Color shift with power adjustment in various monolights

I had some time on my hands so I decided to satisfy my curiosity regarding color temperature shift related to power setting on my Einstein, AB800 and Genesis 300b moonlights. I did not test multiple units just one of each.

I started at full power and reduced power 10 tenths between each exposure. This offered the additional benefit of seeing how accurate the power changes were. I controlled the Einstein and AB800 via a Cyber Commander. The Genesis required changing power via its digital power controller.

I placed a light stand 11 feet from the stand holding Mannie and a Color Checker. All the shots were made with a 5D2 fitted with a 200mm F2.8 L lens. Shutter speed was 1/200. Because of the power range of the Einstein and AB800 is was necessary for me to increase the ISO from 100 to 200 for the extreme low power exposures.

The readings are from Lightroom after it applied my camera calibration preset and a strong tone curve. I have found the strong tone curve results in the gray scale square readings very close to the reference values provided by X-Rite for the white and black squares.

I'm not going to bore everyone with stop by stop readings; rather I am going to give the max and minimum reading from Lightroom for Temperature and Tint.
The RGB readings were within a few tenths of a % in the white square of the color checker.

Changing ISO does not appear to have an affect on color temperature or tint.

Einstein at full power: F 22, ISO 100 5050/+3, minimum power; F2.8, ISO 400 5100/+8
There were a couple of power settings in the mid-range with 5000/+3
Had 8/5" PCB reflector mounted

AB800 at full power; F18, ISO 100, 5500/-6, minimum power; F3.2, ISO 200, 5200/-9
Had 8/5" PCB reflector mounted
Power was controlled via a CSRB Plus.

Genesis 300B at full power:F13 ISO 100, 5300/-7, minimum power; F2.8, ISO 200, 5100/-9
Had Genesis 7" reflector mounted.

Based on this quick test, which seems to be a valid approach, it appears that my Einstein performs within its design spec and that my AB and Genesis lights are well within acceptable boundaries for color shift related to power change.

Just out of curiosity, would you be able to provide the whole set of data as a graph? That way we could see the whole set of data in a compact, yet readable form.

A couple observations that I'd like comments on (either from OP or informed others):

1. The temperature of the Einstein seems very low. Specs state it should be 5600. Any idea why you're getting a lower reading?

2. Your results for the Alien Bee seem "better" than what others have reported (at least regarding tinting). Any thoughts on that?

Is it possible the light was bouncing off a non-white surface and influencing results? Just curious.

Thanks for posting your results.

As technical as I love to be at times, when it comes to color temp/tint shifts I was never really that concerned with the actual numbers as much as I was concerned with what was perceivable to the eye.

In some tests that I had done a while back I found that most strobes that varied by ~200ºK didn't really show that change in many instances, whereas I did find that once that number grew to ~300ºK it did make a difference to my eye and was very obvious in the red channel info.

Its possible to create a graph of the output. That will take some time but is probably worthwhile.
The reason for the warmer than expected readings is probably the contamination caused by nearby ceiling and walls that are a soft cream color.

As Rob points out, the actual readings are less important than the variation cross the power range.

As I mentioned these values are right out of the camera with Lightroom setting the white balance.
If I click on the second gray square from the left, the one Adobe uses as a white balance reference, the variation is even less.

The reason for doing this was to see how my lights perform. I would not take this as a standard for all Einsteins, AB800s or Genesis 300bs.

The temp slider adjusts blue/yellow and the tint slider adjusts green/magenta.

Rob's point about numbers being less important than look is significant. In the end all that matters is do the images look good.

I have seen the forum posts about AB lights having color shift problems. I originally had 5 AB800s and an ABR800. I now have 4 Einsteins, 2 AB800s and an ABR800. I never had a problem with color shift with my ABs. That said, I always shot raw, have the camera set to auto white balance, and shot a Color Checker reference image to fine tune exposure at the beginning of a session. And once I have the lighting set I tend not to adjust too much.

Now that I did the test, I guess I do have at least one pretty consistent AB800.

During a setup, its likely one light will be at a relatively low power setting while another is more in the mid-range, and maybe another kind of high. Theoretically a wide color shift should show up. For me it hasn't been a problem. One major contributor to color cast is the modifiers being used. Since I rarely use a light without a modifier the color shift they may cause is more likely to be an issue. And again, I've not noticed anything worth mentioning.

I know the lights are speced with a color temperature. My guess is these are either calculated numbers or read with a color temp meter in a truly neutral test environment. The chances of me being able to find or create a truly neutral environment are slim.

The gray square readings are probably the more accurate measure of color temperature with power adjustments. The Adobe "as shot" reading is the softwares best estimate based on looking at the whole image. Any wedding photographer knows how that can go haywire.

Here is a graph of the Einstein results;
Both Adobe default white balance calculation and using the eye dropper on the second square. The differences are negligible and impossible to see, even on a calibrated monitor.

HOSTED PHOTO
please log in to view hosted photos in full size.

How many shots at each "power level"? I found the severity color cast to be inconsistent and maybe 1 in 5 or 1 in 7 shots at 1/32 would exhibit a noticeable shift.