I find it funny how folks at dpr seem to always be "discovering" things that I thought were common knowledge (and then depleting a 150 post thread arguing about it). The preference for not only which focus point, but the number of focus points in use, has been documented for years.
QUOTE
Basically, the EOS 300's 35-zone Evaluative metering system takes into consideration which of its 7 AF points have locked onto the main subject as well as the differences in brightness between the main subject and the background. If more AF points are concentrated on a darker subject than a brighter one, the camera will assume the former as the subject and calculates the settings to give optimum exposure for that picture.
Simultaneously, if you concentrate more of the camera's AF points on a brighter subject, the camera will give preferences over it and less on the darker elements. This is fine if you have concentrated the AF points on the correct subject, otherwise you will end up with overexposed or underexposed pictures. Photo H is an example of how the EOS 300 gave a correct exposure on the mannequin biker despite the brighter elements in the background.
complete nine page articleCanon's own documentation at
http://www.canon.com/technology/canon_tech...multi_zone.html also clearly states the linkage to AF points. It's even mentioned (albeit minimally) in the manual!
I didn't bother checking to see if there's similar documention of Nikon behaviour, but the clue might be that matrix is disabled with manual focus lenses.
Maybe someone needs to tell folks at dpr about google

In reality, this is what matrix/evaluative metering is all about - trying to determine the subject and relative importance of the various parts of the image. If you don't want this just put it in averaging mode, or do like I do and use your own brain (manual mode) rather than trusting the camera to guess what you want.
- DL