Serene Life

by garik

submit your photo


Hall of Fame
View past winners from this year

Please participate in Meta
and help us grow.

Take the 2-minute tour ×
Photography Stack Exchange is a question and answer site for professional, enthusiast and amateur photographers. It's 100% free, no registration required.

I got this Sigma DC 18-50mm f/2.8 EX MACRO HSM lens few weeks ago, but I noticed considerable sharpness problems with wide apertures. Until f3.5 (or even f4.5) this lens can't really be used.

I would like to know whether anybody else experienced the same issue? I've heared that Sigma lenses can be very inconsistent. You may bump into a very good lens or a lesser quality one (being the same lens type).

I would also like to know whether Tamron's equivalent 17-50mm f/2.8 lens is any better?

Here're examples of the same image taken few seconds apart. You can see the strange halo effect on the 2.8. Top part of both images is 25% original size (and cropped to fit size here), bottom part is 100% crop of strict lens centre.

50mm f2.8 - 25% and 100% lens centre crop

50mm f2.8

50mm f4.5 - 25% and 100% lens centre crop

50mm f4.5

One can of course see the difference in depth of field as well as focus issues I'm having with this lens. I wouldn't mind if it was soft on edges, but both of these images are lens centre crops.

Is this normal or do you think that my particular lens has issues of its own (which are either general for particular Sigma lens model or are just issues with my lens)?

share|improve this question
2  
Could you post example images, including a 1:1 crop and the picture taking conditions, camera, etc? –  labnut Mar 7 '11 at 13:23
    
You should try the Sigma 20mm f/1.8 :) The last full stop, at a minimum, on this baby is purely there for propaganda reasons, it is more of an f/2.8 if you want any kind of sharpness. –  Staale S Mar 7 '11 at 14:32
    
@labnut: I added example images that I hope will be of much help. –  Robert Koritnik Mar 15 '11 at 12:53
    
I can almost guarantee that the focus is off -- the DoF is the same in front of and behind the chart's focus point in your f/2.8 example shot, and the 14mm line in front is apparently sharper than the 14mm line behind. Whether that's an isolated instance or a general problem with this particular lens (not the design, but the calibration of this one lens) I can't say, but it sure does look more like a focus error than optical mushiness to me. –  user2719 Mar 15 '11 at 18:26
    
Stan Rogers: You're probably right about focus calibration (unfortunately my camera doesn't support lens micro adjustment which would help in this scenario), but that would mean that some other side line would be in focus. But is not. Even though it looks like that the most focused line is 4mm line. But it still has this grey halo... Which irritates me lots. –  Robert Koritnik Mar 15 '11 at 19:10

5 Answers 5

up vote 1 down vote accepted

Yes, this is normal. When a manufacturer undercuts price to give you a wider aperture, it is undoubtedly with a loss of quality, mostly at wide apertures.

I've seen this issue first hand with both Sigma and Tamron lenses that cost significantly less than brand name versions. Usually, I consider such F/2.8 lenses as F/4 lenses :)

You get what you pay for is true for lens, regardless of brands. Sigma has some superb lenses which they charge for accordingly. There is a premium for brand name but it is relatively small compared to cost of the lens.

share|improve this answer
5  
Given the sparse evidence presented it is impossible to draw useful conclusions. Almost ALL lenses are somewhat softer at maximum aperture. –  labnut Mar 7 '11 at 17:37
1  
Yes, all lenses are softer wide open but many Sigma zoom (generally the wider ones) become extremely soft at the 2.8 end. So much that after deleting my first images taken at 2.8, I never shot at that aperture again. At F/4 is became perfectly acceptable and F/5.6 is rather sharp. –  Itai Mar 8 '11 at 0:40
    
@Ital: I'm hardly convincing myself not to cut my fingers off when I set this lens to f2.8 :) –  Robert Koritnik Mar 15 '11 at 12:56

I would like to know whether anybody else experienced the same issue?

To answer this question see the customer reviews on B&H Photo Video for this lens with a Nikon mount.

There are 69 reviews and the average customer rating is 4.5 stars out of 5 stars.

Of the 69 reviews 3/69 noted that the lens was soft at maximum aperture at the edge. 1/69 complained that the lens was very soft and returned it.

The comments make it clear that this is a highly regarded lens. The main issue reported is focus problems - 8/69

I would also like to know whether Tamron's equivalent 17-50mm f/2.8 lens is any better?

To answer this question see the photozone tests conducted on these lenses, both on the same camera, a Nikon D200 (so the results are comparable).

