Poor results from Sigma Ultra-Wide Angle lens plus confusing EXIF data

Messages
1,651
Name
Brian
Edit My Images
Yes
#1
I normally use Canon lenses but as part of "Black Friday" I saw an offer for the Sigma 8-16mm UWA lens for APS-C from one of the major UK retailers, so I jumped in.
I was extremely disappointed with the first example I received. Nothing seemed to be really sharp and I noticed that the focus indicator on the lens barrel frequently went beyond the infinity mark. The most strange thing was that viewing the EXIF information the “Focus Distance” parameter seemed to be 3.53 metres on many shots regardless of the subject. For comparison, the same parameter from a Canon lens shows a different value for almost every shot, which is what I'd expect.
I returned the lens to the shop and was given the option of a refund or returning it to Sigma for repair.
I opted for the latter, but the option later changed to a replacement lens, which I collected yesterday.
The results from the replacement are about the same as the original.
Images are lacking overall sharpness, and although the focus doesn't seem to go beyond infinity, out of 53 test shots I took yesterday the EXIF shows the “Focus Distance” to be 3.53 metres on every one, despite there being a variety of subjects.

I'm using a Canon 80D set to “Single Shot AF” with a single centre focus point, and I receive the focus OK indication before I fire the shutter. These are my usual settings.
I was going to do some more tests (like shooting in “Live View”) but I'm not very satisfied with the results so far and feel that the EXIF information is totally wrong.

As I said, maybe I'm making an unfair comparison, being used to Canon lenses, but I'm quite disappointed with the performance of this Sigma and I cannot understand the EXIF results..
 
Messages
1,373
Name
Chris
Edit My Images
Yes
#3
Many cameras have difficulty in getting accurate focus with AF and wide angle lenses. The problem is that these lenses have such a wide depth of field that it can confuse the AF. For that reason many users of wide angle lenses focus them manually.
 
OP
OP
Brian G
Messages
1,651
Name
Brian
Edit My Images
Yes
#4
Can you post the shots? and list shutter speed ISO and fstop?
Not too sure what that's going to prove, but I will post something later.
Still doesn't answer the EXIF question though.
Many cameras have difficulty in getting accurate focus with AF and wide angle lenses. The problem is that these lenses have such a wide depth of field that it can confuse the AF. For that reason many users of wide angle lenses focus them manually.
I also have the Canon 10-22MM EF-S ultra-wide, and I've never had any focus problems with that.
That's what I'm comparing the results from the Sigma with.
Besides which, if it were a focussing error, something in the frame should be in focus, and nothing is sharp anywhere.
I'm going to try and get out over the weekend and do some side by side comparisons between the two lenses.
 
Messages
7,207
Name
Jeff
Edit My Images
No
#5
I’m with Chris I think your answer is to try it in manual mode , if you cannot then get successful results get back to sigma ... I think this is one of the reasons why so many UWA lenses are purely manual focus

L
 
OP
OP
Brian G
Messages
1,651
Name
Brian
Edit My Images
Yes
#6
Not ideal conditions for testing, but I managed to get out yesterday and did some comparisons between the Sigma 8-16 and the Canon 10-22.
One thing that is immediately apparent is the Sigma underexposes by 1/3 to 1/2 stop compared to the Canon.
I think I'm going to hang on to this example of the Sigma, but it's still not as sharp as the Canon, which surprises me, especially as all the reviews praise it's sharpness.
I don't have my tripod at the moment, which makes manual focus a little awkward.
I did try Live view and I can't see any difference between that and normal AF.
I'm still getting strange EXIF information, where all the Sigma shots are "Focus Distance 3.53 m" whereas the Canon Focus Distance is different with nearly every shot.
I intend to contact Sigma UK to see if they can explain this.
These two shots are fairly close, although they are at 16mm FL. Both straight from camera with no adjustments applied.
The focus point was on the closest bridge support.

Millennium Footbridge_01 Sigma
by Brian Gibson, on Flickr
Millennium Footbridge_02 Canon
by Brian Gibson, on Flickr

There are also two crops of the centre, where I was comparing the red signs over the entrance to judge the sharpness:

Millennium Footbridge_01 Sigma Crop
by Brian Gibson, on Flickr
Millennium Footbridge_02 Canon Crop
by Brian Gibson, on Flickr
 
Messages
1,332
Name
Soeren
Edit My Images
Yes
#7
Not ideal conditions for testing, but I managed to get out yesterday and did some comparisons between the Sigma 8-16 and the Canon 10-22.
One thing that is immediately apparent is the Sigma underexposes by 1/3 to 1/2 stop compared to the Canon.
I think I'm going to hang on to this example of the Sigma, but it's still not as sharp as the Canon, which surprises me, especially as all the reviews praise it's sharpness.
I don't have my tripod at the moment, which makes manual focus a little awkward.
I did try Live view and I can't see any difference between that and normal AF.
I'm still getting strange EXIF information, where all the Sigma shots are "Focus Distance 3.53 m" whereas the Canon Focus Distance is different with nearly every shot.
I intend to contact Sigma UK to see if they can explain this.
These two shots are fairly close, although they are at 16mm FL. Both straight from camera with no adjustments applied.
The focus point was on the closest bridge support.

