10.20.2019

Designing the perfect lens for me. Balancing needs and wants.



Photo review sites spend way too much virtual ink describing, reviewing and deliberating about camera bodies. They spend too little time discussing lenses, and that's too bad because most good photographers think lenses are where the magic is. I remember back when Modern Photography Magazine was still published. It was generally a big deal when a camera got reviewed; especially since new camera body models were rushed to the market far, far less often than they are today. What the magazines concentrated on instead was the reviewing of several, or a handful, of popular lenses along with one or two speciality lenses thrown in just for good measure. Nearly every month. 

Instead of comparing the Nikon Z7 to the Sony A7r4 camera bodies I'd love to see a side-by-side comparison and in-depth review of each maker's "holy trinity" of zoom lenses (14-24, 24-70, 70-200) to see where the whole system stands, not just the naked body. Perhaps a comparative overview of all the different macro lenses on the market. How about a toe-to-toe exploration of zooms in the 100-400mm focal length ranges? And with all the new and pricey 50mm lenses coming out maybe we could see a nice comparison there. 

Just one of the many problems with this idea is that most reviewers are very ham-fisted about how they measure stuff and I'd hate to see a lens with beautiful characteristics get denigrated because its corner sharpness was a few lines per millimeter less than a crappier lens that has a flatter focus field and maybe slightly faster focus acquisition. I'd always like more data than just easy to understand, cookie cutter data.

One area of lens evaluation that Roger Cicala at Lensrentals.com has brought to light is that, because of the complexity of new lens design and the equal complexity of the manufacturing processes involved in assembling modern lenses, there is a lot of sample variation even among identical models of lenses. Centering of the elements is a big issue as is calibration of the overall optical system. Any really meaningful evaluation of lens performance should probably included testing at least three samples of each model.... That's a ton of work and much harder than just deciding that one likes to have cameras with panorama modes or that one doesn't like cameras that can't process their own raw files. As if those "features" matter to any of us...

It's also clear that when popular sites do get around to reviewing lenses they tend to always do the reviews about big fast zooms and almost totally disregard really stellar single focal length lenses, unless such a lens comes with such a high price tag that the price becomes good click bait news. 

So, here is a suggestion to the big review sites: Review some lenses! Review some that aren't extreme wide angle zooms, or boring but fast 24-70s, or endless iterations of the 70-200mm models. Review some portrait primes. Review some normal focal length lenses that are price-accessible for the average camera buyer. Review some slower, standard zooms. But don't just stop with lines per millimeter measurements. Try your best educate your reviewers to understand that different kinds of sharpness profiles and rendering characteristics can also be good and interesting in real world use. 

For years Leica designed lenses for high center sharpness and acutance, and the trade off was lower performance in the corners. Not lenses that measured "well" across the frame but lenses that were loved and in demand by some of the world's most discriminating lens users. Currently, lens designers are trying to make reviewers happy by giving all of us lenses that have the same sharpness characteristics across the frame. The lens makers generally have two choices in the design process: they can make the lens equally sharp across the frame but at a lower overall level of sharpness than the numbers they could get from a lens that had its highest sharpness in the center and allowed the corners to degrade, or.... they could make the lenses equally sharp, across the frame, but only by making the overall lens enormous, costly and very heavy. (Hello Zeiss Otus. Hello Sigma 50mm ART). 

I recently bought a lens that is a counterpoint to what I consider the blunt hammer of current lens design and marketing. It's a small lens made for the L-mount system. It's a normal focal length (45mm) but it's slow, relatively speaking, as it only opens up to f2.8. It's also a bit unusual in that it isn't necessarily blindingly sharp across the frame when used wide open. The center is pretty much perfect but sharpness falls off towards the edges and corners. It's a really nice look if you have something interesting in the middle areas of the frame and want sharpness to fall off a bit at the edges. Stop it down to f4.0 and it's nicely sharp across the frame. Stop it down to f5.6 and it's perfect. 

But it's also small, light, well built, and one of the few lenses made for the L mount system that's under $1,000. 

So, as I mentioned in the title I'd like to design the "big picture" parameters of a lens that's personalized for me; for the way I most often use my favorite lenses. This isn't to suggest that you'll want anything even close to this for your own use. This is just my pipe dream....

I'll start with focal lengths and I am defaulting to a zoom lens. I want something that starts at 40mm and tops out at 110mm. According to the metrics from various Kirk Catalogs of images this is the range I use for 90% of my day-to-day, having fun, imaging. If I want something wider I can bring along a discrete 24mm or 20mm. If I want something longer for a particular I can be like everyone else and default to a 70-200mm zoom instead. 

