Dissatisfied with my mid-range zoom

When I bought a KS2 body I realised that the DAL 18-55 'kit' lens that I had been pretty content with on my K-x was sub-standard for the KS2. Having read several reviews, I settled on the Sigma 17-50 f2.8 EX DC which cost me £329 from Amazon, last June.

This lens seemed to be well regarded, but I have been disappointed with its edge sharpness. Some unsharpness in the corners was mentioned in the reviews, but I find mine much worse than the somewhat maligned DAL 18-55 at the edges. The Sigma lens is certainly good in the mid-field - almost as good as the prime DA 35mm f2.8 Ltd which I bought from a PU member a while ago. It is also nice to have a constant f2.8 aperture, but I am always annoyed to find the edges fuzzy when viewed at 100% in PSE, more or less through the whole zoom range.

Maybe someone who won't spend well over three hundred quid on a lens shouldn't be pixel-peeping, but I began looking at some wide-angle primes to complement the superb DA 35mm macro. To my disappointment, I see that the dxomark tests of the Pentax HD DA 21mm f3.2 Ltd are actually slightly worse than their scores for the Sigma 17-50 that I am grumbling about!

What to do? I am sure that many of you use lenses in this focal length range, and maybe you are content with them, but if anyone can suggest a good wide-angle prime, in the 18-25 mm range, that performs better but doesn't cost a month's pension, let's hear from you!

I wait in hope....

Martin

Are you finding the Sigma soft in the corners even when stopped down?

Sounds like you have a bad copy My favourite reviewer rated it quite highly on a Canon ....

https://www.youtube.com/watch?v=t_rPxWpArNY

-

From the ePhotozine test: "Overall I'm a little disappointed by the lacklustre performance towards the edges at maximum aperture."
https://www.ephotozine.com/article/sigma-17-50mm-f-2-8-ex-dc-os-hsm-interchangea...

pcmag.com : "Soft wide open"
http://uk.pcmag.com/sigma-17-50mm-f28-ex-dc-os-hsm-lens/2588/review/sigma-17-50m...

A generalisation, but Sigma lenses often seem poor at the edges though really sharp in the centre.

Our little 18-55 Pentax's are actually pretty good compared to other marque's offerings with edges great compared to a Canon equivalent. A comparison I did some time ago saw a bar code on the edge of the image could be clearly read on the Pentax but the Canon's was just a blur.

Even beat the Pentax DA* 16-50mm in this test....
https://www.pentaxforums.com/reviews/pentax-16-50-vs-sigma-tamron-17-50mm/conclu...

--

I have and am very happy with my K-S2 with my new favourite lens the Pentax 16-85mm. It might not have the constant f2.8 but I find it an excellent lens. I kept my eyes peeled and picked a new unboxed lens from SRS Microsystems via ebay, saved a few quid. It is certainly better than my DA18-55mm WR and a covers a great range. I also have the Pentax DA 18-135mm which gives nice looking images when viewed at standard screen size with mainly excellent centre sharpness but zoom in and you don't have to move far towards the edges to start seeing an often steep drop in sharpness, the 16-85 has a much better edge to edge performance. I also have the DA* 16-50mm, I pick up the 16-85 in preference. The one thing I can't do is compare it to the Sigma 17-50 f2.8 because I have never had it.
All the reviews I've seen rate the Pentax 16-85mm very highly as do I.

My Sigma 17-70 can also be a bit soft at the edges, but it's a great all-round lens and is my most used one. I think to some extent you have to accept some compromises on lenses that have 'real world' prices.

Regards
David

Many thanks for all your helpful comments and suggestions.
Mike-P: yes the images are still fuzzy at the edges, testing from f4 to f8. Worst at wide-angle (19mm) but still noticeable at the long end (50mm - see my example below).
Algernon: I have wondered if I should have rejected it, as the blur seems a bit worse at the right hand end of most images than at the left hand end. Maybe the glass is slightly off-centre. As for your Youtube reviewer, he heaps much praise on its build quality and good constant f2.8 max aperture, but he does admit to some softness and CA in the corners.
davidstorm: I think you are right. It is a good all-round lens and maybe I am expecting too much from an "affordable" zoom. I am a bit obsessional about sharpness. What surprised me was that the dxo tests found the Pentax 21mm Limited to be slightly inferior in sharpness, and some other respects, to the Sigma zoom. Surely one would expect a prime, and especially a Limited grade, to out-perform a zoom which of necessity will have some compromises.
AbeNormal: I have tried adjusting front/back focus, but found the factory setting to be as good as any.

Below is a composite of a recent photo, in which I show 250x250 pixel clips from the left-hand end, the centre, and the right-hand end of a photo (Whitby West Pier) that I have just uploaded to flickr;
link
This photo was taken at the long end of the zoom range (50mm) and even here the lack of sharpness is evident, especially at the right-hand end. Taken at f5.6, 1/1000, ISO400. I don't have any suitable wide-angle images to use in this way at present, but edge sharpness is worse when one is in 17 to 30mm range.




Maybe I'm just being hyper-critical in 'pixel peeping' at 250 px clips from an image that is 5472 pixels long?

Thanks again for your thoughtful comments.

Martin

Looks a good lens to me Is the problem only at 50mm? If it is zoom out slightly and crop off later.

The easiest way to test the centering is take one shot with say a building on the left edge. One with it in the centre and one with it on the right. AF needs to be off.

The left and right should be the same, but rarely are and the centre should be better than the other two. The Hollywood 28mm f/2 is about the best I've seen for centering.

Here's the same type of test using 5 shots which takes longer...
https://www.pentaxforums.com/articles/photo-articles/how-to-check-your-lens-for-...

It's also worth trying with the SR off although it's usually WA (10-12mm shots) that show blur at the edges.

--

Were they tested with the same sensor, or optically bench tested?

FWIW, and going back in this thread a bit, my very early Sigma 17-70 is doing pretty well - not had any real issues with the edge definition, but then, generally, I'm not a pixel-peeper

The DxO results that I looked at were both mounted on a K3. The Sigma 17-50 was given an overall score of 21, and rated 13 for sharpness. The Pentax 21mm HD f3.2 Limited was given a score of 18 and rated 11 for sharpness. I have no idea of how DxO set up their tests and I have not seen any analysis of performance at various apertures and focal lengths on their website.

Perhaps those are relative scores to other zooms/primes rather than absolute. Also depends on what they mean by sharpness! The Sigma might have one sweet spot that beats the Pentax and they've taken that value. Certainly sharpness won't be consistent over the whole zoom range.

One also has to take into account sample variations, in all probability you haven't got the lens submitted for test and if I was a manufacturer I'd make sure a lens submitted for review was a good'n and not taken at random off the production line. But unless one knows all the facts of the testing it can only be a guide. There's more to lens quality than overall sharpness.

For a rough and ready lens test that actually means something, try pinning up a large newspaper on a flat wall. Make sure the camera is centred and parallel to the wall, on a very firm tripod, focus very carefully using live view and take several runs of all apertures. Use the 2 second delayed action to disable SR and also make sure there is no camera movement. Then examine the clarity of the fine print to see what's what.

The samples shown so far are three dimensional and subject to subject movement. Trees and leaves blow about in the wind. Lack of a tripod invites camera shake.