Holes in the gamut?

[gordo]

I'd appreciate the experts opinion on a theory that was presented to me. Is this a known issue or?

The idea is that one reason for not being able to reproduce a color from one system in another may be that just because one had a similar gamut volume, that does not mean that one has access to all the colours within that volume.*Maybe this is also known but it may be a fundamental cause of frustration for people trying to obtain a match.

The notion is that the gamut is full of holes or probably more like tunnels that go through the gamut volume and colors are not obtainable in these tunnels. If one would do a very comprehensive survey of a gamut volume and make a horizontal slice through it, it might look something like swiss cheese. *The voids would be the cross sectional areas of the tunnels.

The reason for this is that when dealing with pigments or colourants, the spectral curves for these are not limited to specific areas of the spectrum. *So when one builds a color, one can not adjust the reflected light in specific areas of the spectrum and therefore have control of the generated color. *If you have lights that have narrow wavelength ranges, then adjusting their intensity has a greater control of changing the XYZ values independently and therefore control of the color generated. *But with colorants, this is not possible.

Changing a single colorant, changes the whole spectrum and adjusting the XYZ values independently is not possible.

The simplest example of this problem is the generation of black or gray with CMY colorants. *Most people know that getting a pure neutral black or dark gray with CMY is not easy. *Also one can say that dark gray and then less dark gray are hard to obtain. *This demonstrates the tunnel effect since it is probably hard to get a neutral gray all the way up through the gamut volume. *Only at the light grays does one not notice the inability to get the neutral gray due to the lightness but it is probably there. *

One could modify the CMY so that a better black or gray is obtained but that will shift other colors. *Black is a color just like any other color but it is one that is more easily identified as being non neutral when it is. *But one can have a purple or some kind of red that is just not quite right and no matter how hard one tried, it will not match another system's color even if both are within the gamut volume.

best, gordon p

[Lukas Engqvist]

What you are explaining is a more coherent description of the argument against heavy GCR. If one looks at the "web-safe" gamut it is not described as a swiss cheese rather as a quilt of well defined colours. Here the holes do not exist (wich they would if you made space for possible colours which would but do not exist in the gamut) but there are many areas with steps or thresh holds.
A colour has to be either or it can't be nothing when translating to a space... can it? Simplified if a colour gamut can only have red or yellow and no orange, translating and orange to that gamut will be either red or yellow, never undefined (a hole, as in a swiss cheese model).
This is thinking out loud, I may be totally off

[meddington]

Not sure I buy the concept of a gamut “hole” or tunnel that extends significantly beyond the gamut surface. I would be more apt to blame issues of “internal gamut matching” to interpolation...a limitation of the resolution of the color target (lower patch count vs higher patch count-see attached pics) that generated the profile, perhaps coupled with the gamut mapping attributes of the profile/CMM. Possibly a limitation of the profiling target’s use of whole numbers rather than fractional percentages. Lighting/a high metamerism index can certainly play a role as well . In short I think there are a variety of issues to look into before one could conclude that there is indeed in fact a “hole” that makes certain colors in a given gamut simply unobtainable.