Sigma 10-20mm f/3.5 EX-DC HSM

[KevinBarrett]

I think Alpha users have been told to expect this lens in August, but has anybody seen any reviews of this lens in its available mounts? I have read only a few brief opinions from slrgear.com and flickr.

I intend to buy either the Sigma 10-20/3.5 or the Sigma 10-20/4-5.6 within the next month, and I'd rather have a lens sooner than later because I've two weddings to shoot in early August (consequently, they're the reason I'm suddenly able to afford a new lens!).

[alphaomega]

I am extremely keen on reading competent reviews of this Sigma lens as well. I have the current model. Quite happy with centre performance but a bit uneasy with its corner sharpness. I thought of purchasing the new Tamron 10-24 but have been put off by reviews not exactly praising sharpness. Essentially I am more interested in improved corner to corner sharpness and reduced distortion than in the additional speed at F3.5 throughout all F stop settings. If it measures up I am a buyer.

[David Kilpatrick]

I have it here, testing on Nikon not Alpha sorry - that's what they had available for review. Don't bother 'upgrading', it is not an upgrade. The f/4-5.6 will still be made apparently. Like the Tamron 10-24mm, the new Sigma gains its extra maximum aperture at the expense of corner sharpness. It's just about OK on Sigma cameras 1.7X, nt quite good enough on Canon 1.6X, and I feel it falls well short of the required coverage for 1.5X.

David

[KevinBarrett]

Thanks, DK! I had read from a review on slrgear.com that the new 10-20/3.5 suffers distortion that "gets quite extreme as you near 10mm," whereas most who shoot the older version rave about its distortion control. Can you comment on the barrel distortion in this lens?

I like the way Sigma builds its EX (for extra expensive) lenses and this model is as good as the other two EX lenses I own. It can take 82mm filters and one can use the lens cap even when the butterfly hood is in place.


I need top-quality glass for my work but I'm willing to compromise for $$$ savings when I can. I have a Tamron 17-50mm 1:2.8 which is incredibly sharp - a great buy. (It is tested here at SLRGear.) I compared it to this lens, both set to 17mm under very controlled conditions. (First I tested the new Sigma for focusing accuracy. It was short by just a small bit. I made the adjustment in my Nikon D300.)


In overall impression the Sigma is comparable to the Tamron, a bit warmer but with similar contrast. Distortion at 17mm is much less than the Tamron's, but of course gets quite extreme as you near 10mm. In resolution there were large differences. At 17mm, both center and corners, the Tamron outperformed the Sigma for sharpness by a very noticeable margin with fast apertures. The gap starts closing around f/11 and the two lenses are roughly equal by f/22.


The Sigma's bokeh is a bit rough. Specular highlights aren't quite round, have some striations and halo, and are not smoothly monochromatic. None of these shortcomings is a killer, I'm just a perfectionist.


If you'll never print larger than 8x10" this lens might be ideal for you, but it is not truly up to professional caliber. However, neither are any of the other ultrawide zooms, it seems. The exceptions cost two or three times as much and don't go nearly as wide.

[01af]

Kevin, if you're interested mostly in even distribution of sharpness across the field, even illumination, and low distortion, then better look at the Tamron/Konica-Minolta/Sony DT 11-18 mm 1:4.5-5.6 lens. The Sigma 10-20 mm 1:4-5.6 DC has slightly better center sharpness but falls short to the 11-18 in most other aspects. I don't know how good the new Sigma 10-20 mm 1:3.5 DC is going to be ... but generally, lens speed in super-wides always comes at a price, i. e. more vignetting, more distortion, less corner sharpness, and higher price tag. Or none of the first three at a much higher price tag, plus exorbitant weight and size. I wouldn't wait for it unless speed is your only concern.

-- Olaf

[KevinBarrett]

Actually, after DK's post here my mind is pretty much made up--I'll go for the older Sigma. I continued probing out of curiosity and so that the thread might be more useful for anyone else that's still undecided.

[David Kilpatrick]

When I read someone who says 'but of course gets quite extreme near 10mm' I realise that the writer does not know what distortion in this context means. It usually emerges that they think converging verticals are 'distortion' and have been failing to grasp the meaning of barrel, pincushion and compound distortion patterns.

