1. Reminder: Please use our affiliate links for holiday shopping!

Curvature of field demo

Discussion in 'Adapted Lenses' started by addieleman, Aug 11, 2013.

  1. addieleman

    addieleman Passionate amateur

    Nov 13, 2012
    Netherlands
    Ad Dieleman
    Lately there have been some questions and remarks about curvature of field. Curvature of field means that the the field of optimum focus of a lens is not flat, but curved. This can be both ways, optimum focus at the edges is farther out than at the center or closer. I've encountered the latter more, often edges that are closer to the camera than the center are surprisingly sharp.

    Last week I received a Minolta MD W.Rokkor 20mm 1:2.8, a lens that is beautifully sharp in the center even wide-open, but that delivers quite blurry corners due to pronounced curvature of field. Here are 100 % crops from center and corner. Focussing was done with flexible spot in center or corner.

    The total frame:
    i-VjxdTX5-XL.jpg

    i-bcJdWwb-X2.jpg

    i-HtpfjQ5-X2.jpg

    You can see that this lens can be quite sharp in the corners if focussed there. And no, the unsharpness is not caused by the corner subject being closer, depth-of field should have covered this: the other edge farther away shows the same pattern.
     
  2. gio

    gio TalkEmount Veteran

    382
    Sep 12, 2012
    Manchester, uk
    nice demo, good work, but I feel we take too much within the photo too seriously, generally the photo looks good as it was meant to look, ie the full view,
    unless we want to blow up to poster size and have it scrutinised, then that is fine, most if not all of the greatest photos ever taken were spontaneous and did not pay much heed to sharpness or depth of field, they achieved there fame through their content.


    perfection is boringly repetitive, the human touch is a winner.
     
  3. Bimjo

    Bimjo Super Moderator

    Oct 28, 2011
    Washington State
    Jim
    Very nice examples Ad, thanks for posting that up. :)
     
  4. eno789

    eno789 TalkEmount Top Veteran

    720
    Jan 1, 2012
    NoCal, USA
    Brian
    I searched a little bit on the explanation of the term, this article is pretty good with illustrations

    What is Field Curvature?

    This got me thinking, to correctly demonstrate the problem, the center and corner should be both on the flat focal plane, which is not the case here.
     
  5. addieleman

    addieleman Passionate amateur

    Nov 13, 2012
    Netherlands
    Ad Dieleman
    Ideally you should indeed use a flat subject, which is obviously not the case in my experiment. However, depth of field should cover things up: when using a fairly stringent 10 ┬Ám circle of confusion, depth of field ranges from 3.5m to infinity for a 20mm lens at f/5.6 and set at 7m. So this should cover the subject matter, the corner was more than 10m away and the rest was further out. I checked the other corner: it showed the same behaviour and it was at least 30m away.

    Maybe the most important practical consequence is that you should be careful with some lenses to focus-and-recompose, meaning that you focus with the center spot and then frame the picture to your liking. When a subject near the edge is intended to be sharp, you can end up with that subject being unsharp using the focus-and-recompose approach. Another thing is that such a lens is simply not able to render everything sharp at a large distance like in landscape shots.

    Such behaviour can also be used to advantage: you can isolate a subject in a landscape by the blur in the rest of the frame. I really love what my 20mm is doing so maybe someday I'll put it to use this way.
     
  6. eno789

    eno789 TalkEmount Top Veteran

    720
    Jan 1, 2012
    NoCal, USA
    Brian
  7. addieleman

    addieleman Passionate amateur

    Nov 13, 2012
    Netherlands
    Ad Dieleman
    Depends. With my 20mm it's quite easy to see: when I focus on a brick wall I just see the peaking move from center to corner when I focus! I can imagine it gets difficult if you want to quantify the amount of curvature, but I didn't intend to make a fully scientific experiment :).