Prerequisites

Once you have installed the software, you also need

1. A batch of spectra to reduce

The spectra do not have to be normalized, but do have to have cosmic rays removed. The files are ascii files, each of which consists of whitespace-delimited columns (no headers) containing wavelength (in angstroms), flux (arbitrary), and noise (same units as flux), like as follows. Here is an example.

These files have the same format whether you are intending on generating a calibration from them, or applying a calibration to them. If your files are of empirical spectra (and not synthetic), we recommend that you restrict the wavelength range from 3911 to 4950 angstroms. (We found that this resulted in the best normalization, especially given the denser lines at the Ca II K end.)

2. A list of the file names of those spectra.

This is a comma-delimited ascii file. If you are applying a metallicity calibration to spectra, the header includes the original spectrum file name, and additional quantities: subtype, phase, literature [Fe/H], etc. Here is an example. (If you don’t know the quantities other than the original spectrum file name, just leave them blank, but do include the commas ,.)

If you are generating a new calibration based on synthetic spectra, the list of input spectra also needs to include the parameters used to generate those spectra, like [Fe/H], Teff, etc. Here is an example.

3. If a new calibration is being generated: a list of [Fe/H] values based on high-resolution spectroscopy, and low-resolution spectra of the same stars.

This will enable the removal of a systematic offset in ‘retrieved’ vs. ‘true’ [Fe/H] values. Here is an example of a list of [Fe/H] values of a basis set of RR Lyrae stars, which are composite values from several literature high-resolution spectroscopic studies. (This particular list was generated by the script make_high_res_feh_basis.py in the repo, though this script is very specific to this particular basis set and the way the literature data was organized.)