Cross-dispersed grism data sets are sky subtracted, and the object images combined, using the redxspec task by unsetting the linear flag, setting the combine flag, and supplying values for the spectrum, obstype, and zerosection parameters. spectrum is the base name of the combined spectrum files. obstype defines the type of observation in the dataset. obstype=abba indicates that the object was nodded along the slit while frames were recorded in an ABBA sequence. Differencing each AB pair produces positive and negative spectra in the object image that are averaged and output to a file named by appending .oim to the base spectrum name. `A beam' and `B beam' sky images are also formed by averaging the B and A beam object frames, respectively, and these images are output to files named by appending .aim and .bim, respectively, to the base name. Refer to Figure 33 to trace the file name conventions used. The zerosection parameter specifies an image section that is used to sample the inter-order background level in these sky images. This DC level is subtracted from the sky images. Sky images are carried throughout the reduction and can be used to define the illumination profile of the spectrograph slit and to improve the wavelength calibration on long exposure frames. If this is to be done, it is advisable to record DARK frames of the same duration close in time to the sky frames. obstype=soos indicates that the dataset consists of sequences of sky, object, object, sky images, where the sky images are recorded with the slit positioned off the object. obstype=osso indicates that the dataset consists of sequences of object, sky, sky, object frames, where the sky frames are recorded with the slit positioned off the object. Averaged object and sky images are formed for both these observation types and are output to files named by appending .oim and .sim, respectively, to the base name. In both cases, the zerosection parameter specifies an image section that is used to subtract a DC level from the combined sky image. obstype=comparison indicates that the dataset is to be combined to form a comparison observation (e.g., an arc lamp image or a twilight sky image). All images in the dataset are averaged and the result is output to a file named by appending .cim to the base name. The obstype parameter is written to each output file header and is used in subsequent processing of the dataset. No allowance is made for drifts in the position of the object along the slit during the observations.
Figure 33: File name conventions used in the cross-dispersed grism reduction.
A typical redxspec parameter list for combining a single abba dataset in the list file tfiles is shown below. The sky subtracted and combined object image will be output to the file sp150.oim.
I R A F
Image Reduction and Analysis Facility
PACKAGE = caspir
TASK = redxspec
(images = @tfiles) List of CASPIR input images
(spectru= sp150) Base name of spectrum file
(linear = no) Linearize data?
(combine= yes) Combine individual 2D spectra?
(flatten= no) Divide by flatfield?
(fixbad = no) Fix known bad pixels?
(clean = no) Interactively clean additional pixels?
(transfo= no) Straighten orders and subset?
(illumin= no) Correct non-uniform slit illumination?
(extract= no) Extract 1D spectra?
(fluxcal= no) Flux calibrate spectra?
(merge = no) Merge spectrum segments?
(flatdiv= no) Divide by flat spectrum star?
(plot = no) Plot merged spectra?
(bias = bias) Bias frame to use
(dark = dark180) Dark frame to use
(obstype= abba) Type of observation made
(zerosec= [190:195,10:90]) Zero level image section
(flatfil= ) Flatfield frame to use
(badfile= caspirdir$grism) Bad pixel file
(badtype= interp) Type of bad pixel correction
(illfile= ) Illumination frame to use
(reffile= ) Comparison extraction reference file
(fluxspe= ) Flux calibration spectrum to use
(weights= ) Merge weighting function file to use
(flatspe= ) Flat spectrum star file to use
(verbose= yes) Verbose output?
(imglist= )
(mode = ql)