Color standards - jump from another thread.

Just wanted to follow-up on a subject without hi-jacking a thread.

I posed a question...

UncleDoug wrote:
Here is one to ponder....

Why do most of the "Color Pros" perform their color matching with a profiled monitor calibrated to D65 2.2g and view things / have their environment illuminated, under D50 lighing?

StealthLude replied...

Here is my reply and logic...

In some instances yes, this could be their logic.
But for the majority it is because they are trying to adhere to a standard so that they can communicate color effectively and consistently to anyone that adheres to those same standards.
That is ICCprofile standards, ISO and sometimes ANSI standards for viewing images for critical comparison.
That is the larger goal of Color Management.
As has been mentioned many times before, you can tweak things in your environment so that a screen-to-print-to-orignal match can be achieved without conforming to Color Management standards, but then how do you communicate your color to the outside world?

The standards are always being revised or under scrutiny, but over protracted periods of time.

Make sense?
Or am I rambling in a over cafeinated frenzy?

It's almost a trick question: D65 calibrated monitor colors should not match the same colors on a print illuminated with D50 lighting. Or should they match?

I'm looking at a white piece of paper on my desk illuminated by tungsten lighting and I am convincing myself that it is quite similar, if not equal, to the white light on my monitor (D65). You might not agree - but we wouldn't be too far apart in our disagreement because of the eye's tremendous ability to adapt to different lighting conditions.

The monitor to print comparison is a matter of eye adaptation. My print to your print comparison is a matter of standardizing lighting and viewing conditions.

So yes - it's a matter of standards. If we're going to argue, let's at least agree on a standard for projecting the monitor image and a standard for the illuminant that reveals our prints.

Moreover, there is the issue of metamerism (mismatch of apparently identical colors under different illuminants). You can't match colors to all illuminants, but you can insist that they match for one illuminant. D50 seems to be an industry standard illuminant which many artificial light sources attempt to approximate (i.e. the light approximately matches the spectral density distribution defined by the D50 standard). So when the arguing begins, we pull out our D50 lights first before escalating the battle.

UncleDoug,

1. Thanks for starting a seperate thread for this, much appreicated from not jackin my own =)

2. This is a subject thats kickin me in the but. Not becaues I cant understand it, but really because there are soo many variables when it comes to lighting, color management, and what really is RIGHT.

I went out and bought 5000k tube lights from Lows for my room to creat a consistant enviroment. I did take advise from others on this forum and calibrated at 6500k (which i belive is my monitors native white point) sRGB...

My screen seemed me be off in that matter. Room lighting matters so much IMHO. Even the light the customer views the print it. Im no profesional in color, or photography, but I have been using photoshop long enought to grasp everyhting thats going on here.

Is there a certain meathod you guys are taking to have a controled enviroment? Lighting, no windows,.. If YES on the lighting, what are you using, and what is your monitors color temp calibrated at?

I HOPE IM NOT HIJACKING this thread.. just trying to add on to what uncleD here is trying to say.

It's well worth having these discussions, so that we may all hopefully understand better.

As has been said elsewhere, CRT monitors typically perform poorly at D50 - they're usually happier calibrated to D65. sRGB uses a 6500K white point, which is another reason to calibrate to that standard - software that isn't colour managed should get near to the results intended on a monitor calibrated to the sRGB standard (D65, 2.2 gamma).

The standard for viewing prints is D50 illumination. Norman Koren makes some interesting observations (scroll down to the "Monitor viewing conditions (a review)" box). Rather than an expensive viewing booth, he uses a high CRI 4700K light bulb, and comments that his eyes adjust quickly to the difference in white point between the monitor and the illumination. We shouldn't forget the capacity of the eyes to auto white balance - it works much better than the camera manufacturer's algorithms.

Norman Koren also remarks that he has a halogen desk light to hand to look at the results under typical indoor lighting. You do have to think about how a print will look in its eventual surroundings. Metamerism supposedly isn't too bad with modern photo printers, paper and inks, but if you know the target lighting conditions, it seems best to use them.


The illumination in my work room is a horrible mix - the desk light is tungsten, the room light is low energy fluorescent (horrid green casts if it's the sole light source) and there's some sunlight to my left as well. I'll probably replace one of the light sources with something around 5000K with a high CRI and see where that takes me.

There is a standard (ANSI, possibly) for rooms used for graphics work - which dictates the type of lighting used, grey walls and various other things. Colour is perceptual to an extent, and subtle changes in the viewing environment can change colour perception.

The standard for screen calibration is D65. That's just the way it is. If anyone has anything further to add, I'm interested.

David

StealthLude,

Sounds like you are on the right track.
What may be playing games with your situ is the intensity of the lighting.
In theory you want the same light intensity illuminating a print as what is coming from your monitor. (Yet another PITA!)

Personally use Just-Normlicht D50 fluorescent bulbs. Can't remember the CRI of the bulbs, but am pretty sure they are CRI=98. Yes they are "spiky" as compared to the SolLux 4700°k bulbs but I find you get much more even illumination, locally and globally in the studio, from fluorescent than from the small "track-lighting" bulbs from SolLux. Therefore the over all color temperature in the studio is consistent. But there is always a trade off....

Pages 208-222 of Real World Color management (original version) really help to explain the various concepts and theories in regards to all of this.

About the monitor color temperature and sRGB.
David explained the native white point of monitors concept spot-on.
One trouble though with profiles based on D65, AdobeRGB, sRGB, AppleRGB, BruceRGB, etc. is that their reference white, D65, is not the ICC standard reference white and thus when you convert from one space to another a complete re-maping of the ref. white, i.e. the L-axis of the LAB coordinates occurs. May not seem significant but it is another transformation/calcula​tion that has to go on that could introduce errors.


This article gives you a glancing blow of the relevant ISO standards.

Bare with me on this one... since in still new to color and learning some of the tems here... Can you please explane the difference beteween D50 and D65?

Also looks like im going to buy some 5000k bulbs from HomeDept along with calibrate to 6500k on the monitor.

Would to be wise to try and get the lumination levels the screen and lighting about the same using a light meter? Measuring ambient light? To around where my eye level is?

Im trying to keep this is realistic as possible, but for some odd reason, with the small 18" floro i bought to test out my calibration, I had better luck using a white point of 5600k so far... Maby Im lookin at this all wrong.

At the end of the day, I just want my colors to match my prints.

What blows my mind, is that I spent $200 on hardware calibration, and I didnt see much change in my monitor. Am I not applying the profile correctly? And correct me if im wrong, durring the calibration process, I am supposed to be changing the settings from the monitors hardware pannel, NOT the video graphics cards software settings, (which also had contract, brightness, RGB levels ect...

Leave the graphics card settings alone - they have an unknown effect and the profile and LUT can handle anything that's necessary. If you drive the 2005FPW with DVI at the native 1680x1050 resolution, as I recommend, the only setting you can change on the monitor is brightness.

Contrast doesn't apply with DVI (any curve adjustment necessary will be done by the profile and LUT). Positioning should be fully zoomed out (the default - you don't want to zoom in to the centre part of the picture). With Positioning fully zoomed out, you have 1680x1050 pixels in the input mapped to 1680x1050 pixels on the screen, so the different Scaling settings (under Image Settings) have no effect.

Leave the Color Settings at sRGB - don't try to do an adjustment that is really aimed at CRT users on an LCD (on a CRT you really can adjust the white point by varying the gain of each of the guns). My experiments with OPTIX XR Pro show that you get worse results changing the white point on the monitor than changing it solely via the profile and LUT.

D50 and D65 - the CIE D illuminant basically has the spectral distribution of a black body at the given temperature, though there's a slight offset to green side from the true black body curve to match natural daylight illumination (the sun is pretty much a black body radiator, but once the light has come through the atmosphere it's been spectrally changed a bit).

You add two zeroes to the end of the digits after the D to get the black body temperature in Kelvin - D50 is 5000K, D65 is 6500K.


6500K could be anywhere on the 6500K isothermal line on a chromaticity diagram - it could be to the magenta side of black body, or the green side. D65 is a specified point on the chromaticity diagram, as the green/magenta position is specified as well as the 6500K temperature.


If you think in terms of a photograph being processed in Adobe Camera Raw that was taken in light that measured 6500K on a colour temperature meter, you know what the Temperature value should be. However, you have no idea what the Tint value should be unless you measure the spectrum and know how to translate that into a Tint value.

If the light was D65, you know that Temperature should be 6500K, and Tint should be a little way to the green side - in fact my limited experimentation with ACR shows that D65 is Temperature 6500K, Tint around -6.


OPTIX XR allows you to calibrate to a specific colour temperature. OPTIX XR Pro allows you to calibrate to a D illuminant. Can you measure your ambient light colour temperature with OPTIX XR? I know I can with OPTIX XR Pro, though I only played with the feature.

David

David,

You are correct in saying using DVI on my Dell at the Rez i cannot change the contrast settings, only brightness, this got a a little concerned at first but thank you for bringing that up. I will go ahead and calibrate again using sRGB on the monitor and i wont tamper with the color guns.

To answer your question, I CAN measure ambient light with my OPTIX XR unit, it asks me to point it at a light source.. or white wall... not sure which one. Am i correct in saying that. I too have only played with that feature, i think my ambient light is around 3200k.


For some reason my pictures look a lot bringter on the screen, maby I should try setting the optix XR to "Dim" light.

From all this reading, I feel I should build my room into a controled enviroment with the right color temp bulbs. I see 5000k floro bulbs at home depot, would these work well?

Also, as far as matching my monitors brightness to my rooms ambient lighting, do you think i might be able to accomplish this with my light meter?

I feel like im doing everything right, but my pictures done seem to be lookin like my screen. Ill have to make some re-prints and check it out. Hopefully im not leaving something out.

This subject on light is very intresting, so i appreicate everything who has contributed to this thread.

Probably not. Unles they are Philips #F32T8/TL950 or #F25T8/TL950. I've found then at HOme Depot. They have a CRI of 98 and are less spiky than standard fluoros.

You should. Can't remember the calculations/conversio​ns though.

Getting close light intensity readings between monitor/ambient and viewing can be a pain but you can do it without investing in big dollar equipment. If you feel the ambient light is too much. Pull one of the bulbs from the fixture and see if that helps.

Right now i feel like i dont have enought ambient light... The monitor just seems way to bright when i put a photo right by it... I just got my light meter so im learning how to use it. I got the ambient light readings down and I figuted how to see the difference in lumination with the Ev thingy... Now if i can just figure out what to do with it.

Can you please send me a link or 2 to where I can find the bulbs you are refering to? Maby even something from a hardware store you think i can get by with?

I would really like to get some floros in the room and get this animal under control.

Is an on screen/print side by side comparison that realistic? I do most of my editing in the dark, which is quite the opposite of the viewing conditions for a print.

StealthLude,

http://bulbs.com …e=products&inve​ntory=9361

Also try Googling F32T8/TL950.

Print to screen comparrisons can be achieved, to a greater or lesser degree.
That is the fun of Giclée/Fine Art reproductions. You would be amazed at the screen-to-original match that can be achieved. Not perfect, but damn close.