Making a fast wide angle (SLR) prime lens

[daveleo]

For their APS sensored cameras, Nikon, for example, makes several zoom lenses that start down at 16mm at the wide end. This gives a "FF equivalent" view angle of 24mm on their APS cameras.
Many people are waiting for a prime 16mm lens, and there are not even rumors about one coming along.

My question is for the optical tech people out there: Is it really technically difficult to design a 16mm prime lens when you already have zooms at that focal length?

 

[robbieduncan]

How fast would you want it to be? How much would you pay. Canon have the EF 14mm f/2.8 L. OK, it's a full-frame lens so a lot of glass. And it's very expensive.

 

[Jamie Pillers]

But the OP's question still remains: Why don't Nikon/Canon folk produce a 16mm prime? And I'd add: Why can't they produce an INEXPENSIVE 16mm prime?? Afterall, they can produce inexpensive zoom lenses with 16 to 18mm wide end of the range.

My half-baked theory is that they think 99% of us don't want prime lenses and thus they don't think it would be profitable. I think they are out of their minds. Just like I think they've completely blown it by developing only SLRs with WAY, WAY too many bells and whistles incorporated. I don't think they understand that the only reason they're able to keep selling all these big black behemoths (bodies and lenses) is that there aren't any alternatives. Maybe the mirrorless camera sales will finally be the wake-up call they've needed for the past 20 years!!

 

[ronnies]

Nikon already have a 14mm f/2.8. It's also an FX lens.

Ronnie

 

[rbsinto]

But the OP's question still remains: Why don't Nikon/Canon folk produce a 16mm prime? And I'd add: Why can't they produce an INEXPENSIVE 16mm prime?? Afterall, they can produce inexpensive zoom lenses with 16 to 18mm wide end of the range.

My half-baked theory is that they think 99% of us don't want prime lenses and thus they don't think it would be profitable. I think they are out of their minds. Just like I think they've completely blown it by developing only SLRs with WAY, WAY too many bells and whistles incorporated. I don't think they understand that the only reason they're able to keep selling all these big black behemoths (bodies and lenses) is that there aren't any alternatives. Maybe the mirrorless camera sales will finally be the wake-up call they've needed for the past 20 years!!

I"ll bet that their Market Research, unlike your personal opinions have been carefully researched and so, is fully-baked.
Exactly where is your data showing a great need for a 16mm fast prime by enough photographers to make the R&D costs required recoverable, and the sales profitable?
I'll bet if you could twist Nikon's corporate arm they could supply data showing why they aren't making them.

 

[daveleo]

My original question is legitimate and technical, and I THINK that the answer is . . . .

it is NOT technically difficult to make a (let's say) 16mm/f2.8 in APS format and sell it reasonably cheap. I speculate this because Nikon already makes a cheap and (arguably, but let's not) decent kit zoom that goes to 16mm. So I am provoked to ask, why not strip the zoom capability out of that and add some glass to get f2.8? Is that technically hard? I think not, but then again, I am not an optical engineer.

I am not asking a marketing / sales / profitability question.

 

[Godfrey]

..
it is NOT technically difficult to make a (let's say) 16mm/f2.8 in APS format and sell it reasonably cheap. I speculate this because Nikon already makes a cheap and (arguably, but let's not) decent kit zoom that goes to 16mm. So I am provoked to ask, why not strip the zoom capability out of that and add some glass to get f2.8? Is that technically hard? I think not, but then again, I am not an optical engineer.

A fast ultra-wide prime for DSLRs would cost as much and be at least as heavy/bulky/complex as a zoom of similar specification and performance. And have nearly all the same imaging issues. At which point, who would want to buy it over a similarly fast, top notch performing zoom lens? Doesn't make any sense to buy a 16mm f/2.8 vs a 16-35mm f/2.8 if the performance is no different.

The issue, optically, is that with an SLR lens mount register of 40-50 mm, it is difficult to make ultra short focal length lenses that pose any significant advantage over the zooms you can construct. This has always been the case even with 35mm format SLRs, and it is exacerbated with smaller formats and larger maximum apertures.

