"Tony[B]" wrote:
> 
> > To me it looks like the Chromaticity values are the CIE x, y and z
> > tristimuli values, and Luminance is the CIE Y value.
> > To convert these values to RGB, I think you have to multiply each of the
> > chromaticity values by the luminance value.
> 
> I tried that and ended up with some very very weird colors. Leaving them
> as-is works fine.

You are right, multiplying by Y yields too dark colours, but on the
other hand, using the xyz-value gives too little difference between the
colours to justify names like Light Blue, Dark Blue or Strong Blue
(which are the colours I have used for my tests).

I have reviewed the CIE formula, and come to the conclusion, that it is
better to convert back from xyz to XYZ values and use them as RGB. That
is done by multiplying x and z by Y divided by y: X = x * (Y/y), Z = z
*(Y/y). Y we already have as the luminance. 

/Ib

Post a reply to this message