On Wed, 28 Mar 2001 17:20:11 -0800 shAf (michael@shaffer.net) wrote:
> ... Tony seems to be
> under the impression, for those scanners which have been chracterized,
> Vuescan will transform the scanned RGB data into "device RGB".
The raw scan is in an unspecified device space, scanner RGB. Ed's transform,
applied during the production of the Crop file, munges that against his
characterisation and the result is a scan with altered data values within
Vuescan's working space (which I previously said I thought was maybe sRGB, but
as has been pointed out it ain't, it's Kodak's PCD space - too many facts, too
little brain:).
This isn't classical ICC-type colour management (=don't change the data, just
append a tag which provides a map for interpretation of device colour). Vuescan
does change the data, so you can't go backwards - the original scanner RGB has
gone. However you do now have a scan in a known colour space (PCD) which can be
mapped and tagged to any other colour space, sRGB, AdobeRGB etc. It's perfectly
legitimate, nothing is broken - but you should regard VS internal working space
as the device space here. So you don't want to go applying other profiles for
the scanner, from other sources, at that stage, as the image is no longer in
scanner RGB.
>(Tony
> ... correct me if I'm wrong ... I think this is what your
> 'step-by-step' Vuescan method implied. This implimentation of "device
> RGB" makes me itchy, because while it is in Ed's evalutated "device
> RGB" space, it is NOT in the same RGB space as implied by a
> manufacturer supplied, or 3rd party calibration, device color space.
Quite. The whole point of VS is to avoid all that - either because VS does it
better (=design aim) or because the OE scanner software doesn't use ICC at
all.
> To impose (assign) one on top of the other makes me uncomfortable ...
It's an either/or proposition; you should never be applying scanner profiles to
the Crop file. If you want VS to produce legit, tagged files, corrected for the
scanner's deficiencies (and film characterisations too),
use the Crop file. If not, use the raw file, or other software, and DIY.
> I certainly am more comfortable with the scanned image inheriting the
> device space because nothing was done to it (... not implying the
> 'raw' scan' because we are still trying to use Vuescan's cropping
> tools ...) ...)
But VS doesn't implement CM 'properly' - if it did, Ed would have to bundle
his scanner characterisations as ICC profiles and he declines to do that as
everybody would pinch them and not buy VS. For the same reason, you'll find
vanishingly-few PD profiles on the web, and people like Jon Cone making a
living by selling them (for Epson piezos).
The downside is that this approach locks VS out of using any device profiles
you might make yourself for your personal scanner/film combination, to
produce the Crop file. However VS still gives you the wherewithawl to
create and save a raw scan which you can then work with in PS, using
device profiles or anything you like.
Regards
Tony Sleep
- Online portfolio & exhibit; + film scanner info &
comparisons
