next up previous contents
Next: Spectral ranges and units Up: Fit parameters Previous: Apertures   Contents

Background

Besides the instrumental background a constant source background sbg can be added (units [counts/Å]) accounting for eventual continuum radiation or unresolved weak lines (''Pseudo continuum'') originating from the source. If the switch sbgit (on the same line of the parameter file) is activated, then the given value for sbg is ignored, and the source background will be determined by by an iteration of median values, applied to the part of the spectrum under consideration (see below).

The third value on this input line, the bgsmooth switch, decides whether the instrumental background will be smoothed, i.e., that a polynomial of 2$^{\rm nd}$ order is used to model the instrumental background (a polynomial of lower order is used in the case that the 2$^{\rm nd}$ order polynomial fails). Only the part of the spectrum under consideration is smoothed in this way.

More refined methods to estimate the source continuum are provided (see Ness et al. 2003, A&A 407, 347 and 2004, A&A 427, 667). A $\chi^2$ method is programmed using the model parameters given by start values in order to generate a model spectrum. This is added to a source background value to be iterated by minimizing $\chi^2$. This method is a good approach in cases when line wings are broad (thus, too many bins belong to emission lines) and when no lines features in the spectrum are excluded from the fit. With the fourth number on the input line in the parameter file this procedure can be activated. The last entry in that line ($n\sigma$) is a value used for a refined median method. Since the median has to rely on that at least half of all bins have to belong to the background, too crowded line regions can be problematic. This can be solved by removing all bins which obviously belong to lines from the list of bins used to calculate the median value. In this way median values are iteratively calculated after removal of all bins with count values higher than $n\sigma\times \sqrt{{\rm median}_{\rm last}}$. Small values of $n\sigma$ will more critically remove high-count bins resulting in lower source background values. Usage of this parameter represents a parameterized choice of source continuum values by eye.


next up previous contents
Next: Spectral ranges and units Up: Fit parameters Previous: Apertures   Contents
Jan-Uwe Ness
2003-05-23