Документ взят из кэша поисковой машины. Адрес оригинального документа : http://xmm.vilspa.esa.es/sas/7.1.0/documentation/releasenotes/xmmsas_6.1.0_OMGrismInfo.txt Дата изменения: Mon Nov 22 18:06:14 2004 Дата индексирования: Sat Dec 22 15:22:12 2007 Кодировка:

SAS 6.1 AND OM GRISM DATA PROCESSING
------------------------------------

Processing of OM grism data has changed substantially in this SAS 6.1 release.

Although the methodology is exactly the same than in the previous version, the
output products have been simplified and they are now more easily usable and
understandable.

The complete processing chain extracts now by default only the spectrum of the
main target, or object placed at the XMM-Newton boresight, and this is what the
user will find in the final products. Alternatively, the user may select to
extract all objects in the field of view (extractfieldspectra = yes).

omgchain runs in a similar way than before:

1. In case of full frame low resolution (the observing configuration used for
spectroscopy of all objects in the f.o.v.) the four image files forming the
entire field are combined using omcomb. The resulting image file, or the
raw image corresponding to the rectangular window encompassing the main
spectrum, if single target configuration was selected when observing, is
copied with additional header information (omprep). The resulting image is
an intermediate file called g*IMAGE*.FIT

2. The prepared raw image is corrected from modulo_8 fixed pattern noise using
ommodmap. Then, the geometrical distortion introduced by the OM detector is
corrected by using the distortion map, available through OM_ASTROMET CCF.
The undistorted image is rotated so as to align the dispersion direction
with image columns. All this is performed by omgprep, and the result, called
p*RIMAGE*.FIT is the first final product of the chain.

3. A detection algorithm (omdetect) is used to search for all spectra, zero and
first orders, in the rotated image. The detections are written in a ds9
region type file p*REGION*.ASC and also in the file p*SWSRLI*.FIT.

Then omgrism searches through these detections for the spectrum of the main
target (default mode). If a reliable solution is found (zero and related
first order) then the spectrum is extracted and calibrated in wavelength
and flux using the OM_GRISMCAL CCF. If not, then the extraction is performed
using default nominal values for the positions of the zero and first orders.
The results are written into the files:

- p*SPCREG*.ASC a region type file depicting the positions of the zero and
the extracted first order
- p*SPECLI*.FIT the fits file corresponding to the extracted spectrum
- p*SPECTR*.FIT contains the extracted spectrum: wavelength, net count rate,
background, flux and corresponding errors.

The spectral extraction mode is by default of box car type. Some optimal
(signal weighted) extraction algorithms are optional although its use is not
recommended for the moment.

If extractfieldspectra option was used, then the detections performed in the
whole image are correlated to assign to each zero order its corresponding first
order. All successful relations are extracted, calibrated and written to the
described files that will contain now all spectra in the f.o.v. properly
cross-identified in the *SPECLI*, *SPECTR* and *SPCREG* files.

Finally, the extracted spectra are plotted with omgrismplot in PS or PDF
format.


Using OM grism extracted data
-----------------------------

As we pointed out in the previous SAS version, OM grism data can be very
complex. Overlapping zero and first orders of other field objects can
contaminate our object(s). This can now be very easily verified by displaying
the rotated image *RIMAGE*.FIT with ds9, and over-imposing the *SPCREG*.ASC,
which will tell us how each spectrum was extracted.

Also the *SPECLI* file tells us whether each extracted spectrum corresponds to
the main target, or to a default extraction, or to a field object. The keyword
OBJ_TYPE flags this also in the *SPECTR* file

In case of failure or doubt the interactive task omgsource will allow the user
to fully control the extraction regions. This task uses as input the rotated
image already mentioned.

The described rotated image *RIMAGE* contains all corrections proper to OM
grism data. Therefore it can be used within any spectral reduction package or
customized reduction system to perform the extraction of the spectrum, which
can afterwards be calibrated using the contents of OM_GRISMCAL CCF.

It should be noted that no astrometry is performed for the moment in grism
data. The coordinates appearing in the fits headers are the S/C pointing
coordinates, even in case of extractfieldspectra. The next SAS release will
include the determination of the coordinates of all objects whose spectra are
extracted.