I've been thinking about a replacement camera lately, which of course for me means thinking about lenses. I know a fair amount about lenses from an enthusiast standpoint, although not very much from a scientific one.
What knowledge I do have is 80% empirical: try it, look at the results, decide what I think. It's not the most rigorous knowledge. Phil Davis used to refer to me as an "eyeballer," not always meaning it as a compliment. But at least it's practical.
History and background
Years ago, because I had access to an archive of old photo hobbyist magazines and photo annuals, I did a survey of changing attitudes towards lenses among enthusiasts. The personal theory that I developed at the time, which I never wrote up in any thoroughgoing or scholarly way, was that the "cult" of lens sharpness evolved in tandem with the miniaturization of formats. Believe it or not, there was a time when 120 film—what we now call "medium format" because it was the format in the middle—was referred to, once in a while, as "miniature." Certainly, in the beginning of 35mm and really all the way until the 1950s, 35mm was dismissed by many pros and even more amateurs as being fundamentally unserious, suited only for tyros and snapshooters. When Cartier-Bresson adopted the Leica in the 1930s for serious photography, he was ahead of the curve.
What changed the game, I think, was combat photography. It was a lot easier to carry 36 exposures in a little cassette than it was to cart around the holders or backs required to feed a 4x5" press camera.
But for many years, film emulsions were not up to the task of the kind of enlargements needed by editorial professionals…and desired by amateurs, who liked to make prints up to 8x10", sometimes with some cropping. And even though the emulsions weren't much, the lenses struggled to keep pace with them. So for years, there was a great enthusiasm for faster films, "fine-grain" developers, and ever faster, ever sharper lenses.
The late Burt Keppler wrote many times about the discovery of Nippon Kogaku lenses during the Korean War by David Douglas Duncan and other war correspondents, and how much Duncan helped put Nikon on the map. There really isn't a counterpart now of what Duncan was then—an accomplished professional who made his bones in the thick of combat but was also all over the hobbyist publications with double-exposure pictures of flowers and everything else under the sun. I'd venture to say that more amateurs once knew his name than knew who Ansel Adams was. It may also surprise some enthusiasts of more recent vintage that Leica was not always the top name in 35mm. When my grandfather, who lived all his life a rich man and who once advised me to "always buy the best," went on vacation to Europe in the '50s, he returned not with a Leica but with a Zeiss. It was the Zeiss ƒ/1.5 Sonnar, not a Leica lens, that was the sharpest lens money could buy back then, and that committed amateurs aspired to own. Zeiss was top dog, Leica second fiddle. The reason all the war photographers used Leicas was that the Leica shutters were far more durable than the high-tech but finicky Zeiss ones. It really wasn't until the brilliant collapsible 50mm Summicron—to this day a gorgeous lens when used within its limitations, and the lens Cartier-Bresson used for 97% of his work all the way from its introduction until the end of his life—that Leica responded to the pressure—from Zeiss but from Nikon even more—and began to deserve the mantle of the maker of the "best" 35mm camera lenses.
In the digital age
My sense now that the digital age has gotten fully underway is that lens sharpness is now becoming relatively less important. There are two reasons for this. The first, of course, is software sharpening—we can take a digital image file and "add" sharpening to it, so easily that "oversharpening" (a term that would have been nonsensical to our counterparts of one and two generations ago) is a common pitfall. And the second is just that lens technology and manufacture has gotten so uniformly good that many lenses, even some "budget" offerings and even some lenses from decidedly second-tier makers, are nearly indistinguishable, at least on the evidence of the pictures. It used to be true that different samples of the same lens would exhibit performance that varied enough to be visible to the eye, making it make sense for really fastidious photographers to select among a handful of outwardly identical samples. Although sample variation still exists, it's not very significant any more. If anything, the opposite has happened—many lenses are "bunched" so near the top on a graph of optical quality that differences between different lenses are not very apparent to the eye.
Finally, there's the fundamental lack of resolution of the Bayer array. Even with pixel pitches as small as they've become, they're still significantly larger than the grain in fine-grained films, and any strategy that depends on extrapolating detail from a quilt of four samples is likely to limit ultimate real resolution.
I continue to think that fine optics do impart special qualities to pictures, even digital pictures, but I'm not sure I'd want that cherished belief to be put to too rigorous a double-blind test. We can still determine good, better, and best in lenses either empirically or experimentally, but it takes a lot more work now, and the visual differences are not, perhaps, very compelling to non-experts.
Just a property
Subjective sharpness is actually a complex set of related ideas, conditions, and technical properties. Photographers love to decode and deconstruct resolution of fine detail, but lighting, tonality, and edge contrast matter just as much, if not more.
