Lowest distortion Nikon wide angle

[martynas_kundrotas1]

<p>I have a Nikon FM2 camera with a 50mm 1.8 and a 28mm 2.8 AIS. While i'm quite happy with the quality of both, i really hate the distortion of the 28mm. I'm talking about barrel distortion and edge distortion, what other alternatives are there? I really like to put things, people at the edges, so that's kind of an issue, and i often shoot urban landscapes which includes alot of straight of lines. I wouldn't want to go anywhere narrower than 28mm. I also was looking at medium format, the Mamiya c330, but i can't really find any info about the lenses, how much better do they perform? Thanks</p>

[carl_becker2]

<p>The 65mm lens for the Mamiya 7 is very well corrected. See what Bjorn has to say about the AIS 28mm f2. <a href="http://www.naturfotograf.com/index2.html">http://www.naturfotograf.com/index2.html</a></p>

[ShunCheung]

<blockquote>

<p>I really like to put things, people at the edges, so that's kind of an issue</p>

</blockquote>

<p>If you use a 28mm lens on an FX body (including 35mm film bodies) and put people at the edges of the frame, the type of distortion you will see is related to the perspective, similar to the type of exaggerations on the polar regions on a flat world map. In other words, that is not the result of poor optical design.</p>

<p>That is why for group shots, I rarely use anything wider than 35mm on FX bodies. If you insist on using 28mm lenses and put people near the edge of the frame, the most expensive 28mm lens will give you the same type of distoration.</p>

[luis_g]

<p>Shun is right. What the OP is seeing has nothing to do with optical design. This effect can be produced with longer lenses than 28mm simply by moving the subject close to the lens. It is easy to produce what is commonly called "light-bulb head" in a subject with a 50mm.</p>

[ilkka_nissila]

<p>I think it is possible to reduce this problem by simulating a cylindrical projection in software. It would improve the appearance of people towards the edges but then straight lines will be curved which may not be what you want.</p>

[jean_yves_mead]

<blockquote>

Shun is right. What the OP is seeing has nothing to do with optical design

</blockquote>

<p>

Part is - the problem of straight lines near the edges not staying straight. It's a common problem with

wide-angle lenses on SLRs (well, retrofocus designs, really). A reasonably cheap solution, especially if

the OP intends to go wider still, would be to pick up a used Voigtlander R (or earlier screw-mount

rangefinder camera) and the Voigtlander lenses in his preferred focal lengths. As long as he doesn't need fast glass, he should be able to find something to suit his needs, starting from 12mm.

</p>

[akira]

<p>The distortion that annoys the OP is the perspective distortion and not the barrel-shape distortion. So, using a wideangle lens of symmetrical optical design doesn't solve the problem. Ironically, Ai-s 28/2.8 is one of the most well-corrected wideangle Nikkor SLR lenses in terms of the barrel-shape distortion, which kind of exaggerates the perspective distortion of the people depicted at the edges.</p>

<p>I think that the easiest way to milden the perspective distortion is to use a lens that suffers from strong barrel-shape distortion or use the widest end of a 28-XXmm zoom lens (the widest end of a zoom is mostly suffer from the barrel-shape distortion).</p>

[Andrew Garrard]

<p>KR likes Fisheye-Hemi software for cylindrical projections; I've never used it and I don't know how well it copes with a rectilinear source, but his review has some images if that helps. Obviously you don't get to keep straight lines. To get a 28mm rectilinear photo to look right, you have to print big and stand close (to the centre) - there is no "distortion".<br />

<br />

I once heard a discussion about some conference-call hardware that used a fish-eye lens. Someone requested that "the table should look straight" and "the people should be made to look forward". People have funny ideas about how light travels.</p>

[rodeo_joe1]

<p>The stretching of bodies and faces at the edge of a wideangle frame is a real distortion - that's why the old-fashioned photographers who specialised in large groups used a swing-lens camera and arranged their groups in a part-circle.</p>

<p>Modern wideangle lens designs give a false perspective that's more suitable for architectural shots and copying flat artwork. This rectilinear rendering can't actually be seen in real life, and in fact barrel distortion is much closer to a natural perspective, where things get smaller toward the edges of the frame. If you were to arrange a group in an exact semicircle around a fisheye lens, you'd see that this is so.</p>

