Frequency errors in MEM processing

TL;DR: AR spectra generated by the Energy-Weighted (EW) method are shown to have improved performance over AR spectragenerated by the Burg algorithm, including sharper resolution, less frequency bias, less phase dependence, less noise spectral variance and absence of spectral line splitting.

TL;DR: It is indicated that the reduced frequency estimation error noted in the above paper may well be due to the modification of the primary signal segment prior to analytical processing rather than to the analytical processing itself.

TL;DR: The Burg method is compared with established Fourier transform based approaches using both simulation and experimental data taken from a WSI measurement of a step-height sample.

TL;DR: A straightforward non-recursive approach is described which (for sinusoids) not only is shown to be free of these anomalies but also seems to offer superior noise performance.

TL;DR: Using maximum entropy power spectral estimation, the estimate of the frequency of a sinusoid in white noise has been shown to be very sensitive to the initial sinusoidal phase as discussed by the authors, which can be reduced by replacing the real data by its analytic form, reducing the sampling rate by two, and employing the power spectral estimate for complex data.

TL;DR: The maximum entropy method (MEM) for spectral analysis was suggested by Burg (1967) and its mathematical properties have been discussed in detail by Lacoss (1971), Burg (1972), and Ulrych (1972b) who found that MEM in general is superior to the more conventional methods of spectral estimation as mentioned in this paper.