I've been testing furtherly my 14-24 DG DN. In my opinion it has noticeable field curvature issues which need to be understood in order to obtain the maximum potential of the lens for corners sharpness.
What I've been testing is different focusing points to maximize corner acuteness. When I focus on dead center, no matter if close or at infinity, corners are not going to be maximally sharp. Only if I select a point of focus toward middle of the right side then the corners are brilliantly acute. Of course that when practicing this focus method then the center is less sharp than when focus point is selected at the center of the frame. But this is a compromise we have to accept to have corners at high level.
Another method is to focus on the center part of the frame and then zoom in with the focus magnifier assistant to check the corners. If they are soft proceed to manually tweak the focus ring slightly and improve them to get tack sharp acuteness and then shoot. When doing this, the center probably will lose some sharpness, but not very much in my view.
When using the Sigma 14-24 DG DN I have decided to disengage IBIS permanently in the camera body to avoid weird issues of drops in sharpness along random parts of the frame. Always off, no matter if tripod or not. I am starting to be annoyed by IBIS system.
For references, take a look at these reviews regarding field curvature.
And I am also starting to feel that the 14-24 DG DN test at Lenstip is not very realistic as almost all of the sample photographs are not properly focused, since corners are soft in everyone, no matter the f-stop number.
I also feel the same for other test at Les Numeriques where the reviewers where unable to properly focus at close distance the Sigma 14-24 DG DN. Consequently the results are very soft at f2,8-f4, where the field curvature shows its maximum effects. They even explained that the Imatest numbers for sharpness were much higher than what the photography test suggest. This may lead to mistakenly conclude that the 14-24 is soft wide open, which is not true in my experience.
Is your copy well centered across the whole range? The various copies I have tried indicate to me that fleld curvature is very low for such a wide lens, but sample variation is such that FC can look like it is stronger on some copies. Also, unless the reviews you mention use flat surfaces or infinity-only scenes, they will not be much use in assessing FC.
Hi rvh23
I think it is quite well centered. Fred checked out my test shots some time ago, which I am also posting here again. He told me it was a fine unit, slightly decentered at 24 mm but he also commented that it is imposible to get a 100% perfect zoom, which I agree totally.
It is interesting what you point out about FC not being able to show on a flat surface. From my experience it is indeed posible. Keep in mind that what is causing the FC is the rounded external element of the lens and that the plane of focus shifts wavyly along the entire flat sensor plane. What I didn't know is that some copies of a FC'd lens may have more FC than other copies.
In the test of Les Numeriques it is easy to see that they were not able to properly focus for simultaneously sharp center and corners at the same time. It may also be possible that the copy they had were more prone to a high FC degree. By the way, the averaging of 10 copies that Roger Cicala applies to all his tested lens is really a good idea that gives much more credibility to the tests he publishes.
What I am considering is to sell my copy and to get an Sony 12-24 f4 G, which has been tested at Lensrentals and its field of focus is very flat with almost no FC at all. I am somewhat confused now, but I have to experiment more with my Sigma lens before doing that.
Here are the test shots for decentering of my unit:
14 mm:
18 mm:
20 mm:
24 mm:
Best
Agreed, your copy looks fine. It looks close to perfect at 14mm, and only slighlty asymmetrical a the long end. Given we are talking about wide open performance, that's rather impressive for a zoom. By F5.6 it should be great across the whole range (and noticeably better than a Sony 12-24 in the corners).
My previous comment about FC assessment methods was in relation to those lens reviews that use general 3D scenes to assess corner sharpness. If you don't have objects in all four corners at the same distance from the camera (e.g. a flat surface, or everything at infinity) the assessment of FC will clearly be compromised by DOF considerations.
Dec 27, 2019 at 06:54 AM
Previous versions of rvh23's message #15089467 « Sigma 14-24mm f/2.8 DG DN first impressions »
