In checking out the distortion results in v0.7.0, I noticed resulting PSFs getting shifted in ways I was not expecting. These shifts depend on values specified in detector_position attribute. For the SW modules, I would have expected any PSF shifting to be relative to the center of the entire focal plane rather than the center of a given detector.
For instance, setting detector='A1' and detector_position=(1024,1024) shows a well-centered PSF, whereas values of (0,0), (0,2047), (2047,0), and (2047,2047) all show PSFs shifted diagonally by about 0.47". However, SWA and SWB consist of four detectors with the center of the field situated at their four adjoining corners. I would have expected that center of the field to be the preferential position where the PSF is "unshifted." This appears to be a consequence of the fact that the apply_distortion function uses the detector-specific aperture (e.g., NRCA1_FULL), so it's always using the center of the detector as the relative position to determine PSF shifting.
In checking out the distortion results in v0.7.0, I noticed resulting PSFs getting shifted in ways I was not expecting. These shifts depend on values specified in
detector_positionattribute. For the SW modules, I would have expected any PSF shifting to be relative to the center of the entire focal plane rather than the center of a given detector.For instance, setting
detector='A1'anddetector_position=(1024,1024)shows a well-centered PSF, whereas values of (0,0), (0,2047), (2047,0), and (2047,2047) all show PSFs shifted diagonally by about 0.47". However, SWA and SWB consist of four detectors with the center of the field situated at their four adjoining corners. I would have expected that center of the field to be the preferential position where the PSF is "unshifted." This appears to be a consequence of the fact that theapply_distortionfunction uses the detector-specific aperture (e.g.,NRCA1_FULL), so it's always using the center of the detector as the relative position to determine PSF shifting.