14-24 and field curvature

Discussion in 'Nikon' started by fluppeteer, Nov 26, 2013.

  1. Hi all,

    Quick question. I think (and without setting up some test charts I'm not 100% sure) my 14-24 shows significant field curvature.

    I've attached a (shrunken size, since a 20MB shot of my garden is a bit awkward) image that shows what I believe is going on (yes, I need to cut the grass): with the image centre in focus on a wall, the areas nearer to me than the wall on either side of the garden are more in focus than anything in plane with the wall. In real shooting, I've seen subjects at my feet be in focus when the image centre is metres away, but distant objects at the frame edge have been out of focus - the focal plane appears to be bowl-shaped (though at least not wavy, like the 24-70). This image was at 14mm, if that makes a difference to correlating with others.

    Nikon have checked the lens and claim it's in spec (though I did have to explain that I meant field curvature and not distortion), so I just wanted to know whether others have had similar behaviour from this lens. Since there was a report around here of someone with an incorrect flange distance in their D800, I'm a little nervous that there's a strange interaction going on. Until I noticed the sharp close objects, I was convinced that the lens just wasn't as sharp as everyone claimed in the corners.

    Yes, the problem goes away on stopping down (presumably because the depth of field is extending) - though it's still there until about f/7, by which point I'm past the D800E diffraction limit. It won't be a problem for all shots, but if it's fixable, I'd sooner it be fixed. Nikon are asking if I'm happy with their service (and I'm happy that they don't charge if they didn't fix anything!) so it was about time I sought a second opinion.

    Thanks in advance.
    00cBhN-543799684.JPG
     
  2. I haven`t specifically checked field curvature on mine, but I assume it should be there... I think it was Ilkka who advised to use Live View in order to assure corner focus on test shots for this reason. With this method I noticed that the lens is very sharp to my standards... I must recognize that I have never shot a parallel plane, flat subject in front with this lens.
    I use to take people photos with it, quite close and at wide apertures, so I have never been disappointed with the distortion and possible field curvature issues.
     
  3. Thanks, Jose. I'd certainly expect weirdness if I was trying to corner focus with phase detect (the AF alignment issues were apparently worse with wider lenses), but that may be more to do with telecentricity than field curvature. I mostly care about longish distances for landscapes - with this lens I'm using the aperture for light gathering - but I agree that when I use it for portraits it doesn't matter so much. Just checking whether this behaviour was normal!
     
  4. I suppose a bowl shape models the shape of typical architectural interiors very well, explaining perhaps the better than expected results of the 14-24 in such applications ;-) I haven't had field curvature problems with my 14-24 (that doesn't mean it is not there, just that it hasn't been a serious problem for me), however, when I use it for landscape, I stop it way down, e.g. f/8 to f/11 (or even f/16 if water flow needs to be blurred). I know the sharpness of the 14-24 is not ideal at those apertures, but but depth of field in those near-to-far landscapes is more important to me than peak sharpness in scenes where I have close foreground interest and then something at distance. By contrast, for architectural interiors the 14-24 seems to give very good results even at f/4, which makes it practical for e.g. museums and large public buildings where tripod use may be forbidden and also the people visiting those sites and my picture area might not be still for long enough for me to use a small aperture. The focus distances in these cases are quite long if the interior is large and the field curvature behaviour may be different from the relative close distance you have in your example.
    For your case, with landscapes, I would carefully investigate which focus distance and aperture setting yields the best result. I don't believe I've personally ever used a superwide for distant landscapes, it makes the faraway stuff (such as mountains, distant trees) too small and usually there is too much foreground in superwide shots for the far-away subject matter to make an impression, unless it is cut for a panoramic. If you want a panorama, I recommend using a short tele and stitching rather than using a really wide angle and cropping. I tend to use a telephoto for distant landscapes and wide angle for subjects where there is a close by foreground interest (roughly 50cm to 1.5m from the camera position) and then something a bit further away but even the most distant objects in the frame are likely to be less than 20m away to avoid them becoming too small in the frame. With distant landscapes I use a telephoto to make sure the faraway stuff stays big in the image since it was likely what caught my attention in the scene.
    I would like to be able to recommend a superwide angle with minimal field curvature at long distances but I don't know such a lens, at least if the requirement is that the performance should be high at a relatively wide aperture. The Nikon 24/1.4 is relatively flat field at long distances but you should not assume a perfect image at f/1.4 (and it won't solve the problem for 14mm focal length obviously). The Zeiss 25mm f/2 (note: aperture important) is known as a very high performance lens but I don't know how it works for long distances. If you're going to buy the Otus then I guess you don't think shy of putting down the money for a 15mm Zeiss either ;-) However, you should be aware that Zeiss lenses often sacrifice vignetting behaviour to obtain high sharpness towards the outer parts of the frame; if you want to use wide apertures to gather light in landscape shots then this is something to be aware of. Lloyd's site documents all this very carefully.
    Few sites do testing of wide angles at long distances ... actually I think Lloyd Chamber's blog and Zeiss tests (www.diglloyd.com) might be the best place to look for such things, but I'm not currently a subscriber. I recommend to you that you subscribe to both as you might find it the only place to find the information you seek in this case.
     
