Johnston's Lens Ramblings (1999)

[Originally published on MIR.com (Photography in Malaysia) way back in October 1999, under the title "Johnston's Lens Ramblings." It was very popular, Internet-wide, at the time. Cue theme music. When the song came out, it was 1982, and 1999 seemed futuristic. Now, of course, it's so last century.

This was originally an email(!), written to a reader who asked some questions. I should really write a new "Lens Ramblings Revisited," because the landscape has changed quite a bit.]

-

I tend to get in a lot of trouble when I talk subjective lens-talk in public. I love lenses, and maybe half of my photography is involved in simply trying lenses to see what they look like.

I'm no optical expert in the scientific sense, so I'm often left "thinking" that I see something and believing that "maybe" it may correlate to a scientific descriptive term. I'd be much more at home in Japan, where subjective connoisseurship of lenses is a well-established tradition. Here are a few of the things I believe about lenses:

—That the "cult" of lens "sharpness" and quality evolved in parallel to the miniaturizing of film formats. It's very difficult for most lenses to look bad contact printed, or enlarged 3X; the smaller the film got, and the more it was enlarged, the more the characteristics of the lens were exposed; so along with the interminable push for more sensitive and finer-grained films and developers, the "cult of lens quality" emerged. [This continued, amped up even more, with what our friend Michael Reichmann dubbed "pixel-peeping."]

Case in point:

Nobody ever looked at a Daguerreotype and said "not very sharp. Too bad the lens was a dog." If you've seen a decent number of Daguerreotypes you'll know what I mean.

—Nobody knows what anybody else means by the term "sharpness." Resolution of ultra-fine detail? Contrast? At what degree of resolution? Edge sharpness that accentuates boundaries? Some people like one kind of "sharpness," others another.

—Sharpness doesn't matter. Seen from a gestalt perspective, images can be very diffuse and broad-brush and still be recognizable, in the same way extreme abstraction can still be "read": that is, a line and two dots can be a "face," and vague fuzzy patterns of light and shade can be figurative. Ever seen those "pictures" made from thousands of little pictures? Fine detail in that case is little pictures that don't have anything to do with Marilyn Monroe, or whatever the "big picture" is "of." (Forgive all these quotation marks.) Fact is, fine detail resolution doesn't have anything to do with why most pictures work as art--it almost never contributes to greater recognition of the subject, or to meaning. I could provide a list as long as your arm of great pictures in which even coarse detail can hardly be rendered.

Having said all this, why should a lens be "sharp" and what does that mean?

—Most lenses are loaded with aberrations of different kinds, most of which are visible in pictures. But people don't see them, I guess because they don't know what to look for, or what they're looking at when they see it.

Case in point:

Many of our readers never noticed the quality of the out-of-focus blur until we did our articles on "bokeh" (bo-ke, Japanese for "blur"). We know because they wrote and told us. (Some people never wanted to notice it: we know because they wrote and reamed us out, highly indignantly, for presuming to discuss it.) Yet many, many photographs have large areas of out-of-focus blur and always have had and always will have.

It's a technical property of lenses. It's there in pictures. Why ignore it? Similarly, you can see color fringing, linear distortion, coma, spherical aberration, falloff, and on and on, in pictures. Some people don't "notice" flare ghosts or obvious falloff, much less the many, much subtler cues as to how a lens is behaving.

—Some lenses go "well" or "not so well" with different kinds of films. Many color photographers are very preoccupied with the optical property of color transmission. This is not so much a concern with the film I use (Tri-X ). That's a blatant example, but you can get as subtle as you want. Many descriptions of a lens that begin, "This lens looks..." actually mean, "This lens with this film looks...."

—You can't buy wine by the label, books by the cover, etc. In the 1980s, German tycoon Heinrich Mandermann owned both Rollei and Schneider, and Schneider built a set of wonderful new lenses for the Rollei 6000 series. Guess what? People wouldn't buy 'em. They wanted the magic word "Zeiss" on the lenses, and they'd buy older, bigger, heavier, more expensive, worse-performing lens designs to get it. Most of those Schneiders are now discontinued. German lenses aren't "better" than Japanese lenses, Leica lenses aren't all better than any other brand, etc.

—On the contrary, there do seem to be "house tendencies" among optical companies. Zeiss lenses do tend to have lousy out-of-focus blur; Nikon lenses do tend to be super-sharp-looking but at the cost of a certain harshness that can make Caucasian skin look pasty; Canon does favor a smooth, high-res but not quite so high-contrast look that works best with color film (because color can function as contrast: it can help distinguish adjacent areas and their edges. Imagine two areas of equal value, but one red and one green. In B&W, it may all be one undifferentiated gray; but in color, these two areas would tend to, ah, stand out from each other. ZAP). Re Leica, some Japanese savants, I'm told, can wax poetic regarding the philosophies of the era of Mandler vs. the era of Kolsch.

