The purpose of this task is to take a time series and a GTI file, and align the GTIs with the time bin edges.
As initially conceived, the task was restricted in applicability: the only time series input it would accept was the EXPOSU extensions of an EPIC event list; the input GTI series was aligned separately with each of these EXPOSU time series, the respective results being then written to separate extensions of the output file. This behaviour has been retained, but new parameters were added in version 2.9 to allow the task to be applied to generic OGIP-compliant time series and GTI files.
The problem which gtialign is intended to solve is illustrated in the figure below.
| | | | | | |
| 1 | 2 | 3 | 4 | 5 | 6 |
| | | | | | |
___________________________________
| |
| |
_________________| |_____________________
time -->
The top line represents the time boundaries for a given time series. In the case of the EXPOSU extensions the time bins are actually the CCD readout frames: in this case each vertical line represents the end of one frame readout cycle and the beginning of another. The bottom line represents a GTI, as produced for example from housekeeping data. `Up' corresponds to a good time interval. Note that the time-series bins are not necessarily of constant duration.
Looking at time bin 2, we see that the GTI is valid only for part of this interval. To avoid uncertainty about the validity of bin 2, all of it should be excluded. Similarly, all events from bin 5 should be excluded as well. Thus the GTI should be adjusted as follows:
| | | | | | |
| 1 | 2 | 3 | 4 | 5 | 6 |
| | | | | | |
____________________
| |
| |
________________________| |_____________________________
time -->
