Fluorescent White Balance Shift

I

ItsJanosz

Active member
Hey All,

Caught something bizarre while looking at dailies of footage shot inside an office building. We were shooting 24P Advanced, white balanced to the existing lights in the building. They were not strobing, and the color temperature of the lights (which were exposed to blow out) was neutral white. However, in the footage, the color temperature shifts back and forth from slightly warm to slightly cool, oscillating back and forth every few seconds. It appears that the camera is trying to find white balance, or perhaps the fluoros are cycling in some way?

What can be done to avoid this?

Thanks!
 
It's very possible you shot the footage with a high shutter speed, resulting in this strange cycling phenomenon. Keeping your speed closer to 1/60th of a second (I think 1/100 or lower is enough) would have avoided this. The thing is that flourescents cycle, but not the same color each time. It's the rapid cycling that gives us an impression that it's a balanced light. While in phase with the cycle, you can have exposed for only a part of the lamp cycle, and so get this color shift in the final footage.
 
That would make sense, except that we were shooting at 1/48th. We purposely shot with a lower shutter speed to avoid flicker, but did not see the cycling while we shot. It seems unlikely to me that shutter speed would fix this, as the cycling was much slower than that; perhaps cycling every 5 or 6 seconds or so.
 
ItsJanosz said:
That would make sense, except that we were shooting at 1/48th. We purposely shot with a lower shutter speed to avoid flicker, but did not see the cycling while we shot. It seems unlikely to me that shutter speed would fix this, as the cycling was much slower than that; perhaps cycling every 5 or 6 seconds or so.
Click to expand...

It is surely about shutter speed. The 6 seconds difference is not the time of the cycling, but the time it takes for minor differences in the shutter X cycles rythm to take effect and get your almost phased response change the end of the flourescent spectrum. It's like, in 6 seconds you are shooting in phase with the lamps, getting only the orange spectrum, and the next 6 seconds you got a little behind and gets only the blue spectrum. But since your problem happened at 1/48, and I have only seen or heard stuff like that with high shutter enabled, I'm now as clueless as you... :P
 
Ok, thanks, that makes sense to me. Now, anyone have any idea why this is happening to us 1/48th? Don't think we can safely slow down the shutter speed... Any possible solutions?
 
I've seen the problem too at 1/48th with 24p under fluorescents. And that was with a predominantly white set so the color cycling was very noticeable later! Technically anything under or over 1/60th second (60 hz) could cause problems - but not with all types of ballasts. You'll get better results shooting at 1/60th to match the frequency of the lights.

Try this experiment: Set a PC monitor (CRT not LCD) to a 60hz refresh rate or just use a TV. Point your camera at it set at 24p and 1/48th. Notice the severe scan bars? Now set it to 1/60th. The scan bars are greatly reduced or eliminated. The same principle applies.
 
This is making me very nervous. I just shot something 2 days ago and I sure wish I had know this earlier. How likely is this to happen? I shot 24P under fluorescents.
 
Not always a problem...Some fluorescents lights use electronic balasts, secondly, some fluo lamps don't show the wb changes in the afterglow cycle, and a third reason for not having problems (apart from the ideal 1/60 or 1/120 in NTSC with 60Hz mains freq) is the absence of other lightsources like direct daylight or incandescent lightsources. So maybe y're lucky!
 
It turns out this did screw up one of my shots and the client noticed it. And I found out about this 2 days after the shoot but before the editor and the client noticed it. I knew it was going to show up. I'll never do that again.
 
You must log in or register to reply here.
Back
Top