Let's get anal about white balance with flash

Wilt wrote in post #4823666
Save yourself the test shot...Just meter the scene for ambient light, and if the f/stop and shutter speed used for the flash exposure differs from the ambient light reading by more than 2.5EV, the fluorescents are making NO difference in the world as it registers as BLACK!

That was kind of my initial assumption last night. Since I was shooting @ ISO100 and I've shot in the basement without flash before (ISO1600+ required) I figured the ambient wouldn't even come into play...

AinsworthPhoto wrote:
When you measurebate God kills a kitten

Time to get measurbating.

I only like cats when they are served teriyaki style.

RichNY wrote in post #4816942
Hey Pio- Robert and I will be taking an posting images this weekend. I'm in agreement that Bron are the best but I didn't want to put out the cash to buy one for our tests. Likewise the only Elinchrom light in our test will be the cheapest of the cheap light which is only slightly more than Alien Bees. I have no doubt that if we tested their RX and BX lines that we would get better results than the D-Lites.

We've got to grab Mark and go out shooting.

Hey Rich, that sounds interesting! I am about to get a set of lights and I am still not sure which way to go. I'll be glad to see your posted images, thanks for taking the time in doing so.

Thanks to all who have contributed here in this thread, as well as the whole sub forum. For what it's worth, I still don't really know which set to purchase!

Here's a link to a thread in the Paul C. Buff forums. Paul has a post midway through that explains the background of monolights (as he sees it, anyway) and how nearly all of them have the color temperature shift as you reduce the power. I thought you guys might find it interesting:
http://www.paulcbuff.c​om …d474cd95285e67c​b4accdbd15

Hi Michael,

Just wanted to let you know that I'm still going to be doing testing but I had a crazy week and busy schedule, coupled with getting some of the wrong items we ordered.

Since each software application used seems to read and determine the actual color temperature differently, and since different make and model lenses will render color differently, my tests will only show the change in color temperature from maximum to minimum power with everything else being equal.

The strobes being tested are an Alien Bees B800, Photogenic StudioMax III, Elinchrom D-Lite 4, Hensel Integra Pro 500, and a Profoto Acute2 2400 pack and head.

I don't want to commit to a date for publishing results but I do promise that I'll have those results within a week and I'll share them with you asap.

Wow! I feel really inadequate. I usually set up my lights and take a photo of a white/gray card. Set custom WB and shoot.

That's what we all do David.

Some of us like to measurebate as well, but never on a shoot.

I'm beginning the testing process and would like to ensure that we are in agreement about the method used to determine the color temperature.

Without using a color meter, the results are subject to the particular application used to read and display the temperature but, for instance, in Lightroom, if you shoot a RAW gray frame and then use the eyedropper tool to sample neutral gray, I've noticed that the temperature derived will vary from spot to spot on the card. This also happens with an ExpoDisc, so that makes it difficult to sample the exact same spot on each gray frame. Moving the eyedropper slightly can affect the temperature by 50 or 100 degrees so there is too much room for human error.

I could also shoot the gray frame using Auto White Balance and then use the color temperature derived for the whole frame but I'm not sure that is an accurate or correct way.

So, before I dig in to this I would like to sort out a preferred and correct method for deriving color temperature from each gray frame in a consistent, accurate and fair manner.

Robert, I believe a scientific method would be something like this:
1) Click on the card in several places in each image. Record the results of each click and average them. If you clicked each image 6 times or more, you could probably disregard the highest and lowest values and average the rest.
2) To determine the accuracy of any measurement, you would probably want to take six or more identical shots, determine color temp and tint values for each, and compute the standard deviation. This will help to determine how much of a difference is statistically significant.
3) You may find that keeping the gray card slightly out of focus yields more consistent measurements, as the slight changes in the color of the card will get blurred throughout dozens of pixels.

I acknowledge that my understanding of scientific method is limited, and that the methods I used in this OP were not as thorough as what I recommend here.

Thanks Curtis,

Can you think of a reason why I can't use the ExpoDisc rather than photographing a gray card which is subject to much greater variation from surface reflection and uneven lighting?

Regarding the tint values.... I'm not sure I understand how that value comes into play and how to interpret. I'm usually pretty quick with math and science but how do you compute the standard deviation with tint values?

I have never used an expodisc, though I think it should work fine in theory. I don't know if it would be better than a WhiBal card for this purpose.

Standard deviation is just a statistical computation. Excel has a built-in function that will compute the standard deviation of a set of measurements painlessly (I have honestly forgotten how to do it by hand). As long as your method of measuring tint is the same throughout your tests, and that you let us know what the scale is (for example, tint in Lightroom goes from -150 to +150), then you should be able to compile statistically useful data.

TMR Design wrote in post #4842203
Thanks Curtis,

Can you think of a reason why I can't use the ExpoDisc rather than photographing a gray card which is subject to much greater variation from surface reflection and uneven lighting?

Regarding the tint values.... I'm not sure I understand how that value comes into play and how to interpret. I'm usually pretty quick with math and science but how do you compute the standard deviation with tint values?

Robert, wouldn't using the ExpoDisc be a royal pain in the backside for this test? If the ExposDisc has to be placed over the lens, wouldn't you have to stand at the subject location to use it?

Yes Leo, but how does that present a problem. If all I'm doing is adjusting power output and shooting a gray frame it seems that would be quite simple with the meter mounted on a light stand and the camera on a tripod at subject position facing the light.

Am I not seeing something here? Shouldn't that process be just as accurate without all the surface reflection caused by shooting a gray card?

I don't see any practical disadvantage to the expodisk for this purpose. You put the camera on a tripod facing the light, meter it, adjust aperture and shoot.

TMR Design wrote in post #4842359
Yes Leo, but how does that present a problem. If all I'm doing is adjusting power output and shooting a gray frame it seems that would be quite simple with the meter mounted on a light stand and the camera on a tripod at subject position facing the light.

Am I not seeing something here? Shouldn't that process be just as accurate without all the surface reflection caused by shooting a gray card?

OK, maybe I misunderstood how you are going to test. I initially though that you would be setting the white balance for every change in the power level but thinking about it some more, I suppose you'll just do that at full power and then not touch the white balance for the rest of the power levels.

That's correct Leo. I want to see the change when white balance is set at full power and then power is reduced to minimum in whole stops.

Not sure if Curtis saw my question above. Can you (or Curtis) help me understand how the tint figures come into play and how to derive a standard deviation from those figures?