Notes about DECam_crosstalk, mkbiascor, and mkflatcor


  • The "minimal" call is: DECam_crosstalk <input_file> <output_file>. This only splits different CCDs into individual files, with no crosstalk corrections, overscanning, or trimming. Note, the default is to split into 70 files, everything but the guiding chips I presume.
  • -overscan is the "default" overscan setting, does line-by-line simple averaging
  • -overscansample <type>:
    • mean (default), median, mean with min/max values rejected. I confirmed that the overscan region is removed.
  • -overscanfunction <BinSize (max of 50)>
    • cubic spline, line-by-line (default), Legendre polynomial
  • -overscanorder <order (1-6), default=1>
  • -overscantrim <NCols (default=0, must be < 20)>
  • -maxhdunum <N (default=71, have not tested what happens if N>75, I did not see this throw an error in the code.>
  • -satmask ...Doing this takes place inside block if statement for crosstalk correction. Does this make sense? But it does happen before cross talk correction.
  • -xtalk <xtalk_file> ...This comes before overscanning


  • -median(default), -average, -avsigclip <Nsigma>, -avminmaxclip <N_upper/lower>
  • -overscan <type> ...Calls a different overscan subroutine than OverScan in DECam_crosstalk. Fewer options are available: median, 3rd order polynomial, 2nd order polynomial, 3rd order spline, 2nd order spline, mean. All of these are line-by-line.
  • -image_compare <another_bias> ...Reports how many pixels between frames are "significantly different". According to Robert this needs to be verified.
  • There are no bad pixel masks here. Is that usual?


  • Same defaults, overscan, and image_compare options as mkbiascor
  • -bias <bias_file>
  • -bpm <bpm_file>
  • -pupil <pupil_ghost_file>
  • -scale <scale_region> ...Something about the scale (median) of each image. I don't fully understand this.
  • -noscale
  • -flattype <flatcor or tflatcor (twilight)>