[scribus] Colour Profiles - make colour matching worse!
Hal V. Engel
hvengel at astound.net
Tue Jul 28 01:48:22 CEST 2009
On Monday 27 July 2009 02:21:20 pm John Beardmore wrote:
> Peter Nermander wrote:
> >> OK, but that might be small price to pay if it made CMS accessible to
> >> more users. At the moment, complexity makes it less than worthwhile for
> >> most people.
> > It wouldn't make it easier, because you would still need a profile for
> > each output device.
> Yes, but presumably at the moment, for each input device you need a
> profile for each supported internal representation,
The device only has one representation and the device profile describes how
that device representation relates to a device independent color space such as
CIELab or CIEXYZ.
> or if there isn't an
> internal representation, each intended output device ?
No you are not understanding this. There are two basic types of profiles.
Those for input devices and those for output devices. Those for input devices
are only required to have the information to go from the device color space
(like say the RGB values of a camera or scanner) to the PCS and output device
profiles have the information to go from the PCS to the output device color
space (the RGB values of a monitor or the CMYK values of a printer). Profiles
can also have tables that go in the other direction but this is not required.
You need two profiles to do a meaningful transform and the PCS is an
intermediary between the input device and the output device color spaces. I
think you are making it more complicated than it actually is.
> > All colors in CIELAB can't be shown on a monitor, and all colors in
> > CIELAB can't be printed. Some of the colors in CIELAB can be shown on
> > a minitor but not printed, and some colors in CIELAB can be printed
> > but now shown on a monitor.
> > Also, a color that can be shown on one monitor might not be possible
> > to show on another, as well as some printers won't be able to print
> > colors that other printers can print.
> > This is where the rendering intent come into place (and this is the
> > reason Pantone and similar "color indexes" exists).
> Possibly, but presumably these describe a specific colour, but you still
> can't represent them on every output device. Maybe you just buy a bottle
> of the right ink ?
The real purpose of the Pantone system is to sell these bottles of ink.
> > There als also colors that can not be reproduced in i digital image.
> > Gold and silver are two.
> Not sure that gold is a colour. Reflectivity is perhaps more of a
> surface property ?
Gold is most certainly a color and the Pantone system has color patches that
are gold and silver and variations on these. You can also purchase inks for
these colors to use as spot colors.
> > You can not show gold or silver color on a
> > monitor, nor can you print it with CMYK inks.
> No, but you can print the colours in an image reflected off the surface
> of a lump of gold.
> > The same goes for many
> > flouroscent colors.
> Well - you can't reproduce the property of fluorescence in CYMK print,
> but you might have a stab at representing the colour fluoresced.
> Yes - there are loads of things that you can't represent accurately,
> e.g. very monochromatic colours, but as these aren't the problems that
> CMS intends to address, I'm not sure that's either here or there ?
> Cheers, J/.
More information about the scribus