Estimation of the power distribution function

The authors present two estimators for the power spectral distribution (PSD) function associated with a time series. Although both of them are based on a fast Fourier transform, they converge, with probability=1, the true PSD, and they are consistent under the appropriate hypotheses. Monte Carlo simulations confirm the result that the variance of the proposed estimates depends on the inverse of the number of data points. Finally, the PSD's resolution capability is compared to that of the standard Welch procedure. >