Untitled Document

Acording to the ISO 3664 the conditions to observ colors in prints are a color temperature of 5000K. For screens, 6500K. ¿This means if you want to COMPARE colores beetween print and monitor the hall should be 5000K and the screen 6500K or it is about the conditions to see independently? It is. To work with the computer, 6500K, to work with prints, 5000K. ¿It is not a problem try to compare to versions of the same image with different color temperatures?

As I understand, if I want to compare beetween print and screen I should adjust both color temperatures at the same value ¿Why not?

;Paco Rosso. www.pacorosso.com

___________________________________________________________________________

Paco,

I have been playing around with this for years. I started with 6500K since I wanted to use the Web and Windows standard. Personally, I never got a good color match until I went to 5000K on my monitor. At 6500K, my prints looked very yellow (actually, the monitor was too blue, which led me to add more yellow to the image when editing, and these numbers ended up in the printed file). All the papers I print on (including proofs for press) measure between about 4750K and 5200K. Press sheets tend to be around 4900-5000K or so.

I know others like 6500K, but I don't, at least for a good monitor to print match.

Lou

___________________________________________________________________________

Prints should be viewed with light with a color temperature at or around 5000K because they are intended to be viewed in sunny daylight, which has a color temperature around 5000K. Monitors are typically set for a white point of 6500K -- if you set your monitor for a white point of 5200K, the images will look yellow.

The idea is not to view everything with the same color temperature or white point, but to view each thing with the appropriate color temperature or white point. For example, when shooting video under tungsten light, the color temperature (white balance) of the camera should be set around 3200K (same goes for film -- film balanced for tungsten light is intended to be exposed with light having a color temperature of about 3200K). When this video is viewed with a monitor, the monitor's white point is set for 6500K, not 3200K, in order to display the proper color balance. If you set your monitor's white point for 3200K, the images will look orange.

Randy Hoffner

___________________________________________________________________________

Randy-

I'm not sure if this is answering the initial question.

If I am viewing a print under 5000K lighting, and comparing that print against a monitor, it seems to me that the white point of both must match. If the monitor is set to 6500K, the print will appear yellow by comparison.... no?

If anything, shouldn't corrective work be done on a monitor having a white point set to the temperature of the environment in which the printed output it will be viewed?

I have approached it this way for years -matching monitor temp to viewing environment temp- which seems to give me an accurate softproof of the printed output. Louis Dina made this same point in a prior post.

I understand that our eyes adjust to various white points - essentially forcing white to be white. But how does one reconcile the difference when presented with two different white points at the same time, such as 5000K illuminating a print and 6500K coming off a monitor?

Where am I going wrong here?

Les De Moss

DigiGraphics LLC

___________________________________________________________________________

Wrong? I'm not sure you are. Many, but by no means all, people experience a better match to a 5000K lightbox with their monitors set to 6500. If you are going wrong anywhere it is only in assuming that what works for you and your set-up should work the same for everyone. Logically it would seem that it must, but from my own experience and in listening to others I've found that it doesn't. I know some people who like to set their monitors to 6000K. I'm happy with 6500. If I set my monitor to 5000K it looks too yellow to my eyes.

IMHO its partly a matter of expectation. By this I mean that a screen and a print can never *match* and a lot depends on what kind of compensation our minds expect to make.

The scientific explanation I've heard is that our perception of color changes with the amount of light, and the dim light from a monitor tends to look yellower to most peoples eyes than the relatively bright light from a 5000K viewing light.

It's one of those gray areas which makes me think of calibration as more of a personal preference than many would like to admit. My 2¢.

Ric Cohn

___________________________________________________________________________

I think the thing that's missing here is the luminance level of the display. IN GENERAL the higher the luminance the monitor is capable of, the closer you can get to calibrating it to 5000K and have it appear "daylight white".

Much of the (outdated?) recommendation for 6500K was from the days of CRT displays that ranged in luminance from about 80 to perhaps 120 cd/m2. If you tried setting the display to 5000K at around 100 cd/m2, it would appear decidedly yellow. At such a low luminance level, 6500K simply looked "whiter" or closer to daylight. Now days we have LCDs that start around 120 cd/m2 and that go over 200 cd/m2. At that sort of luminance, 5000K looks much better. At the end of the day, you simply need to play with the color temp until you get appears to be nice "daylight" white compared to your 5000K viewing environment.

