Stellar models in IR calibration

For the astronomical community analyzing ISO-SWS data, a first point to assess when judging and qualifying the observational data concerns the flux calibration accuracy. Since the calibration process is not straightforward and since a wrong calibration may lead to an over- or underestimation of the results, knowledge on the full calibration process and on the still remaining calibration problems is crucial when processing the data. One way to detect calibration problems is by comparing observed data with theoretical predictions of a whole sample of calibration sources. By using an iterative process in which improvements on both the calibration and the theoretical modelling are involved, we will demonstrate that a consistent theoretical data-set of infrared spectra has been constructed. This data-set has been used to derive the OLP10 flux calibration of the ISO-SWS. We will shown that the relative flux calibration accuracy of (high-flux) ISO-SWS observations has reached a 2% level of accuracy in band 1 and that the flux calibration accuracy in band 2 has improved significantly with the introduction of the memory-effect corrrection and the use of our synthetic spectra for the flux calibration derivation, reaching a 6% level of accuracy in band 2.