The influence of window sidelobes on DFT-based multifrequency signal measurement

One of the causes influencing the uncertainty of the measured parameter is the mutual influence of the window spectra contained in the multifrequency signal spectrum. This paper analyses the effect of window spectrum sidelobes on the evaluation of multifrequency signal RMS value estimated by two DFT-based techniques (the “mainlobe method” and RMS value estimation using interpolated DFT). The RMS bias caused by window sidelobes when applying twelve different generalized cosine windows is compared. Both simulation results and theoretical relations are included. The presented principle is applicable for all DFT based algorithms, not only for those used for RMS value estimation.