Overview of the Elixir pre-processing
1.1 Scope of the data pre-processing
The scope of the Elixir MegaCam data pre-processing is to provide the user with an image which has been fully corrected for the instrumental signature across the whole mosaic. This means:
The photometry must be uniform across the field of view: a given star or galaxy has to have the same flux in the final image no matter where it falls on the focal plane.
The image must look flat in the case of deep sky exposure (not some star forming region which has intrinsically some large scale diffuse structures).
The bad pixels and the overscan regions must be set to the 0 ADU value.
The initial dynamic of the data must be preserved (no truncation).
If possible, all CCDs in the mosaic are astrometrically and photometrically calibrated (this part is discussed in the "CFHT Data Calibration" section of this site).
The input raw data are 16 bits unsigned data ranging from 0 to 65535 ADUs. Elixir pre-processing generates an output also in 16 bits unsigned but with the BSCALE and BZERO parameters judiciously used to properly represent the initial dynamic of the data on a 16 bits range even after the data have been scaled up during the flat-fielding process. There is no reason indeed to go to 32 bits as the noise in the signal (read noise or sky photon noise) still largely dominates the error budget versus the bit truncation error.
http://www.cfht.hawaii.edu/Science/C...rocessing.html
1.1 Scope of the data pre-processing
The scope of the Elixir MegaCam data pre-processing is to provide the user with an image which has been fully corrected for the instrumental signature across the whole mosaic. This means:
The photometry must be uniform across the field of view: a given star or galaxy has to have the same flux in the final image no matter where it falls on the focal plane.
The image must look flat in the case of deep sky exposure (not some star forming region which has intrinsically some large scale diffuse structures).
The bad pixels and the overscan regions must be set to the 0 ADU value.
The initial dynamic of the data must be preserved (no truncation).
If possible, all CCDs in the mosaic are astrometrically and photometrically calibrated (this part is discussed in the "CFHT Data Calibration" section of this site).
The input raw data are 16 bits unsigned data ranging from 0 to 65535 ADUs. Elixir pre-processing generates an output also in 16 bits unsigned but with the BSCALE and BZERO parameters judiciously used to properly represent the initial dynamic of the data on a 16 bits range even after the data have been scaled up during the flat-fielding process. There is no reason indeed to go to 32 bits as the noise in the signal (read noise or sky photon noise) still largely dominates the error budget versus the bit truncation error.
http://www.cfht.hawaii.edu/Science/C...rocessing.html
Comment