Re: filmscanners: Vuescan color spaces (long)

["shAf" <michael@shaffer.net>]

Mark writes ...

> From: "shAf" <michael@shaffer.net> To: <filmscanners@halftone.co.uk>
> Sent: Monday, April 02, 2001 4:45 PM
>
> > ...  This is
> > speculation on my part, but I don't believe Ed uses a "conversion
> > engine", rather a mathematical matrix transform.  ...
>
> Micheal,
>
> It is of no use to speculate, better ask Ed if you want to know.
Vuescan
> does something to the raw scanner RGB to convert it to the known
color
> spaces like Adobe RGB. It is probably not the usable ICC Color
Management
> procedure. If it works its oke. If it is not Vuesan always have the
"Device
> RGB" option. Thats why this option can be important.

    Fair enough.  Ed informs me "device RGB" isn't really when the
medium is a negative.  Therefore, "device RGB" is truly scanner space
only when the media is a positive.  This news naturally begged the
question, "when its a scan of a negative, then what is 'device RGB'"?
Is it at all useful??  Unfortunately, Ed didn't instill too much
confidence in what he thought "negative/device RGB" would be.  Quoting
...

        If the scanner's device color space is close
        to sRGB, then you'll getsRGB out the other end.
        The farther the device color space is from sRGB,
        the more awful the resulting image will look.

    ... Therefore without a tag or some other knowledge, its
usefulness would certainly require you find a color space which would
match.
    My speculation can only be applied to the Nikon LS-2000 (possibly
the LS-30), for which Ektaspace is a very good beginning, but if it
isn't for your "device RGB", I might have a better match, which should
be considered a better approximation only, and only for the Nikons
mentioned.
    Still, you're probably better off trusting Ed's characterization
for your scanner ("close to sRGB" or not), and simply asking for the
desired color space.

shAf  :o)