How do I create a grid of spectra?

For an overview visual inspection of the spectra in a cube, it is useful to plot many spectra simultaneously, albeit at a lower resolution, in their respective spatial locations. While KAPPA provides the clinplot command to make such a grid, sometimes the sheer number of spatial pixels can make the spectra unreadable and will take some time to plot. Therefore the DATACUBE package offers the gridspec shell script. It has an option to average the spectra in the spatial domain, thereby reducing the number of spectra plotted by several times, generating more practical graphics quicker. See the figure. Below is an example. The -z option requests that the white-light image be shown and a subset of the cube selected with the cursor. The -b option sets the spatial blocking factor. Different factors may be given for $x$ and $y$, the two values separated by a comma.

   % gridspec -b 4 -i ifu_file -z

         Input NDF:
           File: ifu_file.sdf
         Shape:
           No. of dimensions: 3
           Dimension size(s): 59 x 110 x 961
           Pixel bounds     : 1:59, 1:110, 1:961
           Total pixels     : 6236890

         Collapsing:
           White-light image: 59 x 110

     Left click on lower zoom boundary.
     Left click on upper zoom boundary.

          Zooming:
          Extracting:
            Lower (X,Y): 20,61
            Upper (X,Y): 42,82

          Plotting:
            Clinplot: Grid of spectra, averaged 4 x 4

Figure: The gridspec script, operating on a subset of the 3C 27 observation consisting of the central core of the galaxy, averaging sixteen spectra in the cube for each spectrum plotted. The exterior axes indicate the average spatial co-ordinates of each block of averaged spectra.

A useful strategy is to select a blocking factor that gives no more than ten plots² along an axis. Then focus on regions of interest--you either supply an NDF section or select from the white-light image--by decreasing the blocking, and thus increasing both the spatial and spectral plotting resolutions.