- This topic has 54 replies, 6 voices, and was last updated 9 months, 1 week ago by
.
-
Posts
-
… and use the white patch from a colorchecker to have a D65 ref ?
In fact I plan to buy a 6500K Yuji Sunwave to have a D65 ref because the spectrum is rather good.
I’m a little bit exhausted by finding the right observer and trying to find a CCMX procedure. The only observer that can make my 2 monitors closer in term of whitepoint is the CIE 2012 10 deg but if I tailor the CCMX to this I get inaccurate reds. I used the CIE 2012 2 deg and got near perfect results for primaries (Full procedure in attachement)
My last resort is to take the Yuji 6500 Sunwave and also the Normlight on to get a 5000K ref. I wanted to know if it’s a good way to avoid metameric failure. I’ll use the CIE 2012 2 deg observer as it is rather accurate in term of primaries.
Guillaume,
I recently investigated the possibility of using a 6500K fluorescent lamp in my workflow.
Traditionnally. I calibrate my monitor to D50 and view my prints under a D50 simulator (aka 5000K) desktop viewing booth made by Just-Normlicht . In the web offset print shop I work, we also use JUST-Normlicht D50 lamps in the overhead luminaires in each press cabins.But I wanted to investigate the desirability of 6500K? So I ordered a 6500K fluorescent lamp from JUST-Normlicht to replace the 5000K lamp in my desktop viewing unit.
Then I calibrate my monitor to D65.
I placed my usual prints in the viewing booth and compared to the same image in Photoshop, side by side.
To my eyes, the match I got was inferior to the match I get with a D50/5000K workflow. But that is MY experience. Yours may vary.
I know of Yugi lamps but never ordered.
I recently got some lamps from a US company called “Waveform”.
Neither Normlicht or Yuji are “perfect” D50/D65 simulators.
Older CCFL normlicht have bumps in peak emissions. Newer multi-illuminant normlicht led have a lot of peaks with a severe cyan depression. IMHO they are worse than yuji for D50 but: a) they are multi-illuminant b) they have UV
Older (8-10 years) Yuji D50 VTC have no UV and it was low on violet-blue (it falls “too fast” to shorter wavelengths, so colorchecker blue vs display may be darker, but otherwise excelent D50 simulator. Newer Yuji leds seems to be more advanced.I cant remember where I read but the traditional D50 is a slightly lower kelvin than average daylight, to a point that D65 might me a closer representation.
That supports what Vincent said that its not only about the kelvin value but also the implementation.
When dealing with strong metameric failure, I tried alternate CMFs with high emd probes. They helped a bit, but not enough.
Still needed to do a perceptual match to make the displays look the same. For white and primaries.
There is no such thing as a “perfect” D50 simulator except in one spectrophotometer made by X-Rite where, to best comply with the M1 requirement, they chose to creste a mighty good D50 light source. I would trust this instrument measurements with my eyes blind-folded.
At one printing plant I worked, where they used Eizo and NEC monitors as “proofing device” on press, even though the calibration/profiling software said “D50 Zero DeltaE”, I still had to edit the profile to best match our set of hardcopy proofs, moving the media white point a* and b* quite a bit, usually pushing the a* towards ‘red’ . And even then, it was far from perfect. It was a trade-off. I prioritize neutral grays and skin tones. We used AbsCol all the time.
Keep in mind, the Colorimetric Standard Observer (cmf) is still an “average” of 17 individuals who were tested for “normal color vision” : there is no saying how far you and I are from the average. And a test like Farnsworth-Munsell is unable to characterize a person’s color vision in comparison to the 1931 Standard Observer.
Have you seen the current price of Yugi D50 T8 lamps?
Standard Illuminant D50 5000K T8 LED Tube Basic (ISO3664:2000)$357 USD for three lamps?
I don’t have pockets deep enough to place an order for a set of these three tubes. I eventually stumbled upon the Waveform 5000K lamps which are quite affordable and, as far as my own measurements indicate, are excellent in all respect, except they are not very bright and not all that uniform.
@Vincent : You wrote “Neither Normlicht or Yuji are “perfect” D50/D65 simulators.”
Keep in mind the industry is not looking for “perfection”.
Only “God”, apparently, has this technology 😉T8 lamps seems to be used in direct replacement of old T8 tubes, thus this are not exactly D50 sources but D50 sources disigned to be mounted on some existing design.
Yuji has “inexpensive” E27 bulbs simulating D50 about 35euro/each.Regarding whitepoint NEC/Eizo whitepoint visual adjustment in HW cal, although it seems Eizo is better since you can apply it post calibration. DisplayCAL has it too.
Regarding color mismatch after visual whitepoint, Eizo has a 6 axis HSL control for CMYRGB, using a physical & digital (better if matching custom masured colorcheker) colorchecker as reference for CMYRGB… although this approach is possible, it is not recomended bc you are moving to an alt D50 distribution that may not match other people setup for print evaluation. But you can get “yours” to match printed & digital just by an small finetune in RGBCMY HSL controls (usually in blue-cyan for old Yuji, which are nit the same as the ones supplied now, IDNK if newer ones need it at all)
Regarding near perfect D50 simulators is just a matter of time, price and market (people willing to pay that amount for a slightly better Dxx simulation than existing one at YYY bucks). We can have this technology right now, this is unrelated to nature/teology.
On multi-illimunant the N-LED approach by Normlicht seems better “jack of all trades, master of none” bc varying gains on each LED color you got “good enough” Dxx simulations at Y price, although better specific single Dxx simulator can be made at lower price, like Yuji or others.
On single illuminant desk-lamp/DIY project yuji E27, modules or reels are better and chaper. On multi purpose booth (wich has to simulate several illiuminants) Nromlicht seems a wiser approach at this time.On an OLED my eyes were seeing a green tint while the spectro & CIE1931 were seeing magenta. Alternate whitepoint I settled on was about dE 7 away from 0.3127, 0.3290.
My reference was a lower end LED. Would have loved to compare it to a CCFL or CRT as well but none were available.
Here is set of spectral irradiance measurement I did with my i1pro of JUST-Normlicht LED viewing booth at Phoenix, Arizona, during a GATF Color Management conference. Michael Gall was demonstrating the booth to the public for the first time in the US.
If you look at the bottom graph, you’ll see two “lines”. The black one is the measurement while the green one is D50.
I also include the measurements, in case you want to do your own analysis…
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?
