DA 14 vs. Sigma 12-24

[Joseph Tainter]

Having time today I decided to test the DA 14 f2.8 against the Sigma 
12-24 f4.5-5.6 DG at 14 mm. I did this in part to clarify in my own mind 
what is going on when people complain about the DA 14 being weak in the 
corners. I know, I know -- I am comparing a prime to a zoom. No contest, 
right? Well, the Sigma is the only lens I have access to that includes 
the 14 mm. focal length. Also, the last time I compared a prime to a 
zoom, it was the DA 14 vs. the DA 16-45 at 16 mm. -- and the zoom won. 
The subject was, as usual, an adobe (mud brick) wall, with the camera at 
about one meter distant. I used ISO 200, and a tripod. All comparisons 
were at actual pixels, although I went to 200% in one case.

Firstly, the DA 14 becomes progressively sharper through f8 (the 
smallest aperture I tested). The center is noticeably better at f4. At 
f5.6 the corners sharpen up considerably, then improve a bit more at f8. 
Stop this lens down. Now the comparison.

Both lenses wide open: (f2.8 for the DA 14, f4.5 for the Sigma). The DA 
14 is much sharper in the center. It is mostly sharper also at the edges 
and in the corners, although the difference is less. The Sigma may be 
slightly better at the left edge and in the lower left corner. Where the 
Sigma seems better, it is not by much. Where the Pentax is better, the 
difference is very clear.

F5.6 (both lenses): The Pentax is far better in the center, slightly 
better at the edges and in the corners.

F8 (both lenses): The DA 14 is much better in the center and corners. It 
is also better at most edges, the exception being the left edge where 
the Sigma seems a bit sharper.

The DA 14 is clearly the better lens overall. What is interesting is 
that it has an arc of slight weakness from the lower left corner to the 
left edge (about 225 to 275 degrees).

Shouldn't lens performance be symmetrical? I would think that even 
assymetrical lenses should be symmetrically assymetrical. Any thoughts?

BTW, on aperture priority the Sigma consistently exposed the images 1 EV 
less than the DA 14. The Sigma seemed to be underexposing.

Joe