group

 GROUP N1 N2 N3 NBIN METHOD WEIGHTS -- Combines spectra in groups, effectively giving 
                                       higher signal, lower time resolution data.

 GROUP is useful if you have so many spectra that everything is slow and
 plots are so dense that you can't make anything out in them. I added it
 bceause of some HST spectra (12,000 of them).

 Parameters: 
     
  N1    -- Slot number of first spectrum to average
  N2    -- Slot number of last spectrum to average
  N3    -- Slot in which to start 
  NBIN  -- Number of spectra per group
  METHOD - Averaging method. At present supports
           U  -- Unit weights
           D  -- Dwell weights
           R  -- Response weights
           V  -- Variance weights

  WEIGHTS   -- file of weights. Whatever method is chosen, the weights can
               be multiplied by individual weights first. <CR> to ignore
               These can, for example, allow you to weight spectra with
               a certain spin phase or whatever and provide a more flexible
               method than straight selection.


 Response weights are proportional to the local ratio of COUNTS/FLUX
 and thus effectively add the counts together (optimum for photon
 counting data). Variance weights use a single value equal to the
 mean Signal/noise**2 for each spectrum.

 This differs slightly from AVERAGE and PBIN in that an attempt is made to 
 simulate the effect of just having made a longer exposure. Thus compared
 to the other methods, the counts and dwell are multiplied by the number
 of spectra in a bin. This works for response weights. The flux arrays
 are the same in any method.

 Selection is done. Spectra must have same length.

This command belongs to the classes: arithematic and manipulation .