Quantifying Rapid Variability in Accreting Compact Objects

I discuss some practical aspects of the analysis of millisecond time variability X-ray data obtained from accreting neutron stars and black holes. First I give an account of the statistical methods that are at present commonly applied in this field. These are mostly based on Fourier techniques. To a large extent these methods work well: they give astronomers the answers they need. Then I discuss a number of statistical questions that astronomers don't really know how to solve properly and that statisticians may have ideas about. These questions have to do with the highest and the lowest frequency ranges accessible in the Fourier analysis: how do you determine the shortest time scale present in the variability, how do you measure steep low-frequency noise. The point is stressed that in order for any method that resolves these issues to become popular, it is necessary to retain the capabilities the current methods already have in quantifying the complex, concurrent variability processes characteristic of accreting neutron stars and black holes.