Re: Color matching

On Mar 26, 9:44 pm, Tom Lianza <tlia...@xxxxxxxxxxx> wrote:
Hi,

I first did a little simple test to determine if we have a problem with
the measurements. I took your data and expanded out to XYZ. The first
test is to see if the RGB XYZ data sums to the measured white XYZ data.
This implies that the signals can be thought of as additive...which a
display should be for the white point. I summed the XYZ values for RGB
and these are listed as "white as calculated" These values are then
divided into the measured values to calculate an error ratio. Note that
for the reference display we have values of 1.12, 1.03, and 1.32. This
is a concern because it says that the measured white is a different
chromaticity than the white that would occur by simply turning on the
red green and blue channels.

Note in your Test panel, the ratios are .92 .92 .91 . This means that
the panel has the same chroma, but the sum is appears to be brighter
than the measured value. These values should fall between .97 and 1.03.

This test probably indicates an analog problem. I would make sure that
I could see the difference between 192 and 255 and I would make sure
that I could see a real difference between level 1 and level ll before I
made any more physical measurements. Although from the grey scale data
you sent that doesn't seem to be a problem

With an instrument like the CS200, you have to be very careful about
flare. I would put a mask on the display that allowed only a small
amount of image area surrounding the measurement spot to extend into the
view finder of the CS200.

You need to find out what is going on here. If we can't predict white
reliably on the monitor we have just measured, there is no way to build
a solid model to map one display to another display.

Does the panel have some kind of color model built into it (NEC 6 axis
control)?

Do you have an RGBW panel, that is to say Red Green Blue White matrix
panel?

Regards,
Tom.

Ref Panel: Full scale
Lv x y X Y Z
Red 72.54 0.66 0.31 152.78 72.54 5.99
Green 176.71 0.19 0.64 52.63 176.71 46.21
Blue 29.51 0.15 0.07 57.83 29.51 309.84
White 287.78 0.28 0.27 295.64 287.78 478.11 as measured
263.24 278.76 362.04 white as calculated
1.12 1.03 1.32 Ratio of measured:calculated should be 1.000
Test Panel: Full scale

Lv x y X Y Z
Red 118.70 0.67 0.32 251.15 118.70 5.90
Green 247.05 0.20 0.64 77.62 247.05 59.19
Blue 54.21 0.14 0.08 98.86 54.21 544.61
White 385.48 0.30 0.29 392.26 385.48 551.96 as measured
427.63 419.96 609.70 white as calculated
0.92 0.92 0.91 Ratio of measured:calculated should be 1.000

In both cases, the displays are showing some inconsistencies.


Hello,

Thanks again for your replies and the basic test with explanations as
well.

I shall look further into the panel specifications to find out if
there is any inbuilt color model. I do suspect that the reference
panel has some sort of dynamic control that i've been unable to switch
off.

It also seems that i would have to repeat my measurements again. I'm
not too sure about the flaring affecting measurements. i've found a
reasonable repeatability with the measured values, which i believe,
will not happen unless the flares affect the measurements all the
time.

Thanks again for all the help.

SR

.

Relevant Pages

  • Re: Screen profiling
    ... I measure a black patch printed on paper, the spectro is set to D65 ... so as to match the XYZ value mesured in 1) on paper. ... how can I "transform" the measurements in order to ... but I will also to profile LCD screens. ...
    (sci.engr.color)
  • Re: Color matching
    ... The first test is to see if the RGB XYZ data sums to the measured white XYZ data. ... This implies that the signals can be thought of as additive...which a display should be for the white point. ... I would make sure that I could see the difference between 192 and 255 and I would make sure that I could see a real difference between level 1 and level ll before I made any more physical measurements. ... Does the panel have some kind of color model built into it? ...
    (sci.engr.color)
  • Re: 8ms Timer for serial port access
    ... >which is done depended on the measurements I take. ... The OS's serial port hardware buffers ... You can probably display your ... The first bit of every byte is a frame control bit. ...
    (comp.os.linux.development.apps)
  • Re: PowerPoint Centimeters Different from Actual Centimeters
    ... The measurements used in PowerPoint are more of a reference for relative ... sizes during the design process rather then an actual display size. ...
    (microsoft.public.powerpoint)
  • Re: Solar-Heated Workshop (more)
    ... | How effective is the type 3 panel at trapping air flow? ... The other measurements of interest are full-sun noontime flows of ... The best "instrument" to ascertain air flow is a 1/2 ... it is very easy to see whether the hoods ...
    (rec.woodworking)