I completed some tests today on color temperature of fluorescent lamps. As a bit of background, my workbench area is currently lit with two fixtures that both have lamps with a color temperature of 5000 Kelvin (this is also referred to as "sunlight"). I purchased two pair of 3500 Kelvin lamps to try them on for size, based on Gerry Cornwell's excellent commentary on the Model Rail Radio Podcast (#25).

The images in the gallery are of my workbench area with 5000 Kelvin lamps, 3500 Kelvin and a mix (one of each in each fixture). To control the images, I took paired photos using the "Sunny" and "Fluorescent" white balance settings - not auto. This will allow you to see the relative differences in color, though none of the photos are the actual color I'm seeing in real life (one is close). Details of the specifics of the lamps used are in the photo captions.

<!-- m --><a class="postlink" href="http://goodman312.zenfolio.com/p602649692">http://goodman312.zenfolio.com/p602649692</a><!-- m -->

A note about lamps in spite of what the Big Box Store guy told me, within a color temperature not all lamps are created equal. For example, GE makes two similar T8 lamps: F32T8 SP35 ECO and F32T8 SPX35 ECO. Both are T8, both are 3500 Kelvin color temperature, but the SP35 has a CRI (color rendering index) of 78 and 2800 Initial Lumens while the SPX35 has a CRI of 86 and 2950 Lumens. The "X" is also about 75% more expensive (~4.50 vs. ~2.60).

Big Box stores will generally have the cheaper lamps.

GE also sells lamps optimized for other uses (long life, high efficiency, quicker start) that didn't apply to me, but did serve to make this topic more complex than I'd expected. I would suspect that other brands have similar ranges, but I didn't investigate.

As expected, the 3500 Kelvin lamps are quite a bit redder (or depending on your perspective, the 5000's are quite a bit bluer). From a CRI perspective, both lamps I'm experimenting with are equal and the 3.5K lamps are a bit brighter - though I can't see it - so it will come down to what looks good. If you change a lamp color - make sure you leave it in for a few days. Your brain will correct to the new "normal" - white will again seem like white.

My reason for trying the 3500s was partly out of curiosity and partly to amplify the reds, of which there were plenty in the era I'm modeling (1930's) - just about every boxcar had some red in it! As a side benefit, it may help my father, who has a degree of red / green color blindness, by boosting the reds he can see.

Excellent experiment and great food for thought. I can appreciate your dedication to obtaining the optimum lighting for your layout!

For me, when I bought my layout's 35 fixtures from the supply house, they came with lamps, so I am just using those, which are "cool white", don't know the temp. They're probably not optimum, but with 70 lamps needed, I won't be changing them to a different temp anytime soon.

Your experimentation has caused me to re-examine my camera though, to find the various exposure modes. I'll play around with that to seewhat looks best.

Great stuff

Gary S Wrote:They're probably not optimum, but with 70 lamps needed, I won't be changing them to a different temp anytime soon.

Good point - changing temperatures takes awhile because the tubes last so long. When one finally goes and you replace it with a different temp, it can be distracting since it's the odd man out. I know - the side of my basement that will house my layout is a smattering of unknowns, cool whites, 5K's and 6.5 K's and who knows what else!

At some point, I'll be replacing the existing magnetic ballast T12 fixtures with electronic T8 fixtures (quieter and more efficient, though not quite as much light), which is what spurred the interest in color. I've narrowed it down to 5K and 3.5K. 6.5K is too blue for me and 4.1K doesn't look natural. It's all subjective.

Gary S Wrote:Your experimentation has caused me to re-examine my camera though, to find the various exposure modes. I'll play around with that to see what looks best.

If I had taken those photos with the white balance set to auto, the difference in color would have been negligible. Both would have looked "natural". Cameras (at least mine) do an amazing job in this regard. I believe this is the only time I've left auto for a reason other than "fiddling" around! The best case would have been to set a custom white balance outside so that the colors shown would have been relative to sunlight, but alas, it was dark out when I got the urge to do this....

With my camera, I can take several photos of the layout in the same area, and each of them will seem to come out with a different color balance, varying from red to blue. I try to correct for this in my photoshop program. I will experiment with the camera settings to see if it can make the colors more consistent.

Fascinating. I think I like #6 and #8 in your photos best. It's totally a goldilocks decision - too blue, too yellow, ah, just right!! By that I mean it's a pretty subjective choice based on your lived experience in the setting.

I brought home 6500K cfl bulbs for my newly installed track lighting and man is that a 'white' light! The difference between the new layout lighting and the standard 'warm' lighting in our hallway is huge both in intensity and color.

Personally I prefer a bright, white light even though that brings out blues and greens, just because the more light you can put on a scene the better for photography. Easier to dial it down than trying to compensate for too little light. Lighting my photos was by far the greatest challenge in the most recent articles I had published. Now that I've installed the new lighting AND rec'd a new-to-me camera (thanks Dad!) I will have to go through an experimentation period to determine what settings will look best. Having an evenly bright layout is wonderful. You can always get a couple clamp-on spots to vary the color and bump up the intensity for special effects.

Galen

ocalicreek Wrote:Fascinating. I think I like #6 and #8 in your photos best. It's totally a goldilocks decision - too blue, too yellow, ah, just right!! By that I mean it's a pretty subjective choice based on your lived experience in the setting.
Galen

Exactly (subjective)! My initial reaction was that I like the 5K lamps better, but I'll leave the 3.5K lamps in for a month or so to see if that opinion changes. This test may be flawed as the other side of my basement, and a temporary work area for the shelf layout are lit with fluorescent lamps that, on average, are 5K or whiter. My eye will probably always see the workbench lighting as relatively orange.

Since my original posting, I received some additional information the gentleman I referred to in the first post. I'm reproducing it here (lighting is his day job):

Quote:Color Temperature is the color appearance (whiteness) of the light source measured in Kelvin (K). The lower the number, the ‘yellower’ or warmer the light appears. A typical 60W household light bulb is about 2,700 K. A typical ‘Cool White’ old style fluorescent lamp is 4,100 K. Generally, lamps at the extreme values (3,000 K or less and 4,100 K or more), tend to be deficient in the opposite end of the spectrum. That means a 3,000 K fluorescent will not render blues very well, for example. That’s why I like 3,500 K fluorescent lamps because they do a reasonable job of both red and blues.
Color rendering is a measure of the color accuracy of a light source. An incandescent halogen lamp has a CRI (Color Rendering Index) of about 98. An F32T8/835 fluorescent lamp is about 85. A screw-in compact fluorescent lamp is about 82. The number has no units attached to it and it is not a percent. It also has no relationship to daylight or natural light. Finally, there is no meaning to light source CRI values that have different Color Temperature ratings. There is no valid CRI comparison between two fluorescent lamps, one of 3,500 K and the other 4,100 K. It’s kind of like saying this box car is Box Car Red and this other box car is Pennsy.
The comment from one of the fellows about GE fluorescent SP and SPX lamps is true, the GE SPX lamps have higher CRI. All the lamp manufacturers have different lines of lamps. Generally, the more expensive lamps have more lumen output and better color. Also, the better CRI lamps are not often available at the Big Box stores, you need a good electrical supplier.
The best strategy for modelers is to buy one of each and test them, or check what other modelers are using. Color perception is subjective.
Oh yes, the higher quality T8 lamps (also called tri-phosphor) have the lowest UV output. This reduces fading.