Sigma 12-24mm f/4.5-5.6 EX DG Aspherical HSM

SLRgear Review
June 15, 2009
by Andrew Alexander

The Sigma 12-24mm ƒ/4.5-5.6 was announced in October 2003, which still holds the record as the widest zoom lens available for a full-frame camera body. At 12mm, the lens produces an impressive field of view of 122 degrees.

The lens was designed to fit the 35mm film frame, and on cameras fitted with APS-C sized sensors, the lens will produce an effective field of view of 19-38mm (Canon) or 18-36mm (Nikon and others). It 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 change in aperture with focal length:

The Sigma 12-24mm ƒ/4.5-5.6 features an integrated lens hood, and does not take front-mounted filters; gels can be attached to the rear element. The lens is currently available in Sigma, Canon, Nikon, Sony and Pentax mounts for approximately $775.

Sharpness
The 12-24mm provides moderately sharp images throughout its zoom and aperture range, though stopping down only improves performance marginally.

When used on the APS-C sized sensor of the 20D, the lens showed some problems at its widest apertures. Between 12 and 15mm we note some unevenness to the overall sharpness of the image, with images showing between 2-3 blur units. Between 17mm and 20mm this unevenness clears up but we now note substantial corner softness (around 6 blur units) on the left side. Average sharpness at these ranges is around 2-3 blur units with a nice central spot of sharpness (1.5 blur units). Interestingly at 24mm neither of these issues are prevalent, and we are rewarded with a generally even image (around 2 blur units) with just slightly soft corners (3-4 blur units).

Stopped down, there is some improvement in sharpness, but not much. At 12mm you have to stop down to ƒ/11 to get an optimally even image, and at this setting we note image sharpness of just less than 2 blur units. Between 15-17mm you'll want to stop down to ƒ/11 if not ƒ/16 to remove traces of corner softness. 24mm is probably the most controlled aperture of them all, with excellent performance at ƒ/8; a large region of 1-1.5 blur units in the central region of the frame, marred only by slight corner shading (3-4 blur units) in the top corners. Performance fully stopped down isn't bad but it isn't great, at 3-4 blur units across the frame between ƒ/22 and ƒ/29.

On a full-frame camera such as the 5D, the lens shows a bit more 'wild' performance. Wide open at ƒ/4.5 and 12mm, the lens shows average performance throughout the frame in the range of around 2 blur units, but the corners - the extreme corners, mind you - get some impressive softness in the realm of 5-8 blur units. The overall image is quite uneven between 15 and 17mm and ƒ/5 while still being moderately sharp (around 2 blur units) and things get a bit more controlled at 20-24mm, with performance in the 2-3 blur unit range and some corner softness at 24mm.

As in the APS-C sensor body testing, stopping down mitigates these problems slightly, but doesn't cure them completely. The optimal settings remain almost the same: ƒ/11 at 12mm, and ƒ/11 or ƒ/16 at 15-17mm. The story changes slightly at 20-24mm, where the 20mm setting seems to be the most stable; performance at ƒ/5.6 is similar to that of ƒ/8 or ƒ/11. Stopping down at 24mm removes some corner softness, but not all.

Chromatic Aberration
For an exceptionally wide-angle lens, the Sigma 12-24mm controls chromatic aberration quite well. On the sub-frame 20D we noted some CA in the corners in areas of high contrast. According to the tests it's most prevalent at 15mm at ƒ/8.

The full-frame 5D exposes the sensor to more of the lens' edge performance, so it's not surprising to see a higher prevalence of chromatic aberration. It's controlled well at 12mm, where this lens will see a lot of use, and trouble spots on full-frame are at 15mm and 24mm.

Shading (''Vignetting'')
On the subframe 20D, the 12-24mm produces very light corner shading, and then only when the lens is used in its wider angle settings (12-15mm) and wide open. At these points, the corners are about a half-stop darker than the center. As the lens is zoomed out and / or stopped down, this corner shading reduces to the quarter-stop level.

Upper left corner, 5D, 12mm, f/4.5

On the full-frame 5D however, the lens produces some of the darkest corner shading we've ever seen. At its worst points the data is literally off the charts; at 12mm and ƒ/4.5, the corners are 2 1/4 stops darker than the center. Corner shading is always a problem with this lens, with average corner shading being around a full stop darker than the center, ie., any focal length at ƒ/11. Either you'll love the look on full-frame, or expect to spend some time fixing it in post-processing.

Distortion
For all its corner shading, the lens is remarkably free of distortion. On the sub-frame 20D the lens shows some slight barrel distortion (+0.4% in the corners at 12mm) throughout the range as you approach 24mm, where distortion reaches zero.

