Loving ISIS - Confessions of a Former XSPEC User

Loading the Data

The basic ISIS function for loading data is:

     id = load_data("filename");

This function will read the header and then load associated RMF, ARF, and background files, as well as any systematic errors defined in the file. The data, RMF, and ARF will each be assigned unique ID values.

The information for these various files can be seen using the following list functions:

     list_data; list_rmf; list_arf;

To make these list functions as verbose as possible, in my .isisrc startup file I have set:

     Isis_List_Filenames=1;

For convenience I also have taken the above three list functions and placed them in a single function: list_all;

Additionally, my .isisrc file contains two defined functions:

radio = read_radio("filename",#);

id = load_radio(radio);

These take an ASCII file (consisting of frequency in Hz, radio flux and radio flux errors in mJy), create effectively very narrow bins around the frequency points of interest, convert the radio fluxes and errors to counts per bin, and assign them to an ISIS data set. If no RMF and ARF are assigned, ISIS assumes a diagonal response with 1 cm^2 area and 1 sec integration time.

Note that, as we have mentioned, all of the above functions assign ID numbers to the datasets, responses, and ARFs. Owing to the fact that the radio has neither an RMF nor ARF, and that the PCA data only has a combined RMF/ARF, the dataset ID numbers are not identical to the RMF and ARF ID numbers. There are ISIS get_* functions to retrieve all of this information, or one can use the list_all; function.

Thus, for this example, the data loading sequence was:

     radio_data = read_radio("radio.data",50);
     radio_id = load_radio(radio_data);

     pca_id = load_data("pca.pha");

     hexte_id = load_data("hxt.pha");

Use of the list_all; function then produces:

Current Spectrum List:
 id    instrument   m prt src    use/nbins   A   R     totcts   exp(ksec)  target
  1                 0  0   0       2/    3   -   -  1.1119e+00     0.001
  2           PCA   0  0   0     129/  129   -   1  7.5787e+06    10.528  GX_339-4
sys_err:  on
file:  pca.pha
back:  pcaback.pha
  3         HEXTE   0  0   0     256/  256   1   3  8.5617e+05     5.883
file:  hxt.pha
back:  hxtback.pha


Current RMF List:
 id grating detector    m type   file
  1             PCU0    0 file:  pcaresp.rmf
  2              PWA    0 file:  hxt.rmf


Current ARF List:
 id grating detector   m prt src   nbins  exp(ksec)  target
  1            HEXTE   0  0   0      970     5.88
file:  hxt.arf

With four command lines, we have loaded three data sets, including a radio data set from and ASCII file. The radio data can be plotted and fit simultaneously with the X-ray data set, with the X-ray portion of the fit being properly carried out in 'detector space'. (I.e., no unfolding will be involved in fitting the X-ray data.)

Important Note: You might see ISIS throw off a number of warning messages concerning the response files. Unfortunately, it is sadly common for software from many different satellite missions to create responses that are not fully compliant with accepted FITS standards. Without naming names, such problems as negative energy bin boundaries, overlapping bin boundaries, non-sequentially spaced energy bins, etc., have cropped up in various data set response files. Equally unfortunately, it has been the practice of XSPEC to hide this from the user by quietly making internal corrections to the responses.Whereas many of these response issues ultimately have been harmless, the fact that XSPEC silently fixes (and/or ignores) them has removed a lot of the impetus for the software teams to "get it right" in the first place.

ISIS will do some corrections, and then warn you. It is not as tolerant of bad responses as is XSPEC. ISIS will fail to read especially poorly written response files. For these cases, one can "fix by hand" the affected response (e.g., by using FTOOLS or DMTOOLS), or one can complain to the mission that created the response software.

Next up: Grouping Data

This page was last updated Sep 16, 2013 by Michael Nowak. To comment on it or the material presented here, send email to mnowak@space.mit.edu.