Sigma 18-50mm f/2.8-4.5 DC OS HSM
Lab Test Results
November 16, 2009
by Andrew Alexander
At the time of writing, Sigma has five lenses available in the 18-50mm range, each designed to fit the APS-C sensor. Two feature constant ƒ/2.8 apertures; two feature variable apertures of ƒ/3.5-5.6. Sigma's latest iteration of this design is an amalgam of the two, featuring a variable aperture of ƒ/2.8-4.5, and adds optical image stabilization.
Designed specifically for reduced-frame sensor-based camera bodies, the lens provides an effective field of view of 30-80mm on Canon bodies, and 27-75mm on Nikon and others. This lens isn't a "constant" lens, in that as you increase the focal length, both the maximum and minimum aperture sizes decrease. The following table reflects the changes in aperture:
The lens is available in Sigma, Canon, Nikon, Pentax and Sony mounts. The lens ships with a petal-shaped lens hood, 67mm filters, and is available now for approximately $400.
The Sigma 18-50mm is decently sharp, less so on the wide end, but improving as the lens is both stopped down and zoomed in towards 50mm.
At the wide end (18mm), the lens shows significant corner softness - 3-4 blur units in the corners, set against a good results in the center - around 1.5 blur units. Stopping down to ƒ/4 offers a substantial improvement, reducing corner softness to 2.5 blur units, but that's about it - results for ƒ/5.6 through to ƒ/11 are essentially the same as those found at ƒ/4.
The lens likes the ƒ/4 aperture - as the lens is zoomed in towards 35mm, corner softness decreases from what we noted at 18mm, and the center maintains its results of 1.5 blur units in the center. By 35mm the maximum aperture reaches ƒ/4, and we note results where the corners are as sharp as the center, at around 1.5 blur units. Stopping down offers marginal improvements to sharpness, achieving optimal results at 35mm and ƒ/8, at just over 1 blur unit across the frame.
While diffraction limiting seems to set in at ƒ/11, the effects aren't noticeable until ƒ/16, and even at that point we still note sharpness results of around 2 blur units (except at 18mm, where it seems to be more like 2-3 blur units, softer in the corners). Fully stopped-down performance, as in most lenses, is best avoided with this one, offering between 4-6 blur units of uneven focus across the frame.
Chromatic aberration shows up prominently at 18mm, offering a significant red-blue color shift in areas of high contrast in the corners. In the center of the frame it's not bad, and at focal lengths longer than 18mm it's quite well-controlled.
Corner shading is about average for a wide-angle lens: at its worst, the corners are 2/3 EV darker than the center at 18mm and ƒ/2.8: between 24-35mm at the widest aperture, it's around a half-stop darker. At ƒ/4 at any focal length it's around 1/3 EV darker; at any other setting, it's at a quarter-stop or less.
The 18-50mm exhibits the complicated mix of barrel and pincushion distortion that's common in consumer zoom lenses: in the corners, barrel distortion at the wide end, and pincushion distortion at the tele end. Throughout the zoom range there is a generalized barrel distortion, meaning absolutely straight lines are very difficult if not impossible to achieve. At the wide end, we note +0.75% barrel distortion in the corners and +0.4% barrel distortion generally; at the tele end, we note -0.3% pincushion distortion in the corners and just +0.1% barrel distortion generally.
The Sigma 18-50mm ƒ/2.8-4.5 OS is designated as an HSM (hypersonic motor) lens, but it doesn't share the full implementation of what we've come to expect from Sigma's HSM lenses. Specifically, it doesn't feature full-time manual override, so if you want to override autofocus results you must first disengage the autofocus. Due to the lens' short focus throw, focusing is quite quick, at around one second to focus from close-focus to infinity, and the lens makes very little noise in the process.
The lens offers fair macro performance: 0.24x magnification, with a minimum close-focusing distance of 30cm (just under one foot).
Build Quality and Handling
The lens is built with economy in mind, and features an all-black, all-plastic construction. The lens is coated with Sigma's rubberized coating, which offers good traction. There are two switches on the lens, one to activate or deactivate image stabilization, and one to enable or disable autofocusing. The lens features a distance scale, but doesn't offer a depth-of-field scale. The lens mount is metal, and the 67mm filter threads are plastic.
The focus ring is mounted at the end of the lens, 1/2-inch wide and composed of raised rubber ribs. As mentioned, the focusing range of the ring is fairly small, only 45 degrees, making manual focusing a bit tedious. We found that the focus ring has very little dampening, letting it turn very freely. The focusing throw is bounded on either side by hard stops. The lens will focus slightly past infinity. The front element doesn't rotate while focusing.
