Skip to content

plotting CIE chromaticity diagram?

6 messages · ishi soichi, Bryan Hanson, Ken Knoblauch

Original
Bryan Hanson
# 
As it turns out, I am working on this right now, and it's a bit of a challenge to get the color gradient correct.  I don't think this is really the right kind of question for this list, so let's see if anyone has attempted it or has ideas, and then take the conversation private.  I can send you my not quite perfect attempt a little later today.

Bryan

****************
Prof. Bryan Hanson
Dept of Chemistry & Biochemistry
DePauw University
Greencastle IN 46135 USA
academic.depauw.edu/~hanson/deadpezsociety.html
github.com/bryanhanson
academic.depauw.edu/~hanson/UMP/Index.html
On Mar 18, 2013, at 6:12 AM, ishi soichi <soichi777 at gmail.com> wrote:

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Ken Knoblauch
    

    

      
      #
    

  


  

      
ishi soichi <soichi777 <at> gmail.com> writes:

            

      
      
      
    

        
I think that plotting the spectral locus and the line of purples is trivial, 
but that's not actually what you are asking.  Though filling in the interior 
with a color gradient would be a challenging and interesting problem, 
there are those who would consider this a misguided attempt at 
representation (even though it is frequently done) as the CIE coordinates 
provide no information on color appearance, per se.  
They specify what lights look alike, not what they look like.

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Ken Knoblauch
    

    

      
      #
    

  


  

      
ishi soichi <soichi777 <at> gmail.com> writes:

            

      
      
      
    

        
<sermon>
And following up on my previous mail (diatribe), after
having contemplated the image at the link that you
provide, there is another significant distortion to take
into account in the representation that you propose.
Suppose that you are considering the appearance of
isolated, centrally viewed lights in isolation, in neutral
adaptation by an observer whose vision corresponds to
the average observer of CIE1931.  Then, the set of primaries 
that you will be using to generate those colors (on a screen
or on paper) will have a gamut that excludes a significant
portion of the diagram, so that if you take this limited gamut
and stretch it out to fill the diagram, then the coordinates of
the colors indicated will deviate significantly from the positions
in which they are represented (not to even speak of their
appearance).  So, in general, while this leads to a pretty
diagram, there is a lot of potential for misunderstanding
in such a (mis)representation.  </sermon>

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Bryan Hanson
    

    

      
      #
    

  


  

      
I am, unfortunately, well-aware of the limitations that Ken points out (and I do appreciate him making these points).  One can readily demonstrate the gamut limitations by printing the diagram Ishida links to on different devices.  My hope is to get something close and include a disclaimer.  Bryan

        
On Mar 18, 2013, at 7:08 AM, Ken Knoblauch <ken.knoblauch at inserm.fr> wrote:

        

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Ken Knoblauch
    

    

      
      #
    

  


  

      
Unfortunately, a bad idea that looks good never dies.

That said, I still think that it is an interesting challenge
to create this graph by programming.  I can imagine that if
you specified the coordinates at a certain number of points
in the interior and along the spectrum locus, you could use
an interpolation algorithm, to reproduce what is on the
wikipedia page.  Of course, I'm kicking myself for encouraging
this be done given the opportunity it provides for misguidance
and misrepresentation.

Ken


Quoting Bryan Hanson <hanson at depauw.edu>: