Wide angle: compromise, or resolution through a workaround?

[David Lyga]

Simple and difficult question: Are wide angles for SLRs inherently inferior to wide angles for rangefinders (due to the (in)ability of the the rear element to get close to the film plane)?

Theoretically, rangefinders have an advantage in that there is no restriction placed upon how close that rear element is permitted to be to the film plane. But, can we thus conclude that, therefore, rangefinder 'wides' are better? Take, for example, the best Nikon '28' vs the best Leica RF '28': is there an optical difference that manifests with final print resolution? Or has the ability been attained for SLR objective manufacturers to surmount this theoretical problem with an effective and complete workaround?

My question does not end with the more common 28mm focal length. How do other wide focal lengths compare? Given the high optical quality, it is difficult to say with unhesitating and steadfast conviction that a Nikon, or Canon, or Pentax, or Minolta SLR wide objective is inferior to a Leica RF's, or Contax RF's wide. But is that truly the case? Is that quandary the underlying Achilles' Heel for SLRs? Or has the problem been completely resolved with, at least, the higher-end SLR wides?

(NOTE: I am not talking about special SLR wides that require mirror lockup.) - David Lyga

 

[David Allen]

Interesting question and yes, theoretically, it is at least much easier to design a wide angle for a rangefinder.

I suspect that this question will open up a can of worms with everyone defending their favourites.

I used a Hasselblad for all of my commercial work. Whilst the optics were good, my main reason for choosing the Hasselblad was that it was a complete system with the easy opportunity to hire particular lenses as and when needed. For my personal work, I have used a Fuji 6 x 9 with wide angle lens and, for the past 12 years, a Mamiya 7 with a 65mm lens.

From my experience, the lenses on the Fuji and Mamiya outstripped the 60, 50 and 40mm Hasselblad lenses. The Hasselblad lens that matched the quality of the Fuji and Mamiya was the Hasselblad SWC - but that has no mirror so can be designed like a rangefinder lens.

Bests,

David.
www.dsallen.de

 

[Slixtiesix]

David, were it the old designs you used on the Hasselblad or the 50FLE, 40FLE?

 

[David Lyga]

I suspect that this question will open up a can of worms with everyone defending their favourites.


www.dsallen.de

Do not worry: controversy is my forte, David. - David Lyga

 

[Xmas]

If you are photographing buildings you might be worried about distortion.

But you will get 'projection' distortion with any wide ie don't use them on groups of people.

 

[flavio81]

Simple and difficult question: Are wide angles for SLRs inherently inferior to wide angles for rangefinders (due to the (in)ability of the the rear element to get close to the film plane)?

In theory, yes.

But for example even in the early 60s Canon had the widest wideangle, the 19mm f3.5, as a non-retrofocus design (required mirror lock up).
And then, just two or three years later, they computed a retrofocus version of the 19mm f3.5. And this one is said to have even better performance than the former. Now, this retrofocus lens is far more complex than the predecessor. And bigger.

I would say that, all things being equal (same technology, same possibility of using the best glasses, no constraints on dimensions, same mechanical tolerances), the non-retrofocus lens should be better. But i'm not an optical designer.

Meanwhile many excellent retrofocus lenses have been built...

 

[BrianShaw]

I'll be watching this very interesting thread carefully. Technical inferiority of the SLR wide-angle lens may explain the reason why I've never had much compositional success with wide-angle lenses.

 

[flavio81]

I'll be watching this very interesting thread carefully. Technical inferiority of the SLR wide-angle lens may explain the reason why I've never had much compositional success with wide-angle lenses.

Which wideangles have you used? Once you find a good wideangle, it is such a joy!!

For example i own the aforementioned Canon FL 19/3.5R (retrofocus lens) and i also own the Nikon 20/4.0 AI wideangle (retrofocus as well). The Canon is way, way bigger and heavier than the Nikon. The Nikon lens is even smaller than a 50/1.8 AI nikon lens!! Oh so compact and light. Yet the Canon has lower distortion (and i agree with you Brian, only once the distortion is low, one can feel comfortable with a wideangle). The Canon has less vignetting (obviously -- because it has a huge front lens element). Bravo Nikon, thanks for bringing down the size and throwing the optical performance to the toilet as well... Ok, something's has got to give, i guess.

I had a Canon FD 24/2.8 that had very low distortion as well and i used it a lot because of this. Most shift or tilt-shift WA lenses are low in distortion. My Nikon 35/3.5 shift lens (ancient lens) seems to be low in distortion if you don't shift it too much.

 

[summicron1]

define "better."

Without that, the whole discussion is a waste of time. A 21 on my Leica M has the same angle of view as a 21 on my Leica R, while the R lens uses a different optical theory to get the image onto the film (the rear element of an M 21 mm lens is very very close to the film) the resulting image is still there, still sharp.

Different? I dunno -- don't have an R 21. My R-24 looks really nice, though. Is it better?

The M lens is certainly simpler, optically. Retrofocus opens a big can of worms for optical designers, which is why it took so long for really good SLR super-wides to be developed. Now they are. Which is better depends on which one produces images you like.

Which brings us back to the original question. What do you like?

Define "better" and we can talk.

 

[David Allen]

David, were it the old designs you used on the Hasselblad or the 50FLE, 40FLE?

Towards the end of my time doing commercial work, I had the 40mm FLE - which was an improvement (but mainly because it was not so huge).

Bests,

David.
www.dsallen.de

 

[Slixtiesix]

Thanks. I´m asking because I think about upgrading to the 50FLE in the future... Like to come back to Summicrons argument about how "better" is defined. The Hasselblad 40/4 IF for example has clearly better resolution figures across the whole frame than the 38mm Biogon, but distortion is a lot worse. So which lens is better? Rather philosophical question I think...

 

[EdSawyer]

Yes. (RF wide angles are better). Retrofocus was a workaround. When there are no constraints (large format lenses for example), retrofocus is never the design paradigm chosen.

 

[E. von Hoegh]

Yes. (RF wide angles are better). Retrofocus was a workaround. When there are no constraints (large format lenses for example), retrofocus is never the design paradigm chosen.

That's more than a bit of a generalization. My (retrofocus) 35/2 Nikkor O is a better lens than my (non-retrofocus) 35/2.8 Jupiter-12. But, the J-12 was designed in the early 30s.
IIRC the Super Angulons were a bit retrofocus - there's an evenness of illumination advantage.

 

[Dr Croubie]

The one thing that rangefinder lenses (or indeed any other non-retrofocus lens like LF) are not better at is vignetting.
Take your regular biogon-type lens, the centre of the film is maybe 2mm from the exit pupil, and the corners of the film might be more like 16mm away. That extra distance equals less light on the film (I can't remember if it's one or two cos-terms).
Compare that to a more-telecentric design like a retrofocus SLR lens, the centre of the film may be 50mm away, but the edge may be only slightly worse like 60mm away.
That's only 20% instead of 800% difference (even though I pulled those numbers from nowhere, it'll be something like that).

But anyway, all things equal, in general an RF lens will be more symmetrical and have less distortion. In general an RF lens will be closer to the film plane and vignette more. In general the RF lens will be smaller (ever seen a ZF 21mm distagon next to a ZM 21mm Biogon?).
But in general doesn't really exist in real life. For a start, comparing lenses (especially wides) made more than 10 years apart is pointless these days, computer simulations have come so far lately. Look at the performance of Canon's latest 24 IS and 28 IS vs their 20-year-old original designs.
The performance is more proportional to R&D money than flange-distance. Think about how many (d)SLRs are sold these days compared to rangefinders / P&S / compacts. Even if a few years ago the P&S etc market sold a lot more lenses, the buyers weren't as discerning with regards to image quality, only SLR and the very few Leica/ZM/CV shooters really demanded the very best IQ. And dSLRs have outsold Leica/ZM/CV by orders of magnitude for decades. Plus there are more players and more competition in the SLR space, look at the Zeiss vs Sigma Art battle of 50/1.4 and 35/1.4 lenses in the last year. That's not happening in the RF market, there's Leica, Zeiss, and CV, all at their own very distinct price-points and no real competition between them.
Ergo, more money for R&D budgets in the SLR space, and so SLR lenses are just going to have all the better IQ these days, for no technical but a lot of business reasons.

 

[David A. Goldfarb]

In general, rangefinder wideangle lenses should be sharper, but there are some very sharp modern SLR lenses, like the Canon TS-E 24/3.5 L II, which I got mainly for its T/S functionality, but it's also astonishingly sharp and low in distortion.

That said, if I don't need the agility of a small handheld camera, I'd rather shoot wide with a larger format, because a wide lens can take in so much information, and a larger piece of film can render more detail, even if the comparable lens doesn't have as much resolution as the 35mm format lens.

 

[cuthbert]

That's more than a bit of a generalization. My (retrofocus) 35/2 Nikkor O is a better lens than my (non-retrofocus) 35/2.8 Jupiter-12. But, the J-12 was designed in the early 30s.
IIRC the Super Angulons were a bit retrofocus - there's an evenness of illumination advantage.

IMO the Jupiter 12 is one of the best wide angle available for rangefinder...if you get a good one.

I prefer mine to the Summaron 2.8 on the Leica M3.

 

[Sirius Glass]

All lens designs have design trade-offs, which trade-offs makes a difference. One cannot in absence of testing or analysis assume that a rangefinder wide angles is better or worse than an slr wide angle lens.

I have a WideLux 7 [35mm] panorama camera with a swing lens that covers approximately 120 degrees.
I have a Nikon 20mm to 35mm zoom lens. Zoom lenses tend to have more compromises that fixed focal length lenses.
I have the Hasselblad 903 SWC 38mm Biogon lens which is a lens and camera. I also have a 50mm Hasselblad lens.

 

[ic-racer]

? Is that quandary the underlying Achilles' Heel for SLRs? Or has the problem been completely resolved with, at least, the higher-end SLR wides?

(NOTE: I am not talking about special SLR wides that require mirror lockup.) - David Lyga

To compare any two lenses without having them in hand, the MTF curves may be of benefit. In terms of overall utility, the SLR lenses can have features not found in rangefinder lenses, like autofocus, more even field illumination, wider apertures and high quality WA zooms that may sway one's opinion over which is better.

 

[250swb]

The rear element of the Jupiter 12 nearly touches the shutter and is a work art in itself, but it is only a 35mm lens, so by any standards it is not an extreme wide angle design necessity to need such a large rear element.

Equally good lenses have the rear element where you'd expect it to be, yet the difference would be the character of the lens, not the overall performance. So it illustrates that rangefinder lenses offer more design options, and sometimes this feeds into overall performance, sometimes it feeds into 'character', and sometimes both. Now if you are looking at overall performance I don't think there is any difference between the best SLR or rangefinder lenses in a wide angle resolution battle. What the SLR can't do however is adopt the mad Jupiter 12 (it didn't seem mad at the time) idea of design unless the mirror is locked up. This is where I think the confusion between 'better or worse' comes in. You can have Jupiter 12's that have good performance and character, you can buy another type of lens and it will have yet another type of character and the same performance, and the differences can be identifiable in a photograph. This is less so with SLR lenses, but not impossible, but I'd argue you'd need to go to even more extreme wide angles to begin to identify lenses by a photograph made with them.

Steve

 

[grahamp]

Computer aided design, modern glass, aspherical elements, and improved coatings have gone a long way towards offsetting the issues of retrofocus (and telephoto and zoom lenses). Wide angle lenses do have the unique problem of edge illumination fall-off. To a certain extent, lens design seems to have been pragmatic. A lens has to be good enough for the job, but it does not have to be better than that. It seems that as the tools improve, lenses get wider, not necessarily better.

I cannot think of a case where I would change the camera just to get a particular lens, but I am sure some people do. The 38mm Biogon (I think) would be a case in point for some professionals.

 

[flavio81]

Computer aided design, modern glass, aspherical elements, and improved coatings have gone a long way towards offsetting the issues of retrofocus (and telephoto and zoom lenses). Wide angle lenses do have the unique problem of edge illumination fall-off. To a certain extent, lens design seems to have been pragmatic. A lens has to be good enough for the job, but it does not have to be better than that. It seems that as the tools improve, lenses get wider, not necessarily better.

I cannot think of a case where I would change the camera just to get a particular lens, but I am sure some people do. The 38mm Biogon (I think) would be a case in point for some professionals.

I agree with Graham... And i also change camera just to use some lens; happens to me when switching from Canon to Nikon and viceversa!

 

[flavio81]

It seems to be a matter of cost. The more you are willing to pay, the better any design can be, and a state of the art retrofocus lens could be a better performer in most if not all respects compared with a non-retrofocus design. The recent Zeiss "Otus" line is worth looking at here. Zeiss chose to use a retrofocus (Distagon) rather than Planar design for the 55mm. By some accounts these Otus lenses are blowing even Leica rangefinder lenses away. They ain't cheap though.

Excellent observation, i wasn't aware of this.

The Standard lens for the Mamiya RB, the 90/3.8C, is also a retrofocus design and overly complex for such a focal length and aperture; 7 elements in 5 groups. This due to the lens-to-film distance being VERY long on that camera. Otherwise a 90/3.8C lens could have been easily done with just a 5-element-in-3-group, which is what Mamiya did with with the 127/3.8C lens; and with the 80/2.8 for their TLR cameras, with excellent results.

However, despite it being such a retrofocus lens, the performance is excellent in all fronts.

 

[Alan Klein]

I always thought my Nikonos Underwater camera's 35mm lens shot very clear and sharp photos with a lot of contrast, above and below water. The camera is basically a rangefinder. Considering the lens was designed by Nikon to be waterproof and to work above and below, that always impressed me. These were all shot with it. Of course, the 35mm slide film being scanned, it may be hard to confirm. https://www.flickr.com/photos/alanklein2000/sets/72157625526207614/

 

[E. von Hoegh]

IMO the Jupiter 12 is one of the best wide angle available for rangefinder...if you get a good one.

I prefer mine to the Summaron 2.8 on the Leica M3.

I apparently got a good one, and I have negatives from a prewar Biogon to compare with. My Arsenal 1961 Kiev mount J-12 performs like a coated prewar Biogon in this instance, I'm very pleased with it.

 

[Dr Croubie]

Zeiss chose to use a retrofocus (Distagon) rather than Planar design for the 55mm. By some accounts these Otus lenses are blowing even Leica rangefinder lenses away. They ain't cheap though.

True, but don't also forget that the Otus line was specifically designed for future extremely-dense digital sensors, and digital sensors work best with more telecentric designs due to the nature of photon-wells and all that.
Planar designs are great, but they do curve out right to the outer element; my 80mm Biometars have rather bulbous rear elements.
One reason they chose the Distagon design for the 55mm Otus is because it meant they could get more correction elements into the design before they hit the ~45mm flange distance (it's 12/10 from what I see here).
Planars being pretty-much symmetrical means any extra elements would have gone deeper in and hit the mirror.
Their 85mm Otus has a regular Apo-Planar moniker, because with more room between focal-length and flange-distance they can make it a lot more symmetrical and don't need to go retrofocus. (even though it's 11/9 from Dead Link Removed, and doesn't look very much like a traditional symmetrical Planar, it's more symmetrical than the 55mm).
Also, the 55mm being retrofocus means that it would have a much larger image circle than normally needed for 135, in effect it only ever uses the very-sharp centre portion of what it's fully capable, so no dying in the corners like a regular lens. 85mm Planars usually have enough image circl for 6x6, so cropping that to 135 means there was no reason to make it any bigger by also making that retrofocus.

If anything, the Otus line just proves what I said earlier, performance is more closely related to R&D money and price, than by design heritage...