Regards,

Terry Wyse

_____________________________

WyseConsul

Color Management Consulting

G7 Certified Expert

704.843.0858

http://www.wyseconsul.com

http://www.colormanagementgroup.com

___________________________________________________________________________

Yes, I would largely agree with this. Although I may be able to get used to it eventually, a monitor set for a white point of 5000K looks too yellow to me, and one set for 9300K looks way too blue. I tried calibrating my monitor to 5200k once, and I just couldn't stop seeing it as yellowish, even when using a word processor. Preferences vary, of course. Consumer TV sets in the US have white points around 7000K. In Europe, they prefer closer to 6500K. In Japan, consumer TV sets are about 9300K.

On the print, the inherent color and reflectivity of the paper, as well as the characteristics of the illuminant, determine what's white. If the paper is perfectly reflective, and illuminated with a relatively continuous spectrum of 5000K light, as opposed to, say, a spiky fluorescent bulb, then the areas without any ink coverage will appear to be white. We well know, of course, that not all inkjet paper, or darkroom photographic paper, is really white or perfectly reflective. Some inkjet photographic papers, and some darkroom photographic papers, look pretty yellow when viewed with 5000K light, and some look pretty blue. Some look white. What happens in the inked areas depends on the colors and color purities of the inks. (Likewise for the phosphors or RGB filters of the monitor, but we are presuming that the monitor was calibrated with a probe.)

For me, looking back and forth between a print illuminated with 5000k light and a monitor set for 6500K (presuming that the monitor is not contributing to the illumination of the print, and that the 5000K illumination is not reflecting off the face of the monitor) gives me a good basis for comparison, I think, personally.

Randy Hoffner

___________________________________________________________________________

I agree about expectations, and the mental translation we make when anticipating how an image will appear in a different form. We do that from viewfinder to film or capture, from film to-scan-to monitor, and to print... almost without thinking. In general, we know what to expect.

I will say this, I *prefer* the appearance of a display set above 5000K (in fact mine is set at 5600) and would do my corrective work in a cooler environment if it weren't for the visual discrepancy when viewing/comparing printed output.

Perhaps part of the reason this works for our particular workflow is due to the fact that our clients often request to see a comparison between monitor and print, and frankly, we have less explaining to do when the temperature is close.

As far as luminance levels, we use variable intensity GTI light boxes in order to key back brightness to approximate monitor levels. Perhaps that's why color perception based on luminance is less of a factor in our environment.

Les De Moss

DigiGraphics LLC

___________________________________________________________________________

I think monitor temp is one of those things that people will never fully agree upon. There are so many factors at play....room lighting levels, type of monitor, luminance level, gamma, etc.

In my environment, (ambient light level, viewing booth, monitor, etc), calibrating to between 5000K and 5500K always gives me the best match. If 6500K works for others, who can argue? Maybe their variables are totally different.

I will say this though....I always trust a good custom printer profile a LOT more than I do any monitor profile (not that they are perfect). The paper determines the brightest white and the color of that white. There is no backlight, no RGB phosphors, no ambient light issues, or other variables that WE get to choose. The spectro just reads the data. The colors and densities just fall into place based on reflectance. So, calibrating your monitor to settings that match the print is the name of the game, at least where a good monitor to print match is concerned.

