Assigning colour profiles
Assigning a colour profile to an image can be compared to labelling on a packet of food in a supermarket. The label does nothing to alter the contents; it is just information informing you as to what those contents are. Should an incorrect label be assigned then no change to the contents takes place, but your reaction will change when you come to eat the food as it will not taste as expected. In similar vein the colour profile just describes what the image colour space is; nothing is altered in the image. If the correct profile has been assigned then the image appears as it should.
FIGURE 1 If an image is assigned a colour profile that is too wide then only the central range of colours are seen. As a result the image appears to be dull and unsaturated, depending on how wide the original colour gamut was. The wider the incorrect profile, the duller the result. On the other hand if an image is assigned with a colour profile that is too narrow, then the colours are exaggerated resulting in over saturation, most noticeably in the reds. In both cases none of the values in the pixels were altered, only what those values represent.
FIGURE 2 The original image with its correct colour profile.
FIGURE 3 Here we see the very visible effects of assigning profiles that are too narrow or too wide.
When an image does not have a colour profile attached, or tagged to it, then its profile has to be guessed at. It is important to understand that the image already has a colour space, it’s just that there is no tag saying what that colour space is. An sRGB image is still an sRGB image with or without its colour tag, just as a TIFF file is still a TIFF file with or without its “TIFF” file extension. So when guessing at what to assign start with the most likely colour profiles of sRGB and Adobe RGB first.
The basic rule to follow is that if the image is too dull then assign a wider profile, if too vivid (especially the reds) then assign a narrower profile. Go through the list until you obtain the best result. Once the “correct” profile has been found and assigned then the image can have its colour space converted to a different profile if needed.
Converting colour profiles
FIGURE 4 Unlike assigning profiles, converting an image to another colour profile does alter the value of the pixels as they are changed from one colour space to another. The difference can be seen by comparing this illustration to figure 1.
The most and least saturated colours from the original space are mapped to the most and least saturated colours of the destination space. Converting images from a wide to a narrow colour space reduces the range that the values in the pixels represent. If there is a big difference between the wide and narrow spaces then some shift in the colours will be noticable. But broadly speaking the before and after versions of the image look the same, even if some of the subtler colours have been compromised.
Converting images from a narrow to a wide colour space will not improve the image quality, even though the range of potential colours has increased. Any subtle colours not captured in the original space can not be reinvented.
FIGURE 5 Broadly speaking, the before and after versions of the image look the same.
FIGURE 6 The Histogram shows the similarities between the two.
Embedding colour profiles
FIGURE 7 Regardless of whether an image has had a colour profile applied, or had an existing one converted to another, then the profile needs to be embedded within the image.
Embedding the colour profile within an image is a simple operation. When saving the image, use the “Save As” option and make sure the Color Profile tick box at the bottom of the panel is selected. This is labelled either “Embed Color Profile:...” on the Mac system, or “ICC Profile:...” in Windows. This option will be ticked by default, but only if the file format that is being used is able to carry a colour profile. If the image is saved as a BMP file, for example, then the colour profile option will be greyed out as BMPs do not recognise profiles.