'Oh what a difference!': work area lighting

I recently set up an additional Epson 820s in a new room and have

been fighting with mild color correction issues on the thing. I never

recalled having this much trouble with an Epson, and I was sure my

system was in close calibration. No matter how I rolled my printing

profiles I just couldn't get happy with the image quality and got

constant crossover problems with shadows and highlights.

 

So, being a trained in the typical ways of the learned photographer I

choose the proper excuse; 'when in doubt, blame the lighting'. Seems

I was right.

 

I've been real lazy lately in not sticking to basic rules of

production color correction I was brought up on. One of the major

ones is always try to color correct under light that is close to

5000k. I've been using the popular daylight enhanced tungsten bulbs

that are quite popular now that basically have a coating on the

inside to make them more visually appealing; often marketed as 'white

light' and so forth. Yeah, they're better than normal tungsten bulbs,

but no where near 5000K. Not only that, but I've learned these type

bulbs drift badly as the coating slowy gets burned away over time.

 

I went to my local lighting specialty shop expecting some odd looks,

but woulnd't you know it, they had some compact florescents at 5100K,

and 45watt even, and only 20 bucks - in stock! Plugged one in the

work area above by computer and printer and 'bang', night and day in

terms of viewing my prints and my monitor. I could see the white

balance of was off on my monitor, and I was trying to correct my

prints by going accoring to the opposite and incorrect frame of

reference. My printer was actually right, and was delivering the

garbage I was telling it in the first place.

 

The first thing to warn you about color corrected lighting is it's

not a very appealing light source for general room lighting. 5000k

lights inside a house or room appears very 'cool' and not very

flattering, which is why my old labs often looked like autopsy rooms.

In terms of print corrections, and other frames of reference such as

a white wall behind your monitor (that can really screw you up), it's

almost mandatory for proper print balancing. Looking at an ink-jet

print with a lot of color range under 5000K lighting vs corrected

tungsten or daylight florescent is like opening an eye that's been

closed. It doesn't make the print look better, but it places all the

color in proper relation and makes any point of reference that's not

neutral. Saves a lot of migraines I'll tell you that much.

Well, that's great if you want to display your prints in an autopsy room! But what if

you want to hang them in your home?

 

There's no such thing as a neutral inkjet print as the curves cross differently

depending on the light source they are viewed under (metamerism). Not so with

traditional prints.

<I>There's no such thing as a neutral inkjet print </i><P>Cough*, bullshit, *cough. Not all ink-jets suffer metamerism as Jerry Springer and the anti-digital parade is promoting, and I'm not trying to print B/W from an Epson 2000 with encapsulated pigment inks either. Metamerism is also not a spectral problem caused by the light source.<P>Using a 5000K light source to make initial judgements on color analysis is color theory 101, and mandatory with every high quality lab I've worked with, or in. The inks/pigments, color profiles, perceptual monitor settings, film, scanner, etc., are all assuming a common frame of reference, which is 5000-5500k, and not your living room 'lighting by Sears'. This is also why you use a "grey" card to calibrate a color meter or make exposure references and not a 'red card' or paper bag. Because you might hang a print in your living room has nothing to do with calibration or color reference. That's a tweak you make after everything else is in line. <P>I used to have to custom print 'showcase' Cibachromes for clients who would only by lighting them with 300watt halogens, so I used the same lighting when making the initial test prints for them. The result was a print that only looked good under that lighting and was trashed when the show was over. Prints made under 5000K lab lighting looked good under a variety of lighting.<P>You know, when I read posts like this, I no longer wonder why I'm blowing people away with prints off an $80 printer.

<i>"There's no such thing as a neutral inkjet print as the curves cross differently depending on the light source they are viewed under (metamerism). Not so with traditional prints."</i><p>Every colour image Iﾒve ever made, be it an inkjet print, a watercolour, an oil painting, a pastel, a gouache or a traditional frigging darkroom print, has always shown variation in colour depending on the light source it is viewed under. <b>Fact!

'Not all ink-jets suffer metamerism'

 

My experience is limited to Epson dye-based printers - the 750, 870, 890, and 1280/

1290 models. All of these suffer from metamerism.

 

It would seem that Epson have fixed this problem in their latest consumer models.

Great!

