Spectral window in a MEM based spectrum analyzer

In this paper one aspect of a new high resolution spectrum analyzer is discussed. The analyzer is based on computing the sampled auto-correlation function of the signal and employs the Maximum Entropy Method (MEM) of spectrum analysis. The Maximum Entropy Method as applied to this instrument has some drawbacks when analyzing sinusoidal signals. This paper identifies these problems and suggests windowing of the data as a possible solution. Three different types of spectral windows are discussed for this purpose.