XMM-Newton Science Analysis System: User Guide

4.4 Filtering calibrated EPIC event lists

EPIC calibrated event lists must be processed to generate data products including images, spectra and rate curves.

All product creations can be accomplished using the evselect task or (via a user friendly Graphical User Interface; GUI) the xmmselect task. The event list is filtered on the fly according to user-specified selection criteria. The filtered event list can be stored on disk and/or products can be extracted from this filtered event list.

The filtering process is carried out on an event-by-event basis controlled by user-specified selection criteria. These criteria take the form of character strings which can be composed of a variety of elements including numerical and logical constants, operators, functions, and attributes. Arithmetic and logical operations, file-based filtering, new column or arrays construction and symbolic reference to other arrays can be performed.

Events for which the selection expression does not evaluate to true will be marked as invalid or discarded from the data set and shall not be considered in the product extraction. xmmselect (and evselect) support selections based on intrinsic event attributes falling into the following sub-categories.

1. Spatial selections in raw pixel (RAWX/RAWY), camera coordinates (DETX/DETY) or sky pixel (X/Y) space, or in any spatial coordinate system. Special pre-defined region shapes exists such as CIRCLE, ELLIPSE, ANNULUS, BOX, POLYGON. Spatial filtering with mask images or region files is possible as well.

2. Energy selections in the form of one or more interval specifications in either PHA or PI space.

3. Time selections in the form of interval specifications or Good Time Interval (GTI) files created with e.g. tabgtigen.

4. Event selections based on informational and rejection attributes (see evatt package description for details).

5. Event selections based on any other event properties (e.g. PATTERN or CCDNR).

The user should consult the selectlib package description which details the syntax and semantics of the expressions driving the above operations and gives sample expressions to demonstrate typical usage scenarios.

Note, while all of the data products generation can be done on the original event list, it can save significant computing time and memory first to create a filtered event list (e.g. removing high flaring background periods, restricting the analysis to certain regions and energies) and operate after-wards on the reduced filtered event file.