The DxOMark camera benchmark suite is often a fairly reliable indicator of good stills and video capture in a phone (or, indeed, a standalone camera), and results are in for the Nokia Lumia 1020. It benchmarked just behind its Symbian-powered predecessor, the Nokia 808, for stills performance, though the DxOMark people found issues with video capture which let the Lumia 1020 down overall. Some quotes and links below. 

From the DxOMark summary:

Although there have been other ‘Pureview’ models from the firm the 41-Mpix Nokia Lumia 1020 is the successor to rightly popular Nokia 808 Pureview, a high-end smartphone that sits in first place in our mobile rankings with a score of 77 points. The firm may have dropped the Symbian OS and Pureview moniker from the Lumia 1020, but expectations are high.

The down-sampling of such large files from the 41-Mpix sensor worked well with the Nokia 808 Pureview resulting in detailed images with low noise, and that same strategy has been applied to the Lumia 1020. At first sight though, with an overall DxOMark Mobile score of 74 points, the Windows powered smartphone doesn’t appear to perform quite as well as its predecessor.

The Nokia Lumia 1020 comes fourth overall, behind the rival Samsung Galaxy S4 and Apple iPhone 5S but a closer look at the scores reveals the Lumia, in fact, ranks in second place for stills with a score of 79 points (just two points behind its sibling). What is somewhat disappointing is the camera sits in a lowly 12^th place for video, which obviously pulls the Lumia 1020 down in the overall rankings.

Interesting data points, though I'm a little surprised by the video performance in the tests. DxOMark says, of the video capture:

• Strong row noise evident
• Visible blocking artifacts
• Walking movements remain uncorrected by the video stabilization
• Autofocus oscillations are visible during refocusing.
• Over-correction of video stabilization when steadied (on a tripod).

In my experience, 'walking movements' are sufficiently large that they can't be removed by any stabilisation system, so I think this is an unrealistic expectation. A better test of the effectiveness of the OIS is to look at 'wobble' in detail in a landscape scene, especially when zoomed in. You may remember my own video comparison with the Nokia 808, showing clearly many of the Lumia 1020's video strengths:

Comments about autofocus oscillations are fair enough, as you can see from my video, especially when zoomed in, though I expect that this will be tweaked in an update to Nokia Pro Camera in the near future.

The tripod comments may well be valid, but the whole point of the OIS is that the 1020 can be used handheld, so I don't think this criticism should be taken too harshly.

DxOMark went on to say:

In fact, under low light conditions the Lumia 1020 outperforms rivals, producing still images with low noise and detail preservation that are simply the best we’ve ever seen from a smartphone.

Compared with the more recent Apple iPhone 5S, we can see that while Apple is getting the most from its existing technology but they appear closer to their technical limit than Nokia is with their Lumia 1020.

The last sentence perhaps hits the nail on its head. The Lumia 1020's hardware (as with the 808's before it) is tremendously capable and it will probably take an update or two to really get the most from it. 

In the meantime, the Nokia 808 still sits top of the DxOMark league table for smartphone cameras, with the Lumia 1020 in fifth place overall. I have every confidence that a re-test in a few months will see the 1020 significantly higher.

Comments?