Intro

The Sigma 20mm f2 DG DN is a wide-angle prime lens designed for full-frame mirrorless cameras. Announced in February 2022, it’s initially available in Sony-e or Leica L-mounts, but as always we can pray for Canon RF and Nikon Z versions in the future. Sigma loaned me a pre-production sample to try out and in this video I’ll show you what it can do and how it compares to Sony’s FE 20mm f1.8 G lens. As always if you prefer to read a written version, keep scrolling for the highlights!

The 20mm f2 DG DN joins Sigma’s expanding Contemporary I series of compact lenses designed for full-frame mirrorless cameras. There’s now seven models in total covering 20 to 90mm focal lengths, four of which have f2 apertures, and the new 20 becomes the widest so far. 

Positioned on the edge of ultra-wide coverage, 20mm on full-frame is perfect for expansive landscapes, large interiors, dramatic closeups, astrophotography, and also an ideal length for presenting videos to camera or vlogging. Pop it on an APSC camera like Sony’s A6000 series and it becomes equivalent to 30mm, a nice walkaround length for general use.

20mm may not seem the most common focal length around, but there are still plenty of options for Sony owners. Samyang has a slightly brighter 20mm f1.8 for around $499, albeit manual focus only. Tamron has an autofocus 20mm f2.8, one stop dimmer than the new Sigma, but smaller and cheaper at $299. Sigma itself has an excellent high-end 20mm f1.4 lens that’s one stop faster and part of its acclaimed Art series for around $899, but it’s large, heavy, and an adapted DSLR design rather than a native mirrorless model; I wonder if we’ll see a native DG DN version in the future? 

I’d say the biggest competition comes from Sony which also has a pair of 20mm lenses: there’s the e 20mm f2.8, a tiny pancake lens designed for APSC cameras only at $349, but it’s the FE 20mm f1.8 G, that I’ll be directly comparing against the new Sigma in this review.

Here’s the Sigma 20mm f2 DG DN on the left with the Sony FE 20mm f1.8 G joining it on the right. Both are full-frame lenses, and while the Sony is a tad brighter, it’s also larger and at around $899 or pounds, roughly 40% more expensive. I’ve got loads of direct quality comparisons between them but let’s start with their designs.

So here’s the Sigma 20 f2 DG DN mounted on a Sony A7 III body which I’ll use for all my tests in this review. It’s 72mm long, 70mm in diameter, weighs 370g, and shares the same design as previous models in the I series, with a metal body, ribbed control rings, and weather-sealing at the mount, albeit not throughout the entire barrel. My test sample was made in Japan.

Working outwards from the lens mount is a switch for auto or manual focus, followed by an aperture ring from f2 to f22 in one third increments with an A position for body based control; note unlike Sony’s lens, this is not de-clickable. Then towards the end of the barrel is a well-damped manual focusing ring and a 62mm filter thread.

Like other I series lenses, Sigma supplied the 20 f2 with a choice of two lens caps, one a traditional spring-loaded plastic cap, and the other a small metal disc held in place by magnets. The latter is a fun addition which snaps into place quite satisfyingly, but ultimately I ended-up using the plastic cap, which is also easier to fit or remove when the supplied aluminium petal hood is fitted.

Now back to the lens without the hood or caps, before switching to the Sony FE 20 f1.8 G which at 85x74mm and 373g may be almost identical in weight, but both longer and wider; this sample was made in Thailand. Sony positions its AF / MF switch between the two rings with the bonus addition of a custom button above it. The aperture ring runs between f1.8 and f22, again in one third increments with an A position for body control, although unlike the Sigma, a switch on the opposite side declicks it for silent adjustments that will be appreciated by videographers. Then towards the end of the barrel is a rubberised manual focusing ring that feels looser than the Sigma, but still very smooth. I preferred the feel of the Sony manual focusing and also its shorter throw, but you may prefer the Sigma. 

Sony describes the lens as being dust and moisture resistant with a rubber grommet at the mount, while the filter thread is slightly wider at 67mm.

Ok now for my tests starting with autofocus and you’re looking at the Sigma 20 f2 at f2, on the A7 III body in Single AFS mode with a central AF area. Here you can see the usual contrast-based wobble as the camera confirms the focus in AFS mode and it all happens reliably and fairly quickly.