<p>Anyway. The 28mm f/2.8 AiS Nikkor has fairly low distortion as far as rectilinear rendering goes, but it's such a poor performer otherwise (regardless of what K.R. thinks!) that mine lies unused and unsellable. The f/2 28mm AiS Nikkor is a far superior lens, with about the same or better level of distortion but much better edge definition. It also has a slightly wider angle of view.</p>

<p>I'm attaching a composite of the four 28mm lenses I have at my disposal - the f/2.8 and f/2 MF Nikkors, an f/2 MF Kiron and a Tokina 28-70 ATX pro zoom. As you can see the Tokina zoom shows by far the most rectilinear "distortion" and the least angle of view. The Kiron appears to me to give the least distortion and the widest angle; drawbacks are that the colour rendering is a bit cool, and finding one of these lenses in good mechanical order isn't easy these days. The two AiS Nikkors are close in terms of distortion, but the f/2 is a better lens with a slightly wider angle.</p>

<p>BTW, the only real distortionless lens you'll find is a pinhole!</p><div></div>

[dan_brown4]

<p>Zeiss Distagon in ZF mount.</p>

[Andrew Garrard]

<p>Joe - it depends how you define a distortion. If you make a flat print and place your head so you get the same field of view as a wide-angle lens, the rectilinear projection is correct (it's also what a pinhole will do). The problem is, getting the same field of view as a 28mm lens is tricky - you need to be very close to a big print.<br />

<br />

To get a print of a large group that "looks right" at a normal viewing distance just means moving the

camera farther away - but for a big group, that's quite a long way. Cylindrical projection looks right

if you move your head as you look at different parts of the image (especially with the subject

wrapped around the lens), which is probably a good compromise.</p>

[keith_b1]

<p>To basically echo what then others have posted, the Nikon 28/2.8 Ais is one of, if not the most, distortion-free 28mm wideangle ever produced by any manufacturer. You'll never eliminate the football-shaped heads at the edge of the frame(volume anamorphosis) with any 28mm--it's too wide. 35mm is the widest that'll allow you to freely shoot people without football-head worry. The Zeiss 28/2 has noticeable barrel distortion, as does the Nikon 28/2. So does the Zeiss 35/2, btw. Also, that was one of the worst brick wall tests/demos I've seen...too small an area and uneven bricks. The whole point is that the grid of bricks has to be perfectly regular in size and spatial relationships. Brick walls with square bricks are much nicer, though harder to find.</p>

[photo5]

<p>That's the first time I've ever heard someone describe the 28mm f2.8 AIS Nikkor lens as a very poor performer. Rodeo Joe, have you considered the fact that you may have a lens with a manufacturing flaw? Bjorn rates this lens a 5 out of 5 on his <a href="http://www.naturfotograf.com/index2.html">lens tests</a> website, and gives the f2 28mm Nikkor a 4.5 out of 5. I've owned this lens for years and have always seen stellar results from it, and I've never felt the need to have the f2 version. Paid only $90 for the 28mm f2.8 AIS lens in near mint condition, and they now usually sell for around $150-250 on the auction site.</p>

[tri-x1]

<p>Reducing distortion with a wide angle is partially a matter of keeping the camera level. It the lens is tilted down or up there will be noticeable distortion.</p>

[rodeo_joe1]

<p>Oh well, at least I took the trouble to post some pictures. I read a lot of yak in these threads, but don't see many visuals to back any of it up. And I'm sorry that the interesting walls of my house don't meet strict distortion testing standards - I'll just have to move to a boring modern dog-kennel.</p>

<p>BTW, I've read the Bjorn rating and almost entirely agree with his assessment of both the f/2.8 and f/2 28mm Nikkors. Read carefully what he says about the CA and poor corner definition of the f/2.8.</p>

<blockquote>

<p>"For distant scenes, however, corner sharpness isn't that remarkable and ghosting under adverse conditions can be troublesome. There is some corner colour fringing present, too."</p>

</blockquote>

<p>For "isn't that remarkable" read extremely poor at anything wider than f/5.6.<br>

These faults alone ruin the use of that lens for me, and I'd rather have a slightly higher distortion (huh?) with good edges and fringe-free corners. Plus the wider aperture makes for far more accurate manual focusing in this focal length.</p>