Does anyone have a GRACoL compatible CIE chromaticity diagram I can use?

You can make such a diagram as follows:

1. Download this small text file:

http://docs-hoffmann.de/ciesuper.txt

2. Rename as ciesuper.eps

    The diagram is programmed as EPS vector graphic in generic RGB with K-only CMYK text.

3. Choose in Photoshop RGB=sRGB and CMYK=Gracol...

4. Open file

5. Choose resolution sufficiently high, at least 300 ppi.

6. Choose Anti-Aliasing.

7. Crop if required. In my opinion the diagram should reach until 1.0 for both axes.

8. Save as ciesuper-gracol.tif

The colors in these diagrams are meant only symbolically. Some authors are refusing

to colorize chromaticity diagrams at all.

The magenta dots for the CIE primaries are omitted by writing a percentage sign

in the respective line (almost the last lines in the program), which means 'comment':

% CiePrim

Similarily one gets rid of the correlated color temperature graphics:

% CTCurve

% CTText

Best regards --Gernot Hoffmann

Chromaticity diagrams in xy plane are the 2D projections of 3D color spaces in Yxy, isn't it?

http://en.wikipedia.org/wiki/Chromaticity

Chromaticity is an objective specification of the quality of a color regardless of its luminance...

 

On the other hand, some color spaces such as RGB and XYZ do not separate out chromaticity;

though chromaticity is defined by a mapping that normalizes out intensity

Examples are e.g. CIExyY and CIELuv (as shown in the article).

It's often said, that CIELab doesn't have a chromaticity diagram. But why shouldn't we use a top view
as such a diagram?

Best regards --Gernot Hoffmann

It seems, one cannot edit posts in this forum. New attempt:

From "Meausuring Color"

"... Because these ratios of the tristimulus values are incorporated as cube-roots, there can be no chromaticity diagram associated with the CIELAB space, and therefore no correlate of saturation. This is because the definition of a chromaticity diagram requires that any mixture of two colours falls on the straight line joining those two colours: in CIELAB space this line would not be straight because of the cube-root transformation."

I don't have any exact idea ...

I think that ... top views in CIELab look like the projections of the perceptually uniformed and so deformed RGB cubes ... preserving their hexogonal structures.

It is like getting the HLS represantation of a RGB cube by projecting it on a plane as a hexagon,

I was a little bit slow, but at the end, I've started to understand the geometrical connections between different color models.

I still need to clarify some points,

And I need confirmations/corrections about them.

I would like to ask them one by one.

Do the above diagrams correctly represent the spectral locus in xyY?

These graphics seem to be a mixture of different concepts.

http://docs-hoffmann.de/jimcolor12062004.pdf

Page 5 shows threedimensional views of the spectral locus (according to Jim Worthey)

for x-bar, y-bar and z-bar (not x, y and Y). It is perhaps possible to rename y-bar as Y,

because y-bar is responsible for the luminance.

In my understanding, Y over xy is an open cone, unless a finite white point is defined.

Especially for so called Optimal colors (concept by Roesch, Wyszecki&Stiles, Color

Science, chapter 3.7) we find a threedimensional representation with finite whitepoint,

which looks like a wet sack. Optimal colors are surface colors with maximal saturation.

From where did you get these diagrams?

Best regards --Gernot Hoffmann

It's perhaps better to open a new thread.

I copied these diagrams from the program called Color Space ...

Can be downloaded from here: COULEUR.ORG

I think that it draws spectral locus on xyY by just plotting matching x, y and Y values for every interval of monochromatic lights on the spectrum ... just like we draw the locus line in XYZ by plotting matching tristimulus values X, Y and Z.

Because the underlying data for the most saturated colors human can see is CIE XYZ color matching functions, I think that we can draw the line of spectral locus in every model originated from CIE XYZ by just making necessary transformations to the XYZ tristimulus values for the intended model.

3D representation of the entire human gamut is more tricky, I think. I will ask about it, after getting your confirmation about the spectrum locus.

I tried to start a new topic, but for the first time, I got a warning saying "your content was added successfully, but a moderator needs to approve it before it can be posted" ... after waiting three hours, I decided that the moderator does not have an intention for approving my new topic, so I had to write here.

By the way, do you actively use any other forum on the Internet?

Because, sometimes, just a copying and a pasting may require great effort on this forum.

The new thread is here:

https://forums.adobe.com/thread/1571412

G.Hoffmann