The paper deals with a possible realization of eliminating the effect of noise in scalar hysteresis measurements. The measured signals have been transformed into the frequency domain, and, after applying digital filter, the spectrums of the filtered signals have been transformed back to the time domain. The proposed technique results in an accurate noise-removal algorithm. The paper illustrates a fast controlling algorithm applying the inverse of the actually measured hysteresis loop, and another proportional one to measure distorted flux pattern. By developing the mentioned algorithms, it aims at the controlling of a more complicated phenomena, i.e. measuring the vector hysteresis characteristics. © 2007 Elsevier B.V. All rights reserved.