  5. Thanks, Ilkka. I'd vaguely been considering subscribing to Lloyd's blog anyway, so I'll have a look. My sample is at somewhat closer distances than I normally focus for this kind of thing - though a few trees in the range of 5-10m away wouldn't be unusual. The attached image (only at 24mm) was an example of wanting an ultrawide; the result works best printed large, so you can get the immersive effect by squashing your nose on the print. I agree that stitching might have been a better approach, logistics aside. This kind of thing only really works when you're on the edge of a cliff with something decent opposite - I'm not against longer lenses for landscape use in general. Nor do I really claim to know what I'm doing! (This was a holiday snap.) Still, having the centre (distant) in focus, the corners (by my feet) in focus, and the frame sides (distant) out of focus was a little annoying. I'll have a play with hyperfocal options.

    I think there's a difference between "interested in buying the Otus" and "can afford to buy the Otus any time in the next few years". :) That and the 400 f/2.8 are on my list before I retire, but I'd hesitate to associate them with "disposable income" in the near future. My impression is that the 15mm Zeiss is not magic - though th 21mm might be - but I may resign myself to more tripod stitching for the foreseeable future. "Doctor, it hurts when I do this"/"Don't do it, then".
     
  6. Whoops. My machine's being so slow that I forgot to attach the image between writing about it and being able to hit send. Sigh.
    00cBjm-543801484.jpg
     
  7. What you need is better light; something like this ... (the first image in the Thursday, December 27, 2012 post)
    http://grandcanyonprints.blogspot.fi/
    or this
    http://travel.nationalgeographic.com/wallpaper/travel/national-parks/your-national-parks-photos/grand-canyon-sunset/
    when you have contrast created by the low sun and clouds, the photograph gains ooomph that no amount of corner-to-corner sharpness can mimic! :)
    Yes, I know that light may not be there when a traveller is, but that's what landscape photography is ... a lifetime of chasing the right light! :)
     
  8. Ilkka: Yes, I know! Sadly, I was on a three-day tour that covered the Grand Canyon, Zion, Bryce, Antelope, Monument Valley and Horseshoe Bend, and I twisted my knee badly on my way into the minibus. I barely made it to the top of the trail, though carrying a 150-500 "just in case" didn't help - on this occasion, panoramic stitching wasn't an option, although I hope the circumstances were unusual. I'm just glad it had stopped raining. It was always supposed to be a scouting party - sometime I'll go back, with my wife and with more time, when the sun is a little lower (it actually was getting later in the day). A rainbow would be nice, too. Arrange that for me?

    Next time, I'll try to take in Yosemite, and try not to be intimidated by the ghost of Ansel. Assuming the whole area didn't turn to ash since I stopped hearing news stories about it.
     