Lou Dina

___________________________________________________________________________

Paco, it’s a good idea to remember that rules are intended as guides, not laws to be obeyed. For my own work I neither know nor care about my monitors’ temperature settings. They’re whatever the manufacturer deemed acceptable for general use. If, in my own judgement, the monitor image isn’t close enough to the printed image, I change the video card settings until the monitor image does look close enough to the printed image. Then I correct the image files until the appearance on the monitor is satisfactory. Because I deal almost exclusively with fine art and the artists who produce it, I do have to make further test prints until the printed proof meets with the artist’s approval. The best, most accurately adjusted monitor in the world is not going to produce an image that is guaranteed to result in a print that the artist will accept even when the artist has tentatively accepted the appearance of the image on the monitor. As all of us know, it just isn’t possible to get a perfect match between the two. My other problem is not with color temperature settings but with the observed fact that Epson pigments especially can look quite different when viewed in incandescent light vs. outdoor light (outdoors it really doesn’t make all that much difference whether the day is sunny or overcast). If my monitor was set for the proper temperature as decided by a panel of experts, but my prints were no good, I would give priority to the prints. The rules make a good starting point, but should never be considered a final answer. So far as the problem of often noticeable color differences in different lighting conditions, this was also a problem when I had work done by commercial offset printers. I took the samples from the initial prints in a run and viewed them first in incandescent light and then in direct sunlight. The changes I requested from the printers were aimed at a compromise between the two. Using this approach I rarely had any complaints about the final print colors whether they were offset or inkjet, viewed in incandescent light or in sunlight.

But then I’m not as interested in theory as I am in results.

Howard Smith

___________________________________________________________________________

My CRT monitors were always set to 6500K but I had nothing but problems

calibrating after I bought an LCD monitor so I switched to 5400 and the problems just went away. I'm not sure if it's that the lcd's are very blue by nature or if they are just so much brighter. Whatever it is, lowering the colour temperature helped.

Regards, John

John Denniston

www.dirtbikephoto.com

www.johndenniston.ca

www.dennistonphoto.com (Blog)

___________________________________________________________________________

.

"Howard Smith" wrote:

Thanks Howard, I think that THIS is the starting point. And too often forgotten by the most.

Based on my experience and after a lot of discussions with my clients (photographers and advertising agencies), I think that the parameters that influence the evaluation of an image on a monitor are:

1- the monitor color temperature

2- the monitor luminance value

3- the ambient or the light box color temperature

4- the ambient or the light box total amount of light

5- the gamut and color rendering of the monitor.

The white point of a monitor is something that one should set at the same value of the ambient color temperature theoretically but most of people (and me too) see the images too yellow under this condition, especially the clients to which talking about color temperature is like talking about the sex of the angels.

Im my office I have an Eizo CG210 set at 5600K, an HP 2395 set at 5900K and on old CRT LaCie 21" set at 5200K to make the whites comparable among all the monitors.

Is this wrong or right?

Theoretically (and according to ISO) it's wrong but in practice it's right beacause the whites looks very similar among them and, the most important thing, they look very close to the white of the paper of a Dupont Digital Cromalin seen under a Just ColorLicht light box that works at 5000K.

About the monitor luminance I know that ISO 12646 and 3664:2000 Standard for Soft Proofing requests a monitor set to at least 160 cd/m2 and I've also heard that increasing the luminance at higher values can allow to set the color temperature to a value much closer to 5000K without having the images looking too yellow.

I'm in the Italian TAGA committe for Soft Proofing and the discussion in these days is just about this. A special committee had made some tests taking some prints viewed under 5000K and compared to various monitors but the result was unacceptable for any of them until someone began to raise the luminance values near to 250/300 cd/m2.

I was not there at that time but the only idea to work with a monitor brighter than 150 cd/m2 makes me laugh: I should wear all the day my sunglasses unless I don't want to become blind in a week. And what about the color cast introduced by the lenses? :-D

Let's come back with the feet on the ground, there are a lot of rules for the safety of workers that simply don't allow this kind of luminance so, in my opinion, the luminance of a monitor should be set around 100-110 cd/m2 maximum to match the print viewing condition of 5000K.

But the amount of ambient light how should be set? I think that once set the monitor at 100 cd/m2 anyone should adjust the dimmer of the light box until it gets closer to the luminosity of the monitor. This makes sense but, I know this is not according to the ISO 12646 Standard... (shhhhhh...)

Increasing the luminance of a monitor makes the image brighter, the colors more vivid and the shadows lighter but not gray. Increasing the amount of the viewing light simply makes the print brighter because it reflects more light. That's all. This is why I prefer to set the monitor luminance at an acceptable level for working many hours without any eye's disease and then adjust the luminosity of the light box close to the monitor.

Last but not least, what I've found in my experience with ALL the kind of monitor is that when you calibrate any model of them and after having set a good value for the luminance and for the color temperature, the colors never match acceptably.

And THIS is the MOST IMPORTANT POINT.

Look, when I talk about all this I have in my mind this question: I could be statisfied, but I'll work every day with this monitor, after some days I won't simply notice any difference or better I will expect a certain known printed result compared to my monitor. What could think a client involved in a complaint about the colors of my image that comes in my office and see the monitor and the print for the first time? Could accept the difference that I reatained satisfactory?

For a global evaluation, after I've calibrated my monitor, I use some Digital Cromalins with certain (known) characteristics. One print has a sort of cyanish grays with a model standing on a sofa in a very illuminated room with a cold color cast, one print has a model with desaturated skintones and a brownish gray background, one print is a beauty shot full of colors with fine skintones and another one is a black&white photo with a full range of grays printed with an Heavy GCR.

Of course I have made the printer profile of any print so I have the correct profile to assign at the images when I see them on Photoshop, and the files are the same that were used for the prints. After the calibration none of them looks on my monitor acceptably close to the prints so I have to change the gamut or the gain of the single RGB channels of the monitor to make the things look better. I find very useful the Eizo Color Navigator utility that after these moves remakes the calibration to neutralize the grays and leaving the colors as I wanted. Is this according to ISO 12646 standards? Once again, no. But makes me sleep better at night.

But even in this (good) condition the monitors will NEVER match the print. If I set the colors for matching the girl with the cold color cast, the brownish background of the other model goes on the cold side making all the desaturated but still warm colors looking too cold, or the black&white photo becomes too cold. As I try to match the brownish background of the desaturated model the cold color cast of the other seem to be too magenta or too red and again the black&white photo takes a color cast not equal to the one of the brownish background.

So what?

Calibrating a monitor is like a blanket too small: you can try to arrange it but you know from the beginning that will never cover your body entirely.

And if you do all according to the ISO standards the blanket will become smaller.

I think that it's preferable a better printed result than a perfect match, monitor is only for a general evaluation.

I also agree with Howard Smith when he talks about the metamerism of the Epson ink pigments, I've experienced this especially some years ago when Epson changed from dye-based inks to pigment-based inks. So in this case how can you adjust your monitor for matching a print that will be always greener or will never match the offset print?

People care a lot about the color temperature, the viewing conditions and never about the practical choices like a red wall just besides your monitor or your light box or how a calibrated monitor visualizes the colors or how much time can you stand in front of a monitor set at 200 cd/m2.

And what do you think that a client will look aboveall in a print or in a monitor when a discussion take place?

Will only look how the monitor matches the print. Nothing else.

So, let's make things more practical for a better living and fewer problems.

But please, no one tells this to the International Standard Organization. If they'll hear about my habits they will persecute and kick me off from the Soft Proof Italian Committee...

Have a nice (calibrated) work

Alessandro Bernardi

___________________________________________________________________________

.

As I understand it the recommended practice is actually to make sure there is some separation between the 2 images so you are not viewing them "at the same time". Having this separation allows the eye to adjust to the different white points more easily,

Also if I remember Don Hutcheson's practice correctly if you want the closest possible match you need to have a dimmable viewing box where you place a blank piece of the paper you're matching to in the viewing booth and set the light level as close as you can so the brightness of the paper matches the brightness of the monitor. Then you experiment with different white points on your monitor until you find one that gets the closest to the whiteness of the paper.

It may be 5700 or 6700 or something else. But it will depend more on the paper and less on the theory of white point versus white point.

Dennis Dunbar

___________________________________________________________________________

.

¿A dimable table? if you use tungsten lamps the kelvin will whift. If you use fluorescent.... No, there is no dimmable fluorescents. ¿What fluorescent to see colors? I only know the Osram colorproof. CR 95 at 5000K. Of course, not dimable. Of course, no tungsten lamps dimable if you want to maintein the color temperature at a known value.

The issue, I am afraid, is to stablish a conditions in the studio to work with a consistent color observation system. The ISO says: prints, 5000K, screens, 6500K But ¿What when you want prints TO screen? As I understand: It is not desirable to reduce the color temperature of the screen to 5000K. it is easy to rise the hall to 6500K (you can find fluorescents Osram Lumilux deLux with a color type 965 with stands for CR 90 and 6500K). But, I sould like if it is possible to find any recomandation of ICC consortium, ECI or ISO in that way.

Paco Rosso.

www.pacorosso.com

___________________________________________________________________________

There is a potential hazard in "...whatever the manufacturer deemed acceptable for general use." Many computer monitors have a default white point of 9300K, and this will not be near acceptability for photography. That's probably not the case for the expensive photography-oriented monitors, but it is the case for many general computer monitors.

Randy Hoffner

___________________________________________________________________________

.

Francisco Bernal Rosso:

A dimable table? if you use tungsten lamps the kelvin will whift. If you use fluorescent.... No, there is no dimmable fluorescents.

Dimmable fluoresent viewing boxes are indeed available. GTI makes them, for one.

Les De Moss

DigiGraphics LLC

___________________________________________________________________________

On Dec 10, 2007, at 2:14 PM, Francisco Bernal Rosso wrote:

They won't shift in color temp if you use neutral density filters instead of a dimmer. Solux lamps uses this method to cut the lumens without altering the color temp.

Yep, there are. GTI for sure and maybe Just Normlicht as well. They do it in such a way that the color temp shifts little if any.

Regards,

Terry Wyse

_____________________________

WyseConsul

Color Management Consulting

G7 Certified Expert

704.843.0858

http://www.wyseconsul.com

http://www.colormanagementgroup.com

___________________________________________________________________________

Dear all,

one of the point we should make, and we should all agree with, when talking about softproofing IMHO is what do softproof actually means. Since I've been involved in the same italian tech committee previously mentioned, my colleagues and I have been arguing about it a lot of time in our meetings (as did our colleagues in Germany at ECI). One perspective could be to differentiate both viewing conditions (ambient illumination) and consequently monitor calibration depending on "working situations / accuracy level". When a pressman inspects a printed specimen, he usually does it under a big viewing booth which provides a very high luminosity pressure. Noone would check the matching with the printed proof in a dimmer light condition, so we started wondering: why not try to provide the same (similar) condition when looking to a calibrated display in critical evaluations. Obviously it would be insane to let an operator work with a 300/400 cd/m2 monitor all day long, but further experiments gave great results in terms of visual match between a properly illuminated printed specimen and a full blast (hardware) calibrate monitor. This lead us to consider a bouquet of proper ambient light / monitor setups for different users (photographers, prepress, etc) and different needs.

Few simple rules of thumb in calibrating monitors could be helpful (higher the cd/m2, more close to the 5000K the temperature, etc), considering the only ones that permit hardware calibrations (Quatographic, Nec, Eizo, etc) TFT. I'm not a big fan of manual tweaks, since unlike the CRT they've no actual guns gain to raise or lower, and usually I'm pretty satisfied by the calib/profiling results at different luminosity and temperature. When possible, I've found useful the possibility to choose the chromatic adaptation (deep blues are likely to be better rendered with CAT02) and the raise of black point to higher values to compensate the dinamic range . But these details differ depending on the software used, the monitor brand and the also user experience.

Summing up all these thoughts, I could suggest to consider different softproof meanings/situations: I can spot softproof with a professional monitor at insanely high luminance levels with great accuracy (pressman style :-), or softproof all day long with a display at moderate luminance (say 130 cd/m2) coupled with a dimmable viewing booth in a normal lighted working environment (like most of prepress houses I've seen) or in shadows with subdue monitor as photographers are used to work - IMHO a condition that would not exploit the performances of current display. How to fit all this into ISO specs? Under construction...!

Best regards,

Davide Barranca

___________________________________________________________________________

Davide,

Thanks very much for your contribution about this tricky item. I hope you could provide us with updated information when this "construction" gets into "draft".

Kind regards,

Paul Foerts