Does fluorescent light cause back focus?

Does fluorescent lighting cause back focus?

Tonight I was shooting a meeting in a room that used fluorescent lighting for the first time. My other meeting shooting has been in incandescent lighting rooms. All my shots are back focused by about 6 inches or so.

5D, 100 f2, aperture used was about 2.5, subject distance about 20 feet. And this same camera/lens combo is sharp wide open so it's not an aperture issue.

Any thoughts?

I can't imagine that it would cause front/back focus issues but I have seen strange results based on shutter speed. Florescents flicker at line frequency and fast shutter speeds can show up as poorly exposed or dark bands in the picture.

I've certainly never experienced it, and I shot a basketball tourney in a gym completely lit with fluorescent lights and did not see this.


Can you post some examples?

Fluorescent lights do not cause backfocus. They just cause ugly pictures.

Well I'll be dam*ed.. Seems that there is some precedence here. I did some poking around google and found a few interesting links.
http://www.geocities.c​om …AF_Dependency_o​n_KEV.html
And if you look at page 32 of the Canon 50D white paper it says [QUOTE] As on the EOS 50D, the EOS 5D Mark II’s AF system is sensitive to light levels as low as
EV -0.5, and capable of locking on to a moving subject and tracking it across the diamondshaped
AF area. Improved precision over previous AF systems stems from each camera’s
ability to adjust for different light sources while calculating focusing distance. Here’s how it
works: during Phase Detection AF, the AF engine rapidly measures the density of horizontal
and vertical details in a scene and the camera processor uses this data to determine where
the sharpest edges and details are located.When found, the AF is locked on target. But
under fluorescent lighting and some other artificial lighting sources, a rapid, imperceptible
flickering of the light occurs along with sudden color temperature shifts. These tend to
throw off the density measurements and therefore the accuracy of the AF calculations. In both
the EOS 50D and EOS 5D Mark II, the type of lighting is taken into effect, and readings
are averaged and processed with the help of the super-fast DIGIC 4 Image Processor.
The result is improved AF accuracy and speed when shooting under lighting conditions that
could fool the AF systems found on earlier Canon and competitive models alike.[QUOTE]
http://www.usa.canon.c​om …05D%20Mark%20II​%20WP2.pdf

If I read that paragraph correctly it sounds like some of the older cameras, like the 5D, suffer from poor focusing under fluorescent lighting.

gjl711 wrote in post #8535655
If I read that paragraph correctly it sounds like some of the older cameras, like the 5D, suffer from poor focusing under fluorescent lighting.

Yep, and Canon could correct that easily with a 5D firmware update since the 5D2 has the same autofocus system!

HappySnapper90 wrote in post #8535719
Yep, and Canon could correct that easily with a 5D firmware update since the 5D2 has the same autofocus system!

Might be limited to the speed of the processor though.

HappySnapper90 wrote in post #8535719
Yep, and Canon could correct that easily with a 5D firmware update since the 5D2 has the same autofocus system!

Sounds like they'd need to change the Digic processor as well...

DDCSD wrote in post #8535738
Sounds like they'd need to change the Digic processor as well...

Digic is just the name of the CPU such as Pentium, that's all. It's a CPU and nothing more. Firmware runs the camera.

HappySnapper90 wrote in post #8535803
Digic is just the name of the CPU such as Pentium, that's all. It's a CPU and nothing more. Firmware runs the camera.

Yeah, but the number after the "Digic" is the important part. The chip in a 20D is not the same as the two in the 1DMkIII, just as the chip in the 5D is not the same as the one in the 5DMkII. But they are all still "digic" processors.

They are different CPU's. You don't really think they're using the same CPU and just loading new firmware into it, do you?

DDCSD wrote in post #8535861
They are different CPU's. You don't really think they're using the same CPU and just loading new firmware into it, do you?

Of course not. But the 5D has only "long exposure" NR option, so it doesn't have extensive NR that current models have to do and it only does 3fps at 13MP, but the 5D2 needs Digic4 to do 21MP at 4fps. I really don't see a simple AF color temperature correction for fluorescent lighting being something that is horrible taxing on a CPU. Especially if you set your camera, like I did, to fluorescent WB!

HappySnapper90 wrote in post #8535990
... I really don't see a simple AF color temperature correction for fluorescent lighting being something that is horrible taxing on a CPU. Especially if you set your camera, like I did, to fluorescent WB!

There are a lot of factors that we just don't know. Clock speeds, signal amplification, the speed of the transistors on the silicon, speed of the associated hardware used to support the AF chip, sampling speeds and such. If the fix was as easy as a firmware change, I have to believe that Canon would have done so. Clearly they are aware of the problem.

HappySnapper90 wrote in post #8535990
Especially if you set your camera, like I did, to fluorescent WB!

WB setting has absolutely nothing to do with the issues described above. WB only tells your camera what temperature the light is so that it can record white as white. Telling your camera what color the light is supposed to be doesn't affect anything else that the camera does, especially the AF system. The AF system operates by seeing contrast.

I'm reading the description of the problem as posted by JJ to mean that the issue is the color shifts that occur with fluorescent lighting. Often times, half of your image will be one color temp and the other half another. It sounds like this color shift is causing the camera to not be able to "see" the contrast of the subject correctly, or the camera may be locking on to the point where the color is shifting from one temperature to another.


In simple terms, the AF system sees only contrast. Contrast is created when one color ends and another begins. When you have an extreme color shift (as you can often encounter with fluorescent lighting) the camera may be "seeing" the color shift as an area of contrast and locking onto that instead of the actual area of contrast you are attempting to focus on.

The newer processors take this into account, and apparently average it out and compensate for these "false areas of contrast" and make the adjustments necessary. This takes greater computing power, as it needs to do the calculations of the averaging, then figure out where it "really" needs to focus. Much more complicated than simply finding the contrast and bringing it into focus.