hi everyone,
two questions regarding calibration and integration of canon CR2 files:
because i have been lazy, i have not set a custom white balance in-camera for my clip-in CLS filter. however, dcraw seems to be able to figure out the white balance pretty well when told to ignore the camera's white balance info. for this reason i've been debayering the CR2 files using dcraw, and then calibrating, registering and integrating the files. i realize this is wrong - you are supposed to do the calibration on the raw, undebayered files. in order to do this i have to figure out my R,G,B scaling factors for the clip-in filter and fix the colors after integration. i will do this eventually.
when calibrating the debayered files, i've noticed that even though my dark temperature and exposure length are exactly equal to my lights, the dark scaling algorithm is coming up with scaling factors which are very far away from unity. it wants to scale the red channel by about 2x, and the green and blue channels by about 0.5x. is this just a symptom of having "done it wrong", or is it because of the MRS noise estimation failing on CR2 files? in other words, has dcraw totally screwed up the darks while debayering them because it has tried to compute a white balance on such "dark" data?
what will happen if i calibrate the un-debayered files? does the tool still try to compute scaling factors for each individual color in the bayer matrix, or is the image treated as greyscale? i'm worried that if somehow the scaling results in the debayered case are correct, the calibrated un-debayered files will be wrong since the red scaling seems to be ~4x the blue and green scaling.
thanks
