[davidmanze]

Actually, I would like see some evidence how it looks like and how big's the impact. 7 samples are not a lot, I was scrolling through 48 pictures (less than an hour? ha! 20 minutes - so really nothing to call it a test) and simply could not see these problems in real world. But I realized I have less high expectations on sharpness as there is so much to be seen on the picture - I simply didn't bother enough to see the structures of leaves close to the lens' ∞ position.

 

 

  Well the third real world resolution graph tells the story, whether you can compromise across the frame for infinity focus, but then you are fiddling either with; use once AF fine tune settings; or MF in LV  not ideal...

 

     ...and all that for F4!

[JJ_SO]

In a perfect world scenario the focus field is absolutely flat - so e.g. if you take a perpendicular picture of a wall, the focus field should sit exactly on that wall.

I don't know about you guys, but I'm not buying an UWA for a "perfect world" flat reproduction of maps or huge posters.

 

I could not measure the lens I tried - and looking back and seeing the pictures I was inspired to compose, I have to say, "absolutely flat" would be a physical miracle and worse, nothing to improve my pictures. Long ago I thought enough DoF can save any day, any missed focus and also perspectively challenging shots.

 

12 mm on full frame - how on Earth could such a lens bend the laws of optics enough to be sharpest at some flat pane?

 

The pictures in the tower were very complex in terms of involved distances:

 



 

Pictures like these don't need to be sharp on each little brick or scratch in metal, I'd say.

 



 



 

Do you see field curvature? And if so, does it matter?

 

"... and all that for F4". First, it's f/4, second feel free to develop a lens which covers a whopping 36° zoom range (122.0°-84.1°)  with less distortions, field curvature... Apparently I sometimes can adapt to the fact, not to live in a perfect world scenario  ^_^

 

But if you'd ask me "will you buy it to replace the Nikkor?" I'd hesitate. Although it is much better in terms of flare and offers 2 mm shorter FL. I was not very often in the situation to really need the zoom part of the UWA zoom, but then it was good to have it - the big angle of view sometimes bring in distractive things, then it's good to narrow it a bit.

 

It's not a clear decision to make. I think, Sigma did a fairly cool job with it, but I'm curious about the 14/1.8 as well - and both Irix lenses are as well on my list to play with.

[davidmanze]

Great shots JoJu!!

[JJ_SO]

Thanks, I chose them because I thought the lens is

• decently sharp
  
• reasonably priced
  
• better in terms of flare than it's (10 years old) Nikon colleague
  
• clean of CA "wide open"(which is f/4, so should really be no problem)
  
• coming with an attractive zoom range start - not many 12 mm zoom options available with AF
  
• highly adjustable (although that bit is a nightmare to do)
  
• used to get "dramatic", "dynamic""huge overview" pictures.
  

Geometrically precise reproduction is hardly a big strength for any of these lenses. In a way it's good Klaus checked for this (PZ's tests usually don't have extra chapters for this type of flaws, so kudos to do so)

 

Beside field curvature, the corners are really nothing to be ashamed of in terms of resolution. Depending on the subject, I tend to guess this curvature is eventually helpful. But in any case, the lens has a lot of strengths. So, if I would have to choose between that one and the Nikkor today, with the perspective to get a pretty fast 14/1.8 additionally, I reckon to decide in favor of the Sigma. Simply because front and side light is not such a pain in the butt - and in rooms (architectural shots) a tripod is always a good idea, so the f/2.8 not necessarily an advantage.

 

EDIT: A while after the post I decided to upload more images to a sample gallery. Please bare in mind, I was trying to find out if that lens would handle some situations better when I felt being let down by the still more expensive Nikon - or was at least disappointed.

 

If you like to check exposure data, right hand at the botton is an info icon. As well as the three layered rectangles which allows to choose "original" as size. I don't mind if you download some of the pictures to examine, just don't used them to publish without my permission.

 

Couple of shots didn't make it into the gallery because of missed AF - I learnt it's better to switch on LiveView and don't hope for "it will be fine with PDAF". I didn't add extra sharpening, eventually a little clarity. I'm also no fan of white clouds when I know there's structures behind the white and  the sensor did catch it. So shadows and highlights are to my likes. At two pictures I was a bit confused because of little spots which turnt out to be a cloud of little flies at 100%.

