What our tests don't show:
As powerful as DxO Analyzer is, it's very important to note that there are some critical aspects of lens performance that our testing doesn't reveal. (This is where the personal reviews contributed by our readers become particularly important.) Here are key issues that our testing doesn't address:
Far-field performance (Especially for wide-angle lenses)
Given the finite size of our studio and the targets available to us, the camera is always relatively close to the test target. (Especially with wide-angle lenses.) Good lenses generally have behave similarly at medium and long focal distances, but "far field" performance could be different for some models. - Pay attention to the reports and sample photos offered by other readers for this information.
Statistical quality-control information
At this point, we're only testing a single sample of each lens we review. Some unit to unit variation is possible, but we have no way of telling how much. It's possible that other samples of a lens may perform better or worse than the ones that we test. - Again, check the reader reviews for each lens, to see what the majority of readers' experience is like. (And be sure to share your own experiences with any lenses you own yourself.)
Our testing represents rigorous, but fairly limited use of each lens. We're generally dealing with either new lenses, or lenses borrowed from the manufacturer's consignment-loan pool, and only have each lens in our hands for a month or so at the most. This means that we have no way to evaluate the ruggedness or long-term reliability of the lenses we test. Build quality is one important reason why pro photographers pay premium prices for high-end glass. We can comment on apparent build quality, but the only real test will be how a lens performs over the long term, for a large number of users. Again, read the user reviews, and leave reports of your own experience wherever possible.