- This topic has 54 replies, 6 voices, and was last updated 9 months, 1 week ago by
.
-
Posts
-
In fact I asked because I wanted to have a reference to calibrate my monitor to a near perfect D65 whitepoint. Then if I want to set another color temperature I can use f.lux. In fact using f.lux set the goal of monitors and room lighting match. Using a 5000K/5500K monitor white point tends to be suitable for 2700K/3000K room lighting. Using 6500K after looking at a 2700K/3000K environement looks very cyan/blue. And also if the background behind the TV/monitor is orangish the grays appear greenish or cyanish where using a 5000K/5500K white point neutralize grays. But I can’t be confident about this cause I have no true D65 reference and Yuji D65 lightbulb seems to look very accurate but violets are missing.
I used to have two monitors on my desk, a NEC PA271W and an Eizo CG21. I could never get them to appear to match at whatever target white points I tried to use. At D50, to my eyes, the NEC appeared “reddish” while the Eizo appeared “greenish”. Same calibrated chromaticity on both.
I was told that the monitor “probably” matched and it was my difference with the Standard Observer that explain the difference?
The first culprit would be their respective calibration software, lacking or not be able to match EDRs to actual display backlight. We have several samples of wrongdoings of SV2 and CN7 in this forum (Eizo is easier to “patch”)
If they were measured with 10nm spectro and Xrite SDK this is another source od error.Here is a Heidelberg XL106 press console with “D50” LED tubes.
It seems multiled like Normlicht. Do these tubes offer some kind of conectivity to tweak gains so simulate other illuminants?
Same calibration software used on both units.
Please excuse my ignorance but you use a lot of acronyms. What does “EDR” mean?Same calibration software used on both units.
Then likely to be that software the main source of error = more likely, instrument and user not matching std observer may happen too.
Which software?
Basiccolor should have WG CCFL EDR for PA271W. I’m not suer about spectral power distribution on that CG21Please excuse my ignorance but you use a lot of acronyms. What does “EDR” mean?
EDR is a binary format from Xrite equivalent to CCSS files in ArgyllCMS. They store WRGB spectral power distribution of a sample display. i1d3 colorimeter form calibrite/xrite can correct its measurement with that information when measuring the “same” display model or sibling with close spectral power distribution. Since it stores no colorimeter information they are “portable” between i1d3 while 3×3 matrix corrections (CCMX files in ARgyllCMS, but Calman or Lightillusion have equivalent solutions) are not portable between i1d3.
Nec SV2 software lacks of suitable EDR for PA311D and PA271Q. Eizo CN7 software lacks of EDR for “AdobeRGB green” flavor of WLED PFS backlight, but thanks to community we can inject/replace the proper one. Current CN7 applies an CS2730-like EDR (GB-LED) to all CS and CG models, where newer CS and CG are WLED PFS.
Calibrite Display Pro HL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.Note: there were several widegamut CCFL backlights, the bigger gamut used in PA241/271W and low cost counterparts U2410/U2711H… (and asus equivalents).
But there was a previous version where red was like in sRGB CCFL backlights, providing an efectiva native gamut close to AdobeRGB, without native red beyond AdobeRGB/sRGB.
If that CG21 is widegamut and old enough it may be one of these second ones. There is no official EDR for such backlight although it can be cooked with a text editor, maybe some spreadsheet work and “ccss2edr” python tool.Do you know which backlight uses that CG21? If you own an spectrophotometer can you upload a CCSS @ 3nm (high res mode) of that display?
Heidelberg? Adjust gains? You are crazy! 🙂
IDNK what you were talking about
Roger Breton wrote:
Same calibration software used on both units.
Which software? which configuration?
So I come to tell some news. I recieved my 6500K Sunwave and 5000K Normlight some weeks ago. I dimmed the 6500K lightbulb and used Whitescreen android app that had rgb sliders and I matched the rgb values to the lightbulb. I made a visual whitepoint on Displaycal to match the Whitescrren one and I got an xy coordinate of : 0.3203 ; 0.3524 with the 2012 2deg observer. I have the feeling that this observer is too sensitive on the green component leading to a too magenta calibration. LED backlights are gaussian curves so it’s somewhat strange that I get different results between the lightbulb and the monitor. I think current observers are done wrong in my humble opinion. If it was a spiky spectrum like a CCFL backlight I wouldn’t be surprised.