 

(Should anyone wish to test this, make a neutral B&W print on your Epson printer

(using RGB inks). This will involve some fairly complicated curve tweaking and will

take a few hours. Use a traditional print as a reference. Make sure you give the

Epson inks sufficient time to dry down (the inks change colour as they dry). Now

move to a different light source (fluorescent, tungsten, daylight, whatever). Whilst the

traditional print still looks neutral, the Epson print will have coloured highlights and

(different) coloured shadows. This effect is most obvious in a grayscale RGB inkjet

print, but of course is also present in a colour RGB inkjet print.)

<i>���.but of course is also present in a colour RGB inkjet print�</i><p> Moving <b>any</b> colour print between fluorescent, tungsten, daylight or whatever, results in such a profound change in the overall colour of the print that any effect caused by metamerism would be simply not worth worrying about.

Scott; here we use 5000K lighting in the Photoshop and printing area; and the customer area.............You might also want to add another bank of lights; or two; with an alternate source(s)......we do this when we want to match customer swatches; under THEIR weird lighting source.........Having a viewing booth with many types of lighting is abit of a project; but lets the customer choose the proper test printout.........Test objects; swatches do change appearance under different lighting conditions.....Even acrlyic paints; and artist materials change their mood; color; and appearance under different lighting conditions.......The change is easier to see; when the lighting levels (footcandles ) are higher.........<BR><BR>At low levels of light (like for trolls in caves ); the eye is not that great for detecting small color variations and shifts..........With a decent amount of light; the eye is a better hunter in detecting color shifts...........<BR><BR>Even with a fixed color source; varying

the light intensity will cause better observers to "feel" the color shifts abit.......One can have banks of lights in parallel; and shut off one bank; to lower the light level.....<BR><BR>

Scott, you did the right thing IMHO. Because although what people are saying is true - that lighting conditions vary from building to building and of course time of day - the fact is that 5000K is not only a standard, but approximately the temperature where R, G and B have an even presence.

 

I mean, it's logical what you did but we all take our house lighting for granted (though we shouldn't) and I will keep that in mind. I do have a daylight bulb that I play around with but it's a bit bright for my room and I don't do any colour printing anyway. It only cost me AUS$5-7 but was a tungsten unit and rated at 100W.

I have three light sources in my DD. I have a small flexarm OttLight for reflective

viewing, a bank of Vita-light tubes overhead (which is not used when I am on the

CRT) and a Porta Trace light box for slides, which I removed the, so-called 5000K

tubes from and replaced with more Vita-light tubes. I have been using a couple of

RHEM graphic arts light selector patches to judge the light temp but I am not totally

sure that all my lights are 5000°K. The vita-light tubes seem to cause the RHEM

patches to agree, but the Ott-light does not (even though it says on the box its

5000K. I use the overhead vita-light tubes for large on-the-wall-print-at-a-distance

judging.

Scott - Isn't always nice when something you knew full well you should have been doing bites you on the ass?

 

Troy - Ott-Lites are designed to replicate "natural sunlight". This ends up closer to 6000K than 5000. 5000K approximates the color coming from straight overhead at noon in temperate climates. D50 is the US standard for prepress work; Europe is divided between D50 and D65.

 

Great lights for general work are the Philips TL-950 fluorescent tubes. If you get the T8 size, they are even flicker free. These lights have a Color Rendering Index (CRI) of 98, which is higher than the tubes found in most fancy light boxes. The only caveat I have with the TL-950's is that the spectrum is somewhat spiky. If you have a printer strongly predisposed to metamerism (Scott's example of the Epson 2000P fits the bill), the color can be off. For better behaved printers, these bulbs are cheap marvels.

"Metamerism is also not a spectral problem caused by the light source."

 

Someone flunked physics *or* didn't say what they meant.

 

Metamerism is completely the result of light source spectra. The color we see things is (for the most part) found by taking the light source spectra function, multiplying by the reflection coefficient spectra function for the object you are looking at, and then integrating that resulting function over each of the three response curve functions for the red, blud, and green cones in our eyes.

 

Of course, there are some other issues that come up, such as the brains ability to white-point compensate, and how colors can look different depending on what colors they are compared against, but these are perceptual issues and not meterism.

 

Meterism is, by definition, the result of both the material *and* the lighting, and is purely a mathematical concept quarely rooted in physics.