There is quite an industry of equipment and services surrounding the question of ICC profiles. It's based on the notion that we need the profiles to be "accurate". Do we really? What does accurate get us?

Beware of the hype coming from the industry. Unless someone makes a serious mistake, all color profiles will produce colors that are correct and tonality that proceeds appropriately from white to black. As you move from paper white where the actual color of the paper shows through you are looking at darker and more saturated mixtures of ink that soon hide the color of the paper.

All ICC profiles are generated from a pattern of patches (about 400-2500) made on a particular paper with a particular printer with the ICC profile mechanism disabled by a program like Qimage. The profile comes from a comparison of the printed result to an ideal set of values for each patch. The primary goal is for the gray patches to come out neutral (red=green=blue) and for the color patches to achieve the desired colors. Another goal is to for the neutral gray patches to cover the range from paper white (no ink, RGB=255,255,255) to maximum black (RGB=0,0,0) without introducing any non-neutral shifts along the way.

The industry's job is to to convince the public that we can't live without "accurate" ICC profiles and that we need to invest our time and money to get things right. Caveat emptor!

I just spent a couple of weeks testing different ICC profiles and media type settings.

I wanted to see the effect of selecting the wrong media type. The media type selection separates matte from non-matte paper to let the printer use the correct black ink. It's secondary goal is to match the amount of ink laid down to the right paper surface. This matters more within the matte family since there is a greater variety of surfaces available.

Then I picked an appropriate media type and tried a variety of ICC profiles designed for my printer on a mix of papers. I even used matte ICC profiles on glossy paper as well as Epson's generic color profiles on specific Red River, Canson and Ilford Galerie papers. In every case the colors came out the same regardless of the ICC profile used. The only real difference was in the tone curves for a 21-step grayscale. Different profiles and media type selections had a visible and easily measured effect.

ICC profiles from: RR=Red River, Ep=Epson


Here you can see the result of using different ICC profiles on a specific paper where the horizontal scale goes from 0 (maximum white) to 100 (maximum black) and the vertical scale from 100 (paper white) to 0 (the maximum possible black on the paper). None of the profiles actually reach the absolute black limit of the measuring device. But the tone curves for each profile follow different paths even though they all seem to come very close together around middle gray.

Nevertheless, the colors produced for a specific image are all "accurate" to the viewer. Getting the shadows and highlights to work is up to us.

Poorly made profiles from Red River? If I were you I would ask them about that.

Here is a small subset of the profiles I can use:

As you can see, I am using both an Epson P900 and an R2880. I only use the profiles designated for their respective printers.

I have never come across a "defective" profile. Anyone who takes the trouble to create a profile for a particular printer would have tested it before publishing it.

Even the generic profiles like Universal RGB, Printer RGB, Wide Gamut RGB, Epson Standard RGB - Gama 1.8, etc., all work with my Epson drivers on either printer with no color casts or loss of saturation.

There must be a reason why your experience is different from mine besides our choice of printer manufacturer.