Spectrum analysis using frequency-domain adaptive windowing

An adaptive spectral windowing algorithm has been developed to solve the leakage problem that occurs when the discrete Fourier transform is used to estimate spectra. In contrast to fixed windows presently employed in spectral analysis, the algorithm uses actual spectral data to yield an optimum window. It is derived by minimizing the instantaneous bin spectral power subject to quadratic constraints to maintain the response to signals not at the exact bin center frequency.