By limiting the focal length range I think a good lens design team could design these focal lengths into a decently small package even though I would want a constant aperture f2.8 for my zoom. But they needn't give me something that's super sharp even into the edges, wide open. I'd be happy with a lens that consistently delivers a very sharp image in the central 2/3rds of the frame and then delivers average or adequate results in the corners. The corners and center definitely don't have to match! 

I also don't really care if the lens has absolutely zero distortion. As long as what distortion there is can be corrected with mild measures, by software, in post production. This would keep the size and weight manageable while still providing me with a performance profile that would work well for my usage. 

I want a real manual focus setting ring, delivered via the pull back clutch mechanism that Olympus has on their 12-100mm Pro and their 35-100mm f2.8 Pro lens. It's nice to know exactly where infinity and also the closest focusing point is on my lens. This is also something that the L-Mount Panasonic 70-200mm has...which moves me to consider that lens as a potential buy too. 

Finally, I am not interested in having image stabilization in the lens. I would depend on the image stabilization in bodies like the Fuji X-H1, the Pentax K-1 and the Lumix S-1 to stabilize my jitters for me. This also reduces the needed size and complexity of my "perfect" lens. 

I'd love to have this built as well as the 45mm f2.8 from Sigma and I'd be willing to pay up to $2,000 to own such a lens. I'd be even happier to spend about $1200 instead!

The lens I'm describing would be for full frame cameras but you can shrink the focal lengths proportionally and offer it as an APS-C lens with a max aperture of f2.2 or in micro four thirds (shrunk once again...) with a maximum aperture of f2.0. Olympus was on the right track back in the pre-micro days with their two fast zooms for the 4:3 system. The 12-35mm f2.0 and the 35-100mm f2.0. The later was way too heavy and big but it was a killer optic. The 12-35mm was just about perfect. ...

So, the 40-110mm f2.8 Kirk-O-Flex lens has now been described in detail. I just need to sit back and wait for Sigma or Panasonic to make me one. You can have one too. I hope they make lots of them. 

If you could design a personal lens, just for your own use, not considering whether it would be widely sellable, what would it look like and what would it do? Just curious how far off the mainstream I might be.....

Thoughts?

24 comments:

Kirk Tuck, Photographer/Writer said...

Before you get the keyboard fired up, I am aware that Leica produced a 50-90mm f2.8, Manual focus R lens back in the day. I owned one. It was a great focal length range for a portrait guy. It was big and clunky and the 90mm Summicron was a sharper alternative. Still, it was a move in a good direction...

Gordon R. Brown said...

Did you mean the Angenieux 45-90mm f/2.8 made by Pierre Angenieux under license from Leitz? I worked with a photographer who loved his.

Kirk Tuck, Photographer/Writer said...

yes. that's the one!!!

Eric said...

I've never used one - but I'm very intrigued by the new-ish Tamron 35-150mm f2.8-4. Seems to hit all the right spots for me, maybe you too? Covers the split where I'd usually have to change lenses or switch bodies. I think if I buy a lens before the end of the year - this will be it.

Anonymous said...

I'd be very happy with a reasonably sized 35-80 (full-frame) zoom or its equivalent focal length for crop sensors. F/3.5 or f/4 could work, as anything faster takes it into the 24-70 f/2.8 heavy lens category. Sharp in the middle 2/3 is good enough for those focal lengths. - Ken

James said...

21mm-50mm. It's one step closer to normal than the 17-35s out there. In general, zooms should overlap, so that if you bought every other one in the progression, you had it all covered. You described the half step above the 24-70, I described the half step below it. There is some hope in this regard, Nikon released a zoom half a step above the 70-200. Thom explains why it is a good idea here: http://dslrbodies.com/newsviews/nikon-2019-news/september-2019-nikon-canon/do-we-need-a-120-300mm-f28.html

I think it is for all the APS-C shooters who bought 70-200s and then found those lenses too short when they upgraded to full frame.

Ash said...

My ideal lens would be a redesign of the fixed 23mm 2.0 lens in the Fuji X100.

I owned an original X100 and loved it dearly, but the lens had some major functional flaws.

First, there is an air gap so the lens 'breaths' as it focuses. This allows dust to be sucked in, right onto the sensor glass. Not a big deal in an interchangeable-lens camera, but with the X100 you can't open it up to swab the glass.

Second, the focus is slow. Also not a big deal, except that Fuji has released much faster lenses in recent years.

Third, the lens is not weather sealed. Fuji's top of the line cameras are all weather sealed now, would be nice for the X100 to follow suit.

In short, I would like a new version of the X100 with a redesigned lens that is enclosed with internal focus, fast autofocus and weather sealing. I would pre-order one and take it travelling.

Ash said...

p.s. forgot to add that I do not want Fuji to make it sharper or majorly change the optical format at all.

