Given the rather ho-hum resolution performance of Sony's native E 30/3.5 Macro lens, particularly on the NEX-7 camera, we were quite curious about performance of the Sigma lens. After all, this is the first true competitor to the Sony's native lenses, and actually a very attractive competitor both in terms of price as well as (paper) characteristics.
After running the lens through a series of tests, it's safe to say that it does perform better than Sony's E 30/3.5 (at least in resolution department). Not significantly, but still better. If we look at the side by side comparison, Sigma edges a victory around corners, where it is consistently better than the E 30/3.5 Macro. But, if we look at the lens in isolation, its performance is, while decent, not necessarily jaw-dropping. Center sharpness is superb throughout the tested apertures and both on NEX-5n as well as on NEX-7. Border image sharpness is fairly good too, but clearly lags center in quality. Moreover, while borders improve with stopped down apertures, improvements are rather marginal - depending on the point of view, you can argue that performance is very uniform (which would be a 'glass half-full' point of view) or you can argue that there is no improvement at lower apertures (which would be a 'glass half-empty' point of view).
But the improvement in resolution over Sony E 30/3.5 seems to be quite eluding when it comes down to print quality. If we look at the SQF charts produced by the Imatest, we can see that the the improvement is not sufficient to produce higher quality 24in prints - both NEX-5n as well as NEX-7 are expected struggle, particularly around borders.
The imaging target corner crops below demonstrate what we have to work with. Sharpness on NEX-5N looks fairly good both wide open as well as stopped down, but quality at f/2.8 degrades on NEX-7, making corners look smudgy.
Before making our final conclusions about the sharpness of the lens, let's take a look at the usual infinity shots done with NEX-7 camera (uncorrected for artifacts). We can see that there is basically no difference in the center between f/2.8 and f/8 that can be detected visually. Borders, on the other hand, do show a little bit of softness at f/2.8, which is in line with what Imatest seems to be thinking. Overall though, image quality does not seem to be particularly bad and is likely going to be perfectly acceptable to majority of users.
Color & Rendering
There is nothing really extraordinary that can be said about Sigma's 30/2.8 EX DN lens in this section. The lens performs reasonably well with basically no major flaws, but nothing really that would make this lens stand out in a league of its own.
Color palette is well balanced and neutral (assuming you're shooting in RAW and Adobe RGB color space, which gives you a little bit more color gamut coverage). Global contrast levels are also average throughout most aperture levels, although there is a little bit of degradation at close focusing distances. Tonal reproduction is fairly decent - again nothing that would make this lens stand out but sufficient to give you enough flexibility in post-processing.
This is actually an interesting trend that has started to develop lately. A whole range of new, lower priced lenses like Sigma 30/2.8 delivering fairly solid performance that would besuitable for the majority of users out there. We clearly have not reached a breaking point in lens prices, given the recent hikes from Sony, Nikon, Leica and other manufacturers, but it seems that a new segment of reasonably prices, but still performing quite well lenses is under development. Think about it. At $199, Sigma managed to incorporate a linear AF motor, add a double-sided aspherical element and apply multi-coating ot the lens. A similar entry level lens from Canon, the 50/1.8 Mk2 lens costs even less - $155 after a price hike a few years ago. Or Samyang's 85/1.4 and 35/1.4 lenses, which created their own share of uproar due to their high optical performance. Bottom line is that this is a welcoming trend, and hopefully it will continue to expand so that more users can experiment with a larger variety of fairly decent, yet inexpensive lenses. Digressing....
Going back to the Sigma - as mentioned in the paragraph above, the company has added two aspherical surfaces (single element) to the lens design in hopes of reducing various forms of chromatic aberration. Despite that, the lens does produce some lateral CA though. In the Imatest benchmarking, Sigma clocked CA of ~1.2px on NEX-5n and ~1.6px on NEX-7. This means that you should expect some color fringing around high contrast areas - those would be visible under high magnification, but are unlikely to cause issues with prints. Longitudinal CA on the other hand is minimal to non-existent. There is a little bit of light bleeding that can be noticed in higher contrast areas, for example in night shots with brightly lit bulbs, but this artifact is also quite benign.
DOF & Bokeh
Given its moderately fast aerture, Sigma 30/2.8 would not dazzle you with a creamiest OOF rendering out there. For those purposes, get something faster, with shorter MFD or longer FL (or all of the above). Still, on an APS-C sized NEX cameras, the lens gives you a DOF of 1.02cm at the MFD - not too shabby for most purposes. Bokeh feel with this lens is fairly neutral - OOF highlights retain even illumination and there are no noticeable harsh edging to highlights. As mentioned, contrast levels seem to be a little bit lower at the MFD, but that sort of becomes a non-issue in practical use.