The shorter lens mount register of rangefinder and TTL electronic cameras changes this: with a short mount register, it is far easier to produce a fast, short focal length lens that outperforms a zoom lens of similar specification.

 

[Aristophanes]

2.8 is tough, but f/4 in an almost-pancake, APS-C only, is not:

http://www.pentaxforums.com/forums/lens-clubs/86234-15mm-limited-controls-my-mind-club.html

If you want a 14/2.8:

http://www.pentaxforums.com/lensreviews/SMC-Pentax-DA-14mm-F2.8-Lens.html

As others have said, a 16/2.8 is covered by profitable zooms.

 

[daveleo]

Thanks for your answers, esp the technical issues.

Boy that Pentax 15mm looks nice !

My big dislike is physically large lenses (I consider the Nikon kit 16-55 "large" and it is generally "small" by other people's standards). Those super-wide zooms with 77mm filter rings are just not my cup of tea. I'd almost go down in sensor size to M4/3 format to keep the total package size reasonable.

I have two zooms and 5 primes for my APS Nikon, and I almost never use the zooms because they are too long for me. I love the 24mm/f2.8 and the 50mm/f1.8 (the 52mm version) for their size. I was hoping for a 16mm/f2.8 or f4.0 along those lines. (I will probably buy an old 20mm/f4.0 at some point - the Nikon 16mm is a fisheye.)

Regarding lens sizes, the rangefinder concept is far better than the SLR concept.

Again, thanks for the replies.

 

[rxmd]

My original question is legitimate and technical, and I THINK that the answer is . . . .

it is NOT technically difficult to make a (let's say) 16mm/f2.8 in APS format and sell it reasonably cheap. I speculate this because Nikon already makes a cheap and (arguably, but let's not) decent kit zoom that goes to 16mm. So I am provoked to ask, why not strip the zoom capability out of that and add some glass to get f2.8? Is that technically hard? I think not, but then again, I am not an optical engineer.

I am not asking a marketing / sales / profitability question.

Well, but you are!

It's not that difficult to build a 16-ish lens with a small image circle; Pentax does it. So either you assume Canon/Nikon/... are evil, or you assume they are stupid, or you start asking marketing/sales/profitability questions.

As it happens, Pentax is also the one and only major SLR manufacturer that is not invested into full frame at all. This is nice for Pentax because it allows them to build smaller, APS-only lenses. It is also a problem for Pentax becaus full frame is the new medium format, and everything smaller will sooner or later be dominated by mirrorless camerass, so the Pentax SLR line will sooner or later go the way of the dodo, hopefully (from Pentax' point of view) to be replaced by some Pentax mirrorless system (but for that, in turn, the lenses are too big, because of the legacy SLR lens register; compare Fuji X primes to Pentax primes, for example)

Canon, Nikon, Sony have few incentives to build high-end APS primes that you can't use on their high-end bodies. So they build medium-end zooms that will do the job, and that also, as it happens, aren't much bigger than a prime would be, because of the lens register issues mentioned above. Case closed.

 

[robbeiflex]

I"ll bet that their Market Research, unlike your personal opinions have been carefully researched and so, is fully-baked.
Exactly where is your data showing a great need for a 16mm fast prime by enough photographers to make the R&D costs required recoverable, and the sales profitable?
I'll bet if you could twist Nikon's corporate arm they could supply data showing why they aren't making them.

Market research is typically a great way to prove to yourself that your customers want more of the same, while ignoring what you might actually sell to those who are currently not your customers. It does, however, CYA in case of failure. 😉

Cheers,
Rob

 

[daveleo]

Well, but you are!

It's not that difficult to build a 16-ish lens with a small image circle; Pentax does it. So either you assume Canon/Nikon/... are evil, or you assume they are stupid, or you start asking marketing/sales/profitability questions.

. . . . . .