The 'flaws' inherent in the lens add charm to the images, and that look should be preserved if possible.

Kristian Wannebo said...

> "I'd hate to see a lens with beautiful characteristics get denigrated because..."
I've found that Dustin Abbott doesn't:
https://dustinabbott.net/category/lens-reviews/

( ... as e.g. in his review of the Zeiss 50mm f/1.4 Planar T* ZE.)
And his lens selection is quite interesting and somewhat unusual.
- * -

Kirk,
your wish lens comes close to mine - except that I'd prefer an IS one - only I'd be using it as an APS-C lens.

DCish said...


I'm very fond of the Olympus ยต43 25mm f/1.8. Now on sale for US$250. A very satisfying little lens.

Zave Shapiro said...

I own a Mazda MX-5. The reviewers have pushed on the easy measures like speed from 0-60 or skidpad g's and complained about body roll. I, on the other hand, want a supple suspension on varied roads and transmission ratios that are "right" in the real world. And I want an ashtray, even if it's for parking meter change. I have the pancake 27mm Fuji lens just because of the form. Thanks for your article.

Gato said...

As I was reading I kept expecting you to mention the 24-105 and 24-120 lenses you've talked about recently. I'd like to see a side-by-side comparison on the various brands.

During my time with FF Nikon the 24-85 3.5-4.5 became my favorite zoom, even over the 24-120. A constant f4 or 2.8 in that range would be very interesting.

As an m4/3 user I'm pretty happy with the 12-60 Panasonics, either version, though a 12-50 or 12-45 2.8 would get my attention if the size and price was right.

Michael Matthews said...

I’ll take a 23mm version of the 18.5mm lens affixed to the Fuji XF10. Leave all the other FX10 camera features alone, except for replacing the 15 fps 4K video with 24/30 fps 4k. Give the lens a set of filter threads to accept a neutral density filter for shooting daylight video or a polarizing filter for shooting stills. Or just give the existing 18.5mm lens threads so that can accept those filters and the option of a very mild telephoto adapter to narrow its field of view to that of the 23mm when desired. Offer it at the FX10 price. I’d buy one tomorrow.

tnargs said...

I think Tamron have reently been releasing constant f/2.8 zoom lenses for the Sony FE syatem, that are surprisingly small and probably make the kind of tradeoff that you are promoting, Kirk.

Andrea Bellelli said...

At some point in my life I had a small canon point and shot with a 35-135 (equivalent) zoom, which I found very convenient. Today I habe two olympus micro four thirds bodies one with the 9-18 mm, the other with the 40-150mm. However, I would gladly replace these with a 10 mm prime and a 17-70 mm zoom if these existed and were as lightweigth as my current lenses (my whole kit weigths less than 1,5 kg).

Patrick Dodds said...

I have high hopes for the Nikon 85mm 1.8 S...

Ronman said...

One of the reasons I was drawn to the Fuji was I liked the 16-55 f2.8, giving me a 24-83mm equivalent . I shot with a Nikon 24-120 for several years and found it convenient, but rarely used it beyond 90mm . I'm content with a 24-70, but having an additional 12 mm is very nice .

Anonymous said...

"Any really meaningful evaluation of lens performance should probably included testing at least three samples of each model"

A great idea, but the downside is, do consumers have the same choice when purchasing, especially when purchasing online? Would also suggest that the testers give results for each of the three- consumers would like to be aware of the existence and extent of variation.

Additionally, shouldn't there be a standard range of acceptability, and if an item is not in spec/ out of tolerance, and hence sub-par, shouldn't manufacturers, being aware of unit variability, just not ship it? It should be the manufacturers' responsibility, not the consumers, to minimize unit variance.

Final thought- No doubt lens assembly is very complex, but unless there is hand-fitting (and I don't know that there isn't), consumers would be happy if say possibly even more complex equipment- automobiles, for example.

Really really final thought- lens cost and assembly today may be simpler than lens assembly 40 years ago, simply because lenses in large part no longer need to be so optically correct, since a lot of the correction is performed in firmware/software.

I guess this is a long way of saying: Lenses today are not cheap. Consumers should demand consistent quality.




Kirk Tuck, Photographer/Writer said...

Anonymous, must be nice to live in a dream world where manufacturers scrupulously inspect every item, watch out for the consumer and minimize sample variation. The real world is different. We try to buy as many of our lenses as possible from a local bricks-and-mortar store specifically because we know there is sample variation and they have been very good about trading out poor performing lenses for good ones.

Today's lenses require lots and lots and lots of hand assembly. More so with zoom lenses. The most with complex, fast zoom lenses that feature image stabilization. So much has to be just right for the lens to perform correctly.

