A Square Peg in a Round Hole
The term "gamut" describes the range of colors possible for a given device or color space. So if your document contains colors not within the gamut of your printer, where do the leftover colors end up, and does it matter that not everything says it will fit?
The world is a complicated place, and the technology we use to describe it isn't perfect. If someone could make a printer capable of printing every possible color they'd sell a lot of them no matter what the price. Unfortunately such is not the case. Any real printer can only reproduce a limited, albeit quite useful, range of colors. Cameras aren't perfect either of course. But the capability of any good modern camera still exceeds the gamut of current printers and paper. And therein lies the problem I want to talk about this week.
Color management can seem a formidable topic to many, at least in part because of the terminology involved. The concepts themselves aren't too difficult to understand, but the ways they are generally communicated seem to be only intelligible to those who have an advanced degree in something that only others who have such a degree would even know what the degree was.
When trying to convert from one color space to another with a smaller gamut there are two basic approaches. Think of the problem as being much the same as trying to get a square peg to fit in a round hole. Either you can whittle down the sides of the peg until it fits, or you can compress the peg to make it small enough to fit. If you apply these two strategies to the problem of mismatched color space gamuts, cutting off the edges would be analogous to the Relative Colorimetric rendering intent while squeezing to make things small enough corresponds to the Perceptual intent. As I said, the terminology is confusing even when the concepts aren't.
Strategies for converting color spaces are known as rendering intents. Think of this as describing the reason why you are converting profiles. That is, what is your intent in doing so? The Relative Colorimetric intent maps every source color to its closest destination color. This implies that colors that fall within the gamut of both profiles remain essentially unchanged while those that fall outside the gamut of the destination get mapped to the closest possible in-gamut color. In other words, the gamut mapping basically just chops off the edges where necessary and the out of gamut source colors get truncated to force them into gamut. This can have the potential consequence of causing banding in what should be smooth gradients in the extreme edge colors. By contrast, the Perceptual intent tries to avoid this problem by keeping gradients smooth throughout the image but at the expense of forcing more change to colors that otherwise wouldn't need to change in order to fit. The entire source gamut gets contracted inward proportionately until the whole thing will fit. This might sound preferable since it avoids the problems that Relative Colorimetric can cause, but it is not without potential problems of its own. If the two gamuts differ markedly in size, the compression required can diminish saturation making the result look a bit washed out.
If you're lucky, the square peg that is your source profile gamut will fit in the round hole that plays the role of your printer profile in this analogy and the choice of rendering intent will be a non-issue. If not, for any given image one or the other of these strategies will typically work well and all you have to do is select the one you want. Once you do, the color management system built into your software and operating system will do the rest.