Everything else you said is true and knowledgeable, but this is wrong. I was not asking "why" Nikon, Canon do not produce such a lens (I was pretty sure it was a marketing decision). I literally only wanted to confirm that it is NOT technically hard to do.

 

[x-ray]

Canon has a 14 f2.8 and a 17 tilt shift but they're not cheap. They also have a 10-22 zoom. I'm guessing there's not much market for a prime 16mm. Zeiss has an 18mm and voightlander has a 12 and a 15mm for RF. I think the 15 CV was made in a Nikon mount that required the mirror to be locked up and Nikon made two 15's at one time. I had the later one and sold it for the 14 which was much better. They also made a 13mm which was very large and extremely expensive.

 

[f16sunshine]

Everything else you said is true and knowledgeable, but this is wrong. I was not asking "why" Nikon, Canon do not produce such a lens (I was pretty sure it was a marketing decision). I literally only wanted to confirm that it is NOT technically hard to do.

Producing a retro focus wide is not hard to do. It's hard to do well.
Any level of excellence is going to be very expensive. Coming from the Contax SLR system before the digital era. I still own a few of the Contax wides.
Lenses like the Distagon 2.8/21 and 2.8/25 are still considered top of their class. Outperforming The current Wide angle zooms often by a wide margin (punny).
It was never cheap to build or buy those lenses and no matter how good they were they are far from perfect.
Field illumination is not even, distortion is present, corner fall off etc... all the traits of a retro-focus wide angle Yet, still very expensive and not performing as well as their RF cousins (for often much less money).
Those were for full frame at 21/25 (28 etc..)
For Apsc, 14/16/18mm would be the FL to achieve the field of view you seek.
It can be done but it won't be cheap. Canikon would not go to the effort of making a prime without trying to exceed the IQ/performance of their zoom offerings.
There just won't be any cheap fast wide primes coming probably ever.

 

[Godfrey]

2.8 is tough, but f/4 in an almost-pancake, APS-C only, is not:

http://www.pentaxforums.com/forums/lens-clubs/86234-15mm-limited-controls-my-mind-club.html

If you want a 14/2.8:

http://www.pentaxforums.com/lensreviews/SMC-Pentax-DA-14mm-F2.8-Lens.html

As others have said, a 16/2.8 is covered by profitable zooms.

Absolutely. F/4 or f/5.6 are certainly doable. f/2.8 grows the lens far too much, as does anything faster, if you're looking for compact.

This is because SLR lenses in this focal length range must be retrofocus designs to clear the swinging mirror. To make a retrofocus lens fast, the size (and cost!) of the lens' front elements grows enormously as you work towards larger apertures.

The Nikon F lens mount has a register distance of about 45mm. The swinging mirror comes within three or four mm of that register distance. What that means in practical terms is that the shortest symmetrical design lens you can make that won't foul the mirror is around 45mm focal length.

By comparison, a Leica M-bayonet register is about 28mm and there's no mirror flipping around behind it. This means you can insert the rear of the lens much deeper into the camera's mount cavity and can produce symmetrical lens designs down to 18-20 mm, theoretically, without mechanical obstruction. There are some downsides to this sort of design, particularly for digital sensors, which retrofocus designs again solve. But you don't need as strong a retrofocus design to achieve what you're trying to get to so the resulting lenses can be smaller, simpler, and better performing than SLR lenses of similar focal length.

I owned the Pentax DA14mm f/2.8 for a while. It was a great performer, but certainly a large and heavy lens physically (approximately a pound all by itself IIRC). I ended up carrying it very infrequently as it was so bulky and shooting with the far more compact DA 21mm f/3.2 Limited.

(I'd exited using Pentax gear by the time the 15mm Limited became available.)

 

[Speedfreak]

My half-baked theory is that they think 99% of us don't want prime lenses

That is what marketing told them based on consumer surveys. A few nerds on RFF or dpreview wont affect Nikon / Canon much.

