I appreciate your input Eric. I've read LensRentals blog for a long time and value a lot what Roger has done.
Concerning the dependance of the focusing system upon the color of light, I agree with the fact that a dependance is likely to exist. Note that some Canon have focusing algorithm that take that into account and therefore correct it. Nikon may have the same kind of correction. All of that is theoretical, and although I value highly theory, there is nothing as good as experience. And experience shows that:
- Before going to Melville, my D800 did not have that dependance. I've fine tuned it under exactly the same conditions, and there was no difference in between the AF Fine Tune measured under natural and artificial light. The way I measure it gives me an accuracy of 3 AF Fine Tune units, and I measure here a difference of 10 AF Fine Tune units! I would be interested to see any other user of a D800 showing me such a difference.
- My D200 does not show this color dependance problem. The way my test is done (measuring the distance in between the focus point and the expected focus point), is resolution independent. There is no reason a D800 should fail a test passed successfully by a D200.
I agree with you that they are error bars that manufacturers can't shrink with a given technology, at a given price of production. All I want is a D800 that works as good as the D800 I sent to Melville. It is not the case, and it obviously fails at tests that are passed successfully by my D200, tests that are resolution independent.