I’m referring to the step, in ISIS, where a Miles (or other) standard, is applied to the reference. Assuming the Miles reference is atmosphere corrected, shouldn’t our standard also be corrected before it is divided by the Miles?
Edit: Or to put it another way, as I understand it, the sequence in ISIS is:
Creation of IR:
1. Process the IR spectra with AOD configured and enabled.
2. Create an IR curve of the processed spectra using a published reference and the continuum shaping tools in ISIS.
Processing target spectra:
1. Enable the AOD function (assuming you have a web connection and have correctly specified object name).
2. Specify the atmosphere corrected IR curve (from above).
ISIS will then remove the atmospheric effects before applying the IR correction.
The alternative method is to take reference spectra of a nearby star at every run. If a target and reference star are very close together then they will be equally affected by any atmospheric effects (attenuation, pollution) and as one is divided by the other those effects are self cancelling. The IR is processed as above (ie, with AOD enabled in order to correctly compare to a published reference) and then applied to the target with AOD disabled during processing.