 

[gavinlg]

The reality is no one does an APS-C prime because the APS-C/DX sensor in terms of digital SLRs is dying out. Mirrorless cameras now use the aps-c/DX sensor size as their standard, and DSLRs are going to migrate to full frame. Check out the coming Nikon d600 which is full frame and rumored to be coming in at 1500us brand new. Why would anyone buy a 16mm DX prime lens for $500-600 when you could buy a 24mm nikkor lens full frame lens which is faster, smaller and cheaper for the d600. There's also persistent rumors that Pentax will be entering the full frame market with the smallest full frame DSLR available this year.

The point is, APS-C is a bad compromise in the dslr world, mirrorless cameras can use the same sensor and be half the size, and still be as good in areas like ergonomics and AF, whilst 35mm DSLRs take the stage for highest IQ and versatility.

 

[BobYIL]

A few technical points need to be explained:

- From the production point of view the inexpensive zoom lenses for crop sensors are way different than their costly (pro) zooms, no matter what the focal length zone seems to be similar. For the first one the lens elements are produced rather by molding whereas the latter one requires complex grinding/lapping operations. Mounts & assembly also differ considerably, for the former one generally plastic components are employed with over 2/100th mm tolerances whereas their expensive cousins are built mostly by metal components with stricter tolerances.

- Any lens wider than 75 degrees view angle (shorter than 28mm in FF terms) requires complicated formulas against fall-off and sharpness loss toward edges but also distortion. Add to these the complication of having to design it for f2.8 or larger apertures, with zooms this is stretched to extremes. That's why the pro-zooms (any 24-70 for instance) having f2.8 max. aperture is bound to have huge front elements. Also the cheaper zooms exhibit horrible distortion figures.

- Register distance plays an important role: The conventional (RF) type register distances always result in far more compact, better corrected with less distortion and optical deficiencies so consequently much better wide angle designs compared to those for SLRs. However this is true with film only; for digital applications wide RF lenses demand offset microlenses to take care of oblique rays if not also native correction-programs to reinforce it.

It's really hard to make any generalization however based on sensor-lens combination some reasonable "pairings" can be suggested. My humble tests with the Nex-5N sensor indicated the followings:

15/4.5 Heliar and 20/2.8 Nikkor are fine but not brilliant. 25/2.8 Biogon is outstanding, better than any 24mm SLR lens I tried. I found G28/2.8 Biogon performing just a little finer than the Distagon 28/2.8. And if you want the best 35mm for the APS-C then the later (E55) Summicron-R 35/2; better than any other Leica or Zeiss lens I tested on this sensor (and I tried a lot..)

 

[BobYIL]

Check out the coming Nikon d600 which is full frame and rumored to be coming in at 1500us brand new. Why would anyone buy a 16mm DX prime lens for $500-600 when you could buy a 24mm nikkor lens full frame lens which is faster, smaller and cheaper for the d600.

An insightful recommendation against the complications of wide angles with the mirrorless bodies as long as "size factor" is not your priority. Take the new 28/1.8G for example for $699; very hard to duplicate this quality with this price.

 

[rxmd]

Register distance plays an important role: The conventional (RF) type register distances always result in far more compact, better corrected with less distortion and optical deficiencies so consequently much better wide angle designs compared to those for SLRs. However this is true with film only; for digital applications wide RF lenses demand offset microlenses to take care of oblique rays if not also native correction-programs to reinforce it.

Throat size also plays a role here. The wider the throat, the easier it is to build fast lenses, including wideangles. This is why fast Nikon RF lenses tend to be front-heavy, and why there are generally faster SLR lenses for the Canon system than for Nikon. With the new mirrorless systems with their large throats relative to the lens register, it removes a number of constraints.

15/4.5 Heliar and 20/2.8 Nikkor are fine but not brilliant. 25/2.8 Biogon is outstanding, better than any 24mm SLR lens I tried.

And where you have really outstanding SLR lenses in 24mm, they tend to get really big; as remarked elsewhere the 24/f3.5 TS/E II for Canon is just about the best lens in that focal length that I've tried, but it's a big heavy lens with 82mm filter diameter (its image circle would fit a Leica S2 and most medium format sensors, though).