The final step revisited

Given `calib' as the calibration spectrum, no matter how it was generated, it is applied as follows:

   ICL> spflux spectrum=obs calspect=calib output=calobs

which calibrates a spectrum called `obs', creating a resulting spectrum called `calobs'.

At the risk of being obvious, you should be aware of what you have created here. Each element of `calobs' is a sample at a particular wavelength of the continuous flux density function; it is not a measure of the total flux within a wavelength bin of finite width, which the original spectra were. Apart from anything else, adding two such spectra has the magic (and totally unreasonable) effect of doubling the magnitude of the object! Generally, this is not something to worry too much about, but it is disconcerting to have a spectrum like this generated with a non-linear wavelength scale. (One has a gut feeling that if the bin covers half the wavelength it should have half the value, and that is not true for these spectra.) That is the reason for the warning about not using unscrunched data.