# comp.graphics.algorithms

## Subject: Re: Using colour to indicate response intensity

Thanks, very useful these documents in your page.

Regards,

H.

hoffmann@fho-emden.de wrote:
> Hector,
>
> you're right. The complete explanation is here:
> http://www.fho-emden.de/~hoffmann/ciexyz29082000.pdf
>
> Short explanation:
>
> >From color matching to XYZ:
> In the CIE color matching experiment, RGB values are
> found for spectral colors. A matrix multiplication maps
> RGB to XYZ. If we draw XYZ in cartesian coordinates
> then we have three non-orthogonal base vectors for RGB.
> Whatever XYZ means - each triple describes a PHYSICAL
> color.
> The matrix transform is tricky - Y alone describes the
> luminance.
>
> >From XYZ to xy:
> XYZ is mapped by a perspective transform onto the plane
> X+Y+Z=1 . The center of projection is in the origin.
> Plenty colors are mapped to the same point in this plane,
> all of them have the same hue and saturation. They differ
> only by luminance.
> Finally, the plane is viewed in negative Z-direction from
> top.
> The resulting 2D-diagram is renamed x,y or xyY and called
> chromaticity diagram.
> Purists are saying that this 'horseshoe-diagram' shouldn't
> be shown colorized, because the luminance is missing,
> because spectral or highly saturated colors cannot be
> shown by monitors and because they cannot be printed.
> If we understand the colorization symbolically, then it's
>
> Best regards --Gernot Hoffmann