The zoom ring is the larger of the two, 3/4-inch wide, also composed of raised rubber ribs. There is around 50 degrees of rotation in the zoom ring, and a nice level of resistance to the ring; not too tight, and not too loose. There's no evidence of zoom creep, and Sigma is confident enough in their design that they didn't include a zoom lock.
Optical image stabilization is included in this lens, making it an obvious improvement over the previous 18-50mm versions. We test this feature separately, but in informal testing we can say that it does work well.
The included lens hood is of the circular design, adding a further 1 1/2 inches to the overall length of the lens. The interior of the hood is deeply ribbed to reduce the impact of any stray light entering the front element, and the hood can be reversed onto the lens for storage.
Canon EF-S 18-55mm ƒ/3.5-5.6 IS ~$150
Canon produces one of the better kit lenses, matching the Sigma for sharpness - if not being a bit sharper in some focal lengths. CA is also better at wide angle, but after 18mm, the Sigma takes the lead. If straight lines are important to you, there's a point of zero distortion at 35mm on the Canon lens.
Nikon 18-55mm ƒ/3.5-5.6G AF-S DX VR Nikkor ~$175
The Nikon kit lens we tested showed obvious signs of de-centering, though even with this head start it's reasonable that the Sigma shows sharper results. CA is controlled very well by the Nikon glass, corner shading is about the same, and distortion is controlled better by the Nikon, offering a point of zero distortion at the 35mm mark.
Pentax 18-55mm ƒ/3.5-5.6 AL II SMC DA ~$160
The Pentax kit lens is also very good. Sharpness is on par with the Sigma, if not better - at 43mm and ƒ/5 it's tack-sharp across the frame. CA is well-controlled, better than the Sigma at 18mm, and comparable across other focal lengths. Distortion and corner shading are slightly better on the Pentax, again showing a point of zero distortion at around 35mm.
Sony 18-55mm ƒ/3.5-5.6 DT SAM SAL-1855 ~$200
The Sony kit lens is on the same level of sharpness as the Sigma, marred only by an odd spot of corner softness on our review sample at the 24mm mark. Stopped down to ƒ/8, the Sony is one of the sharper kit lenses we've tested, and sharper than the Sigma. CA is a bit high at wide angle, but not higher than the Sigma, and on par with the Sigma at other focal lengths. Corner shading is similar to the Sigma, and distortion is better controlled, with significant barrel distortion that reduces to zero at 55mm.
Tamron 17-50mm ƒ/2.8 XR Di II VC LD Aspherical IF SP AF ~$700
Offering a 17mm wide angle and a constant ƒ/2.8 aperture, there's a lot to like about this popular Tamron model, now modified to include Vibration Control image stabilization. The price tag is a bit higher and at the time of writing it's only available in the Nikon mount, but for the money you get better performance in sharpness and chromatic aberration.
The Sigma performed well in our tests, with good results for sharpness wide open, and even better when stopped down to just ƒ/4. CA is a bit high at 18mm and distortion is a complicated mix of barrel and pincushion, but nothing overly objectionable.
So who's this lens made for? Image stabilization is more important for Sigma, Nikon and Canon, where it's not built into the camera body; it seems to be a good replacement for the standard 18-55mm kit lens, offering a slightly larger maximum aperture (ƒ/2.8-4.5 instead of ƒ/3.5-5.6). The Nikon and Canon kit lenses are fairly competent, so it's going to come down to personal tastes which are beyond the scope of this review.
The VFA target should give you a good idea of sharpness in the center and corners, as well as some idea of the extent of barrel or pincushion distortion and chromatic aberration, while the Still Life subject may help in judging contrast and color. We shoot both images using the default JPEG settings and manual white balance of our test bodies, so the images should be quite consistent from lens to lens.
As appropriate, we shoot these with both full-frame and sub-frame bodies, at a range of focal lengths, and at both maximum aperture and ƒ/8. For the ''VFA'' target (the viewfinder accuracy target from Imaging Resource), we also provide sample crops from the center and upper-left corner of each shot, so you can quickly get a sense of relative sharpness, without having to download and inspect the full-res images. To avoid space limitations with the layout of our review pages, indexes to the test shots launch in separate windows.
