THE CLEANEST FOURIER SPECTRUM

A new method of removing false peaks from a power spectrum is introduced, the CLEANest algorithm, together with several variations on the theme. It is shown to be an effective technique for detecting and describing multiperiodic signals, and at least in some cases is capable of divining the true nature of a multiperiodic signal, even when the single strongest peak in a periodogram occurs at a spurious frequency. The method is capable of describing long-term time evolution of period and amplitude. These techniques are applied to artificial test data, and to visual observations of variable stars from the International Database of the American Association of Variable Star Observers (AAVSO).