Color Balance Changes with Fluorescent Lighting - Explain this to me..

Rocketman766

I photographed a cheer event this past weekend and decided to use a different lens this time. I used a 100mm f/2.0 USM due to the low amount of light (gym lights). Using my 1D MK III, I shot in AV mode, 2.2 and had set the WB to what I thought was correct, after taking a dozen or so test shots. No flash was used, although I am working on this for next season. Aside from missed focus on some shots (I am sure it was operator error), can someone tell me why the WB seems to change from one shot to another? Or is this just an exposure issue? What did I do wrong? Also, the shots came out much darker than I expected. My shutter speed wasn't much faster than the last time I shot in this gym using a 2.8L . The first of these is shot at 1/640 ss and the other was 1/500. I know these are not great shots, I will deal with each issue separately. Thanks.

adbsgicom

Perhaps she moved directly under a light so that she's getting more of the greenish gym light directly on her?

It doesn't look to be just exposure since bumping the exposure on the first just gets a nicely tanned girl, and the other has a green cast to it. Your shift in exposure may be the result of where you are metering. If you are center-weighting your metering or spot metering, and hit on white it will expose darker than if you hit on the maroon because the meter wants to go to grey. If your lighting is pretty consistent, I'd just set your metering in manual, get the histogram looking right and keep it consistent. Don't let the Av change your exposure depending on where you point on the uniform....

If you shoot raw, and overexpose a bit, you can still pull it down in post and not loose the highlights.

rainbow

Many gyms have lights that cycle (very quickly below our perception ability) through different color casts. If you take a continuous series (eg -- 6 FPS) you can see different color casts (or WB) in the series.

Don't know if that is the cause of these in your pics, but should be considered as a possibility.

Scott_Quier

Hg-vapor (produces very blue light), Na-vapor (a very orange light), and flourscent lights both cycle on and off at the frequency rate of the power supply. In the US, this frequency is, nominally, 60Hz (60 cycles/sec). The thing is, the power doesn't cycle per a square wave, rather it's a nice sine-wave pattern. This means the power voltage is constantly changing. As the voltage changes, so does the color of the light.

To avoid color cast variations from one shot to the next, one needs to be shooting such that they capture light from at least 1/2 of the power cycle .... meaning that one needs to be shooting no faster than 1/120 and should be shooting at one of these: 1/120, 1/60, 1/30, 1/15.

To test this for yourself, find yourself a flourscent tube and place a sheet of white paper under that light (oh, and turn on the light ). With your camera on a tripod (makes it easier to achieve easy framing reproducability of the subject matter) and your WB set to something other than AWB, take a series of shots, varying only the shutter speed from 1/10 to about 1/500. In shots at shutter speeds other than noted above (and maybe even at those shutter speeds) you will note all sorts of color cast changes, color banding, etc. Most of it will be pretty ugly.

pathfinder

Great answer Scott. This has been discussed a number of times here, but it keeps surprising folks.

I shot the kids in the Special Olympics in the gym at Indiana State University last spring, and although I tried to keep my shutter speed longer than 1/120th, you can clearly see white gym walls as magenta and green in different frames, due to the color cycling of the fluorescent lights. I did set my color balance to fluorescent so it was not a malfunctioning AWB setting. You may need to include a white spot in your image to color balance in RAW during your post processing.

Sad but true!