EXTRACTION OF THE SPECTRA

• Principles
• Use
• Options description

• Principles

• For every lens (spatial element), for every wavelength, the optical parameters of the spectral stage of OASIS are used to find the CCD position of this wavelength in this spectrum. If the Optimal extraction is not selected, the CCD column pixels are then summed up over a total width of W_spec centered on this position. If it is selected, the effects of the inter-spectra pollutions, of the CCD pixel sampling of the cross-dispersion profile, as well as the varying S/N ratio of the five pixels across this profile are taken into account by :
  • Using a 3-gaussians fit of the focal reducer PSF.
  • Using a theoretical cross-dispersion profile obtained through a modelling of OASIS + telescope optics.
  • Adjusting its parameters on the actual cross-dispersion profile of the 30 central micropupils.
  • Obtaining the best overall fit to the actual cross-dipersion profile of this particular spectrum.
  • Computing the weighted sum of the W_spec pixels centered on this particular spectrum ridge at this particular wavelength. The weights are computed as 1 over the total signal variance over this pixel, using the photon noise of the fraction of the signal on this pixel (from the cross-dispersion profile obtained previously, integrated over the pixel) and the known CCD readout noise.
  The value obtained is taken as the value of the N^th spectrum for this wavelength, if N is the number of the lens involved.
  The user is given here the additional possibility to have the optimal extraction Handle inter-spectra pollution. It is of course strongly recommended to let this option ON, and this is the default. But this extraction algorithm is quite young for us; we got a limited background on the subject, and the "do not take care of pollutions" option is a relic of debugging work. It should disappear from subsequent releases of XOasis; but, just in case...
  
• All the [lambda,N_1...m] sets are arranged in a datacube, with two spatial dimensions (the coordinates of the N^th lens), and lambda, which will later be more finely calibrated.
• As there is only a single mask for all the exposures of one configuration (in the same run), usually obtained with the telescope in a zenithal position, one must take care of any slight displacements due to flexures or small uncertainties in the repositioning of the beam steerer used to shift the grism null deviation to the central wavelength of the filter. The program thus computes the precise offset between the mask and the object via a correlation between the calibration exposures associated with the continuum and with the object exposures respectively. It is therefore essential to select the right calibration exposure; and remember that the object and associated calibration exposures must have been obtained with the same telescope position without any change in the configuration (save the exposure duration...) between them. The first thing to check is that they bear consecutive exposure numbers ! If not, this must be explained in the Comments of the file header, to be read with the Reduction logbook function.
  When you run the spectra extraction, check that the computed offset between the mask and the object is reasonable; you should get a subpixel value in x (less than 1 pixel) and a few pixels in y (less than 20 pixels). Larger values indicate a problem...
  A warning is issued if the calibration frame and the object frame are separated by more than one hour. This may happen to the most well-intentioned astronomer, if the integration time on the object is one hour.
  



• Use

  • Input frame : the name of the preprocessed frame from which you want to extract the spectra. It may be an object exposure, or a GUMBALL continuum flat, or a sky flat, or a wavelength calibration exposure (neon, argon, Perot-Fabry, ...). You can type it in directly, or use the browse icon at the end of the field, or drag and drop it from the Reduction folder.
    
  
  
  • Input wavelength calibration frame : the name of the wavelength calibration exposure associated with the input frame; that means obtained in exactly the same TIGER configuration, same telescope position, obtained just before or just after, with no beamsteerer movement between the two. If the input frame is already a wavelength calibration exposure, it is associated with itself, and the same name is to be input in the two zones. Same input possibilities as above).
    
  
  
  • Extraction mask table : the name of the mask table built during the Mask creation reduction step. Same input possibilities as above.
    
  
  
  • Output datacube : the name of the file which will hold the tridimensional data. Do not specify any extension, a ".tig" will be added to the name you give. This is a special format, specific to TIGER data. Same input possibilities as above.
    
  
  

• Options description

  • Summation full width : the width across which pixel values are summed up to obtain the spectrum intensity at this point. Default is 5, it may be lowered in some special cases to get rid of any possible inter-spectra pollution; 7 is the extreme width possible, as this is the spectra separation on the CCD. It is usually wise to stay with 5, which has always given good results.
    
  
  
  • Optimal extraction : Checking this button switches the algorithm from brutal cross-dispersion summation to clever weighted summation (see principles for a brief summary). You must use that is you are afraid of possible inter-spectra pollution; they are usually at a very low level, due to the spacing of the spectra on the CCD, but may become noticeable in some cases (strong spatial gradients and huge emission lines). It is supposed, too, to improve slightly the local S/N ratio.
    
  
  
  • Handle inter-spectra pollution : This option is ON as a default, but may be deactivated for debugging purposes. See principles for details.
    
  
  
  • Debug : This switches the program to verbose mode, and more informations are recorded into the history file (see Getting started). In debug mode, you are allowed to extract a single spectrum, identified by the Lens number, to check the quality of the result, for instance regarding possible pollution by neighbours in a high gradient area.
    
  
  
  • Save values : All the input values (files names, coordinates) are saved, and become the new default values for this user. They can be recalled at will, and are used each time the Extract spectra window is opened.
    
  
  
  • Recall values : The values (files names, coordinates) saved by the user, are loaded to the various input fields.
    
  
  
  • Default values : The input fields are set to the general defaults values; for instance, the file names are set to blank.
    

Back to top of chapter.

Back to OASIS home page.