Infrared pollution and the Canon sensor

If using NDs on the 5D2 to create shallow focus does not turn out to be a fool's errand (i.e., the camera laughs at you and still refuses to open up, changing the shutter speed and ISO instead just to watch how mad you get), filmmakers will be using them just like they do on any other motion-picture camera -- which is to say, a lot. Probably even more liberally than normal filmmakers do, as it may take 1.8 instead of .9 to persuade the 5D2 Exposure Dictator to give you something close to the aperture you are seeking.

(An Arri will stay at ASA 100 and 1/48 shutter speed no matter what. Putting on a 3-stop ND lets you open up exactly 3 stops. The Canon 5D2 is unlikely to be as cooperative.)

Infrared pollution is a problem associated with heavy ND use. The NDs block the visible spectrum but not the IR spectrum. In still photography, this becomes an issue mostly in specialized situations, as it's somewhat rare for a still shooter to resort to a really heavy ND to control light; increasing shutter speed makes more sense in most situations. When you're shooting movies with a fixed shutter speed, however, NDs are a routine fact of life for daylight shooting. If you want to knock that f/22 down to f/4, and you probably do if you want this to look like a movie instead of a video, you need an ND 1.8+ in front of the lens. IR problems generally are considered to start at ND .9. That's when the ratio of IR to regular light becomes high enough to overwhelm the internal IR filter, if any.

The RED One camera has IR issues. Result:

http://www.reduser.net/forum/showthread.php?t=9412

There are solutions, none of them great.

Does anyone know what the past performance of Canon sensors has been with regard to IR pollution? (I'd test it myself, but right now I only have an ND .6 that fits my Canon lenses.)

Just updating to say I've noticed no IR pollution problems with the 5D2 yet.

I have only had the camera a couple days, but the first thing I did was slap a B+W ND 1.8 on the lens so I could use a wide aperture outdoors. That's 6 stops of ND (probably the most you'd ever need to use -- it knocks f/16 down to f/2) and I saw no problems with the footage at all.

Haven't tried multiple tests in different conditions, but I actually had done so with my 5D1 and also noticed no problems with IR pollution. Stills I took with ND 1.8 were not substantially different in color rendition than stills I took without any ND or with ND plus an IR-cut filter. In fact, the IR-cut filter was the only filter to introduce problems (apparently because it was correcting for infrared pollution that wasn't there).

So we may have another reason to look down our noses at Red.

Will update if I do see any IR pollution in any shots. Apparently certain fabrics (especially in black) can be the worst culprits, but I don't know what those fabrics are.

You can also say it the other way around : People hoping to use 5D II for infrared movies will be quite disappointed. I happened to have an IR filter flying around and you have to go up to iso6400 and point the camera directly in to a Lamp to get the slightest bit of visible information. The IR shield is very strong.

Pay attention to the reflection off the woman's dress. There's your IR pollution right there.