I think you are right: When scanning negatives with the Coolscan scanners
the backend tells the scanner it is doing a negatives and the
scanner adapts its exposure range to the right levels for negatives.
That's why you don't get the same image when scanning a negative
using "positive" as a backend-option and inverting the image yourself.
I havn't tested the quality of this exposure adaptation throughout
a significant image base of negatives yet, but I didn't have any problems with
it either.
Perhabs this is different for the HP-Photosmart?
So without doing multiple scanning there is no way of getting more
than 8-Bit out of the LS-20, 10 bit out of the LS-30 or 12 bit out of the
LS-2000 and there doesn't seem to be much neccessity to change
exposure values to get these 8/10/12 bit.
Can the other users of these Scanners confirm this hypothesis?
Apart from that I still think it is a neat idea to have a
calibrated system where you can figure out the optical density
from the RGB values (but I know it is difficult to achieve).
When scanning color images, optical density allone is not enough,
we need to know what color (wavelength) was used to do the
scanning and what is the detector efficiency as a function of
wavelength.
We may need to say what is "standard Red" what is Green and what is Blue
and try to do a color space conversion from the scanners
color to the standard space.
I don't know too much about the subject, but I think there must be
some standards allready defined that we can use.
Who does know more about this.
Andreas
-- Source code, list archive, and docs: http://www.mostang.com/sane/ To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com