Switch to Continuous AFC mode and the camera employs phase-detect autofocus with a more confident focus pull from near to far and back again. The actual focusing speed itself is similar, but by not wobbling to confirm, the overall process takes less time.

For comparison, here’s the Sony FE 20 1.8 at f1.8 on the A7 III in Single AFS mode where you can see it’s visibly a little faster, snapping into focus more swiftly than the Sigma. There’s still a wobble, but it’s less noticeable.

Switching to Continuous AFC and the wobble goes, although I wouldn’t say it’s any quicker than the Sigma in AFC mode. So from this first test I’d say the Sony lens is faster in Single AF mode, but in Continuous autofocus, they’re roughly similar.

Next for a similar test but for movie autofocus which on the Sony cameras is always continuous. So starting with the Sigma lens at f2 and filming in 4k on the A7 III, you’ll see it refocusing smoothly and confidently. I’ve also left the ambient sound from the internal mics on, where there’s no evidence of the lens audibly focusing.

Next up, the Sony 20 1.8 at f1.8 and again filming 4k video on the A7 III, and again a similar result, smooth, confident and essentially silent even when using the built-in mics.

Before I dive into my optical comparisons, a note on distortion. Like most modern mirrorless lenses, especially compact wide models, the Sigma 20 f2 DG DN employs a software profile to correct for geometric distortion.

To apply this to JPEGs on a Sony camera, you’ll need to go into the Lens Compensation menu and ensure Distortion Compensation is set to Auto. If you’re shooting RAW, the profile should be applied by the converter, although as an early test sample mine was not yet supported by Adobe Camera RAW. This will come in an update, but in the meantime I found the existing profile for the Sigma 24 f2 DG DN was pretty close if you didn’t fancy applying manual corrections to RAW files.

Either way, distortion correction will incur a small crop, so before going any further, here’s a quick before and after. First, Brighton Pier with the Sigma 20 f2 DG DN without the profile, and now after Distortion Compensation set to Auto. Note the reduced field of view at the sides. So how does this compare to the Sony lens?

Here’s the Sigma 20 f2 DG DN at the top and the Sony FE 20 f1.8 G at the bottom, both with Distortion Compensation set to the default off, so there’s no geometric correction being applied to either lens. I’ve cropped the sky and beach so you can see both shots at the same time, but presented their full width here to compare the coverage. Looking at the sides shows both lenses capturing roughly the same field of view as you’d expect, but notice how the building in the middle of the Sigma image looks larger than the Sony, bulging-out a bit due to barrel distortion.

Next here’s both lenses with Distortion Compensation set to Auto, where both have applied a small crop to perform the correction, but Sigma’s is more significant, resulting in a slightly reduced field of view compared to the Sony. Everything is looking correct now though.

Here’s another example, starting with the Sigma 20 f2 DG DN with Distortion Compensation set to off, where the barrelling effect is more obvious thanks to the straight lines in the composition. 

And now with Distortion Compensation set to Auto where the geometry has become fully corrected, albeit again at the cost of a small crop. If you’re not used to seeing this, it can look dramatic, but I’m personally at peace with software corrections on relatively compact and affordable lenses, as for me the end result is what counts. But the bottom line is the more distortion you have to correct, the greater the crop will be.

For comparison, here’s the Sigma 20 f2 DG DN on the left and the Sony FE 20 f1.8 G on the right, both with Distortion Compensation set to Off, where it’s clear that while both lenses suffer from some barrel distortion, it’s much worse on the Sigma.

And now here’s both lenses with Distortion Compensation set to Auto, both now looking much better in terms of geometry, but again the more significant correction required by the Sigma comes at the cost of a greater crop and a slightly reduced field of view compared to the Sony.

The Sony lens clearly has less geometric distortion to start with and if you’re not shooting straight lines, you can actually get away without applying the correction at all, and thereby enjoy the widest potential coverage. While no-one’s stopping you from also shooting the Sigma without correction to match that coverage, I personally feel it needs it, so for all my tests and results to follow, both lenses will have their full profiles applied, and therefore the Sigma will look a little less wide than the Sony. It’s still comfortably wider than a 24, but a little tighter than the Sony when profiles are applied to both lenses.

Ok so let’s start with that landscape view, with the camera angled so details run into the corners where lenses typically struggle the most. You’re looking at the Sigma 20 f2 DG DN at f2 on the Alpha 7 III here with Distortion Compensation set to Auto and the lens focused in the middle. Taking a closer look at the centre shows a well-corrected image with lots of sharp details with the aperture wide-open, and only minor benefits to sharpness and contrast by closing it. 

