Join Date: Aug 2006
Location: Hobitville, New Zealand
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 64 Post(s)
Easy to test for differences between a profiled and un-profiled meter. The issue is will you see a difference let alone measure it.
It would require a reasonable amount of data.
you need to look at the accuracy and precision, however because of the influences of display variability and possibly source variability accuracy is confined to the noise of the system and ability of the meter being tested.
Select your reference meter, run the simple color checker routine. output the data to a spread sheet. repeat until 5~6 sets of data.
Select the colorimeter Un-profiled, run the simple color checker routine. output the data to a spread sheet. repeat until 5~6 sets of data.
Select the colorimeter profiled, run the simple color checker routine. output the data to a spread sheet. repeat until 5~6 sets of data.
you could start with the primary and secondary colors, but I'd advise the color checker to look at internal color volume and thus avoid anywhere where the display could clip.
The concept of 5~6 sets of each color and greyscale is to look for repeatability, 5~6 sets is the minimum to be statistically relevant, which is under the umbrella of accuracy and precision.
We would first look for precision
Avoid dE as this adds another equation, just use raw xyY, u'v' or XYZ.
First plot the reference data into a xy chart, look for clumps of 5 readings.
If they are tightly grouped you have reasonable precision to work with, you should find each colors standard deviation. You kind of need a 3 d chart for xyY.
But you can run average and standard deviations on x,y and Y, XYZ, also plotting each component individually you could identify precision differences.
Then repeat with the Un-profiled data and profiled data.
Then compare the data against the reference, which is closer?
If the clumps of data points over run each other you can get close to no difference, especially if the average of the data point plus the standard deviation crosses into the other meters average and standard deviation.
Masterpiece Calibration Ltd