—Lens tests are meaningless. [c. 2021 there's one exception: Roger's.] They almost never give you data, they almost never sample enough of the many parameters under which a lens can be used, they almost never adjudge sample variation, they almost never consider how well a lens will remain in spec over time, and they almost never contribute to your understanding of what the lens will actually do, with your film, in picture-taking situations typical for you. (The one single thing they can offer is a comparative grading to make shopping easier.) Most especially, lens "tests" which purport to reduce the quality of a lens to a single number, grade, or ranking, are always invalid.

Always. There is no exception to this rule. Most of them don't even tell you how the disparate performance parameters are weighted! For instance, what if money is no object to you, and image quality is of paramount importance, and a lens tests weighs "value" highly in its final rating? What it if rates distortion as unimportant, and you're an architectural photographer who needs straight lines at the edge of the picture to look straight? What if it penalizes a lens for severe falloff wide open, yet every time you shoot wide open it is in "available darkness" with nothing but blackness and no important subject content anywhere near the edges? What if it rates ultra-fine detail resolution very highly but you use a grainy film with limited resolution? What if it "tests" the lens by photographing a test target four feet away at ƒ/8, but you're an aerial photographer who needs to know how the lens does wide open for subjects half a mile away?

###

Well, there's a lot more to it than this, of course. Where the fruit of all my investigations have led is that I sort of understand what properties I personally tend to value in a lens for my own work with my own film, and how a large number of the available lenses compare with each other within the focal-length ranges I work with. I.e., I don't know much...and the more I find out, the less I find I know.

But here are my recommendations:

1. Believe the evidence. If you love a lens but it's not "supposed" to be good, believe the pictures before reputation, published test results, or the status of the brand.

2. Don't believe one or another property should be important to you unless it is. If a lens is universally admired because it is super-sharp, don't accept this uncritically as being a good thing. Maybe that lens looks harsh to you; maybe your work needs a softer look.

3. New isn't necessarily good. Manufacturers in many fields typically expend a lot of effort and engineering expertise learning how to suck value out of a product—that is, to make it possible to make a "good enough" or an "almost good enough" product that can sell more cheaply and/or have a higher profit margin. With one lens design I've investigated thoroughly, I can virtually trace the bell curve as the makers first learned how to improve it and pour value and performance IN, and then as they subsequently discovered how and where they could cut corners and suck value back OUT! I.e., how they could reduce the number of elements, where they could get away with planar rather than spherical surfaces, which surfaces could be single-coated or left uncoated as opposed to being fully multicoated, how they could cheapen the barrel. The lenses at the top of the bell curve perform best, and some of these were manufactured thirty years ago.

4. Good isn't necessarily that good. Science can make much better lenses than any photographer will pay for, or than can be purchased for use on any camera extant.

5. Bad isn't necessarily bad. There are two parts to this rule. First, some very inexpensive lenses are surprisingly good; and second, even many poor lenses are good enough...for some use or other. Artists are people who can take tools and materials, perceive the properties of these tools and materials, and apply them to good effect. Good photographers can make, and have made, great photographs with really, really "bad" lenses.

6. Never be blinded into thinking that good tools = good work. The world is full of photographers who churn out sharp but wretchedly poorly-seen pictures. They can break their own arms smugly patting themselves on the back for owning the latest apo-this or aspherical-that, but regardless, Johnston's eighth law still holds: crap is crap.

7. Despite the high popularity of testing lenses and the great relish with which photographers argue the topic, ever wondered why photographers have no enthusiasm for conducting simple surveys with pictures? Actually, much of the seminal research into lens quality and exposure (in the 1930s and '40s) was done in just this way. Pay attention to what viewers of your pictures notice and tell you. Interestingly, the only lens I've ever used that got much in the way of compliments from non-photographer viewers was a 40mm ƒ/2 Rokkor-M.

8. Finally—this is again apposite to the foregoing point—most viewers can't tell. Mushy feelings of delight at the optical prowess of this or that lens squelching up from within ourselves is something we photographers pursue for ourselves and for other photographers. Viewers of pictures just look at the pictures, not how grainy the film is or how luminous the bloody shadow detail is said to be or how many lp/mm the lens allegedly resolves on a test bench. How would they know anyway? They have no standard of comparison. So, much of lens connoisseurship is akin to masturbation: something one does for own's own diversion and gratification. Best not to confuse it with something done for the sake of another, or, for that matter, with anything important.

Lastly,

9. Never sell a good lens ! The three best lenses I've ever used are gone with the wind, swept out along with the sundry detritus of this acquisitive hobby. Woe is me.

Hope this helps—

Mike Johnston

[ TOP is on hiatus this week. I'm hard at work on a new
opportunity I hope I can tell you about soon.]

