Ok, I've checked this with Phil Green, the ICC technical secretary and with Ronnie Luo, who most of you already know.
The typo is in Lam's paper. The reference Phil gives is Hunt, R. W. G. 'Reversing the Bradford chromatic adaptation transform', Colour Research and Application 22 pp355-356, and this gives the same value as that used in the ICC spec. Ronnie corroborates the coefficients in the last row should add to 1.0. Regards Marti Quoting marti.ma...@littlecms.com: > Hi, > > Thanks for reporting, I am bringing this issue to the ICC. > > Regards > Marti > > Quoting Esben Høgh-Rasmussen Myosotis <e...@phaseone.com>: > >> Hello there. >> >> I am looking into the pit known as chromatic adaptation and noticed >> something strange. >> >> In the ICC specification, Bradford adaptation is based on the >> following XYZ/cone-space conversion matrix: >> >> [ 0.8951, 0.2664, -0.1614 ] >> [ -0.7502, 1.7135, 0.0367 ] >> [ 0.0389, -0.0685, 1.0296 ] >> >> This is also what LCMS use. >> >> This method is, presumably, taken from K.M.Lam's thesis "Metamerism >> and Colour Constancy" (what he calls KING1 method I think). However, >> when I look in the thesis, it specifies the following matrix: >> >> [ 0.8951, 0.2664, -0.1614 ] >> [ -0.7502, 1.7135, 0.0367 ] >> [ 0.0389, 0.0685, 1.0296 ] >> >> The numbers are the same, but the sign in A(3,2) is positive instead >> of negative. The matrix is written in the same way 3 times in the >> thesis (unless the OCR missed something; I did not actually go >> through all 426 pages). >> >> Any ideas why there is this discrepancy? Are we using a buggy method >> for adaptation due to a typo? >> >> >> Esben H-R Myosotis >> Software Engineer >> >> >> >> >> ------------------------------------------------------------------------------ >> _______________________________________________ >> Lcms-user mailing list >> Lcms-user@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/lcms-user > > > > ------------------------------------------------------------------------------ > _______________________________________________ > Lcms-user mailing list > Lcms-user@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/lcms-user ------------------------------------------------------------------------------ _______________________________________________ Lcms-user mailing list Lcms-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lcms-user
