Fallacies About Lenses

Dry Lagoon SP11-2

Recently, I ran across an old article by photographer Mike Johnston over at The Luminous Landscape, where Johnston makes several claims which some might find rather startling:

Lens coatings are of critical importance to modern lenses; virtually all zoom lenses and many highly-corrected multi-element lenses would be useless for general photography without them. Often, coating is what makes the most difference between an average lens and a very good one


Around the web, in forums and on blogs, I’m constantly running across photographers who, whether explicitly or implicitly, deny this. They insist that the “optical formula” of a lens is all that is really important, and that lens coatings hardly matter at all. This emphasis on optical design also reinforces another mania: namely, the belief that the most important quality of a lens is sharpness, and that great lenses are great because they are sharp. There are people out there who to rate lenses based on how well they score on resolution tests at photozone.de, lenstip.com, and (worst of all) dxomark.com. But this is largely an illusion. As Johnston notes:

But resolution of very fine structures seldom helps pictorial photographs much, and, in my opinion, is an overrated property where lens quality is concerned.

In short, to much is made of sharpness, of “resolution,” at the expense of other qualities of a lens (particularly those qualities dependent on the coatings, such as contrast and color). Part of this results from a mania for quantification. There are people out there who think the only things worth knowing are what can be measured. If it can’t be measured, it doesn’t exist. This sort of nonsense has taken over far too many lens reviews, which have in consequence misled thousands upon thousands of consumers. What a lens contributes to making an excellent image can only be partially and very inadequately measured. Lenses contribute to color, contrast, microcontrast, luminance, clarity, the quality of blur (i.e., “bokeh”), and the overall impression or “gestalt” of an image. None of this can be adequately measured. It is merely perceived and appreciated. It can no more be measured than the beauty of a sunset or the foulness of a dog’s breath can be measured. Yet there are people out there (the photographic forums are bursting at the seams with them) who honestly believe that a lens can be judged by how many lines it can resolve on a test chart, or by the width of its chromatic aberrations! Recently, dxomark.com published information that claims, based on numerical tests, that Zeiss lenses weren’t really any better than their Nikon or Canon equivalents. This created a bit of a stir, since Zeiss lenses routinely cost almost twice as much as their Nikon or Canon equivalents. What was forgotten in these numerical tests and meretricious comparisons is that Zeiss lenses normally produce more beautiful images than Canon or Nikon glass, featuring bright, vivid colors that leap off the photographic image.

One of the worst consequences of this tyranny of numbers is the homogenization of lens manufacture. Lenses are now made on the basis of computer designs, which depend on numerical algorithms directed toward getting high scores on numerical tests. Lens making has become a technique, rather than an art. Even worse, there are people who seriously believe that this constitutes progress, and that the lenses of today are vastly superior than the lenses made even ten years ago, let alone twenty or thirty years ago. However, in contradistinction to this, some of my most complimented images were taken with 35 year old glass. Consider the following image, the closest I have to a “best seller” (okay, I’ve only sold a few copies, but that’s more than any other of my photos):

Cascades SM11-137

I can take very little credit for this image. It was photographed from a small viewing platform. It’s impossible to take the photograph of this waterfall from anywhere else, as the falls exists in a steep and narrow canyon, inaccessible accept by the trail to the viewing platform. Hence every photograph of this waterfall (which is Toketee Falls in Oregon) is very similar. You can change the orientation of the image; you can zoom in or zoom out (or use a longer or wider prime lens); you use a higher or lower megapixel camera; and you can shoot in a variety of light: but other than all that, there aren’t many options. The credit for this image mostly has to go to the lens I used, the Pentax K 28 f3.5 prime lens. This old lens, while plenty sharp, is certainly not as sharp as the top glass from Canon and Nikon. It would not produce an outstanding score on a numerical lens test. The Quantifiers would sneer at it. They would wonder why anyone would bother with a lens that lacks autofocus and doesn’t allow automatic light metering. But I know of no lens that could render the colors of Toketee Falls quite like this. No Canon or Nikon lens would ever quite capture that special hue of dark aqua that ripples over the water, or the wonderful hues of green that deck the foliage. Some photographers, of course, would complain that these colors aren’t realistic. And of course, they aren’t. They don’t copy nature, they improve upon it, and in so doing, transform a mere photograph into a work of art.

The K 28 f3.5 was produced before the age of computer design lenses, when lens making was an art, rather than a technique. The top glass from Nikon and Canon is undoubtedly very very good; but it is all very good in a generic kind of way. It has no character, no soul, and, in some of the most critical facets involved in producing beautiful images, remains inferior to the best lenses of older vintage.