Previous: Spectrum extraction and calibration
Up: Spectrum extraction and calibration
Next: Arc subpackage
Previous Page: Spectrum extraction and calibration
Next Page: Arc subpackage

Flatfielding

In order to take out small scale variations in the sensitivity for the dispersed image, it is necessary to flat-field the data using a white light image. A routine MAKEFF exists for the purpose of generating the necessary flat-field, which can then be divided into the data image.

This routine works by dividing each row within a box by the marginal distribution obtained by summing along the columns in that box. This takes out the variation of sensitivity with wavelength but does not account for features which are constant in wavelength for all positions along the slit. If it is believed that genuine features of this type exist, it will be necessary to smooth the marginal distribution before dividing it into the data. This technique will artificially flatten data in boxes affected by vignetting. This could affect the flux calibration if the object is not in the centre of the box. The original image should have uniform illumination of each slit and a signal level high enough to avoid injecting noise into the flattened data.