 

I found shooting at 12 mm challenging - but never because of the lens, always because it's hard to see in finder or on screen all the little details I can look at later on on a big screen. And I know how easy it is to miss a nice detail or to show too much non-important things.

[Klaus]

JoJu ... you have qualified as the defender of the faith now ... 

 

- the field curvature is an issue STARTING at 18mm - not across the range. You are probably aware that we use a slicing approach to determine the best spot on the focus field. If the center and corner slice differ by "5 slices" I wouldn't bother too much - yes, you will see that on a wall but - as mentioned in the review - who really cares. At 24mm f/5.6 the Sigma was at up to 9 slices and at the edge of what we covering (ca. +/-12 slices). The Canon 11-24L didn't have this issue.

- remember that 50mp are magnifying any issues substantially. Yes, I will provide the 21mp figures soon (and that will, as usual, ease the pain)

- a comparison with the Nikkor 14-28mm is meaningless in the Canon section. The benchmark is the 11-24mm L here. The 11-24L flares as well, of course, but not quite as much as the Sigma. 

- if your sample is different - good for you - yet this was a brand new lens and passed Sigma's QC. The center was quite fine centering-wise. From there on it is not my problem anymore. PZ does not exist to defend manufacturers. It is worth remembering that any criticism is also meant to tell the manufacturers that they have to push more QC-wise and quality-wise. If you want to read hallelujah ratings ... well, there are many sites out there that do that.

[JJ_SO]

I do know that you do series of focus samples, so I understood why there are differences between the resolution charts and the field curvature chart. I also don't want to defend or show faith in Sigma although I admit they have my sympathy, more than Nikon does for their lenses. As I said I have doubts to decide for that lens, and part of it is of course your findings with the field curvature.

 

In the set of pictures I took there were a lot in narrow space, perspective distortions (not only by the lens), other things hardly impressed me - like AF failing in few pictures although I could not recognize why. Purposefully I don't say the lens failed to AF - at first, it was not at all calibrated. And I'm unclear if calibration would be easy or helpful afterwards.

 

But in practice I can imagine some shots blown due to field curvature and others saved because of it - would the lens be entirely useless at the borders, I'm sure you would have showed that. I also understand that doing more than unusual "border shots" to get the peak performance is annoying. But it was worth the effort, and it was also great to point that out. Otherwise people buy it, do their usual brick wall tests and complain about PZ  :lol:

 

One of the pictures I took I was a bit puzzled to check afterwards. It was wide open @ 24 mm and from the door in the background to the white chess figures in the left corner the sharpness surprised me. Now I know, that field curvature played some part in it. That's what made me asking "is it in the wild such an issue like it doubtless is on a test chart?" Anyway, that brings the "3D-effect" to an entirely different meaning  ^_^ Next time I use my 3D-glasses...

 

Oh, and the first guy who brought me that Nikkor to attention was a Canon photog, who praised it very much and used it with adapter. 8-10 years ago, that was. The Nikkor was for long a lens Canon shooters were lusting for, so I reckon I can refer to it. Forgive me tho post Nikon samples in a Canon forum on PZ, if you would have put more and more critical shots in your samples, I would not habe been so tempted to do so 

[Klaus]

Here's the "field curvature" at 21mp.

 

Again - this looks MUCH WORSE that it is. If I would take these readings with a +/-2 slice horizon the chart wouldn't look that extreme anymore.

As mentioned I am therefore reluctant to providing these figures.

 

So at 21mp, the f/11 figures is livable again.

[Klaus]

If I chose the +2 slice (outer region) the chart would look like this AT 21mp ...

 

This gives probably a more realistic figure what to expect in the real life (at 21mp, it will still be much worse at 50mp).

 

Again ... I don't like to play that field curvature game but it is part of the whole picture.

[Klaus]

FWIW, 50mp at +2s. I replaced the chart in the review with this one - that's more fair.

[Klaus]

I just updated the review with the 21mp data.