Actually, the new 10-20mm is pretty good and uses a clever type of compound distortion which leaves verticals placed right next the edges of the frame very straight and clean. Unfortunately it is at the expense of barrel distortion closer to the axis which changes to pincushion 'out of view'. This means that a horizon placed 2/3rds down the shot on a vertical pic looks pretty straight, but if you place a horizon about 2/3rds down a landscape view, it will be wavy. The Sigma will do pretty well on distortion figures measured by using a standard like 8mm from the axis, I think.

I need to do a bit more shooting with it, but distortion is not a major negative point. The softness to the corners wide open is what you will notice compared to the less ambitious 4-5.6 model.

David

[01af]

Actually, the new Sigma 10-20 mm [...] uses a clever type of compound distortion which leaves verticals placed right next the edges of the frame very straight and clean. Unfortunately it is at the expense of barrel distortion closer to the axis ...

I am not sure if I'd call this 'clever.' It may be clever from Sigma's point of view, as it helps avoiding bad results in tongue-in-cheek distortion tests and thus will lead to good apparent performance. But as a photographer I'd rather have a lens with ordinary barrel distortion (preferably not too much of it).

-- Olaf

[KevinBarrett]

Just when I've almost decided on the Sigma 10-20/4-5.6, I look over all of my reviews again and notice that while the Sigma gets excellent test-chart scores, Kurt Munger and many at Dyxum prefer the Sony/Tamron 11-18. Focus reliability is cited most often, and now the distortion thing is bothering me. Does the Sony/Tamron 11-18 have a more natural and correctable distortion?

[bakubo]

Kevin, if you're interested mostly in even distribution of sharpness across the field, even illumination, and low distortion, then better look at the Tamron/Konica-Minolta/Sony DT 11-18 mm 1:4.5-5.6 lens. The Sigma 10-20 mm 1:4-5.6 DC has slightly better center sharpness but falls short to the 11-18 in most other aspects.

I have the Sigma 10-20mm f4-5.6 in Canon mount, but when I got the A700 last year I decided to get the Sony 11-18mm f4.5-5.6 for it instead of the Sigma. Naturally, I didn't want a dog, but this is going to shock and horrify some here. The main reason I got it is because it is the lighter of the two. Weight is important to me for travel. Even though they both have 77mm filters the Sony lens hood is about 8-10mm larger in diameter. I wish it was the same size as the Sigma. The depth of the 2 lens hoods is the same though. The length and diameter of the lenses is almost identical. I haven't done any testing or comparisons, but both seem okay to me.

Personally, for this type of lens I don't care much about the maximum aperture. Don't need f2.8 or f3.5. I never shoot a super wide angle lens wide open anyway. Usually I have it stopped down to f8 or f11. Most shots are at the shorter focal lengths too so I don't worry too much about the longest focal length. 10-20mm or 11-18mm is not such a big deal, IMO.

[David Kilpatrick]

If anything the Sigma old 10-20mm has the simplest type of barrel distortion at 10mm, and just needs a Photoshop Distort > Lens Correction tweak. But if you don't correct it, the barrel distortion is strong enough to be visible. The Tamron 11-18mm has a different projection which leads to less barrel distortion but like the new Sigma and Tamron designs, and the CZ 16-80, and the CZ 16-35 (etc etc) it is not completely rectilinear and is best corrected using a 'lens map' type fix - as and when LensFixCI, Panotools or DxO Optics Pro choose to add the lenses to their databases. But the 11-18mm has less distortion than the newer lenses and it is hardly ever necessary to correct shots - it would have to be a sea horizon in the wrong place or something. Room interiors and buildings generally look straight enough.

I've used both the Sigma 10-20mm original and three different versions of the KM-Tam-Sony, and it really comes down to price. I would be happy with any of them but I got a good deal on all three of the others at different times - KM closeout sale, Tamron journo discount, Sony mint s/h. The new Sigma is now £650 from Warehouse Express and that's a hell of a price for an APS-C wideangle which does not have absolutely PERFECT rendering.

I think the oversize lens shade issue is down to Tamron cost saving, it is exactly the same hood as the 17-35mm full frame (same part number) and they just made one w/a petal hood to fit all. Even the new 10-24mm seems to take a variation on the same hood.

David

[KevinBarrett]

Ah, now that complicates matters! I would be using the UWA zoom first to shoot indoors at a wedding reception, so it would almost appear that the Sigmas is a better choice. Considering all the news of new Sony bodies coming out though, I think I'll hang onto my money until I see what Sony is going to do with the system this fall.