Room lighting for photo editing

[jldominguescosta]

Dear all


My question today deals with the room lighting setup for photography editing. Please bear in mind that I will just do photo editing. Occasionally I will make some prints but not at home (so printer proofing should be taken out of the equation).

I am interested to know if I still should get special light bulbs (such as 4700K Solux, which are really hard to find here in Canada). If so, can you recommend an alternative to Solux 4700K? OR, in case decent, "normal" lamps are enough, could you please recommend as well?

Maybe I am confused by the complexity of color management, but I remember reading that if you are printing then the color temperature of your room lighting should match as much as possible that of the monitor. I just wonder if the same requirement applies when you are not printing.

In case it helps your answer, I have a Dell U2413 which I plan to calibrate to 6500K and Gamma 2.2.

Thanks!!

[digitaldog]

IF you can get Solux, do so, they are awesome. Forget the one's at CCT 5000K, they cost more, last less. And no, the CCT values of the bulbs and what/how you calibrate your display don't have to match. The correct value for a display WP is the one that produces a visual match, YMMV.

http://www.luminous-landscape.com/tutorials/why_are_my_prints_too_dark.shtml

[MarkM]

I am curious if anyone thinks that the smooth spectral output from bulbs like the Solux is really necessary for an application like this. I imagine you want the ambient light to be the right chromaticity but you don't really need more than that if you're not evaluating prints.

I've been experimenting with the Philips Hue LED bulbs in my workspace, but the jury is still out. They do not have a silky smooth SPD like the Solux bulbs, but I can specify a very specific xy chromaticity and brightness. After several months of use, I feel pretty good about them, but so far it's entirely subjective.

[PeterAit]

The room lighting is, I believe, irrelevant when you are editing on a monitor. Why would it make a difference? It should be dim, of course, and not interfere with your seeing what is on your screen. Criminy, I have heard that some fuss-budgets even make hoods that fit over their monitor and head to prevent even a single naughty photon of outside light from getting into their eyes.

Viewing physical prints is another matter, of course.

[MarkM]

Peter, your eye's idea of what is neutral changes based on the lighting environment. In theory you want the chromatic adaptation state of your eyes to be close to the white point of your monitor so that the monitor's white and gray look neutral. A monitor calibrated in the 5000k - 6500k range is quite cool and can look blue in an environment dominated by tungsten light. This will influence the way you see colors and can bias your color correction. You can, of course, work in very subdued light so that the monitor is the dominant light source, but that's not a great solution for everyone.

[Tim Lookingbill]

Keep in mind the brighter the room light the more it will affect the differences you notice between the color cast of your display which will look neutral calibrated to 6500K (due to the only light the eyes are focused on), but will constantly start to look too pinkish, bluish, greenish or brownish if looking away from the display to the room's ambient color cast that may be far off in either direction of too cool or too warm and especially if it's much brighter or the same brightness as the display.

For me I work in a sort of dim room that has mixed window light filtered by a closed blind with two 600 lumins daylight balanced flotubes, the GE Sunshine & Philips Natural Sunshine T8's installed into a lamp behind me and to my left.

If I start seeing this green cast in my display it usually induces me to edit too much red into skin tones and golden yellow sunsets. Everyone's eyes may or may not adapt the same way or at the same degree. For me I do notice the dimmer the room the less my eyes notice the display change neutrality.

For instance I know my calibrated display's LED is noticeably greenish compared to daylight flash off my camera which records scenes with a slight reddish warm bias. I do notice after calibration with the Xrite Colormunki Display a grayramp that looked neutral mainly in the mids and shadows will turn slightly warm (similar to what I get with camera flash) when the calibration LUTs are loaded in the video card.

I can't see a color cast change in the highlights which illustrates my point about overly bright room ambience, but I know when I focus only on a white page on my screen, the rest of the grayramp remains neutral which is a testament to the value of hardware calibrating a spiky spectra display.