Millennium Footbridge_01 Sigma
by Brian Gibson, on Flickr
Millennium Footbridge_02 Canon
by Brian Gibson, on Flickr

There are also two crops of the centre, where I was comparing the red signs over the entrance to judge the sharpness:

Millennium Footbridge_01 Sigma Crop
by Brian Gibson, on Flickr
Millennium Footbridge_02 Canon Crop
by Brian Gibson, on Flickr
Ehh? Focussing on a closer object and then looking at a far object for sharpness?
 
OP
OP
Brian G
Messages
1,651
Name
Brian
Edit My Images
Yes
#9
Ehh? Focussing on a closer object and then looking at a far object for sharpness?
In case you don't understand, a comparison involves comparing two similar things.
Besides which, at f8 the depth of field more than enbraces the comparison range.
The Canon lens is clearly sharper here.
There is (According to the Flickr pics) almost a full step difference in shutter speed which may explain the under-exposure.
Yes, that's right - but why?
 
Messages
1,373
Name
Chris
Edit My Images
Yes
#11
It might. Micro adjustment helps when focus is reliably misplaced by a certain amount. These very wide angle lenses however can cause the AF on some cameras to be very unreliable, i.e. repeated focusing on the same thing keeps focusing the lens differently. I have this 8-16mm lens, an upgrade from my previous wide angle favourite, the Sigma 10-20mm. My previous two cameras (Sony A350 and Sony A500) were unable to get any reliable AF out of the 10-20mm, so I ended up focusing it manually, which greatly improved results, especially with the A550, which had much improved manual focusing aids which made the MF a lot easier. Then I broke the 10-20mm and decided to buy the 8-16mm rather than have it repaired again. Because it was even wider I expected the 8-16mm to be even worse at AF than the 10-20mm but to my surprise it was better -- still unreliable in the sense of repeated AF on the same thing giving different results, but better in that the range of unreliability was over a much narrower range, often well within the DoF, so the unreliability was mostly irrelevant. I was also pleased to find that my 8-16mm was both a bit sharper in the centre of the image than my 10-20mm had been, and much better behaved at the edges.

Then I upgraded my camera body to my current camera (a Sony A77). That has such excellent manual focusing aids that if I'm not a hurry I often manually focus just to check how the AF is performing, which has improved my understanding of when I can rely on AF and when I can't. It's also got more accurate AF with all my lenses, including being able to AF in much dimmer light than my previous cameras could ever manage.

In other words, in my experience you can't generalise about the AF performance of a specific lens in the abstract. It's a combined function of the lens and the specific camera's AF. Testing AF required requires careful control of the variables which can affect sharpness, i.e. camera on tripod, use remote shutter release or timer to fire the shots, be careful not to nudge the camera off exact position between shots.
 
OP
OP
Brian G
Messages
1,651
Name
Brian
Edit My Images
Yes
#12
If I were experiencing misfocus then that would have been taken care of by focussing in "Live View" but I cannot see any difference between a shot taken in Live View and one using "conventional" focus, so I don't think micro adjustment is likely to make much difference, besides which, any misfocus would be completely swallowed by the massive depth of field.
On Saurdays results I'm prepared to accept this example as acceptable (it's the second one I've had) but what bothers me most is the strange figures recorded in the EXIF data. On the first occasion I tried this lens I took 53 shots and every single one reported a "Focus Distance" of 3.53 metres using normal, single shot focus on a variety of different subjects.
On Saturday the images taken with the Canon lens recorded a different "Focus Distance" for virtually every shot, whereas the Sigma was still showing 3.53 metres for every normally focussed shot, but the Sigma images recorded using "Live View" did show a variation in the Focus Distance, which is both puzzling and worrying.

I intend to contact Sigma UK and put these results to them and see if they can offer an explanation.
 
OP
OP
Brian G
Messages
1,651
Name
Brian
Edit My Images
Yes
#14
If it's any consolation you're not entirely alone.
https://www.dpreview.com/forums/thread/2924346
Thanks for that link, I'd not seen that before.
This is the second example I've had of this lens. I returned the first one to be sent back to Sigma UK, but I actually received a replacement.
When I suspected a focussing problem with the first example, I was deliberately changing focus manually and forcing the lens to AF.
I noticed on a number of occasions the first example locked focus beyond infinity, which does not seem to happen with the replacement, which does seem to perform slightly better in terms of overall sharpness.
As I said, I'm going to contact Sigma UK and see what they say.
 
Messages
2,173
Edit My Images
Yes
#15
I noticed on a number of occasions the first example locked focus beyond infinity, which does not seem to happen with the replacement, which does seem to perform slightly better in terms of overall sharpness.
The infinity position on any lens is a movable feast - it is dependent on both ambient temperature and air pressure. The infinity position in sunny weather in January will be different to the infinity position in poor weather in August - that is why infinity is usually marked with a sideways line.
 
Last edited:
Messages
2,173
Edit My Images
Yes
#16
Not ideal conditions for testing, but I managed to get out yesterday and did some comparisons between the Sigma 8-16 and the Canon 10-22.
One thing that is immediately apparent is the Sigma underexposes by 1/3 to 1/2 stop compared to the Canon.
Lenses don't expose at all. It is your camera that is underexposing. The only part that the lens plays in exposure is adjusting the aperture. As both photos have the same aperture (f/8), the lens is behaving properly as regards exposure. Your camera body is changing the shutter speed for no good reason and is failing to achieve focus (the auto-focus system is in the camera, not the lens). Your real question is why your camera does not behave with the Sigma lens. If the camera body will not work with a Sigma lens, changing the Sigma lens for another Sigma lens will make no difference.
 
Top