Dans la même rubrique :
AMBER Data Reduction Software
AMBER bad pixel maps (BPMs) and flat field maps (FFMs)
AMBER image interferences report
Spectral Shifts Solved by ATF Run
13 octobre 2008, par
People wanting to reduce AMBER data are warned that the calibrations of the spectral displacement between the photometric "channels" and the interferometry, as they are implemented today version 2 of amdlib (version also used in the ESO pipeline), can be very wrong depending on the spectral setup and the S/N of the calibration. At least some observations taken in the last years have been made with insufficient S/N in the spectral calibration files (taken at the beginning of the P2VM), and the data reduction will be unable to assess for the spectral displacement based on these calibrations files alone.
The ATF have shown in their report that in fact this displacement does not appear to have changed at all between the openings of the spectrograph dewar.
Additional work by Le Bouquin and Millour enabled to precise the shifts for all periods and all spectral configurations. The new amdlib version 3 software by JMMC automatically corrects for the good spectral displacement.
The following Table gives these values as a function of spectral mode and julian date. Note that observations made beween September 2007 and January 2008 without FINITO in LR could be particularly affected by the optical setup. See ATF report for more details.
| MODE || MJD < 54352 || 54352 < MJD < 54526 || MJD >= 54526 |
||3, 3, 1
||3.4, 3.4, -0.2
||4, 4, -1|
||0.873, 2.064, 0.472
||2.000, 2.576, -0.263
||3.008, 2.850, -1.319|
||-1, 1, 0
||1, 2, -2
||0, 2, -2|
Although the values quoted above are precise, there is a caveat : a subpixel change is still possible depending on the variations of the beam shape. It is recommended however to use integer values (say, +3,+3,-1 ) wich should suffice in most cases.