The scene posted below I constructed to illustrate what I described above concerning the affects of a standard daylight appearance between my camera flash WB (the scene on the right) along with Xrite's definition of daylight (warm grayramp) compared to the actual color cast given off by my 5000K daylight balanced flotubes.

When I stare at the scene on the right lit by bounced camera flash (left at "As Shot" in ACR) and switch to the upper left cloud image it induces my eyes to see a match to the print shown in the lower left flotube scene. However, when I stare at the flotube scene and switch back to the cloud image it looks bluer and NOT A MATCH.

I would suggest you get the flotubes I use and set it up the same way as shown in the image below.

[digitaldog]

In addition to what Tim wrote, darker is better because any ambient light can strike your display and affect it's perception of black (light pollution if you will). The ambient light can't really be too low, unless you run into a wall as you enter <g>. It can be too high.

[Schewe]

The room lighting is, I believe, irrelevant when you are editing on a monitor. Why would it make a difference?

Well, you would be wrong...the entire viewing environment of your display has an impact of how you see what you see on screen. The ambient lighting should be dim...less than 1/10 of the brightness of the computer display with a white screen–measure the Lux of the display white and the incident reading of the ambient light. The surrounding area should be a neutral gray (50-60% gray)and no direct light should be falling on the display (and yes, if needed, use a hood). The white balance of the ambient lighting should be +- 200)K of the display's white point. These specs are included in ISO 12646 Displays for color proofing–characteristics and viewing conditions.

Your viewing your display is highly impacted by the surrounding environment because your eyes are very sensitive to differences.

[Ernst Dinkla]

This is a site with a very wide spectrum of lamp tests, all done thoroughly:
http://www.olino.org/us/articles/category/overview/light-measurements

Peter Ait's comment is what I had in mind too for this request. Nice if you can match the calibrated monitor light in color temperature as much as possible but lamps that have more spikes in their spectral distribution can be used too. So LEDs would be my preference. The display conditions later on of the images edited should tell something about the relation between the room's lumen level and the illumination level of the monitor you edit on. I guess the image editing is done for the web and then even the average of the display conditions does not tell anything, the illumination level of tablets, phones, etc adapts (or not) to an ever changing environment.

--
Met vriendelijke groet, Ernst

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
December 2012, 500+ inkjet media white spectral plots.

[PeterAit]

Peter, your eye's idea of what is neutral changes based on the lighting environment. In theory you want the chromatic adaptation state of your eyes to be close to the white point of your monitor so that the monitor's white and gray look neutral. A monitor calibrated in the 5000k - 6500k range is quite cool and can look blue in an environment dominated by tungsten light. This will influence the way you see colors and can bias your color correction. You can, of course, work in very subdued light so that the monitor is the dominant light source, but that's not a great solution for everyone.

I don't disagree with what you say, but I am not convinced it is relevant in this situation. If the OP has the option of changing bulbs, as was implied, then surely he also has the option of dimming the light - no? Also, less important than the color balance of the ambient light is that it be constant - the same for each editing session. Then, the photographer will soon learn if he has to compensate as a result.

[Tim Lookingbill]

This is a site with a very wide spectrum of lamp tests, all done thoroughly:
http://www.olino.org/us/articles/category/overview/light-measurements

Could you direct me to the photos on that site that show how these lights affect certain colors and that show the quality and characteristic of light output similarly to what I just posted? Charts and graphs don't show me how those light's color cast and spectra are going to influence my perception of color through adaptation during long edit sessions.

I'm glad there's sites like the one linked showing a variety of newer lights on the market. Competition is good. It appears to be a very valuable site of information but I've found from all the different lights I've bought online and locally that claim specs through analytic procedures that the data often doesn't jibe with what I see with my eyes. I then return it for a refund.

A photo speaks a thousand words and reduces if not ends a lot of return shipping charges and refunds from happening after the purchase because the lights didn't work as the spec data claimed.

