next up previous contents
Next: 8.5.4 Multiple Gaussian Extractions Up: 8.5 Low Dispersion Extraction Previous: 8.5.2 Rectangular Extractions

8.5.3 Gaussian Extractions

 A series of procedures exists for re-extracting fluxes from the low dispersion line-by-line files using a Gaussian fitting technique. This technique can improve the signal-to-noise in the continua of weak spectra, over that given by the current IUESIPS extraction. When the technique is applied to high signal-to-noise spectra, the results agree with IUESIPS to within a few percent. These programs use a symmetric Gaussian approximation to the point spread function (PSF) perpendicular to the dispersion direction. Tests with asymmetric Gaussian PSFs indicate that the IUEDAC software can result in an error on the order of 2 percent compared to the more complex PSF.

Basically, a three step process is used in which the location and width of the spectrum is found as a function of wavelength, and then, with the spectrum located, the flux at each wavelength is found by fitting the observed data with a Gaussian of specified width and centroid. The data may then be calibrated. The procedures are:

GEX1,IMAGET,WMIN,WMAX,DELS1,wave,width,center,centord, $
/silent,aperture=aperture  

GEX2,IMAGET,WMIN,WMAX,DELS2,w,f,err,q,/silent,/fwhm, $
/noplots,outname=outname 

GEXCAL,OUTNAME,h,w,f,err,q,file=file,/noplots,row=row, $
exps=exps,center=center,thda=thda  

Please use DOC_LIBRARY for more information about these programs. Rather than running these programs separately, you may run PREGEX. This allows you to execute the GEX procedures in batch mode (on VMS and UNIX systems). It writes a script file with the above commands in it for each line-by-line file you wish to process:

PREGEX,FILE,wmin,wmax,dels1,dels2,center,/nocal, $
aperture=aperture,outdir=outdir,login=login,/prompt
 

FILE This can either be a vector of names of files to process, a search string for the ``findfile'' command, or a dummy variable (in which case you will be prompted for file names). PREGEX will process lists of files with one command.
wmin Minimum wavelength of region to be extracted. If a wavelength range is not given via wmin and wmax, defaults covering much of spectrum will be used.
wmax Maximum wavelength of region to be extracted.
dels1 Bin size for GEX1 (which determines the location and width of the spectrum). If not given the default value is ten.
dels2 Bin size for GEX2 (which determines the flux at each wavelength). If not given the default value is one.
center Central order for GEX1. If not given, defaults appropriate to the type of line-by-line file will be used. See Table 8.2 for a list of default central lines for various line-by-line file types.
nocal If set, GEX1 and GEX2 will be executed but GEXCAL will not. Note that for IUESIPS data certain items (like the exposure time) must be set ahead of time in order for GEXCAL to execute. Thus, if the NOCAL keyword is not set, PREGEX will attempt to run HFIX on each IUESIPS file.
aperture Since NEWSIPS line-by-line files include both large and small aperture data, the user must specify which aperture to extract. If this keyword is not set to ``L'' or ``S'' (large or small), the user will be prompted. This is ignored for single aperture spectra and all IUESIPS spectra.
outdir Normally, PREGEX will assume the output files should be written to the current directory. The user can override this by specifying a different directory via this keyword.
login In order to run the gex routines in batch mode, PREGEX must write a script file in the user's login directory. If you have commands in your login.com file that change directories, you need to enter that directory via this keyword. Otherwise the gex programs will search fruitlessly in sys$login for the script file. This keyword is for VMS systems only.
prompt If set, PREGEX will prompt the user for any information it needs. Normally, if something was not entered via the calling sequence (or something that was entered was unsuitable), PREGEX uses default values.

The simplest way to use PREGEX is to simply type: pregex,0 and let the program prompt you for the file names. If you elect to run in batch mode you then need to wait for the job to run. If you wish to run the routines interactively (e.g., if your system doesn't support a batch mode), you may simply type @gex.txt at the IUE_IDL> prompt. Output files may be read with IFITSRD (see example in section 8.5.1. To run the job in batch mode at a later time, you will have to submit the file !iuer.exe:gex.bat with a command appropriate for your operating system. On a VMS system the command would be:

IUE_IDL> spawn,'submit/notify '+!iuer.exe+gex.bat

Don't forget that on VMS systems gex.txt must reside in sys$login or the batch job will not be able to find it!


next up previous contents
Next: 8.5.4 Multiple Gaussian Extractions Up: 8.5 Low Dispersion Extraction Previous: 8.5.2 Rectangular Extractions

2/9/1998