Studio Tests

The 10-20mm exhibits significant curvature of field, which means that the region of sharp focus is not flat across the frame, but bowl-shaped with the corners curving towards the camera. This is a problem when testing a lens using flat charts, as it's impossible to achieve correct focus across the whole chart area at the same time. We have therefore in this case chosen to present results focused for maximum sharpness towards the edge of the frame (approximately 60% from the centre, corresponding to the outermost star in the chart representation in the widget above). This gives the best overall impression of the lens's capabilities, although not the highest achievable central sharpness. Click here for a screenshot showing a fairly typical comparison between the published data (left) and that focused for maximal centre sharpness (right).

Test results are reasonable, if unspectacular. Sharpness results are rather inconsistent, especially in the middle of the range, and distortion quite high. However chromatic aberration is impressively low.

Sharpness Sharpness is generally high in the centre of the frame, but is less impressive towards the edges. The lens is sharpest at the long end, and while it's also pretty good at 10mm, performance is weaker in the middle of the zoom range, especially towards the corners. Optimum results are generally achieved on stopping down to F8 or F11.
Chromatic Aberration Chromatic aberration is extremely well controlled, indeed this is the standout feature of this lens. When present it tends to be mainly of the visually less disturbing blue/yellow type, and even then is only really visible in the extreme corners of the frame at wideangle. Very impressive indeed.
Falloff We consider falloff to start becoming a potential problem when the corner illumination falls to more than 1 stop below the center. Maximum falloff is around 1.3 stops at all focal lengths, falling slightly on stopping down, but it's still around 1 stop in the extreme corners at F11 (again at all focal lengths). This is unlikely to cause any real problems in actual use.
Distortion Distortion is a little high, especially when compared to the other wideangle zooms we've tested recently. There's fairly strong (2.4%) barrel distortion at wideangle, with a rather unusual and complex pattern, localized mainly towards the corners of the frame (and therefore difficult to correct fully in software). This changes to pincushion distortion at longer focal lengths, peaking at -1.4% around 14mm.

Compared to the Tamron SP AF 10-24mm F3.5-4.5 Di-II, the Sigma is perhaps starting to show its age; the Tamron is significantly sharper across most of the frame at wideangle, although it has serious problems wide open, and the extreme corners tend to be rather soft. Towards the long end, though, the Sigma edges ahead; but it's probably fair to say that wideangle zooms aren't particularly bought to be used in this range, which is shared with the typical kit zoom. The Sigma also shows notable control of lateral chromatic aberration, but against that the Tamron has significantly lower distortion and falloff. This makes the choice between the two less than entirely clear-cut, however the Tamron probably edges it overall, unless you intend to shoot a lot wide open, or really need the Sigma's HSM focusing.

Extrapolation of test results to Four Thirds Format

The Sigma 10-20mm F4-5.6 EX DC HSM is available in Four Thirds mount, and it's possible to predict its expected performance on this system based upon the MTF50 data for the DX format. The Four Thirds frame (17.3mm x 13mm) has approximately 82% of the height of the DX frame (23.6mm x 15.8 mm), and the diagonal is approximately 75% of its length. Therefore, the expected data for Four Thirds can be attained by multiplying the MTF50 data (y-axis) by 0.82, whilst considering only the data out to 75% of the frame diagonal (i.e. the x-axis of the graph).

Overall this suggests that the 10-20mm F2.8 will be a reasonably competent performer on Four Thirds, but not outstanding. Compared to the Olympus Zuiko Digital 9-18mm F4-5.6, the Sigma is simply outclassed in practically every regard, with only its excellent control of chromatic aberration to boast about. In contrast the Olympus is significantly sharper, especially wide open, and has much lower distortion and falloff, making it the clear winner in this contest (although it is also rather more expensive).

Macro Focus

Wide angle zooms aren't really meant for closeup work, and the Sigma follows this trend. Maximum magnification is 0.14x, achieved at 20mm and a closest focus distance of 25cm, giving working distance of 12cm from the subject to the front of the lens; this is slightly below Sigma's specification.

Image quality is quite reasonable; it's sharp in the middle but soft towards the edges, with curvature of field presumably playing a role here. Best results are, as expected, achieved between F8 and F11.
Macro - 163 x 108 mm coverage
Distortion: Slight pincushion
Corner softness: Low
Focal length: 20mm (30 mm equiv)

FX (Full Frame) Coverage

The Nikon, Canon and Sony fit versions of this lens will all mount on full-frame DSLRs; on Nikon cameras (D3, D3X, D700) DX crop mode will be automatically selected (and the camera will therefore shoot at reduced resolution). The lens's image circle doesn't fully cover 35mm full frame at any focal length, as shown below; it's therefore unsuitable even for emergency use on this format.

