The Barkhausen noise emanating from the surface of thin foils of nearly pure iron and nickel was obtained by digitizing the noise signal over a complete hysteresis loop. This digitized signal is used to analyze various attributes of the noise such as the autocorrelation, power density spectrum, jump amplitude spectrum, and jump amplitude correlation. For thin foils (25 μm) the power density exhibits a peak in the low-frequency range, contrary to what is predicted for a series of statistically independent Barkhausen jumps. Return maps and plots of jump amplitude versus time between jumps show no evident deviation from random noise behavior. Thus jump correlations or clustering do not explain the power density spectra, and an additional mechanism for the loss of low-frequency power is required.