  9. I believe that if you were perpendicular to the subject, in the case of your example, the building, you would not have the issue. You should retest and let us know. I have never checked mine but this issue could happen with any lens if you are not perpendicular to the subject as the distance to the left and right side of the image would be different to that of the center of the frame.
     
  10. Elliot: I was as perpendicular as I could make myself by trying to be (standing in the middle of the garden and shooting straight at the middle of the wall); I admit to not having anything to align the film plane mechanically parallel with the wall. I'm reasonably convinced I wasn't far off, though it was a concern. Particularly, the fence on either side of the image - several feet closer to me (at least along the optical axis) than the house wall - was sharp, as was the house wall in the centre. If I'd just been tilting the lens to one side - or if the lens was decentered - I couldn't have the fence in focus on both sides if the wall was also in focus. If there's anything wrong with that logic, or if you can suggest a better way to align the image, please tell me!

    Edit: Sorry, the labels on the fence saying "sharp" on both sides are a little hard to read. I originally sent this annotated image to Nikon at full size, and I'd not allowed in my font choice for downsizing to pnet's 700 pixel limit!
     
  11. I suggest you do a controlled test where you are perfectly perpendicular to the subject. That would be a starting point to finding out what/where the issue is with your lens.
     
  12. Was your camera set up square to the house? It looks tilted backwards to me.

    Also if you fairly close to the wall, since the center area of the wall is closer than the left or right edges at 14mmi might
    expect to see some apparent geometric distortion.

    How are you processing your NEF files?
     
  13. Elliot: Certainly, although I don't think I can explain the regions of sharpness any way other than field curvature. Um. How do I set up a test where I'm perfectly perpendicular? I could try tying some quite long bits of string to the fence posts. I'm open to suggestions that involve less geometry. I really can't believe either that I was far off square, or that the shape of the sharpness zone I'm seeing could be explained by being off-square, though. I was only doing this experiment because I've seen weird behaviour in the past! The roof lines do make it look like I was some way off, but I'm in line with the planks of my decking and the centre of the frame. If anything, that combination freaks me out a bit, because I'm not sure what's going on with the geometry of my house and the result seems asymmetric. Still doesn't explain the sharp bits of fence, though.

    Ellis: I am prepared to admit that I might not have had the camera perfectly vertical. I think I did (using the electronic level), but it's a little hard to judge - my grass slopes away from the house - and I was paying more attention to being horizontally centred. There's quite a large region of central frame that's sharp, so again I wasn't way off. (I'm trying to get away without finding somewhere to put the full resolution image, but I think I'm struggling.)

    The lens should, optically, be mapping from the sensor plane onto a plane in the real world. I'm not expecting perfection, but I believe I'm seeing a bit more distortion from flat than I'd have expected given that I don't remember seeing this mentioned in many reviews yet (until I've subscribed to Lloyd's reviews, at least). Fully-corrected lenses should not be bringing a hemisphere centred at the camera into focus. I was about 5-6m from the wall, I guess? The decking is about 3-4m deep.

    I think that particular image was a JPEG straight from the camera (and may include distortion correction). I have the raw file; I'll try to have a look at that, too.

    This image was rushed together on a morning before I dropped my camera off with Nikon for cleaning, so it may not be my finest effort. I'll give it another go, but - for a reasonable subject distance - papering my garden with test charts is a bit awkward. In the meantime, if anyone else with a 14-24 has noticed a problem (and I'm talking "not all the horizon is in focus", not something obscure) - or the absence of one - then I'd like to know, preferably before I tell Nikon UK whether I'm happy with their lens repair service. :)
     
