Factors Affecting CMM (Color Management Module)

As we all know, when we talk about color management, we will think of the conversion relationship between two feature files, and the two feature files are the characteristics file of the source device and the feature file of the destination device. The characteristics file of the source device contains what kind of color information is in the file, and the device's characteristic file contains what kind of color information can be copied on the target device. The color information returned from the equipment space is not the desired color information in the source file, and a heavyweight color module is needed for control. The color management module (CMM) is also called a color engine.

The CMM is the most basic and important component of the color management system. It provides color management systems with color conversion methods from the source device color space to the PCS and then from the PCS to any target device space. In fact, this is achieved by modifying the color information sent to the destination device's spatial RGB or CMYK values. As for how to amend it, and by what modifications, how the revised value is defined, this is only the mindset of the various manufacturers who make the CMM, which is why there are so many manufacturers of CMM (adobe, agfa, Apple, heidelberg, koda, x-rite, etc., use the colors produced by the CMMs from different vendors to create deviations. There are two main factors that affect it: Interpolation algorithm, white point adaptation, the following discussion of these two aspects.

1. Interpolation algorithm. Interpolation algorithm is to use the function of a certain number of points in a certain area of ​​the function value, to make appropriate specific functions, take known values ​​at these points, in other parts of the interval using the value of this particular function as an approximation of the function; CMM interpolation The algorithm is also based on the principle of this algorithm. In simple terms, it is defined according to a specific set of algorithms by selecting points of representative color matching values ​​on RGB or CMYK nodes to achieve the purpose of simulating the entire color space. .

Example: Under normal circumstances, each of the three channels of the RGB device contains 256 discrete values. This way, the RGB color space contains a total of 16.7 million PCS color definitions for RGB color values, if each PCS definition Using three values, each calculated by one byte, then the amount of information for such an RGB profile would be huge, and then embed this huge amount of information in an RGB picture, or one of multiple such pictures. In the large version of the document, the space occupied by the document will be intolerable. If this is true, then who will use such a thing that cannot be seen? Based on this data volume and transmission speed considerations, the CMM difference algorithm has emerged as the "cost-effectiveness" is high enough.

Ideally, we should do interpolation calculations in the CIE LAB color space, use LAB as PCS, at this time, for the CMM, the only thing it needs to do is to obtain the corresponding LAB value in the input profile, in the output The closest PCS value is found in the device properties file and the corresponding output value is found. If the reasoning is based on this idea, then the CMM is not very useful, but the actual situation is indeed the case, if you follow this approach, you can't get an ideal result in practical applications, because according to this algorithm comes out As a result, LAB is not as uniform as we imagined, and there will be unexpected results, such as the fact that the actual value does not correspond to the desired value. It is based on this consideration that different vendors will use different difference algorithms. The difference algorithm of different manufacturers is mainly embodied in the selection of nodes and the optimization in calculation. Generally, they can be divided into three categories:

A. The first type of manufacturer uses some mathematical formulas and uses clever algorithms to compensate for the lack of LAB color space to correct its inaccuracy. It uses a special method to achieve uniform and smooth LAB color space to “cheat” the eyes of color users, but the actual effect is not as perfect as it looks.

B. The second type of manufacturer does not use the LAB color space, but instead converts the color to other color spaces for interpolation so that the characteristic file looks more perfect and complete.

C. The third type of manufacturer is to use a special color space in addition to any one of LAB or XYZ when creating a characteristic file, so as to ensure uniform and smooth color space. This approach first violates the open and exchangeable nature of the feature file, and most importantly, its actual effect is not as perfect as it may seem, although it does make up for the inconsistencies that some feature files create through interpolation. .

Different manufacturers' interpolation algorithms are not the same, which also leads to different results when using different vendors' CMMs. For example, for some highly saturated colors, different CMMs will result in different results. It will be different, and what is more, it will transform the blue sky into a purple one.

2, white point to adapt. The eye adapts to the ability of the white field to change color, while the white field is the brightest white color that the device can reproduce and its light intensity. Compared to the fact that we pay more attention to the density of black dots than to the color, for white dots, we are more concerned with its color. This is why we may compress some of the whitefield colors when doing monitor calibration. The reason for its brightness.

When the human eye performs white "treatment", it will automatically adapt to it, and then judge and adapt to other colors based on this white color. It is precisely because of this that when we do color conversion, the first thing we must do is The white point of the source device's color space is converted to the white point of the target color space. But sometimes when we perform this kind of conversion, we find that when the white point of the source device is on the target device, it is not the white we want. It may be that the relevant color information appears on different color channels. When another color management software was found, it was found that there was no such situation. This is the difference between the methods used by different vendors' CMMs to process the white point, and the resulting different results (this has nothing to do with the four mapping methods). The difference in the processing of the white point directly leads to the fact that the white point of the characteristic file is not on the same main axis, resulting in the entire neutral gray being biased to a certain hue, and all the colors produced by it will be deviated.

In our process of color conversion, CMM is everywhere, PC and Apple systems, and all application software related to color. How to use it, how to choose it is what we feel is impossible to start. Although we will have many choices of color management software, although there are related CMM modules on color-related application software, although ADOBE has now freely opened its own CMM module, in actual use, we want to With only one CMM conversion module, the purpose of doing this is:

1. Avoid unexpected results when you use multiple vendors' CMMs to perform multiple conversions, and may cause color inconsistencies due to inconsistencies in the conversion methods of different vendors;

2. Even if something goes wrong, we can have enough ways to find the cause of the problem, not because of the use of multiple CMMs.

3. Even if the same CMM is used in different operating systems or application software, deviation problems after color conversion will occur as little as possible.

4. More importantly, after using only one CMM, we can have enough experience and actual data to judge the advantages and disadvantages of this CMM.

In the colorful world of color, the CMM is destined to be an "underground worker" who is obscure but has worked tirelessly. What it brings to us is to ensure the accuracy of color conversion between different systems and different devices, and at the same time, it It is the "culprit" that makes the colors intricate. We can't control its destiny (only the manufacturer can grasp its power to kill and kill), but we have the right to choose to use it, to grasp its internal laws, so that it really takes on the "positive role" of color conversion.