Flat field creation is a simple procedure. You should already have loaded the relevant file names for red and blue flats into the proper fields, and set the red/blue transition row in the Preferences pane. If you're not sure what this last sentence means, refer back to the "setups" section of this wiki for a refresher.
Then, click "Make Flat". The code will first combine the red and blue flat fields in 2D space, using a smooth weighting function to spread the transition over a span of ~150 pixels. After masking out pixels not in the slit, a bspline is used to fit out the wavelength dependence of the flat field, taking the slit curvature into account (this is why arc solutions must be generated before flat fields). The map model is then divided out in 2D, leaving a ration image centered about 1.00 with departures indicating either bad pixels or gain variations. The final flat is presented for viewing if requested in the Preferences pane:
If you do not like the transition region, you can change the transition row and try again. Otherwise you are finished and ready to move on to object finding.
NOTE: at present the code does not support co-added sequences of flats. We will add this very soon, there is no real technical challenge, only there are other higher priority items. In the meantime if you absolutely must have combined flats, we recommend you use IDL or another utility to combine your flats in 2D and save out to a temporary file in your Raw directory. Then load this temporary file into the Red or Blue flat box and proceed as before.
The image below shows an example flat field. The zero pixels are outside the slit boundaries; inside the boundaries look for pixel values near 1.0 +/- 0.05 or so, except for the occasional dark pixel.
