Confusion about monitor white and white point????
AE lover
I am characterize a monitor and I am very confused with the terms
"monitor white" vs "white point".
Are they the same or different?
i.e if I use a spectroradiometer to measure the CIE XYZ values of
three primaries RGB of my monitor, I will get the monitor white. Is
that also the white point of the RGB system of my monitor? If it is
not, what is the white point?
I often see they give a white point in the form of chromaticity
coordinates such as D65: x=..., y=..... So what are the CIE XYZ value
of D65?
One more thing, is it true that the CIE Y of a white point is always
equal to 1. Why when I read from spectroradiometer I get the resulting
values in the range from 0 to 100, and sometimes even higher than 100
(such as 100.3, 100. 5). I guess these values are the normalized
values with some certain values. Does it mean that we need to consider
what the white point is set in the spectroradiometer?
Last thing, when we say about CIE XYZ values, do we need to care what
the white point they set when they measure these values?. do CIE XYZ
values with different white points have the same chromaticity?
A lot of dummy questions, please help me out.
Thanks
Dieter Michel
> I am characterize a monitor and I am very confused with the terms
> "monitor white" vs "white point".
> Are they the same or different?
"White" is the color the adapted visual system recognized as "white".
There is an effect called "color constancy" which means that an object
of known color (here: white) is perceived as being white, even if
the spectral structure (and measurable color) of the light it sends
to the eye varies. So, what the human visual system perceives as "white"
is also a result of adaptation and is not a physically defined property.
So you can e.g. define a white point in your color system and
check whether the measurable white of your monitor corresponds
to that. Ideally, they should be identical, if not, the monitor
might need to get calibrated or is designed for a different color
system.
> i.e if I use a spectroradiometer to measure the CIE XYZ values of
> three primaries RGB of my monitor, I will get the monitor white.
No, you will get three sets of color coordinates for your primaries.
No easy way to determine the white point from these because there
is more than one color that can be perceived as "white".
Of course, you can maesure the color coordinates of the "white"
your monitor displays for R=G=B=100%. That's your monitor white.
> Is that also the white point of the RGB system of my monitor?
> If it is not, what is the white point?
Nomally, the white point of the RGB color system is chosen
beforehand depending on what you plan to do with the monitor.
The monitor white is then adjusted to match that predfined
white point in order to be able to display the other colors
(within the monitor gamut) correctly.
Widely used for computer monitors is sRGB. The corresponding
primaries are:
Red Green Blue White point
x 0.6400 0.3000 0.1500 0.3127
y 0.3300 0.6000 0.0600 0.3290
If you want to use the sRGB color space and the monitor white
you find in a measurement differs (significantly) from the
coordinates given above, the monitor should be adjusted.
You can do that (strictly only for one luminosity value)
using the RGB-controls on some monitors. Better use a
color sensor and calibration software
> I often see they give a white point in the form of chromaticity
> coordinates such as D65: x=..., y=..... So what are the CIE XYZ value
> of D65?
The chromaticity coordinates for sRGB happen to be the same
as for D65, so they are: x=0.3127, y=0.3290
> One more thing, is it true that the CIE Y of a white point
> is always equal to 1.
Only when you normalize it to 1. The CIE chromaticity diagram
cannot display Y because that would require an additional dimension.
Normally, the white point coordinates are not given in X,Y, and Z
but as a color locus (coordinates x and y only, without luminosity).
> Why when I read from spectroradiometer I get the resulting
> values in the range from 0 to 100, and sometimes even higher
> than 100 (such as 100.3, 100. 5).
With spectroradiometers, the Y output is usually given
in cd/m2 (candela per square meter) and is not normalized
to 1. Most often, the Yxy or Yu'v' color space is used
because you can directly see the luminosity (luminance
in case of monitors) and the color locus.
> I guess these values are the normalized values with some
> certain values.
Usually, a spectroradiometer gives you an absolute reading.
You only need the color coordinates of a reference white
if you want to convert into color spaces like La*b*, because
that is used to describe colors seen under defined lighting
conditions (you need to define the light used for illumination).
> Does it mean that we need to consider
> what the white point is set in the spectroradiometer?
Normally not.
> Last thing, when we say about CIE XYZ values, do we need to care what
> the white point they set when they measure these values?. do CIE XYZ
> values with different white points have the same chromaticity?
CIE XYZ coordinates do not have a "built-in" white point
or are given relative to a white point. Of course, you
can say that X=Y=Z=1 is your white point.
For that white, the Yxy coordinates would be Y=1, x=0.3333, y=0.3333
That is a bit different from the white point of sRGB,
for example, which is x=0.3127, y=0.3290 (a little bit
more blueish).
It's the RGB color space, where the white point comes
into play. In RGB, white is R=G=B=100% - and the white
that you see in that case, is the "monitor white"
(or "projector white")of your output device.
So, RGB-coordinates will render different chromaticities
depending on the primaries and white point the respective
RGB color space is based on. That's why there are more
than one of them, such as sRGB, Adobe RGB, CIE RGB and the like.
Best regards
Dieter Michel
AE lover
Your answers are very helpful. I am just stepping in color science (my
backgroud is image processing). It is really helpful to learn from you
guys in this group.
Thanks again, Diater.