10mm 12mm 14mm 17mm 20mm

Specific image quality issues

As always, our studio tests are backed up by taking hundreds of photographs with the lens across a range of subjects, and examining them in detail. This allows us to confirm our studio observations, and identify any other issues which don't show up in the tests.


Control of flare is a critical feature of a superwide lens; with such a broad view of the world, bright light sources will find themselves in the frame on a regular basis. The Sigma 10-20mm generally deals with this reasonably well, but can run into problems in more difficult circumstances.

With the sun placed directly in the frame at wideangle there are generally few really objectionable effects, but as is often the case, stopping down for sharpness or depth of field substantially increases the visibility of flare patterns. In these situations you'll have to choose your aperture carefully, to balance sharpness against flare.

A bigger problem, though, is seen when shooting towards the sun at longer focal lengths; the example below right shows diffuse color patterning across most of the frame (which in some areas could easily be mistaken for blotches of chroma noise). In these situations it would be advisable to provide more shading to the front element than provided by the hood.

10mm F13, sun in corner of frame 20mm F5.6, strong backlight
50% crop, upper left 50% crop, lower center

Chromatic aberration

Chromatic aberration is notably low on this lens, and the examples below illustrate that even in the worst case scenario, there's really very little to worry about. The only slight hiccup is the rather complex shape of the CA profiles at wideangle, which means that the color of the fringing varies across the frame (being red/cyan towards the edge, but blue/yellow in the extreme corner), making it somewhat difficult to correct fully in software if required.

The sample below illustrates this, on the left we have a RAW shot processed with no chromatic aberration correction, and on the right the corresponding JPEG from the Nikon D300 (this camera's automatic CA correction serves as a useful proxy for the kind of result you might expect to achieve in software using Photoshop's 'Lens Correction' filter, or any similar tool which relies on simply re-scaling the color channels). In this case the red/cyan fringe on the left side of the frame is near-perfectly corrected, but the blue/yellow fringe in the top right corner is not. As the blue channel CA profile is highly non-linear, correcting the fringing in the corner would simply introduce new problems elsewhere in the frame.

Nikon D300, RAW + ACR Nikon D300 JPEG
10mm F11
100% crop, left edge
100% crop, top right corner

At longer focal lengths, there's a relatively low level of red/cyan fringing towards the corners of the frame, which diminishes progressively on stopping down; the example below is about as bad as you'll ever see. Again a Nikon D300 JPEG is included, and in this case shows how effective a simple 'channel-scaling' type correction can be at reducing fringing when the CA profile is relatively simple.

Nikon D300, RAW + ACR Nikon D300 JPEG
20mm F5.6
100% crop, top left corner

Corner/wide open softness and curvature of field

Our studio tests show that the Sigma is not a particularly sharp lens at many settings - wide open the lens is distinctly soft towards the edge and corners, and at intermediate focal lengths (12-14mm) this persists even on stopping down. The samples below illustrate the effect this has on 'real world' shots; the Sigma is never able to achieve the biting central sharpness characteristic of the Tamron 10-24mm F3.5-5.6 Di-II, but neither does it quite exhibit that lens's propensity for extremely soft corners.

At the wide end of the zoom range, central sharpness is good at F4, only improving slightly on stopping down. However as we move towards the edge of the frame; the story changes, with sharpness (and to a small extent chromatic aberration and vignetting) problematic at F4, but showing dramatic improvements at F8. This really is a lens you don't want to shoot wide open if you can avoid it; the good news is that you'll rarely have to.

10mm F4 10mm F8
Nikon D300
100% crop, centre
100% crop, upper right
100% crop, bottom right corner

The inconsistency is even more marked towards the middle of the zoom range, and in the example below we can also see the effects of curvature of field creeping into play. Central sharpness is high wide open, but the edges and corners distinctly soft. Close down the aperture two stops and the upper left stays soft, but the lower left corner comes into sharp focus (and looks rather sharper than our chart tests would suggest). This despite the depth of field at in this situation being more than sufficient, in principle at least. Overall this kind of behavior makes real-world results from the Sigma somewhat mixed.

14mm F4.8 14mm F9
Nikon D300
100% crop, centre
100% crop, upper left
100% crop, bottom left corner