[jldominguescosta]

Dear all
Thank you very much for your input and the time you took to address my question.
I can see that this is topic is not entirely consensual. Colour management is indeed a very intricate and complex subject.
My conclusion from this discussion is that it is important to plan your room lighting setup carefully even when only photo editing is in question.
I will go for a pair of fluorescent tubes as Tim suggested but maybe Ernst could be a bit more specific about which LED bulbs he could recommend. I will also have in mind what Schewe mentioned about the ratio between ambient light and display brightness (although 1/10 with a display set to 120 cd/cm2 seems to be too dim indeed, maybe I misunderstand...)
Once again thanks for the input and all the best.
João Luís

__________________________
http://www.flickr.com/photos/jldominguescosta/

[Schewe]

I will also have in mind what Schewe mentioned about the ratio between ambient light and display brightness (although 1/10 with a display set to 120 cd/cm2 seems to be too dim indeed, maybe I misunderstand...)

Well, the ISO suggests 120 cd/cm2...I run my displays at 150 120 cd/cm2 which allows the ambient light to be higher...but yes, the ambient light should be dim enough to not influence your eyes while working.

[Tim Lookingbill]

Dear all
Thank you very much for your input and the time you took to address my question.
I can see that this is topic is not entirely consensual. Colour management is indeed a very intricate and complex subject.
My conclusion from this discussion is that it is important to plan your room lighting setup carefully even when only photo editing is in question.
I will go for a pair of fluorescent tubes as Tim suggested but maybe Ernst could be a bit more specific about which LED bulbs he could recommend. I will also have in mind what Schewe mentioned about the ratio between ambient light and display brightness (although 1/10 with a display set to 120 cd/cm2 seems to be too dim indeed, maybe I misunderstand...)
Once again thanks for the input and all the best.
João Luís

__________________________
http://www.flickr.com/photos/jldominguescosta/

I live in the U.S. and got the Philips at Home Depot and the GE at Lowes. If you can afford the Solux and the lamp to drive it, that would be better in color rendering than the flotubes but far less energy efficient. The flotubes are less than $10 each with the lamps about the same price give or take. Only get the 18" T8's because the larger and longer versions aren't very accurate. I checked them out. Not good at all.

The Philips and GE are pretty much alike in color temp appearance but each have slightly different filtered spectra to help reduce their inherent spikes with the Philips providing a magenta phosphor to its recipe that I use to balance out the GE's slightly greenish yellow hue it imbues to orange and skin tones. That's why I scooped up that lamp pictured above I got for $2 at a thrift store that holds two T8 18" flotubes. With the two together (about 1200 lumens) they provide just enough light for a 10x12' room so you won't be stumbling in the dark corner to corner at night.

To give you an idea of what spiky spectra does to objects lit by these types of lights below is an example of the Philips showing its magenta phosphor effect on a yellow banana compared to the Alzo Digital 5500K CFL's rendering of yellow. The Philips provides the correct rendering which I checked against viewing the banana under direct sunlight.

[Ernst Dinkla]

True, not a nice site for searching but this is the shortest route to the lamp measurements, tap again on the lamp measurements at the top right for further searches, do not try the backward routes:

http://www.olino.org/us/ov/lamps

Enough measurements there; the CRI measurement has a target included and a percentage is shown per patch. There is a link to comments on CRI in general so it all becomes relative. There are plots of the spectral distribution + the fall off to lower color temperatures at the widest angle edge. There is the S/P ratio number with the plots for rods and cones sensivity shift for that lamp if you really want that studio to be below 50 lumen on the walls. I would count in the Kruithof curve, the lower the ambient light to the monitor illumination the warmer the ambient light can be in my opinion. With 6500K on your monitor and near 140 cd/m2 I would not go too low with the studio lighting and keep it at 4000-5000K at most, you will not find nice lamps that match 6500K and have good CRI anyway, LEDs below 4000K can have quite good CRI numbers for your purpose. Wear a grey shirt or black and white striped one and make sure no reflections are on the monitor. In your case no viewing light nearby helps then. The Olino lamp measurements have all the other things like efficiency, PAR, etc etc that may be of less interest to you. Get one Solux lamp for a viewing light for the few times you have to check prints.