  14. Past night I have been reading about aberrations in one of my "bibles" about optics and photography physics, and specially about field curvature.
    To make it short, looks like (as use to be) there is a trade off (a big trade off) between field curvature and other corrections.
    Correction of first or second order aberrations are necessary, but once we are beyond this level, the trade off will become too high, to the point that e.g. a bit of field curvature is interesting to avoid a loose in sharpness... so if the lens is very sharp, it could be "unavoidable" (or even "desirable") to have a bit of field curvature.
    Yes, it`s odd... the first that came to my mind are process and macro lenses... extremely sharp and flat field... well, I`m not an expert. Maybe they are designed for highest performance at close distances, or to work at small apertures. But, what about the 105/2.8AiS, which is a better "at infinite" performer...? Or this only applies to wide angle lenses and/or corner sharpness?
    Obviously I don`t know how "acceptable" the curvature is on the 14-24... to me is just an outstanding lens for people and close photography. The few times I have used it for architecture (indoors&outdoors, mostly for other`s requests), results were satisfying, but I have never used it for long distance or flat front planes! I have somewhere some tests on the D800.
    As I use to think, compared to optics, rocket science is a children's play... :)
     
  15. Jose: Agreed. Besides, apart from GPS, rocket science is mostly just Newton's laws, and Newton got dispersion wrong. (I'm reasonably convinced that brain surgery is actually difficult, though.)

    If my 14-24 isn't flat-field but does behave as expected for a 14-24, I'll live with it - for now, I just wanted to know whether I should still be trying to get it repaired! As for trade-offs, both the 105 micro VR and the 150 Sigma macro OS seem to be weaker at infinity than at close range (as, supposedly, is the 200-400, and even - slightly - the 400 f/2.8). You can't correct everything, though if we customers accept that the nthe optical engineers will stop trying. :)
     
  16. Andrew, I think that you gave earlier a link to Ultrawide comparison in 3d-kraft.com.
    Some of given test images, especially the last ones, show that Nikon 14-24 seems to be doing better in the image corners - when those corners are rather near.
    A "bowl" curvature description given by Ilkka suits this situation well.
     
  17. Lloyd Chambers did observe marked field curvature toward the camera in the corners in the 14-24/2.8, which may help with detail in the lower corners of some landscape subjects (at the expense of distant or infinity sharpness at the middle of the sides of a frame or in distant upper corners). This might be quite bad for subjects where everything is equally distant, but photographic equipment always has limitations and compromises. The 28mm/1.8 G lens also has this sort of field curvature, but I find it very helpful for vertical orientation landscape photos, almost like a built in tilt for Scheimpflug effect for the foreground.
     
  18. Hmm. Thanks, Kari - I'd missed that in my original reading (although it's talking about all the lenses rather than the 14-24 in particular). I'll definitely read Lloyd's articles, although since I've just paid an electrician a lot of money, I may have to put the extra $250 for a subscription on hold for a month or two!
     
  19. When I got my 14-24 new in 2008, I went out and shot some sides of buildings and stuck test charts up on the exterior wall of an outdoor carport, distance about 14-16 feet.
    Typically, at wide open the center was sharp on the focused flat target, but the far edges were a slight touch soft. I took it to the LA Nikon service place, and they were nice enough to shoot tests on their interior wall with test charts stuck on it, and supply me with jpegs of the frames they shot. Theirs were shot at a closer range than mine, maybe about 8-10 feet away. At that distance, it looked slightly more even, center to edge.
    They said "it's fine". I took it home and haven't worried about it since. After all, if I want any wide angle lens to produce the maximum detail, I'm going to shoot at f/8, not wide open.
    Still, at f/2.8, the 14-24 outperforms most other 14-15-16mm lenses available. Enjoy your 14-24.
     
  20. Thanks, Keith. I was only really worrying because - while I get decent results at f/7-ish - that's beyond the point where I'm losing some detail to diffraction on a D800e. If I could have got to f/5.6 and had sharp edges (which, apart from the weird position of the focal plane, they probably are) then I'd have been slightly happier. But no biggie. If I know others are seeing the field curvature, I'll live with it - so long as it's not something I should be getting fixed!
     

Share This Page