Pete, I am not an expert , but as far as I have understood the matter
regarding the colors space is directly affecting the printing process and
its rendering of colors.
IMHO the screens cannot reproduce anything wider (certain blue, certain
green and so on) than sRGB gamut. It is something relatd to phosphores and
the 3 colors added each other with certain intensity.
For RGB printers (and more for other color rendering technologies) this
limitation is not any more valid.
What I have understood , please correct me if I am wrong or saying no-sense,
is :
RGB printers 3, 5 , any number of inks ... can combine the colors in a way
that is not achievable (by some extension) by the phosphores of the screen.
As a consequence we have :
looking to any picture represented by a physically limited device like the
screen ... no matter wich color gamut we have encoded because it may fall
outside the numerical domain representable by the screen itself ... but when
looking to a printed output then the difference is there because the
printers can reproduce / make some of those colors outside the domain of
''screen reproduceable'' colors even if they likely cannot reproduce all the
possible colors, their domains depending from their own/peculiar technology
and its implementation.
Then ... you say:
> AFAIK, any translation from one colour space to another can only throw
away
> colour information, not add it.
This is what I should like to understand better ...
So ... if I have understood ... I start with a wide gamut = input to PS from
scanner
1... I convert to sRGB = I HAVE LOST INFORMATION
2... I convert to a different wide gamut = I loose information , but this is
affecting ''less'' my results provided that the gamut of arrival is defined
for a domain ''compatible'' with the capabilities of my printer.
3... I convert from sRGB to a wider gamut ... how can I add informations if
the information are not there ? ... is this process of passing from a
''narrow'' gamut to a ''wide'' gamut doing something similar to the
interpolation used when passing from a resolution to a higher resolution ?
Sincerely.
Ezio
www.lucenti.com e-photography site
----- Original Message -----
From: "photoscientia" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Saturday, January 13, 2001 12:27 AM
Subject: Re: filmscanners: Fw: Color Profiles for Printers was scanners
> Hi Shaf
>
> shAf wrote:
>
> > > The only time you need or can use Wide Gamut RGB is if you
> > > are going to output your digital files back to neg or
> > > transparency.
> > > ...
> >
> > Somewhat correct ... but you're ignoring archiving all the RGB
data,
> > and the usefulness of highbit editting before you convert to a color
> > space which is more appropriate for hardcopy.
>
> Say what?
> I still don't see how any colour detail is gained from a so called
wide-gamut
> space.
>
> As I see it:
> The scanner outputs numbers between 0 and 255, or 0 and 4095.
> The scanner hardware knows sod all about colour spaces, and the computer
doesn't
> care either.
> The image file might as well be a recipe for chicken soup as far as the
hardware
> is concerned.
> Those numbers only become meaningful as images when they're converted to
> voltages that pull electrons from the colour guns of my monitor, and
that's the
> only colour space that's visible to me, or meaningful to me when I'm
editing an
> image. I call this colour space "my monitors gamut colour space", but
> unfortunately that profile isn't listed by Photoshop, the nearest it gets
to it
> is sRGB.
>
> Please explain how calling that colour space another name can add any more
> saturation, brightness, or subtlety to the colour represented by those
numbers.
> I'm not trying to pick an argument here, just looking for a clear
explanation,
> preferably in engineering terms, not pre-press jargon.
> I understand Munsell charts, tristimulus values, colour temperature and
white
> points, but I'm afraid I cannot follow the logic of working in notional
colour
> spaces at all.
> AFAIK, any translation from one colour space to another can only throw
away
> colour information, not add it.
>
> Regards, Pete.
>
>
>
