is dependent upon the character of the data and the objective of the analysis. Effective filter shapes for rectangular, Hanning, flat top, and Kaiser‐Bessel windows are discussed along with the concept of filter amplitude response and effective noise bandwidth. Appropriate calibration for rms, power, power spectral density, and energy spectral density spectra is explained.