icc profile

It is essential in color management to understand what an ICC profile is : it is the color ID card of a device. They are mainly used to display the "right" colors of a file. It is created when calibrating a device. Two things are important and we will take an example to explain it : when calibrating a screen, we want to know if it displays the colors correctly and how many colors it is able to display. All this information can be found in HIS ICC profile. Same for the printer/paper. The software that manages its ICC profiles will therefore use them to display or print the colors of our beautiful photos correctly.

An ICC profile, calibration tools and conversion engines - CMM (Color Management Module) - are the cornerstone of good color management. For a color reproduction device to do this correctly, you must first know how it does it by default. This is to take into account its possibilities and of course its color characteristics, as we have seen in the previous pages. That's what we're trying to find out during its calibration. All this information will be contained in a special file : its ICC profile. All you have to do is use it with the color management tools to keep the same colors throughout the graphic chain!

What is an ICC profile ?

It is therefore a small file linked to a color reproduction device. It's his color ID card. Each device, in color management, must therefore have its ICC profile obtained during its calibration. It contains a lot of information about its colors of it.

It's a gamut AND a record of its "defects"

It is first of all a color space (seen on the previous page) - so what we call a gamut - but a little special because it is dependent on a device. It is in a way the colour identity card of an image or a colour reproduction tool, as we have seen on the previous page, since it contains the colour characteristics of the latter and in particular

All colours reproducible by a device or contained in an image. In this sense it is a colorimetric space and therefore a gamut. In other words: is this or that (saturated) color reproducible by this device?

But also to which color L*a*b* (therefore its "true" color) corresponds each RGB value. This is the same as knowing if the device is reproducing the colors correctly for short.
These correspondences are contained in a special table: a matrix correspondence table (in other words, an ICC profile).

This table therefore contains even more values (columns and rows) as the ICC profile creation software works with finesse. On the other hand, the finer it works, the more information the table will contain, so the ICC profile will be heavy. In absolute terms, the matrix table would contain 8 million matches for the 8 million colours in the L*a*b* space.
In practice, no device is able to "work" eight million colours like the human eye and, moreover, there is not much interest in analyzing all the colours that a device can see, display or print. One color from time to time is enough to create a good ICC profile! So these are its colorimetric characteristics, in other words, its defects and therefore how to correct them !

An ICC profile is therefore directly linked to a device - unlike a pure color space -. It is the colorimetric characteristic of an instrument measured by a given tool under precise and known calibration conditions. The process of creating the profile is called characterization. The ICC profile created therefore depends directly on this measurement tool, the profile creation software and of course the calibrated device.

The ICC profile is dependent on a device

Either we have to deal with the ICC profile of an input device (camera, scanner...) or output device (printer...) then it is on the one hand, the description of all the L*a*b* colors that it is able to reproduce here to see - its Gamut - AND on the other hand, the transformation rules of the original file so that the colors can be transcribed or displayed correctly. It is a matter of "eliminating" as much as possible or correcting the deformations induced by the device.
Its Gamut is smaller than the L*a*b* space and therefore fits into it. But as much as the L*a*b* space is absolute because it does not depend on a device, so much the space of the device corresponds to the results of a measurement and therefore depends directly on its quality. We saw on the previous page that the same RGB value ( x, x, x, x) will not give a neutral grey without calibration.

It is very clear that even the large DonRGB is much smaller than the L*a*b*. Today, it corresponds substantially to the gamut of an inkjet printer on glossy paper.

In concrete terms, what is going on ?

Thanks to the ICC profile of this device, software such as Photoshop knows that when the image file contains an RGB value for example 128, 140, 128 therefore with a predominant V value (hence this greenish aspect), the scanner has scanned a neutral grey (128,128,128) and it should have as RGB value 128,128,128 and therefore display a neutral grey. Well Photoshop will read the matrix table of the ICC profile of this scanner and then it knows that when it has to display the RGB value 128,140,128 it must actually display the color L*a*b* 54,0, 0 (equivalent in RGB to 128,128,128) and it will then send a corrected RGB signal to the screen in style 128,116,128 to take into account the defects of the scanner in green in this example.

An ICC profile is therefore used to display the "true" colors of the image without taking into account the "real" RGB values read because it depends on a device that had defects. An RGB value of 128, 140, 128 will still be displayed as a medium grey and not a greenish grey here on our screen.

If the difference between the ideal and reality was particularly impressive in the days when scanners were used a lot, it is true today, in the digital age from the moment the picture was taken, that devices WITHOUT calibration are much closer to the right colors.

The mechanisms for creating ICC profiles as well as the rules for managing ICC profiles will be discussed in other parts of this color management guide. In particular, the following page will discuss the concepts of ICC profile assignment : assigning an ICC profile - 7 / 10

ICC profiles

What is an ICC profile ?

It's a gamut AND a record of its "defects"

The ICC profile is dependent on a device

Does an image contain a profile or a space ?

In brief, what is the purpose of an ICC profile ?

In concrete terms, what is going on ?