Lenses are much more complex today because 40+ megapixel sensors require 10x the resolution and calibration accuracy as did lenses designed to deliver about 6 megapixels worth of data on film. If even that much. A zoom lens I bought yesterday features over 20 glass lenses elements in 15 groups. That's much more complex that just about any zoom being designed even ten years ago. And the designers have to work around the variability of I.S.

I think the big lens makers generally do a great job in the factory but where everything falls down is subtle damage in shipping and, generally, consumer abuse of product. Study after study has shown that American consumers in particular are brutish in their care of gear. They toss stuff around, use excessive force on the controls, drop stuff, etc. and then hastily blame the makers.

If consumers "Should" demand consistent quality then manufacturers are right in demanding that consumers treat the gear as somewhat delicate, precision equipment that must be taken care of correctly. Also, current top-of-the-line lenses are much more expensive to make and the overall market is subsiding which means that economies of scale are less generous. If consumers want quality they should be ready to pay for it and then take care of what they've paid for instead of blaming makers for their own cavalier mis-use of said gear. Just because some lug head on YouTube puts his "test" camera or lens in the shower to show off it's water "proofing" doesn't me subjecting optical and electronic gear to "survival tests" is something that makers should cover.

Really final thought, adjusted for both inflation and quality lenses are cheaper than they've ever been. They just seem expensive because real wages haven't risen in the U.S. since 1991 (adjusted for inflation).

Anonymous said...

Interesting insight, out of curiosity, I tried to find new lens prices from an arbitrary time in the past- here's one from 1981 (hopefully I can post a link here- http://forum.mflenses.com/some-1980s-legacy-lenses-prices-t77129.html)

"According to the Bureau of Labor Statistics consumer price index, prices in 2018 are 176.25% higher than average prices throughout 1981"

So we can multiply those prices by 1.76 to get today's rough equivalent.

Another thought- in those days US retailers were still able to discount, and advertise discounted prices.

Kirk Tuck, Photographer/Writer said...

It's a bit disingenuous to compare a 6 element, 5 group manual focus lens with a 15 element lens that also includes image stabilization but I get your point. It's pretty hard to compare like for like. Same as in cars. Airbags, ABS brakes, higher efficiency --- would the prices be roughly the same? I don't think so....

Mark the tog said...

Back in the 70's I used to read all sorts of reviews of the most pedestrian lenses because the photo magazines would review them all. They did check centering and resolution on one sample and then did IQ tests with Kodachrome 25. Every lens tested was different in the lab but uniformly delightful in the field.
My friends and I never really worried about the IQ just whether the lens was ultra wide or a zoom. Back then those were the two dream lenses we wanted.

Anonymous said...


The other thing, the older manual focus lenses mechanically and structurally were different: they had metal bodies, they had to provide mechanical linkages to the camera body for auto-aperture and for metering, they had to incorporate very precise, buttery smooth manual focusing rings, with just the right amount of resistance, just the right amount of give, and repeatability, as well as manual aperture setting rings. Additionally, there was more glass used and no, or almost no, plastic in lens elements. And all this mechanical complexity had to stand up to the wear and tear of normal use, year after year.

So absolutely correct with the car analogy. So much more convenience and safety and features today, which makes it difficult to compare apples and apples.

Jock Elliott said...

I posted this on cameraderie.org:

An excellent article! He says stuff that I don't see articulated elsewhere.

And I mostly agree with him . . . not about what qualifies as my perfect lens, but about the approach that review sites are taking.

It strikes me that camera bodies are mostly excellent these days, as are the sensors they contain. Further, photo processing software is also pretty darn good, making it easy to improve a photo and get it close to where you want it to be. But the best camera body in the world won't deliver the goods if you put a deficient lens on it.

His key paragraph, I think, is this one:

Instead of comparing the Nikon Z7 to the Sony A7r4 camera bodies I'd love to see a side-by-side comparison and in-depth review of each maker's "holy trinity" of zoom lenses (14-24, 24-70, 70-200) to see where the whole system stands, not just the naked body. Perhaps a comparative overview of all the different macro lenses on the market. How about a toe-to-toe exploration of zooms in the 100-400mm focal length ranges? And with all the new and pricey 50mm lenses coming out maybe we could see a nice comparison there.

"to see where the whole system stands" -- that's the heart of it, isn't it? You're not just buying a camera body, you're buying a system, whether you realize it or not. And the lenses deliver the light to the sensor, so if they are not doing what you need them to do, or you can't get the lenses to meet your changing requirements in the future, there's a problem.

There's another problem as well, and that is that you probably need to know a fair amount about optical measurements to be able to write sensibly about lenses and what you're likely to see in photographs taken through them.

In all, I thought it was an excellent article, and it points out a serious flaw in the photo equipment review process.

Cheers, Jock