Bug #1815
Diffuse isotropic model pull distribution looks skewed
| Status: | Closed | Start date: | 07/12/2016 | |
|---|---|---|---|---|
| Priority: | Normal | Due date: | ||
| Assigned To: |  Knödlseder Jürgen | % Done: | 100% | |
| Category: | - | |||
| Target version: | 1.1.0 | |||
| Duration: |
Description
Could this be due to the change in the effective area class?
Recurrence
No recurrence.
History
#1
Updated by Knödlseder Jürgen over 8 years ago
- Status changed from New to In Progress
Removing the effective area limitation recovers the original pull factors. This explains at least the effect. Still need to correct for it.
#2
Updated by Knödlseder Jürgen over 8 years ago
- Status changed from In Progress to Feedback
- % Done changed from 0 to 100
It turned out that the problem came from a mismatch in the Monte Carlo and function evaluation in
GCTABackground 3D. There were also subtle issues related to the solid angle computation and the Monte Carlo sampling. Now all issues have been fixed:
- the function evaluation and Monte Carlo sampling is restricted to the validity range of the DETX-DETY plane
- an spectral oversampling (factor of 10) has been added to the spectrum computation
- the Monte Carlo method now draws uniform positions within a circle that encloses the valid DETX and DETY values, and the DETX and DETY values are computed from theta and phi; DETX and DETY values outside the validity range of the DETX-DETY plane are excluded
- the class was corrected so that it works on any ordering of the information in the response cube
- the solid angle of all background rate pixels are now computed by dividing each pixel into 8 wedges (similar to
GSkyMap)
Merged into devel and ctools science verification of isotropic model is now successful. Still need to do a full science verification run.
#3
Updated by Knödlseder Jürgen over 8 years ago
- Status changed from Feedback to Closed
Science verification successful, close the issue now