Comments

[Alan]

Very well stated Mike. So true and elemental foundation of creative photography.
Look at Hidos work, Todd Hido.
I believe in an interview, he made a statement that HALF his pictures he took were through the windshield of his car. Imagine that....

https://www.vincentborrelli.com/pages/books/113312/todd-hido-alexander-nemerov/todd-hido-bright-black-world-deluxe-limited-edition-suite-with-15-archival-pigment-prints-signed

And here we are, where ever you go on the internet, how sharp is this lens, or this lens, this one not so much, on and on and on

And you have this highly recognized photographer using a Windshield as a lens element for pretty much most of his landscapes.

[alan]

I've just bought a Sony 24mm f2.8 G. It seems just about perfect apart from epic barrel distortion but this is gone when you tick the box and in doing so load the lens profile so does it matter? Maybe not.

The resistance to flare, contrast, sharpness and colour all seem outstanding but all this excellentness got me thinking that maybe lenses can be too good. Maybe the film era primes I have and use do still have their charms even in these days of excellent and just about fault free lenses.

[Henning]

A wonderful article! One that will never be posted on camera/lens sellers' websites, but probably should be, and one that should be mandatory reading before any 'test site' is revealed.

Since I am more of the 'quick to buy, slow to sell' photo acquisitor lineage, I have retained some lenses that I've been told by numerous websites are no good. Fortunately I checked what pictures I've taken with some of those lenses, and find that many of my favourite pictures ('best' is another can of worms) have been taken with those derided lenses. So now I happily shoot with those 'bad' lenses as well as 'good' lenses, and I can calmly stay away from those 'super' lenses that I can't justify buying anyway.

[marcin wuu]

The bit about Zeiss lenses being particularly bad is kinda stunning. Sonnars were always some of my favourite portrait lenses, first the 180 2.8 on the clunky Pentacon Six, then the 150 2.8 on Hasselblad. And particularly because of the beautiful out of focus areas rendition. It's gotta be the case of to each his own...

[Yeah, that stuck out to me too. Dunno now why I thought so back then. It was a long time ago.... --Mike]

[William Giokas]

The famous Leica Look is why I got my M240 with the nifty fifty 50mm f1.4 . I could not justify paying over 14000 for the Noctilux
lens.My 50 year 35mm f2.8 Summon still produces wonderful images.

[Phil Martin]

Very interesting article and very interesting comments. I'll certainly remember the Todd Hido quote next time someone is moaning about the negative effects of protective UV filters.

It's also interesting to hear about the negative response to bokeh. Theses days I often see images that are about little else other than bokeh, so small and insignificant is the alleged subject matter and narrow the plain of focus. It also irritates me greatly, when I see classic and wonderful lenses being dismissed because their not sharp wide open, when they were simply not designed to be used in that manner. And for that matter lenses with modest maximum apertures, being dismissed in favour of those with higher, when the modest lens is often superior when used in the f/5.6-f/11 range.

[Alan Whiting]

Perhaps an extreme example: I've gotten photographs I'm quite happy with from a Brownie. Right, a single-meniscus lens with all the aberrations you could ask for. And as far as lens tests go, I had trouble working out the range of focus, because the processing place I used to print my pictures automatically used sharpening software. It was really hard in some shots to tell where things started to get fuzzy.

[Albert Smith]

The missing intangible: Testing done with perfect technique, flawless light, steady tripod, scrupulous focusing... resulting in terrific results. Then someone buys that lens and handholds the camera at a less than optimal aperture in awful light and then wonders how the lens got such a good review.

Conversely, a lens made to be used wide-open, handheld, in bad light can test bad when the test is a conventional two dimensional chart, where flatness of field, evenness of exposure, and edge to edge resolution are prized. Proof, check out the 1994 Popular Photography Magazine's test of Leica's trio of M mount Summiluxs (35, 50 and 75mm) which caused an uproar in Leica world when two of the three lenses received "F" ratings. This upset people that had been using them for years with great results.

Regarding the advice to never sell a good lens, absolutely agree! I have bought multiple 35mm and 105mm Nikkors and 5 50mm Summicrons after telling myself that they were past their prime ( pun, get it?). Today, I buy a new lens and keep the old one too.

[Herman Krieger]

Photos made with a single element meniscus fixed focus lens in a box camera-
"Along the Riverbank Bike Path with a 'Bent and a box",
http://members.efn.org/~hkrieger/bikepath.htm

[Richard John Tugwell]

Ah... mir. Still a great resource for Zuiko / OM system

http://www.mir.com.my/rb/photography/hardwares/classics/olympusom1n2/shared/zuiko/index.htm

[SteveAitch]

My late uncle, an electronics engineer, once told me he thought it stupid that guitarists paid lots a money for amplifiers that gave a clean sound when all their songs (that he had heard) had a distorted guitar tone. I may have changed his mind by pointing out that some songs may need a clean sound and others a distorted one. A distorting amplifier can't produce a clean sound, but the clean sound of another amplifier can be distorted by the use of an effects pedal.

