Документ взят из кэша поисковой машины. Адрес оригинального документа : http://xmm.vilspa.esa.es/sas/7.0.0/doc/merge/node9.html
Дата изменения: Wed Jun 28 23:25:43 2006
Дата индексирования: Sat Dec 22 07:25:38 2007
Кодировка:
Поисковые слова: внешние планеты

XMM-Newton Science Analysis System


merge (merge-0.12) [xmmsas_20060628_1801-7.0.0]

Meta Index / Home Page

Comments

• 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.

• As this task is in its initial stages, 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.

• Input is welcome as regards whether it is sensible to merge certain types of files, with certain keyword setups. It would be nice for the task to be as unrestricted as possible, though in many cases, care will be needed as regards any subsequent use of a particular merged file.

• 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.

• 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.

• See future developments.