Conventional moving-window analyzers based on Fourier transforms sometimes lack the resolution required to separate each of the modes in a seismic wave-guide. It is possible to enhance the resolution of a moving-window analyzer by using a maximum entropy power spectral estimator to approximate the spectrum of each windowed segment of a trace. Barrodale and Erickson have developed a suitable maximum entropy algorithm which can also be applied to estimating the parameters required in the fast recompression of in-seam seismic arrivals.The Barrodale-Erickson maximum entropy algorithm appears to need a sample rate of approximately ten times the Nyquist rate in order to generate meaningful maximum entropy spectra. The required increase can be achieved in the laboratory by applying an accurate interpolator to field records. Noise captures the maximum entropy spectrum if the input signal-to-noise ratio falls much below 10 dB. Use of a maximum entropy spectral analyzer aids in both identifying modes in a waveguide system and estimating the group velocity-frequency characteristic parameter.