Websites of Note 1: Nikkor Lens Evaluations


The Norwegian photographer Bjørn Rørslett has fascinating site devoted to his evaluation of a large number of Nikon lenses, both old and new. While the site obviously will appeal most strongly to those photographers who use Nikon cameras and lenses, Rørslett’s general point of view as a reviewer of lenses is refreshing and full of wisdom. For example, consider what he has to say about resolution tests involving MTF number crunching:

People have repeatedly asked me to provide an overview of these and other lenses to show their image-producing potential. True, a number of excellent resources for evaluating lenses are found on the Net. Many are based upon subjective evaluation of the items in question, but only a few sites can provide a truly encompassing overview based on a single person's experience. On the other hand, sites exist that overwhelm you with MTF plots purportedly providing indisputable facts. MTF methodology has a genuine scientific foundation and there is nothing "wrong" with MTF as such. I even understand the mathematical equations. However, such statistics basically are as helpful as knowing the mass of a lens — on its own, MFT testing cannot predict the pictorial outcome of any lens. Thus, MTF tests will not show all problems from field curvature, colour fringing, flare and ghosting, the variability in performance that arises from near or distant focus, the subjective 'feel' of the images and in particular the out-of-focus rendition (given the buzz word of 'bokeh'), the way a lens handles under actual use, and so on. MTF data can just indicate there is a problem with a lens, or that a particular lens might be an excellent piece of glass. All of this information can be obtained as easy (but likely not as fast) just by shooting pictures with the lens. Averaging MTF numbers to arrive at a single value in order to rank lens quality is simply impossible and largely a waste of time.

Rørslett also takes aim at another vastly over-rated shibboleth in photography—namely, internal focusing:

I have come to the conclusion that many modern lenses, during their computer-optimised design stage, have had certain performance criteria maximised to the disadvantage of other parameters. The optimised parameters aren't necessarily optimal for a specific end user, however. For example, the clever trick of making lenses focus very close by shortening their actual focal length, a principle frequently applied to 'macro' lenses, will give a short working distance as an undesirable side-effect. When a lens changes its focal length, it becomes awkward to work from a fixed position, such as from a tripod. Sooner or later the tripod has to be moved because it becomes impossible to get sharp focus without changing image magnification and thereby altering the picture composition. This is the very reason I won't touch some recent 60-105 'macro' designs with a ten-foot pole...The internal focusing (IF) principle may unfortunately also introduce some colour fringing outside the plane of sharp detail. This can give rise to disturbing red, green, or purplish fringing of unsharp highlights. Lack of concurrent focus for the primary colours also leads to a certain fuzziness in the depth-of-field (DOF) zone. In the focused plane, colours may coincide more or less perfectly to diverge in front or back of that plane of maximum sharpness. Thus, many wide-angles will never attain anything like the theoretical DOF due to their residual colour aberration. Even very expensive lenses clearly exhibit such problems. I often find older lenses to behave much better in this respect - although their peak sharpness may be lower than the modern designs, they more than make up for this with improved rendition of the out-of-focus areas.

I suspect Rørslett is at least partially right about this. I have an internal focusing zoom lens which seems to feature less DOF, even when stopped down, than equivalent prime lenses. Other issues I have found with internal focusing lenses include poor edge to edge sharpness at longer distances and proneness to decentering issues.