Followup (Tuesday Nov 9). I was mistaken about the M9 being responsive to true infrared, though it certainly is responding to something much more than the Nikon D3x does. I can see a trace of purple in the pack, and the D3x faithfully records what I see. The Leica M9 renders it with a magenta cast, though it’s not due to true infrared, since a B+W 489 infrared blocking filter has little or no effect. However, the spectral cutoff of the B+W 489 is gradual, and this still magenta cast could be due to near infrared (650nm -750nm). The D3x has garnered kudos for its color accuracy, and it matches what I see by my own eyes. The M9 does not.
Update Nov 10: I also tried a B+W 486 UV-IR cut filter (interference filter). No difference.
I also tried the B+W 420 UV-Violet cut filter, a very strong UV-cut filter that also kills off deep violet visible light. Still a purple pack.
For those speculating: this purple cast is not due to moiré (there is some moiré but it can’t cause this huge patch of purple, images shown are scaled down).
New York pro James Knight writes:
This is a common problem when shooting fabrics with digital or film. The specific weave and dyes that were used to construct the fabric have widely varying responses to light. I used to shoot fabrics for DuPont on 8x10 Ektachrome. Often several fabrics were shot in one frame. The ONLY way to get an acceptable result on film was to shoot the scene and check the results. Some fabrics changed color and would have to be photographed by themselves and stripped into the final photo or eliminated from the group. There was NO visible way to know before shooting which fabric would change. Try replacing your gray pack with a gray card and a different piece of gray fabric and see what happens.
DIGLLOYD: I gray balanced using the Spyder Cube seen on the pack at right. Since the effect is probably due to light in the 600-750nm range, a way to "know" is to shoot the pack with an infrared-modified camera and loo for bright stuff. That’s my best theory at least.
I agree that it’s fabric-specific; this is an area I explored when researching my Guide to Digital Infrared Photography. Fabrics like Cordura are highly reflective of infrared and near-infrared.
UPDATE: there is something odd about this pack. I was mistaken, it is my gray (not black) pack. But the fabric is strange, it appears to shift color with angle of light, and there actually might be a faint purple cast (not easy to notice though, no one would call it anything but gray at a glance). I need to wait for some sunlight tomorrow to shoot again. At any rate, my eye most definitely does not see the this kind of purple, and I can shoot it with both my Nikon D3x and the Leica M9 and that will be instructive.
I was having fits getting the color right in some Leica M9 photos, really it was driving me nuts, when I happened to open a documentary-style image from my trip. I was shocked to find that my gray pack had turned purple. This is in sunlight, and white balance is always variable with DNG and Leica files (still an area of confusion for me even after thousands of files).
I thought that the infrared leakage problem had been licked with the Leica M9, but this result bothers me, and I’m going to have to study it when sunlight reappears. See my report on the Leica M8, which had a terrible problem with infrared leakage, which turns the image purple.
At a glance, this pack is gray, no color at all. Image shot with the new Leica 35mm f/1.4 Summilux-M ASPH in sunlight. Could the 35/1.4 be transmitting more infrared? I’ll have to cross check with another lens and against the D3x, because I had checked for infrared with the M9 before, and not found a problem. I can also use a B+W 489 IR-blocking filter to see what difference, if any, there is.