So I come to tell some news. I recieved my 6500K Sunwave and 5000K Normlight some weeks ago. I dimmed the 6500K lightbulb and used Whitescreen android app that had rgb sliders and I matched the rgb values to the lightbulb.
Variation in xy vs Voltaje or current (dimming) as many of these LED spec show
+
1st statistical mismatch a random human observer vs WhiteLED / AMOLED / variety of LED widegamut on direct light from bulb!!! vs screen1
+
“user error” adding some extra to some sliderI made a visual whitepoint on Displaycal to match the Whitescrren one and I got an xy coordinate of : 0.3203 ; 0.3524 with the 2012 2deg observer.
2nd statistical mismatch a random human observer vs WhiteLED / AMOLED / variety of LED widegamut on screen1 vs screen2
+ “user error” adding some noise when picking xy
I have the feeling that this observer is too sensitive on the green component leading to a too magenta calibration.
The pipeline of your test has enough statistical errors in the chain to even consider, if certain std observer is a very good statistical mean for humans or not.
LED backlights are gaussian curves so it’s somewhat strange that I get different results between the lightbulb and the monitor.
Depending on the LED hillside ofr each peak, moved in wavelength to the right or left vs the hillside of observers can cause this. Both on your eves and on device (small innacuracies of i1d3 frimware data or small displacement +- wavelength on spectrophotometer + “averaging” to 10nm)
+ all statstical errors accumulated in our test.
I think current observers are done wrong in my humble opinion. If it was a spiky spectrum like a CCFL backlight I wouldn’t be surprised.
I do not see that way, as explained.
In fact the TLDR would be that our eyes work like a camera where 2 light sources of the same measured xy CCT will not look like the same. So which D65 source should I take as reference ? The Yuji sunwave or none ?
In fact I double, even triple, checked if the color slider were accurately set. I bought 2 lightbulbs and they were the same.
None unless you can measure it (for example with Babelcolor CTA now that is free and an spectrophotometer)
Usually the Yuji products aiming to simulate some CIE D65/D50 state a set or chromaticy coordinates (not CCT) and I just look for it in Sunwave 6500K and this information is missing.
Also these Yuji products aiming to graphics arts in their spec they show how they are expected to vary in xy when you change V or I… and that is mussing to.
You bought a more simple product than the ones sold by Yuji for the task you want.Hence these D65 bubs could be not better D65 (as xy coordinates, not as SPD) match that a 6500K preset on some random monitor.
If you wanted a near true D65 simulation from Yuji, you should have got one of these products, like this:
https://cdn.shopify.com/s/files/1/0344/6401/files/P3230007-2.0.pdf?v=1709793291
but it’s very expensiveor these oens which are afordable, but they sold only in 2800K or D50
https://cdn.shopify.com/s/files/1/0344/6401/files/Yujileds_NormLite_D50_5000K_LED_bulb.pdf?v=1706695293As any manufacturer Yuji product have “segments”, we see the same on cameras, photo editing monitors, laptops… the D65/D50 ones are more expensive.
Anyway, since those sunwaves should be a “good” 6500K daylight simulators and I do not remember if you had an i1Pro-equivalent for babelcolor, measure it with ambient light cap of an i1d3.
Get ambient light xy and try to aim in DisplayCAL to that alternative white… but on ambient light i1d3 is running without an external CCSS for that SPD so measurement xy may vary from actual ones.As I have an I1 Pro 2 I can use it with a laptop I have at hand and use HCFR in high res mode to test. I also bought two D50 Normlight lightbulbs
Try to evaluate with Babelcolor CT&A
https://babelcolor.com/download.htm
Using “spotread -a -x -T” provides Ra info and CIE xy, but now that CT&A is freeware you may find this more helpful to evaluate other aspects of their SPD