Sigma 18-50mm f/2.8-4.5 DC OS HSM User Reviews
9 out of 10 points and recommended by Alexsoox (1 reviews)Cheap,SharpNot full range f/2.8
This is a great lens for two hundred bucks. The construction is very nice. Very sharp at over f6. Compare to the Nikon 35mm/1.8 at ~25-40mm f6-11 is very hard if not impossible to tell the differents at the computer. I was a little worry ordering this lens since a lot of people complain about front and back focus. I been a film guy for over 20years and this is my first Digital SLR lens. Very nice contrast and color too. Wish it came with the Sigma lens bag.reviewed March 19th, 2010 (purchased for $200)
10 out of 10 points and recommended by zoro02 (4 reviews)Affordable, sharp at F4 - 8, OS worked, closeup/macro is a bonus @ 20cm, non rotating zoomSoft @2.8, nothing else ATM until I run a field test
I just got the lens and took some picture with hot shoe flash at night, this is what I think about the new Sigma 18-50 F2.8-4.5 OS HSMreviewed March 15th, 2010 (purchased for $270)
1) Build - Can't be compared with Tokina but acceptable
2) Zoom ring - damped well controlled
3) IQ - pretty sharp even at f4. Very sharp at f8
4) Bonus - closest distance approx 20cm, can be use for closeup/macro which was NOT advertise as a feature. When I add 2x teleconverter, this will give me macro effect.
Will post image in flickr sometime this week. Overall I'm very satisfied especially $$/quality features.
9 out of 10 points and recommended by ksbuehler (4 reviews)light, inexpensive, 18mm @ f2.8 with stabilization, HSMa bit of Chromatic aberration, focal field curvature
I concentrated my tests on performance at F2.8 and 18mm, as that is what I care about. I have a 45mm TSE 2.8 and 70-200mm F4L IS to cover other focal lengths.reviewed September 15th, 2009 (purchased for $300)
For the price, this lens is great. I wanted a wide lens for use in low-light without a flash on my Canon 20D. This is the least expensive f2.8 18mm lens on the market with stabilization.
Weighing only 400g, it is also the lightest stabilized 18mm f2.8 lens on the market. Size and weight are big factors for me. I fly internationally for business quite a bit, and prefer to travel carry-on only. Between dress shoes and business clothes, not much room for my photo gear.
As a bonus, it uses the same 67mm filters as my 70-200mm f4L IS.
The 18-50 lens does not rotate or extend during focusing and zooming. Great!
Build feels pretty solid.
The HSM is NOT full time manual. I've been pampered by my Canon L lens so this was unexpected. I'll get over it.
I used a few USAF 1951 test targets to measure resolution performance from f2.8 through f11. Sharpness is nearly as good across the frame at f2.8 as f8.
Be aware there is noticeable field curvature at f2.8 and 18mm. (I'm talking about the curve of the focal plane, not barrel/pincushion distortion). If you focus on a perfectly flat wall at f2.8 @ 18mm, either the center of the image will be in focus, or the corners, not both. In the real world, a curved focal "plane" is not usually a problem. Most of my shots are not of a perfectly flat object that is perfectly parallel to the sensor plane. I'm not sure how slrgear, or any other site, can account for this in their tests. My method is to step through a few focal distances to check quality of focus in the corners.
The optical stabilization works well at 18mm. I can shoot at 1/15 with nearly 100% of shots being very sharp. At 1/8, it's more like 30~50% of shots very sharp. It seems to depend even more on how much coffee I've had so YMMV.
At 18mm, there is chromatic aberration present in the corners at all apertures tested (f2.8 to f11). It appears to be consistent across all corners, indicating that it could easily be removed in s/w. Further, the CA seems fairly consistent across apertures from f2.8 to f11. I have not tested software correction of the CA yet nor have I checked CA performance at other focal lengths.
There is about 1/2 stop vignetting near the corners at 18mm and f2.8. From reviewing test data for other lenses, this is better than the Canon 17-55 2.8 IS, Canon 16-35 2.8, Sigma 18-50 2.8 EX DC. It is about the same as the Tamron 17-50mm 2.8 XR Di II.
Macro performance. I put on a 12mm extension tube. Working distance is 0" at 18mm, 1" at 28mm, 2" at 35mm, and 4" at 55mm. The frame width is 1.3" at 18mm, 1.4" at 24mm, 1.5" at 28mm, 1.75" at 35mm, and 2.1" at 55mm. Using Canon 20d, followed by Canon 1.4 II TC, followed by 12mm ring, followed by Sigma 18-50 2.8-4.5 gives a 1" frame width at 55mm. 20d+1.4x+20mm tube+Sigma gives 0.75" frame width at 55mm.
Next things on this lens to test, not necessarily in this order:
1) 28mm performance and Optical stabilization performance
2) optical stabilization in 55mm macro configurations
3) 45mm performance vs my 45mm TSE