Which color temp lights to use to view prints.

[digitaldog]

The idea that using a CCT 3500K viewing setup and then trying to match it on a display will always result in muddy yellow is a huge generalization mostly based on the behavior of very old, CRT's where you'd be hard pressed to get more than 100cd/m2 out of em even new.
Or that modern LCD's don't have superior backlight technology well beyond CCFL let alone what was found in CRTs. One advantage of some of the newer GB-R LED technology is more control over setting white point. GB-R LED allows more precise control of whitepoint (vs. CCFL). Some may suggest they flicker and maybe that's why they can't get a good match to their 20th century CRT?
Then there's the idea that you can and should calibrate a display or a lamp to D50 or similar. Measure the results with a Spectrophotometer and show me it matches actual daylight SPD.
Then there's the idea that a bulb labelled 5000K actual produces 5000K on the exact area the lines of correlated color temp state it should exactly lay.
Then there's the idea that once you match your display to it, it's measured WP also exactly lines up with the above.
Then there's the idea that you can't or shouldn’t use a differing CCT light course or even a Fluorescent booth next to the display, get that visual match to the print then hang the prints anywhere you desire with a totally different light such as the CCT 3500K Solux and there's going to be an issue (no, there isn't; our eyes adapt to the illuminant).
What's clearly wrong IMHO, is a blanket statement that CCT 4700K or 5000K Solux bulbs is always the right answer. For display to print matching. There are plenty of Solux customers who will tell you this fact.

[digitaldog]

Apparently, Doug Gray and I didn't get the memo; we both find that matching the monitor to 3500K lighting stinks.

Where did he specifically say that and unlike you, unable to produce a match to a print and specifically what kind of display and bulb?
Maybe with that old CRT you use that presumably doesn't flicker, you're unable to produce a match. But that in no way implies it's not possible; it is. Just not for you it seems.
You of course once again failed to read the OP or the facts:

other well respected photographers like John Paul Caponigro recommend using SoLux 3500K for proofing,

No, you didn't get or read the memo.....

[Daverich]

Apparently, Doug Gray and I didn't get the memo; we both find that matching the monitor to 3500K lighting stinks.

That’s certainly been my experience.

[digitaldog]

Reply #22: "For that matter, ever try profiling your monitor to 3500K? It looks pretty horrible!" It's possible, but it stinks.

I profile to a value that matches the illuminant to the display not some fixed value that doesn't and which defines a large range of possible colors. Maybe that's why for you, it looks horrible. The video referenced that was helpful to the OP might be helpful to you but its fact filled so I doubt that will be the case.

[digitaldog]

That's exactly what I do. When I use 3500K and adjust the monitor I get a match, but the monitor looks grungy and yellowish; when I use 4700K or 5000K I adjust the monitor and get a match, the monitor looks good in both cases, but I prefer the 5000K setup.

Of the 10 people participating in this thread so far, 3 hate the looks of the monitor when matched to 3500K lighting and we don't know if the others have even tried to do that.

And more don't which dismissed your statement "definitely 4700/5000K". For you. Definitely.

[digitaldog]

You don't know if more don't. Critical thinking! Three out of maximum ten don't like it, but according to you it is utter nonsense.

Never said that. The utter nonsense came from your own keyboard starting with "definitely". That isn't correct for everyone. Think about that critically if that's even possible.

[digitaldog]

Reply #17:
Quote from Frans Waterlander:
"The OP clearly stated that he wants to match the monitor to the print viewed in the digital darkroom and in my experience the 4700K or 5000K give a way better match than lower values; 3500K for instance forces me to lower the color temp of the monitor so much that I end up with a dingy yellow looking monitor."

Reply by digitaldog:
"Utter nonsense. As other's have reported. Myself, Art Wolfe, JP; no such issues."

Yes what you wrote is utter nonsense! I told you in post #35 READ!

[digitaldog]

Reply #35 is mine.

Lost again:

[Doug Gray]

I don't know what universe or time warp you are in, but in my world reply #35 reads as follows:
---------------------------------------------------------------------------------------------------------------
Reply #17:
Quote from Frans Waterlander:
"The OP clearly stated that he wants to match the monitor to the print viewed in the digital darkroom and in my experience the 4700K or 5000K give a way better match than lower values; 3500K for instance forces me to lower the color temp of the monitor so much that I end up with a dingy yellow looking monitor."

