Green Cast over partial photos

I recently observed something I never had before. I was shooting photos in a room that had wonderful fluorescent light. I was using the canon 5dmk3 with the 50mm 1.4 and shooting almost all open. My settings were approximately f/1.8 1/250 ISO1600.
I was shooting with Auto WB in the setup. I noticed that about every other photo had an anomaly in that part of the image had a green cast to it, whereas the following photo wouldn't. My camera was still registering 3500K as the white point for all photos.
What could cause that issue? Is it an issue with my camera? Or was I shooting at a speed faster than a bad light may have been pulsating (sort of like photographing an old tv screen).

Its the cycling of the lights. Anything less than 1/60 of a second and you only catch part of the cycle. You can get partial color shift, or even if you catch the flicker right a drop in exposure.

Fluorescent light is probably the problem. See the links in Limitations of Fluorescent Lighting

And AWB is pretty crappy indoors, too. Some people swear that AWB is OK outdoors, but I ran a test indoors & have never used it since. Notice that the very last exposure in the 2nd image of tests was of a gray card, & it's way off.
https://photography-on-the.net/forum/showthre​ad.php?t=54281

+1 to the above. It is the florescent lighting.

gonzogolf wrote in post #16023872
Its the cycling of the lights. Anything less than 1/60 of a second and you only catch part of the cycle. You can get partial color shift, or even if you catch the flicker right a drop in exposure.

A 1/120 second would capture a full cycle of fluorescent light color and intensity (in places using 60Hz power). That is a half-cycle of the power.

There are very specific shutter speeds that should be used with fluorescent lights that are driven at power line frequency. For 60Hz power, they are 1/120, 1/60, 1/30, 1/15, etc. For 50Hz power, they are 1/100, 1/50, 1/25, etc.

I realize, of course, that these precise shutter speeds cannot be achieved with all cameras.

So lesson here is to use a flash or something else to overpower that awful light source..

wdwpsu wrote in post #16027030
So lesson here is to use a flash or something else to overpower that awful light source..

Yes. Because in addition to cycling which can be controlled via shutter speed you have other issues related to white balance. The white balance of FL tubes change as they age so its not as simple picking the fluorescent setting on your camera, they can vary a lot from tube to tube. Some tubes are not full spectrum lights, so they are deficient in part of the color spectrum, and you cant balance what isnt there.

wdwpsu wrote in post #16023850
I was shooting photos in a room that had wonderful fluorescent light.

mike_311 wrote in post #16027962

I concur. "Wonderful fluorescent light" is a complete oxymoron and explains the whole problem!

The OP's error is a common one in today's photographic world. So many, in fact, almost all, of today's "photographers" understand almost nothing about light. They believe that if you can see it, you can capture it. Sorry, cameras may be a lot smarter these days, but they don't equal the human brain yet!

Well, they don't equal most (or, perhaps, many) human brains. They might equal or even exceed some.

The problem here is a lack of understanding about lighting in general and about gas-discharge lighting in particular. There is a lot more to understanding light, and photography as a whole, than the ability to press a shutter button.

SkipD wrote in post #16026570
A 1/120 second would capture a full cycle of fluorescent light color and intensity (in places using 60Hz power). That is a half-cycle of the power.

There are very specific shutter speeds that should be used with fluorescent lights that are driven at power line frequency. For 60Hz power, they are 1/120, 1/60, 1/30, 1/15, etc. For 50Hz power, they are 1/100, 1/50, 1/25, etc.

I realize, of course, that these precise shutter speeds cannot be achieved with all cameras.

One more small point: if you're taking a long exposure under fluorescent lights you should be OK at any speed slower than 1/15 second, simply because the partial frame is a small part of the complete exposure. Slower than 1/15 second you'll have at least 8 half-cycles plus the remainder.

-js

SkipD wrote in post #16026570
A 1/120 second would capture a full cycle of fluorescent light color and intensity (in places using 60Hz power). That is a half-cycle of the power.

There are very specific shutter speeds that should be used with fluorescent lights that are driven at power line frequency. For 60Hz power, they are 1/120, 1/60, 1/30, 1/15, etc. For 50Hz power, they are 1/100, 1/50, 1/25, etc.

I realize, of course, that these precise shutter speeds cannot be achieved with all cameras.

Actually, Skip, any shutter time that can be divided evenly by twice the line frequency will work.

For 60Hz power: 1/120s (1 lamp cycle); 1/60s (2 lamp cycles); 1/40s (3 lamp cycles); 1/30s (4 lamp cycles); 1/24s (5 lamp cycles); 1/20s (6 lamp cycles); 1/15s (8 lamp cycles); 1/12s (10 lamp cycles); etc.

For 50Hz power: 1/100s (1 lamp cycle); 1/50s (2 lamp cycles); 1/25s (4 lamp cycles); 1/20s (5 lamp cycles); 1/10s (10 lamp cycles); etc.

The trick is to have an integral number of lamp cycles in the exposure.

And as you can see, we with 60Hz power systems have the advantage.

20droger wrote in post #16028594
Actually, Skip, any shutter time that can be divided evenly by twice the line frequency will work.

For 60Hz power: 1/120s (1 lamp cycle); 1/60s (2 lamp cycles); 1/40s (3 lamp cycles); 1/30s (4 lamp cycles); 1/24s (5 lamp cycles); 1/20s (6 lamp cycles); 1/15s (8 lamp cycles); 1/12s (10 lamp cycles); etc.

For 50Hz power: 1/100s (1 lamp cycle); 1/50s (2 lamp cycles); 1/25s (4 lamp cycles); 1/20s (5 lamp cycles); 1/10s (10 lamp cycles); etc.

The trick is to have an integral number of lamp cycles in the exposure.

And as you can see, we with 60Hz power systems have the advantage.

Quite true. I just hadn't thought of the odd counts (other than 1, of course).