I know people say this is a filter you want to use when using ND with a RED or cmos sensor. I've shot with ND on both 7D and RED and never really noticed a problem. What are the issues really and will we experience similar issues with the AF100?
The issue is that certain fabrics (especially manmade fabrics) will reflect infrared light, and some cameras have IR filters that are stronger than others. On a camera that has a weak IR filter, you'll see your black fabrics showing up as brown or purple.
It's far more likely to happen under tungsten lighting (because tungsten lights throw off a massive amount of infrared radiation, which is, basically, heat).
In daylight it's not that likely to show up unless you're using lots of ND filters. The more ND you throw on, the more you change the relationship between transmitted light, and underlying IR. In normal circumstances with no ND filters, the visible daylight is much more prevalent than the IR, so everything looks fine. But as you put up more and more ND, you start blocking more and more of the visible light -- but IR sails right on through; ND filters don't slow down IR at all. So there becomes proportionally more and more IR in the signal, the more ND you use.
Why are some cameras more sensitive than others? It's not a CMOS vs CCD issue. It's that the tighter they make the IR filter, the more visible light gets cut too, on the low end of the spectrum. In the ever-present race for more sensitivity, people got a little more lax with their standards for IR filtration. The result is situations like the EX1, where the IR filter is pretty weak and lots of IR contamination gets through. But, it's pretty sensitive to light. It's a tradeoff.
I would have to gather together some IR-reflecting fabrics and blast some tungsten light on them to see how the AF100's IR filtration works.
Earlier this year I had to color correct a short film that had a major ND/IR issue shooting in daylight. They shot outside in bright sunlight and threw on a ton of ND. I believe it ended up at 1.2 or 1.5. One of the actors was wearing a dark navy blue sweater. It ended up being a very light purple. All the green leaves on trees and bushes were brown. The blue water in the pool was green. It was a mess. It was shot on Red too and they let it go because they thought it was just the monitor (NEVER think it's just the monitor). Even with the Red Raw there was no way to fix it perfectly. It was a nightmare color correcting that scene, but I managed to get the sweater blue (it was a lighter blue though) and most of the other colors right, except there was no way to get the trees and bushes green again due to the color of the building and the flesh tones. Maybe with more time I could have done it, but they were on a tight deadline and I only had a day to color correct the entire film. In case you run in to a situation like they did, you're better off taking the ND off and closing the iris. Or always have an IR Hot Mirror handy just in case.
We have a full module on this in the RED101 DVD. I am not sure if it's online or not. I'll have to check. The origional RED Mystirium sensor had some pretty bad IR issues, but the NEW MX sensor came with a stronger filter, so it's not as bad. IR contamination can be hell in post, because it alters your black level in a very unique way. I'll see if the module is online.
With my 5D I use the 5 and 7 stop Tiffen white water IR/ND filters. On my EX1 I use the Tiffen T1 filter that seems to fix the problem when certain black fabrics turn brown or redish, regardless of using ND or not.