I suppose the equivalent of an effects pedal for an 'over-sharp' lens would be a filter of some sort. Or Photoshop*.

* other photo-editors are available. :-)

[Rick F]

I think our culture has a fixation on the quantifiable. Go look at online forum for near any hobby and you’ll find a good number of people obsessing over that what can be measured with numbers.

We base decisions at near every level of society on mainly measurable things and gear head photographers and photography media have escalated this to near art (ha see what I did there?). We’ve made the grave misjudgement that if we can find the gear that puts up the best numbers, we can use it to make quality work, or further that only such gear can be used to make quality work. This in spite of enormous amounts of evidence that says otherwise.

A big part of this is the driving factors of capitalism: camera companies must be seen to produce things of value if they are to continue to exist; so they use measurable quantities such as sharpness, aberrations, distortion, etc etc to attempt to differentiate their product from others’ and further to sell us on the idea that the right quantities of each will let us produce quality work, and even further that lack of those correct quantities is an anchor weighing us down. The tangible and measurable makes for easy marketing.

Go to the comments on any recent DPreview article on the re-realeased Pentax limited lenses and you’ll see the chasm between photographers who care about mtf numbers and the like, and those who who also put stock in the ineffable qualities that attract people to those lenses. Pentax is a company that, while they’ve shown they can compete with the big guns on mtf charts, must by necessity also rely on an aging back catalogue of lens designs. In part because of this they often refer to non-measurable qualities in marketing materials for many of their lenses.

This often leads the MTF chart/ brick wall brigade to essentially accuse them of making things up and trying to ascribe value to old designs for marketing purposes, but really it’s just that many of us are conditioned to simply disregard non-measurable qualities as either non-existent or too subjective to deserve any weight.

Some good reading on this is the late philosopher Robert M. Pirsig’s two novels.

[Daniel]

"You can't buy wine by the label, books by the cover, etc"

If this were always true we would not have Harlequin romance novels.

[Reminds me of a bit a comedian did once...his point was that of course you can buy the book for its cover, because it's what tells you the title of the book and what's in it, and "that's why books have covers." He said it a lot funnier than I just did, of course! --Mike]

[Not THAT Ross Cameron]

Re the Lens Rentals (LR) lens tests, I’d caution that the results tell you only what LR tests for, which may not be what the photographer is looking for, and even Roger states that in many of his posts. Even still, I agree that LR is mostly excellent reading, and on my go-to list.
I also remember reading Thom Hogan’s guide on what to look for when buying lenses, mainly from the perspective of what aberrations can be corrected for in post processing (and the degree & difficulty in doing so).
Both useful, both written from very different perspectives, and both come with their own caveats.
Joys of the subjectiveness of humans :~)

[Maciek]

Ha! As of e-mails becoming articles:
Many years ago I asked you (in e-mail) how do you process Tri-X in D-76 and your answer became one of Sunday Morning Photographer column (at least it was on Fotopolis.pl). I remember being very stressed and wondering if it was appropriate to ask you about it :-) And then so happy reading your answer! And I remember that thrill every time the new issue of the SMP column or 37-th Frame magazine came out :-)
The good old days :-D
All best from Poland!

[Rick Denney]

My own definition of "sharp" is when the viewer never loses the sense of endless detail--the illusion that the only thing preventing more detail is the ability to get closer to the print. That is a demanding standard, but a realistic one--for things intended to be sharp and detailed, seeing it go blurry as one approaches the print violates expectations.

For me, though, it limits print size more than lens selection.

Of course, we often intend parts (or all) of images to be blurry. And then the quality of the blur perhaps becomes important. We technical types try to quantify and describe, the way painter technicians might quantify and explain the effects of different brushes. We hope, though, that the effects being measured are utterly inconspicuous to the viewer--something that works entirely beneath conscious notice, but that still works. So, objective description is not necessarily masturbation, but calling it that can undermine the important, simple fact of objective measurement.

Subjective sharpness is a state of mind--an illusion. But without some degree of objective sharpness, subjective sharpness doesn't survive.

And, yes, that Zeiss Jena Sonnar 180/2.8 for the Pentacon Six is perhaps the most beautiful medium-format portrait lens in history. I have a couple, adapted to most of my cameras of 6x6 or smaller format. But it is not nearly as sharp as more modern lenses of similar image circle and focal length. (The CZJ Sonnar 180/2.8 was formulated in the 50's and updated in the 60's, but the basic design goes back to the 30's.) For portraits, that little bit of residual spherical aberration that allows such pretty faded-edge bokeh is a feature, not a bug.