I have 105-110 CD/M2, 5000K on the NEC 2690, CFL lamp of 5000K or Osram 4500K Cool Blue 51 halogens (measured 4100K though but Solux like technology) for viewing lights (+ odd ones for other display conditions), hardly any ambient light but some 3000K halogens 20 watts at the end of the studio, some daylight coming in at the same side. The printers are in the other space with Philips 950s at the ceiling. Larger Just Norm viewing light there with fluorescents. I want to replace the ambient light halogens with good GU 5.3 12 volt LEDS of 4000K but see no good ones yet. I am a printer so it is harder to get it right. Waiting for the quantum dot LEDS with this technology:
http://news.vanderbilt.edu/2012/05/quantum-dots/

You have to check yourself but this LED strip should be good for 3000K in my opinion:
http://www.olino.org/us/articles/2013/04/21/proled-finland-ledstrip-ledvalonauha-8-3wm-high-cri
At the same color temperature but Edison socket:
http://www.olino.org/us/articles/2012/10/21/philips-dimmable-led-light-bulb-with-e27-fitting
Or 4000K tube 150cm:
http://www.olino.org/us/articles/2012/12/19/lemnis-lighting-asia-led-tube-light-150-cm-glass-cover

There are more + some 4000-5000K LEDs too in ceiling fixtures but the Red patch drops considerably in color fidelity.


--
Met vriendelijke groet, Ernst

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
December 2012, 500+ inkjet media white spectral plots.

[David Eichler]

Any opinions on the Eiko Colormaster Bulbs for editing room lighting?  CT of 5500 and CRI of 93. Fluorescent, so low power consumption and low heat.

[AFairley]

In addition to what Tim wrote, darker is better because any ambient light can strike your display and affect it's perception of black (light pollution if you will). The ambient light can't really be too low, unless you run into a wall as you enter <g>. It can be too high.

Except that the lower the ambient light the more you have to lower the brightness of your monitor, and with at least cheaper monitors lowering the brightness too much can screw up the color rendering from what I have read. 

[digitaldog]

Except that the lower the ambient light the more you have to lower the brightness of your monitor, and with at least cheaper monitors lowering the brightness too much can screw up the color rendering from what I have read. 

No, that brightness is based upon the print viewing booth (condition) which should not strike the display when viewing both. With that light on, you're going to add to the existing ambient light which again, should be as low as possible.

[digitaldog]

Any opinions on the Eiko Colormaster Bulbs for editing room lighting?  CT of 5500 and CRI of 93. Fluorescent, so low power consumption and low heat.

Haven't seen nor measured them but take both values, CCT and CRI with a bit of a grain of salt. There was a recent series of posts here about Solux bulbs that discuss the values (or lack thereof) of the specific numbers.

http://www.luminous-landscape.com/forum/index.php?topic=79604.40

[Tim Lookingbill]

Any opinions on the Eiko Colormaster Bulbs for editing room lighting?  CT of 5500 and CRI of 93. Fluorescent, so low power consumption and low heat.

I missed David's entry.

I bought the Eiko bulb and they pretty much render closely to the Philips/GE combo except a bit more balanced overall. (See the yellow Skittles bag under artificial daylight comparison test below.)

However, Eiko's electronics seem to be subpar in my experience as it was with the Eiko brand Solux task lamp I bought directly off Solux site that came with the Solux 4700K bulb already installed. The task lamp's  power converter block that plugs into the wall failed after 40 hours of use spread out within 2-3 years and the Eiko CFL bulb quit just about the same amount of usage as well.

Considering the odds that I've never bought or ever heard of Eiko brand products in my entire life prior to my purchases and having two of their separate types of circuitry electronics fail is something to consider before you purchase.