I'd expect any sort of diffusing material would work, so long as it doesn't introduce a color bias (and maybe it would work even then; you'd just end up setting different numbers on the dials). If you can't find anything else, though, PM me; I've got a non-functional Philips color analyzer with its diffusing disk that I might be willing to part with. Shipping would be cheapest if you're in the US.
Another alternative is to take readings on parts of the image with a constant color. This would work best if you take a photo of the same object, with a fairly neutral color, on each roll and in each lighting condition. For instance, if you photograph a gray card at the start of each roll, you could take your readings on it. Assuming your lighting doesn't change much through each roll, this should work perfectly well; however, if your lighting does change (daylight to tungsten, tungsten to fluorescent, etc.), this procedure won't work too well unless you take another gray-card shot.