How to read a gamut "graph"

<p>http://www.luminous-landscape.com/reviews/printers/x300.shtml</p>

<p>In this article and also in oodles of other books describing printers there are these very pretty and elegant 3D color graphs e.g. "midtone gamut comparison"</p>

<p>How does one read and interpret these graphs? If you have a link/pointer because it is too complex to describe that is fine too.</p>

<p>Of interest also to me is how did they even produce this graph. What was the input stream and what program could draw this. I've seen output from programs like Mathematica that have this capability but it begs the question what they are using for input.</p>

<p>Just curiosity questions that I have never seen this explained anywhere. All the authors seem to assume you can read and interpret the graphs and what they are saying is obvious.</p>

Short version: Colors are normally modeled in three variables and like three-dimensional space the variables can be

thought of as coordinates. They can be things like RGB, or Hue, Value, and Chroma, or in the case of CIELAB space

lightness, and two color channels called a and b. These color channels are color opponent channels. The a channel

represents green vs magenta, and the b channel represents blue vs yellow. CIELAB is theoretically able to model all the

colors visible to the human eye. When you plot all these colors they form a 3 dimensional shape in the space with axis

at L, a, and b. Outputs like monitors can't reproduce all the colors available in lab space—they have a smaller gamut.

Because of the way the receptors in the eye work it is theoretically impossible to reproduce all the colors with only three

color primaries. So if you were to plot all the colors a particular output device could create as specified in a color profile

of the device, you would end up with a solid in LAB space. The charts you are looking at are comparisons of the solids

formed by the gamuts of various devices, which allows you to see the relative size and shape of their gamuts. You can

see that one printer might give you more saturated greens or less deep blues. To make it a little easier to understand,

people will often take a slice of the shape—a cross section—at a particular value of L—just like slicing a ham. This is what you are seeing in the

2D graphs. They are even kind enough to tell which lightness value the slice came from.

The graphs are made from ICC profiles. I'm not sure what luminous landscape is using to make the graphics, but a lot of

people use ColorThink Pro http://www2.chromix.com/colorthink/pro/. Apples ColorSync Utility also makes nice 3D graphs

of profiles too. You could almost certainly do it in mathematica too, if you like that sort of thing.