4 Replies Latest reply on Oct 13, 2013 2:33 PM by TurgayOrhan

    Cone sensitivity curves

    TurgayOrhan Level 1

      Can I ask questions about CIE colorimetry in this forum?


      As we know, in color matching experiments, we can get three color matching functions ... depending on the selected primaries.

      And, they are called color matching functions, it is ok ...

      But there is an another set of curves called cone sensitivity curves.

      They are different from the color matching functions.

      And, horse shoe human gamut in CIE chromaticity chart is decided by these curves.

      My question is about these cone sensitivity curves ...


      How are these cone sensitivity curves created?

      Are they mathematical transformation of color matching functions?

        • 1. Re: Cone sensitivity curves
          G.Hoffmann Level 3

          Very difficult. Please let me refer to my own document



          and then to p.3 for two different sets of cone sensitivities.

          [1] refers to R.W.G.Hunt, Measuring Colour

          [3] refers to Wyszecki + Stiles, Color Science


          These are two 'bibles'. Of course we have now many reliable docs in the web.


          The color matching functions r-bar, g-bar, b-bar are valid for the

          so called CIE-primaries, but calculated from measured results for different

          sources as linear combinations. 


          The cone sensitivity functions should be linear combinations as well, but

          negative lobes are not permitted for biological sensors.

          Therefore, a unique set of these functions cannot be achieved without


          Hunt [1, 1987,1996] explains some attempts – 'still a matter of some debate.'


          The color matching functions x-bar, y-bar, z-bar are special linear combinations

          of r-bar etc. with the funny feature that the luminance belongs entirely to the

          y-channel and x,z  are mathematical constructs which don't contribute to lumi-

          nance. Mathematically simple, but physically really odd...


          I hope this helps a little.


          Best regards --Gernot Hoffmann


          Yesterday I had to tell Adobe my birthday data for log-in – I didn't.

          If this should happen again then you may contact me directly:


          • 2. Re: Cone sensitivity curves
            TurgayOrhan Level 1

            Thank you Gernot ... indeed, I was trying to understand when and where the cone sensitivity curves are used in CIE system.


            I think that my confusion is coming from the nice applets on the page http://graphics.stanford.edu/courses/cs178-11/applets/applets.html#color ... There is an applet showing spectral locus as a result of the cone sensitivity curves.


            But now, I think that ... in CIE system, spectral locus is drawn from the x-bar, y-bar and z-bar data ... not from the cone sensitivity curves.


            The cone sensitivity curves are just an approximation, a trial for modelling human cones ... and they are derived from original color matching functions ... they are not directly used in the CIE colorimetry.

            • 3. Re: Cone sensitivity curves
              G.Hoffmann Level 3

              Thanks for the feedback.


              By the way, I thought your first name (opposed to family name) were

              Orhan. Is this correct?


              This stanford doc is interesting, especially because of interactivity.

              The color matching part seems to be somewhat buggy, but I didn't

              investigate the issue further.


              You might be as well interested in this doc:



              It shows 3D spectral locus graphics (not only projections onto the

              axes) and refers especially to the color scientist J.W. (name in the doc).


              Best regards --Gernot Hoffmann


              Very strange – log-in doesn't ask anymore for my my date of birth...

              • 4. Re: Cone sensitivity curves
                TurgayOrhan Level 1

                Thank you for the doc ...


                Yes, it is correct ... Orhan is my first name.

                Turgay is my middle name.



                Thanks again,

                Best regards.