Reply by digitaldog:
"Utter nonsense. As other's have reported. Myself, Art Wolfe, JP; no such issues."
--------------------------------------------------------------------------------------------------------------

Aside from the numbering issue that you apparently have in your neck of the world, you clearly called my statement utter nonsense, so man up and admit it. Two other posters agree with me, so I guess that makes their inputs also utter nonsense.

I've noticed on other threads that some posts references seem off by one. I wonder if there is some problem in the way different browsers render the site?

Here's what post #35 looks like on my browser:

[digitaldog]

I've noticed on other threads that some posts references seem off by one. I wonder if there is some problem in the way different browsers render the site?

Here's what post #35 looks like on my browser:

Could be but don't let the vortex of confusion from Franz use it to digress from what he wrote and my screen shot shows; another tactic seen from him that doesn't protect him from what he incorrectly writes here. 

[JRSmit]

I used to work with Solux4700, but now several years i work with fluorescent tubes of Philips for my print-studio, the ones they make specific for the graphical industry. The main advantage over solux is the even illumination of my observation table. I can now look at a full size print, instead of only a small portion due to the limited area of illumination with Solux 4700.

I also tested a JUST box, but whatever i did, the light of the JUST box was a bit purplish to my eyes. As i did not experienced this with my Philips fluorescent lights, i returned the JUST.
The intensity of my EIZO 248-4K is adjusted so that the print on my observation table w.r.t. lightness appears the same.

Since a year now i also have the white point of the monitor adjusted to visually match with a paper of a known whiteness, without any OBA's, but a close to proper white as possible. I ended up with 5300K for my monitor, yet measured with CT&A the light on my observation table is ~5000K.
Obviously spectrally different, but now i can interact with my customers on the common issue " why is the print too dark" and color differences.
Much to the pleasure of my clients, finally having a clear benchmark. (note i care less about the numbers measured, i care only the visual match)

For papers with different white-point, and no OBA's, i use the softproof in Lightroom or Photoshop if needed to see if it still matches the print.

My experience is also that when the print matches what is shown on my monitor, it will also look fine under lighting conditions one can expect in living rooms or galleries.
With respect to color temp of light and how we as humans perceive this, just look up "Arie Andries Kruithof" , I have added a slide of my workshop presentation, be it in Dutch, but the essence is clear. To this date i have not found research that contradicts this phenomenon. And serious museums do know an use this study.
In simple words it shows the capability of our visual system in perceiving white as white under different lighting conditions. A truly remarkable capability.

[digitaldog]

My experience mimic's JRSmit's closely as I use Solux, Just and a GTI booth depending on the needs. With the GTI which we are told is a "D50 viewing booth", CCT 5150K works out to produce a good visual match on my SpectraView using specific software, and instrument along with specific other calibration targets. I like the SOFV-1e not specifically due to the spectrum (it's like all Fluorescent illuminant's, spiky) but because I can set the intensity digitally and I can collaborate with other's using the same booth and display remotely without any ambiguity in how to produce a match IF all the other attributes are specifically defined. It runs cool, unlike Solux. It's consistent in terms of how bright it appears where the Solux varies of course depending on the distance from the print being proofed. It's far more energy efficient. But it's not ideal in other respects. But there's no issue getting a very close print to display match with good ICC print profiles (their soft proof table plays a role) and again, the numbers that work for that match, far more often than not, do not match the numbers of what is 'reported' by a manufacturer for the light source. The same is true with Solux and Just; the numbers claimed are not the numbers measured, the numbers to produce a match to a display more often than not, don't match the calibration targets used for a match. YMMV. No one can or should state what values to use without exact specifications for the booth, it's setting if possible, the display and it's back light technology, the software used for calibration, the instrument used for that calibration and so forth. Someone with the same brand/model of display, booth, colorimeter, and specified settings can match very closely what I see (ambient conditions and the surround play a role too!). Otherwise, take with a large grain of salt anyone who tells you want to set without such specifics. IOW, 'definitely use this' without such specifics is ambiguity, assumption, and just bad advise.

