CIE (1931) XYZ vs xyY

CIE (1931) XYZ vs xyY

Post by Bruce D. Beck » Mon, 08 Oct 1990 16:17:17



|Recently purcased the book "Illumination and Color in Computer Generated
|Imagery" by Roy Hall.  Great book.  In this book, he gives the color
|coordinates of each square of the Macbeth ColorChecker(r) chart using
|CIE 1931 XYZ coordinates.  I ran out & bought one of these charts at
|my local camera store, and started writing a program to calibrate my
|computer's display.  The chart came with a leaflet which lists the
|color coordinates of each square on the chart, but it contains different
|numbers!  It also claims to be using the CIE 1931 system, but the three
|columns are labelled " x y Y " instead of " X Y Z ".  Anybody know what's
|going on here?

        The "x y Y" system is a normalized form of the
        "X Y Z" color coordinate space.

                        X
                x = ---------
                    X + Y + Z

                        Y
                y = ---------
                    X + Y + Z

        Note that  x + y <= 1.0 in this system, and that the
        locus x = 1/3, y = 1/3 exhibits equal energy contribution
        from each part of the visible spectrum. The Y component
        in both systems represents the luminosity function, which
        is a measure of perceived brightness independent of color.
        Because the "x y Y" system is normalized, Y can be scaled
        to any arbitrary reference value.

--
  ,u,    Bruce Becker   Toronto, Ontario

 `\o\-e  UUCP: ...!uunet!mnetor!becker!bdb
 _< /_        "I still have my phil-os-o-phy" - Meredith Monk

 
 
 

1. CIE (1931) XYZ vs xyY

xyY and XYZ are two slightly different color coodinate systems.
Fortunately, there's a simple transformation. I'll show it now, but
you ought to check a good book on color like Wyszecki&Stiles "Colour Science"
(John Wiley and Sons, 1967).

Y = Y;  x = X/(X+Y+Z); y = Y/(X+Y+Z)

or backwards:

X = (x/y)*Y;  Y=Y;  Z=((1-x-y)/y)*Y

As for interpretation, XYZ values represent the amounts of three
specially defined hypothetical primaries (not physically realizable).
The Y primary was specifically defined to be roughly the human monochromatic
response, so a color image given in XYZ, can be shown in Black and White
by just mapping Y to intensity.

xyY is a kind of normalized form, where Y is that same monochromatic
intensity and x and y are chormatic coordinates which don't include
the intensity. This is useful when defining a color without knowing
the intensity as just (x,y).

The book I mentioned gives all the background you'll need to understand
color science. There are many other wild and wacky color spaces out there.
Some are non-linear and are much better at modeling human color vision.

I hope I helped....

2. amazon paint question

3. CIE xyY -> XYZ

4. javarazor - button Applet

5. CIE 1931 extended matching functions

6. declares in povray

7. CIE xyY to RGB

8. page tools

9. Colors and normalization, xyz, xyY and YIQ

10. Is CIE XYZ colourspace available in BMRT?

11. measuring screen color values - cie xyz

12. RGB to CIE XYZ and back