So what IS the right white point for a monitor used for printing?

[tony22]

I didn't want to hijack Hening's thread, but I saw postings there regarding monitor color temp and it got me wondering about it again.

I've read through many messages over the years where there seems to be rationale for 6500K (which is what I've been using for a long time), 6000K, 5500K, and 5000K. And I believe I've seen information suggesting that (seemingly for flat panel monitors) there needs to be some consideration for "native white point". I'm not quite sure I understand that, but as I am still on a good old CRT (Sony GDM-FW900) my question has to do with what white point would be "correct" for a display of this type if I wanted it to be optimized for printing on my photo inkjet. I use a range of papers, from satin and matte to glossies with high OBAs. I've settled on using a monitor luminance of 100 cd/m2 to come close to the papers which have the highest brightness values.

I do understand that the answer in part has to do with the expected viewing conditions for the prints. In my case that would almost always be daylight lit rooms, no explicit directed lighting on the work. Of course in the evening a variety of artificial light might be present, but I'm targeting daylight viewing really.

[digitaldog]

The correct WB target is the one that produces a visual match to the print next to the display. Otherwise the numbers are pretty meaningless.

The only way to get 6500K from a display (and keep in mind, that’s a range of colors) would be to heat to heat it until you had a molten pool of plastic on your desk <g>.

Native WP is useful when you get a visual match to the print or you are working with an 8-bit per color path and excessive banding occurs calibrating away from this native WP.

[tony22]

Hmm, but isn't "daylight" supposed to be in the 5000-5500 range? If so, would it make sense to have the display adjusted to this range to better match the expected viewing conditions?

[digitaldog]

Hmm, but isn't "daylight" supposed to be in the 5000-5500 range? If so, would it make sense to have the display adjusted to this range to better match the expected viewing conditions?

First off, its a range (a pretty big range). Second, what counts is a match. So the question then becomes, what is the print viewing conditions?

[Adam L]

I was curious about the answers to this question.  My thought was that the white point should match the paper used for the print?

[tony22]

First off, its a range (a pretty big range). Second, what counts is a match. So the question then becomes, what is the print viewing conditions?

I guess my original description couldn't be clear enough because I indicated I'm targeting daylight viewing conditions (I can see where that's a little vague). So to me that would be (normally) indoor viewing in a room lit by natural sunlight (through windows and skylights). I appreciate that brings a lot of variables into the equation (maybe even the color of the walls!), but I'm hoping at least in crude terms that would mean "no that's not really 6500K, it's more like xxxxK" or something like that. I'm looking for just enough guidance to target the best probable starting point.

I suppose I could pull out my old Minolta light meter and read the color temp of the incident light in the room, but I'd have to find silver oxide batteries to power the thing!

BTW, assume the walls are in the eggshell white category of color.

[Czornyj]

Measuring the CCT (Correlated Color Temperature) of the ambient light won't help at all. There's a disconnection between the appearance of print under light that has certain CCT and a display calibrated to the same CCT. It's complicated and has something to do with light UV and OBA paper content, different viewing condition and so on - anyway there's no simple rule here. Personally I find calibrating a display to D65 a good starting point, EIZO recommends D55, UGRA recommends 5800K - and so on.

[digitaldog]

I was curious about the answers to this question.  My thought was that the white point should match the paper used for the print?

There’s more to this match than just the paper, there’s a light source and all the interactions between that and the paper. The value that produces a visual match is the correct value.

[digitaldog]

EIZO recommends D55, UGRA recommends 5800K - and so on.

Right which is part of the issue users have. They believe that the recommendations are correct and its kind of silly to provide such recommendations when two important pieces of the puzzle are totally unknown: the paper and the illuminant. The same thing happens with luminance setting recommendations (set your display to 120cd/m2). If the print illumination is totally a mystery, what makes that value a good recommendation? If the manufacturers would refrain from this, users would be better off but that’s water under the bridge after so many years of making these recommendations.

[tony22]

Well it's certainly a complex issue (to me . Let's take a very specific example. I have my display calibrated via i1Pro to D65, gamma 2.2, luminance 100 cd/m2. I used Atkinson's 1728? (memory escapes me at the moment) patch chart to generate an i1Match profile for Epson Photo Paper Glossy. Used Perceptual to print through CS3. After allowing sufficient time to dry, this paper shows a slight magenta cast in neutral mid-luminance grays when viewed in a daylight lit room. So, assuming I didn't do anything wrong in the profile creation workflow, let's say I wanted to set up a print workflow so that the grayscale perception of this paper in that sort of viewing condition was improved. Wouldn't that include a consideration for adjusting the color temperature of the monitor? If so, is downward (warmer) the right direction?

[Czornyj]

Is the paper white similar to the monitor white? Try to cut the monitor gain for green channel to reproduce pink cast of the print - can you achieve visual match? If the pink cast occurs only in part of gradation, then it's rather a matter of profile. How long did you let the print dry? Or maybe you scanned the targets too early? Try to scan the pinkish part of the grayscale with the spectrophotometer - what L*a*b values do you get?

[tony22]

I'll have to address some of this (including the white comparison) when I get home. But in the meantime...

How long did you let the print dry? Or maybe you scanned the targets too early? Try to scan the pinkish part of the grayscale with the spectrophotometer - what L*a*b values do you get?

I'm using a Canon Pro9000 dye sub. I normally let the target prints dry at least 36 hours before scanning to generate a profile. And the test print I used afterwards was dried about 48 hours before I took a close look at it. I haven't done a spot read on the mid-gray values yet.

[tony22]

Okay. When held up next to the monitor showing a white field, the Epson paper visually appears to have a slightly warmer tone to it. I will measure it in a moment. BTW, my office is lit with daylight Ecosmart 5000K bulbs, which is ironic because I thought when I bought them they were the 6500K bulbs.

Using the Spot Measure Tool, when set for D50 Illuminant, the gray tone reads 64, 2.1, 0.7. Selecting the D65 Illuminant, it reads 64, 0.8, 1.2. The a and b values read in PS are 0, 0 for the same area, although when using Soft Proof and the correct profile the image appears on screen to be slightly green in the neutrals. The Epson paper white reads... 97.2, 1.3, -1.5 with D50 selected, which I guess might explain what I thought was a bit of a warm (reddish) tone.

BTW, the color shift appears to be not just through the grays, although it is very slight.

Hmm, in the past with my Spyder I would never touch the RGB Bias controls on my display, since the Spyder software didn't resolve the calibration changes too finely. While I was using i1Match I noticed the software's RGB bars could resolve changes much more finely, so I used the Gains and the Biases to make very fine adjustments. The R and G Biases are both set to 50 (default), but Blue is at 49. Maybe I should not touch the Bias values. I've heard this is generally not recommended or needed. The Green Gain is already set about 8 points lower than Red and Blue, but my display measures out quite well above 20% luminance. R, G, B chromaticities are near dead on from 20-90%.