On the Envelope of Spectral Power of Stochastic Processes

Given a partial covariance sequence of a real stationary zero-mean stochastic process, one method of estimating the power spectrum is that of Maximum Likelihood. The Maximum Likelihood estimate represents an envelope of maximum spectral content corresponding to stochastic processes which may be complex. We determine the envelope of spectral content which is consistent with the assumption that the underlying stochastic process is real.