In the robust AR(1) estimation of the MTM spectral analysis, changing the AR(1) best fit model (1 = linear, 2 = log power) has no effect on the fits produced. Regardless of whether the linear or log-power fit is selected, the program appears to be generating a fit to the linear-scaled median-smoothed spectrum.
In the robust AR(1) estimation of the MTM spectral analysis, changing the AR(1) best fit model (1 = linear, 2 = log power) has no effect on the fits produced. Regardless of whether the linear or log-power fit is selected, the program appears to be generating a fit to the linear-scaled median-smoothed spectrum.