The numbers were borne out by examining the individual color channels of images from the 1D, where I found remarkably little noise in the blue and red channels. (I have to say though, that there is a possibility that some psychovisual effect is coming into play here.