The results are reproduced below (all measurements in LW/PH ) enter image description here
From this it can be seen that at 17(18)mm and F2.8 the Sigma lens is much softer in the corners, but comparable elsewhere. At 17mm(18) and F4.0 the Sigma lens is also somewhat softer in the corners but is comparable elsewhere. In all other respects the lenses are comparable.

In summary then, customer comments indicate the Sigma lens is a highly regarded lens that has some focus problems. Measurements show that the lens is rather soft in the corners of the image at maximum aperture and short focal length. At other apertures and focal lengths it performs well and is very similar to the Tamron lens.

Just to clarify things, the image below shows what photozone mean by their terms centre, border and extreme: (blue=extreme, green=border, pink=centre)
enter image description here
Image on photozone.com web site.

share|improve this answer
    
Oops, in my table I give a focal length of 23mm when it is in fact 24 mm. –  labnut Mar 8 '11 at 13:37

I would also like to know whether Tamron's equivalent 17-50mm f/2.8 lens is any better?

We may compare their resolution tests made by the same people (photozone.de):

Resolution wide open (f/2.8) at short end, in line width per picture height:

        Tamron @17mm   Sigma @18mm   Pentax @16mm   Canon @17mm   Nikkor @17mm
center  1923           2075          2085.5         2476          2114
border  1842           1882.5        1403.5         2169          1759.5
corner  1690.5         1254          1326.5         2055          1565

and wide open (f/2.8) at long end (50 mm or 55 mm), in line width per picture height:

        Tamron         Sigma         Pentax         Canon @55mm   Nikkor @55mm
center  1883.5         1804          1958.5         2261          1952   
border  1732           1709          1158.5         1983          1560.5

So Sigma is not the sharpest of them, but it is comparable to the rest of the competition, and should be as usable as most of the other similar lens. Tamron, in particular, seems to be sharper at wide angle.

P.S. As noted by @mattdm, the numbers cannot be compared between systems directly, because the size of the image is different. I think, that to compare them between each other, we need to multiply numbers on Canon APS-C system by 1.068, and on Pentax APS-C by 1.006, because sensor height on Canon is smaller.

Adjusted resolution numbers for short end wide open, as if all the cameras had the same sensor size as Nikon APS-C:

        Tamron @17mm   Sigma @18mm   Pentax @16mm   Canon @17mm   Nikkor @17mm
center  2053           2075          2099           2643          2114
border  1966           1882.5        1412           2315          1759.5
corner  1805           1254          1335           2194          1565
share|improve this answer
1  
Are higher numbers better? Are the numbers on a linear scale, so e.g. the Canon's 2476 is 16% ((2476-2075)/2476) better/worse than the Sigma's 2075? –  drewbenn Mar 7 '11 at 20:39
1  
Good reply. We need more thoughtful replies like this based on solid evidence. @Drewbenn, note that photozone rate anything above 1650 as Very Good and above 1950 as Excellent. So both the numbers you quote are Excellent. The main overall conclusion from the photozone review is that the Sigma lens is soft in the corners at both 18mm and F2.8 –  labnut Mar 7 '11 at 20:50
1  
@labnut, I'm not sure you can compare the numbers across cameras quite like that. If you look at multiple reviews, you will find that the rating scale varies. In the Tamron review, anything above 1850 is Excellent. –  rm999 Mar 7 '11 at 23:31
2  
In fact, the site specifically cautions that you can't compare across the numbers across camera systems. The Tamron review is on Canon APS-C, and therefore can only be compared to the Canon lens; likewise, the Sigma numbers are from that lens on the Nikon system, and on this chart can only be compared with the Nikon lens. –  mattdm Mar 7 '11 at 23:51
1  
@mattdm and @jetxee, another cause of variation is the AA filter. There are differences between Nikon, Pentax and Canon in the way they use an AA filter. –  labnut Mar 8 '11 at 13:10

As others have said, all lenses are less sharp wide open. What do you mean "can't really be used"? Post some samples.

I have this lens and find it quite good all around. Many of my pictures from this album are wide open. Be warned, many are a bit out of focus from what would be ideal. EXIF is on the right. I also took all pictures from my one-lens challenge with this lens. The shots can be found here, though I believe most of those are not wide open.

Also, you might find Pixel-Peeper.com useful in comparing lenses (both ones you own and ones you don't own).

share|improve this answer

Modern lenses from third parties are generally pretty good these days, if you are unhappy with sharpness check around for samples from other people and then go to the manufacturer (or exchange it for another wherever you bought it from).

If other test images look sharper, you should contact Sigma support and send it in, it's a good idea to send in test images with it. It could be the case it's off-center, or that it's back/front focusing - try shooting a focus chart.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.