On the full-frame 20D, distortion is a bit more prevalent, but even here it's remarkably well-controlled. Average performance (distortion throughout the frame) is very good, at just +0.25% barrel distortion at 12mm. In the corners, you see a bit more distortion: +0.5% barrel distortion at 12mm, evening out to zero at around 18mm, and then entering the pincushion realm at 24mm (-0.2% pincushion distortion). These aren't huge amounts, but just enough to turn your straight lines into curves.

Autofocus Operation
The 12-24mm ƒ/4.5-5.6 uses Sigma's HSM focusing technology, allowing for a screw-less design that works on modern camera bodies. The lens focuses quickly and silently, going from infinity to close-focus and back to infinity in less than one second. Micro-changes in focus are done very quickly. Autofocus results can be overridden at any time by just turning the focus ring.

Macro
The 12-24mm makes a poor macro lens, with just 0.14x magnification. Minimum close-focusing distance is quite close, at 28cm (just under a foot).

Build Quality and Handling
The Sigma 12-24mm ƒ/4.5-5.6 is compact lens, with plastic construction and a grippable black texture. The lens itself isn't overly heavy, just over a pound, and mounts well with both our sub-frame (20D) and full-frame (5D) test bodies. The lens mount is metal, and while there are no filter threads to speak of (the lens hood is integrated into the design) Sigma thoughtfully includes a small clip over the rear element to allow you to slip in colored gel filters.

Only one control feature exists other than the focus and zoom rings, a small switch to enable or disable autofocus on the lens. The lens is equipped with a recessed and windowed distance scale, showing distance information in feet and meters. Depth-of-field measurements are provided showing marks for at least ƒ/5.6 and ƒ/8 (Sigma's press images show a depth-of-field marking for ƒ/11, but our sample of the lens doesn't have this). An infrared index marker is not present.

The zoom ring of the lens is the smaller of the two rings, at a half-inch in width. The ring is composed of raised rubber ridges running parallel to the body of the lens. Forty-five degrees of turning action runs the lens through its zoom range, and given that the lens extends about a quarter-inch at the tele end, zoom creep really isn't a problem. The zoom ring has a nice feel to it; not too tight, and not too smooth.

The focus ring uses the same texture as the zoom ring, raised rubber ribs running parallel to the lens body, but the larger size (one inch wide) means you won't be confusing the two. The ring is smooth to turn, and there are no hard stops on either end of the focus spectrum. The ring will keep turning even at close-focus or infinity focus, but a slight increase in tension will alert you that you've reached the end.

The integrated lens hood is petal-shaped, adding 3/4-inch to the overall length of the lens. As with most ultra-wide lenses, the juggling act to actually make the lens hood useful is a tough one; when used at wide angle, you are getting so much of the scene into your photo that to keep the sun out is a challenge.

Alternatives

When considering alternatives, the field is fairly wide: are we considering ultra-wide angle lenses for sub-frame sensor bodies? Or wide-angle competition for full-frame bodies?

Sigma 10-20mm ƒ/4-5.6 EX DC HSM ~$450
There's not much point getting the 12-24mm ƒ/4.5-5.6 if you have a sub-frame dSLR; consider instead Sigma's 10-20mm, which performed very well in our tests. Overall image sharpness is very good, CA is perhaps higher than we'd like, and distortion is noticeable. Sigma has also announced a constant ƒ/3.5 version of this lens.

Most manufacturers produce a ultrawide-angle zoom lens for the APS-C format.

Nikon 14-24mm f/2.8G IF-ED AF-S ~$1,800
Nikon's probably the closest competitor in the category of full-frame ultrawide zoom lenses, with this lens; at three times the price, it's faster (with a constant f/2.8 aperture) and razor-sharp to boot. The only area the Sigma has any kind of edge is in distortion, where the Nikon shows more barrel distortion.

As well, most manufacturers produce a wide-angle zoom lens for the full-frame body, in the 16-35mm or 17-35mm configuration.

Conclusion
While it's not the sharpest lens in the bag, the Sigma 12-24mm showed fairly decent results for sharpness. For optimal performance the lens must be stopped-down significant, to at least ƒ/8 or ƒ/11. CA is well-controlled and distortion is remarkably so. On sub-frame cameras corner shading isn't much of an issue, but on full-frame the lens showed the most corner shading we've ever seen. Whether you'll love or hate this look is in the eye of the beholder.

Despite this mixed bag of performance, our overall impression of the 12-24mm is good, and it's hard not to enjoy the true ultra-wide performance on full-frame. For the money you can't beat the convenience, and with a little work in post-processing the resulting images are excellent.

Sample Photos

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.

Still Life shot

VFA target

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.