OM grism data are complex, and a grism image can contain a lot of sources, zero and first spectral orders, in addition to complicated straylight features affecting the background and the spectra. This makes the automatic extraction process very difficult and prone to spurious detections and extractions. The user must check carefully the produced spectra. The source list of extracted spectra produced by omgrism is also represented in a ds9 type region file (SPCREG) which can be overlaid on the rotated image to verify the correct extraction for the target spectrum or for all spectra in the field.
The SPECLI file tells us whether each extracted spectrum corresponds to the main target, or to a default extraction, or to a field object. The keyword OBJ_TYPE flags this also in the SPECTR file.
In case of problems or doubts, it is recommended to run the interactive omgsource after omgprep. This task uses as input the rotated image already mentioned.
The described rotated image RIMAGE contains all corrections proper to OM grism data. Therefore it can be used within any spectral reduction package or customised reduction system to perform the extraction of the spectrum, which can afterwards be calibrated using the contents of OM_GRISMCAL CCF.
It should be noted that coincidence loss corrections are currently not applied to grism data. Also, there is no correction for flux in the wings of the cross dispersion profile if the user selects interactively an extraction swath narrower than the base width of the spectrum.
Astrometric corrections are applied in grism data. The coordinates of the targets in the field of view corresponding to all extracted spectra (extractfieldspectra=yes) are written in the spectra source list and in the headers of the spectra files. It should be noted that these coordinates are less accurate than the ones obtained in normal image data using the lenticular filters (UVW2, UVM2, UVW1, U, B and V).