[Doug Gray]

I used to work with Solux4700, but now several years i work with fluorescent tubes of Philips for my print-studio, the ones they make specific for the graphical industry. The main advantage over solux is the even illumination of my observation table. I can now look at a full size print, instead of only a small portion due to the limited area of illumination with Solux 4700.

I also tested a JUST box, but whatever i did, the light of the JUST box was a bit purplish to my eyes. As i did not experienced this with my Philips fluorescent lights, i returned the JUST.
The intensity of my EIZO 248-4K is adjusted so that the print on my observation table w.r.t. lightness appears the same.

Since a year now i also have the white point of the monitor adjusted to visually match with a paper of a known whiteness, without any OBA's, but a close to proper white as possible. I ended up with 5300K for my monitor, yet measured with CT&A the light on my observation table is ~5000K.
Obviously spectrally different, but now i can interact with my customers on the common issue " why is the print too dark" and color differences.
Much to the pleasure of my clients, finally having a clear benchmark. (note i care less about the numbers measured, i care only the visual match)

For papers with different white-point, and no OBA's, i use the softproof in Lightroom or Photoshop if needed to see if it still matches the print.

My experience is also that when the print matches what is shown on my monitor, it will also look fine under lighting conditions one can expect in living rooms or galleries.
With respect to color temp of light and how we as humans perceive this, just look up "Arie Andries Kruithof" , I have added a slide of my workshop presentation, be it in Dutch, but the essence is clear. To this date i have not found research that contradicts this phenomenon. And serious museums do know an use this study.
In simple words it shows the capability of our visual system in perceiving white as white under different lighting conditions. A truly remarkable capability.

Good approach!

This is similar to what I do. Basically, I have two approaches. One for all papers w/o OBAs OR Papers that will be viewed with little uV content. But another for papers with OBAs.

My preference is to start with D50. In the first case I initially profile the monitor for D50 and set the luminance to match the view box but slightly higher. Then, using soft proof using a uV cut paper profile with show paper white selected, I carefully adjust the RGB curves so the tint matches the view box. Then I use Bruce Lindbloom's site to calculate the xy coordinate and plug that into the monitor profiling s/w and re-profile the monitor. This now gives a good visual match to the paper white using soft proof and requires no changes at all with different papers/printers. However, this only works with u/v cut spectros and papers w/o OBAs. This is my main approach where I prefer paper w/o OBAs.

The second approach works with anything and is pretty much the same as Andrew describes in his excellent videos. I  deselect show paper white in view proofing and adjust the xy coordinates and luminance to match the specific paper being viewed. This is easier to do and works with OBA papers but has to be done with each paper that has significant differences in L* max, white points, or uV response.

I have done some work using Luminant A (2800K). It takes a bit to get used to the low color  temp and is best done taking efforts to make the viewing box and display look as close as possible to each other.  One thing I did that seemed to work reasonably well is to take a picture of the view box and embed the working image as a layer inside that image. It kind of fools your mind so you don't see the monitor in quite the same way. Adjusting as above provides a reasonable paper white match. However...

Turns out there is another difference that impacts side by side matching at low CTs. The monitor profiling software adapts colors to illuminant based on the Bradford transform. The print, OTOH, doesn't know about transforms so just reflects light based on the characteristics of the CYM ink. There is some divergence as the CT get further from D50. I got some improvement by using I1Profiler's ability to use non D50 spectra when creating profiles. I1P also uses a Bradford transform and produces a somewhat better match.

That said, there aren't many reasons to be hard/soft proofing with low CT illuminants and it's something I've done more out of curiosity than anything else. My preferred operating point is D50 with xy adjusted to match the white of my view box. This typically comes out close to something along the 5000K CCT line when in mode 1 (using show paper white) and a bit higher for most papers that have negative b* when in mode 2. Mostly I stick with mode 1 since I will always used OBA free paper where possible.

[JRSmit]

Andrew, you give me too much credit ;-) I followed your videos and explanations, which i thank you for.
It works!