The Big Deception
How often have we seen the following graphic?
Shown is the famous 1931 CIExy Color Space horseshoe (colored for screen display with RGB colors) containing the triangular plot formed by connecting the chromaticity (x,y) points of a typical RGB phosphors used in color display devices. Sometimes, but not always, we are advised that this is not an accurate gamut representation, but is only shown to illustrate a concept. I have used this illustration a few times myself in several of the presentations on this web site.
The horseshoe shape is colored by RGB combinations, therefore, the triangle can't possibly be the RGB gamut limit since by definition it should contain all combinations thus leaving nothing to color the rest of the horseshoe.
Less misleading would be setting the areas outside the triangle to a neutral gray, once again with the understanding that we have no way to visually represent the real colors present at these xy coordinates, none of which would be the gray. As an after thought, the least contradictory color to represent these areas would be true black (RGB=000) since the means by which CIExy space is calculated excludes black! This graphic itself is not a true representation since it was not created by crowding in all the color combinations originally falling outside the triangle. What then is true?
This graphic was created by the color matching practice program specially modified to plot CIExy instead of 1976 CIELAB. All 16,777,216 sRGB combinations were converted to xy and plotted on the graph (except black which has no xy, X=Y=Z=0 thus X+Y+Z=0 so x=X/0 and y=Y/0 are undefined) starting with dark and proceeding to white. What results is not all RGB colors since the lighter ones at each chromaticity plot over the darker ones. After all this we are still no closer to representing the true shape and nature of sRGB space! What to do?
Because of its non-uniformity and inability to easily represent lightness, 1931 CIE space is not a good choice for any further discussion. 1976 CIELAB, a dramatically more uniform space with more natural lightness presentation will be used hereafter. The Color Matching Practice program replotted the same xy data from before only this time after converting xyY to LAB.
Shown here is the back side orthographic view of the sRGB color solid resulting from plotting the data in L*a*b* space.
This is the related top view of that same data.
Now the related right side view of that same plot.
Shown here is the orthographic view looking from the bottom.
This is the related front view.
Now the related left side view of that same plot.
Here they are all at once showing how they relate to each other making them easier to visualize.
By now it should be pretty obvious that the original triangle that started this whole discussion was a very misleading, over simplification of the true shape of sRGB color space. Now for the last visual that ties it all together - an isometric, 3D representation of all this.
The color matching practice program available for down load from this site plots all the LAB plots shown here.
your questions, suggestions or other comments.