As you head into the corner, the image becomes darker due to vignetting with the aperture wide-open and details become a little softer. As you close the aperture, both gradually improve with vignetting mostly gone by f4 to f5.6 and the corner nice and sharp at f5.6 to f8.

Let’s briefly return to the wide-open corner crop at f2 which again was with the lens focused in the middle. Now let’s switch it for a version still wide-open at f2, but with the lens now focused on the corner instead, where the details have become much sharper. So the lens doesn’t have a particularly flat field, but it will deliver sharp results wide open on the subject you focus on, or will capture a fully sharp image across the frame when stopped-down to around f5.6 to f8.

Ok now for a comparison between the Sigma 20 f2 on the left and the Sony 20 1.8 on the right, both crops from the centres of their images, both wide open at f2 and f1.8 respectively, and both focused in the middle. Both look good, but the Sony image has slightly crisper details and a little extra contrast. Note the subjects on the Sony look a little smaller due to it delivering a fractionally wider field of view.

Head out to the corner on the same image and you’ll again notice a small difference in the field of view, but it’s clear the Sony on the right has a flatter field and sharper details as a result. As you gradually close the apertures, both lenses improve in the corner, but the Sony remains consistently ahead and by f4 is looking very crisp.

Moving onto portraits, the 20mm focal length is wide but still effective for environmental or contextual shots. Here’s the Sigma on the A7 III filming 4k video wide-open at f2 from arm’s length where there’s mild but attractive blurring in the background. 

And now for the Sony 20 f1.8, wide-open at f1.8 and again held at arm’s length. Notice the slightly wider field of view on the Sony when corrections are enabled on both lenses, which coupled with the slightly brighter aperture has actually delivered a similar amount of background blur. 

Side by side, beyond the subject looking larger on the Sigma on the left, the actual style and rendering looks quite similar. I’ll have some walking vlog tests for you later, but first the rendering of bokeh balls from close range.

Here’s the Sigma 20 f2 at f2 from near to its closest focusing distance, where it’s possible to enjoy more background blur than you may first assume given the wide focal length. As you close the aperture, the cats-eye shaped bokeh balls towards the corners gradually become more circular.

Now back to the Sigma wide-open at f2 on the left versus the Sony 20 1.8 wide-open on the right, both from the same distance. As you might expect the slightly brighter 1.8 aperture of the Sony is delivering slightly larger blobs, but the most striking difference is how much more rounded they are from the start, avoiding the cat’s-eye effect on the Sigma. Looking closer you’ll also see that while both lenses exhibit some textures within the blobs, the Sony’s on the right are visibly cleaner. Ultimately while it’s down to personal preference, I reckon the Sony looks better in this comparison.

At the other end of the aperture scale and focusing distance, here’s how the Sigma 20 f2 handles sunstar diffraction spikes, between f11 and the minimum aperture of f22, where the nine aperture blades render 18 crisp spikes. Here’s the Sigma on the left and the Sony on the right, both at f22 for comparison and other than the orientation of their nine aperture blades, the result is similar.

And now for that vlogging test I promised you earlier, with the Sigma 20 at f2 and held at arm’s length on the A7 III with sensor stabilisation only. I really like the 20mm focal length for vlogging as you’re smaller on the frame than a 24 at arm’s length, while showing more of the surroundings, and as you saw earlier the aperture does allow for some blurring.

For comparison, here’s the Sony 20 1.8, at f1.8, and again the slightly wider field of view after both have lens corrections applied. If you’re holding the camera at arm’s length, the Sony definitely has the advantage here being slightly wider with corrections enabled, and for video you may get away with turning them off altogether on this lens.

Next for a focus breathing test, starting with the Sigma 20 f2 closed to f22 and focusing from the furthest to the closest distance and back again. You have to turn the focusing ring quite a lot for this range, so forgive the wobbling. Notice how the field of view reduces as you focus closer due to breathing.

And now for the Sony which starts a little wider due to lens corrections, but ends up with a similar magnified field of view at the closest-end of the focusing range. So both lenses suffer from quite noticeable focus breathing, but the difference is greater on the Sony. I’ve noticed this before though: focus breathing is an artefact Sony seems happy to tolerate on many of its lenses.