> We are not speaking about the same thing. The fact that you want to change
> some channels in some color model does not mean that the internal
> representation of images must be based on this color model.

True, it doesn't.

> You need tools
> for generating a proper CMYK representation of you image, suited to your
> printing shop hardware, but that does not mean that the image you handle
> must use this representation from the beginning.

I agree, it doesn't have to, yet it is convenient to use such
representation when the image is to be send to print house.

> RGB is the internal representation of the images.

That's untrue.

> It's also a color model,

I'm not sure if it's right to try to set apart "internal
representation" and  "color model".

> not especially suitable for manipulating colors.

I agree, but also it's easy to use in digital environment.

> Thus you need tools that handle the image in more convenient color spaces.

Meaning: tools that convert image to different color space for the
time of their application. Which, by the way, can be quite tricky

> When you define a color
> using the color chooser, I suppose you work in HSV, not RGB?

In fact, most of the time I use CMYK color chooser. Choice of color
model often depends on your "notion" of that model. I prefer to use
the same color model in my chooser that is the color model of my
image. Using e.g. HSV chooser for RGB image can be deceitful since
some color values can be silently changed by CMS. I prefer to use e.g.
Lab for color adjustment and after I'm done with my modifications
convert image to destination workspace. That way I've got full control
over things that happen to my colors during conversion.

Anyway, I believe we should be able to use different color models used
to represent color samples stored in images not because it's "just a
nice thing" but because there's one silent assumption about every
color model: every color model is bound with some workspace not
necessarily bijective in regard to other workspaces.

Best regards!
Gimp-developer mailing list

Reply via email to