摘要：Context. Recently our ability to study stars using asteroseismic techniques has increased dramatically, largely through the use of space based photometric observations. Work has also been performed using ground based spectroscopic observations and more is expected in the near future from the SONG network. Unfortunately, the intensity observations have an inferior signal-to-noise ratio and details of the observations do not agree with theory, while the data analysis used in the spectroscopic method has often been based on overly simple models of the spectra.Aims. The aim is to improve the reliability of measurements of the parameters of stellar oscillations using spectroscopic observations and to enable the optimal use of the observations.Methods. While previous investigations have used 1D models, I argue that realistic magnetohydrodynamic simulations, combined with radiative transfer calculations, should be used to model the effects of the oscillations on the spectra. I then demonstrate how to calculate the visibility of the oscillation modes for a variety of stellar parameters and fitting methods. In addition to the methods used in previous investigations, I introduce a singular value decomposition based technique. This new technique enables the determination of the information content available from spectral perturbations and allows this content to be expressed most compactly. Finally I describe how the time series obtained may be analyzed.Results. It is shown that it is important to model the visibilities carefully and that the results deviate substantially from previous models, especially in the presence of rotation. Detailed spectral modeling may be exploited to measure the properties of a larger number of modes than possible via the commonly used cross-correlation method. With moderate rotation, there is as much information in the line shape changes as in the Doppler shift and an outline of how to extract this is given.