Here's an example of a picture that seems to have a good subjective impression of "sharpness" that actually isn't sharp at all. The man is out of the depth of field, although not too much, and the woman is motion-blurred to a slight extent. Plus, I added blur in Potatochop to help make the point. Yet the picture is subjectively clear. The sense of clarity comes from the tonality, the lighting, and edge contrast, chiefly in the face in profile. Actual resolution is poor—click on the picture to see a larger version and look, for example, at the backlit hairs on the woman's head. They're hardly any clearer than the backlit hairs on the man's head, and he's not even in focus.
Fortunately for photographers, the actual resolution of real detail is not a critical property of very many pictures (even if you prefer more of it than there is in the picture above). Landscape photographers often like lots of real detail, the proverbial "distant tree branches" being the most commonly cited feature where it's said to be desirable. For real working photographers, too much detail is just as often a problem as too little—before Photoshop, fashion photographers had a whole repertoire of tricks to minimize excessive detail, from deliberate overexposure of slide film to drawing black dots on a plain glass filter with a Sharpie marker. (Nowadays excessive detail is just retouched away in post.)
How much does fine detail really matter? Just to make the point, I applied a Photoshop filter called "Paint Daubs" to this picture, which effectively obliterates all of the actual detail. Yet no one has any trouble "reading" the picture, and people who know the subject have no trouble recognizing her.
"Sharpness" is actually just one more property of photographs that can be used and handled judiciously and intelligently—like any number of other properties, properties that engage the attention of tyros much less. Sharpness in the form of limited depth of field can provide editorial emphasis. Many pictures get by with a sharp "accent," which is enough to suggest detail and sharpness to the eye in a picture that otherwise has very little of it. Tonal clarity and color contrast can help substitute for real sharpness. Finally, with digital pictures sharing such a ubiquitous "look," with less and less distinctive technical properties to distinguish them, it can be said that almost all pictures today are sharp enough.
The fine lens
Still, photographers persist in preferring to own fine lenses. The distinctions might be narrowing, but even the small differences continue to matter to us. It's easier to throw away sharpness than it is to concoct it when it isn't there (even if we can now come much closer to doing the latter than we used to be able to), and, even in digital, fine lenses suggest subtle distinctions in pictures that please us out of proportion to the picture's need for them.
Kent Phelan, Nikon D300 with Zeiss ZF 25mm ƒ/2.8 lens. The high optical quality of the original file probably isn't a requirement of the subject matter, but it's pleasing nonetheless.
The reason is probably because continued use and accumulating experience sensitize us to the various telltales of optical quality. Over the years I've learned how to detect evidence of many of the optical aberrations just from pictures, for instance, and for a time after I've learned each one the quality "leaps out at me" whenever I see it—it's a sort of "golden eyes" syndrome that's equivalent to the "golden ears" of the audiophile. I have to say it drives me mildly crazy. The reward of a really good lens is not only that good qualities are often evident but also that unpleasant surprises are minimized.
Sharpness is certainly one of these qualities, although sometimes what pleases us is sharpness under certain specific conditions (a lens that can do this at ƒ/1.4, for example, is truly impressive, the more so the more lenses you've shot at ƒ/1.4, even though the unusualness of this achievement won't be obvious to the non-photographer viewer). I have to say, however, that it's been a long time since sharpness was the most important property of a lens to me (although I did go through that phase). I've cycled through enthusiasms for many properties—when shooting black-and-white film, for a while it was shadow contrast (winner: a particular sample of a Schneider-Kreutznach 80mm ƒ/2.8 breechlock lens for the Exakta 66); with digital it's been color transmission and saturation, something that I was literally blind to for many years (while shooting black and white, naturally), generally a function of coatings. And so it goes. Along the way I've concocted a laundry list of personal preferences, the specific properties and qualities I like to see in a good lens. Being an expression of taste, my list could well differ from that of another aficionado.
While the assertion or insistence that a good-quality lens makes one's pictures better than someone else's is artistically bankrupt, and frequently obnoxious, and brands the offender as an egregious Philistine (not to mention that it will likely blind one to great pictures that don't happen to be "sharp"—and they are legion), it's still true that finding a lens or lenses we like and trust is still important to most of us. Some people employ the strategy of "buying wine by the label," others the strategy of "you get what you pay for," and sometimes both those strategies work pretty well (although they don't always). Sometimes people just learn to use and appreciate what they happen to have. The best way, of course, is the hardest—learning what lenses actually do and how they actually behave and then finding one, by trial, that meets your own requirements. That's an ongoing process for most photographers, in part because it's a moving target. Arguing over what's best is fruitless and pointless, but understanding what's good and why can be fruitful and rewarding. And finding a lens that you really like is still a goal for many photographers—and one that continues to make sense, even in the digital age.
Copyright 2008 by Michael C. Johnston