Before any processing on the input frame is done, the REDUCE/CCD command will first collect all resources needed for the calibration of the science frame(s). These include the master calibration frames, the overscan offset, and scaling parameters. So, at this point no operations are done yet. This is done for efficiency reasons: all standard calibration arithmetic on the input frame is done in one go. As an example, suppose the science frame is supposed to corrected for dark current and to be flat fielded. From the reduction table the command first identifies the name of the dark and flat frames, and checks for their existence. If they are not present they will be created. Next, the master dark and flat field will be checked on its processing status. If they have not been processed, that will first be do by by another (recursive) run of the REDUCE/CCD command. In this second run also the scaling factors ( i.e. exposure times and the mean of the flat field frame) will be determined.
The standard calibration operation is done by a large COMPUTE/IMAGE with the following input:
Here, out is the output frame, in is the input frame, scan is the overscan bias value or frame, biasfrm is the master bias, darkfrm and darkscale are the mast dark frame and scaling factor, and flatfrm and flatscale are the master flat field and the mean value of the flat field. In the COMPUTE biasfrm and darkfrm can also be constants.