Документ взят из кэша поисковой машины. Адрес
оригинального документа
: http://xmm.vilspa.esa.es/sas/7.1.0/doc/merge/node9.html
Дата изменения: Mon Jul 9 02:02:30 2007
Дата индексирования: Sat Dec 22 07:27:55 2007
Кодировка:
Поисковые слова: m5
Care should be taken in using any
output. Cases have been found whereby new output file attributes need to be
calculated on the basis of input file attributes, and no doubt, these cases
will continue to crop up. In many cases, care will be needed as regards
any subsequent use of a particular merged file.
The task is now able to correct both individual input event files for
small errors in attitude, prior to merging. Very thorough testing of this feature
has not yet been performed. Position angle is measured clockwise on the sky.
Problems can occur when one of the input files contains GTI
extensions (not STDGTI) and the other one doesn't. Everything seems ok
but evselect can't handle this case and will calculate the wrong exposure
time for files extracted from the merged event file.
The GTI merge assumes that only one set of GTIs exist in each input
file. If more than one set exist then the final result will probably be
wrong.
In the event file merging, the BADPIX extensions are not merged.
It may be that the default imagesize may change, as merged
images are likely to cover a large area of the sky than single event file
images.
Out-of-time events modeling in task esplinemap
does not work
with merged eventsets.
When merging two files (C, A say) from different exposures, the two
files will be automatically ordered so that they are merged in time order
(AC). If one then wants to merge a third file (B) to the first two, this third
file data being from a time in-between the first two, then problems may
exist
with subsequent tasks performed on the resultant merged file (ACB) as the data
will not be time sorted. It is advised therefore to merge all data in time
order (i.e. first A with B, then AB with C).
It is advisable also, because of the large file sizes involved, the
time-consuming nature of the reprojections and the fact, in merging e.g.
four event files together, that three merge
runs are
necessary (A+B, AB+C, ABC+D, done preferably, remember, in time order, and
not for instance in file size order), that the event files are first
filtered to only the good events one is interested in.
