On Comparative Characteristics of the Statistics of Spectral Densities of Stationary Random Processes

The authors offer an asymptotic investigation and comparison, for finite sample sizes, of various statistics of the spectral densities of stationary processes. The standard deviations are compared both in relation to the local property of the spectral density and in relation to strong perturbations and noise at adjacent frequencies. A new type of statistic, which possesses properties of strong noise immunity together with some other conveniences, is investigated.