Wednesday, January 9, 2013

Nikon D5200 Sample Image Review: Noise at High ISO

Photography Blog has sample images from the Nikon D5200 here. For reference, I am comparing them to their Sony NEX-7 samples. Unfortunately, Photography Blog wasn't using the same bookshelf test target for the D7000 two years ago, so the nearest equivalent would be their Sony NEX-5N review. This should be a good baseline to compare the new Toshiba 24mp sensor versus the known quantities of the Sony 24mp and 16mp sensors. (However, a really frustrating thing is that the D5200 samples, though shot at the same f/8 aperture, use a tighter focal length). I'll also be discussing their D3200 samples images to compare Nikon to Nikon, and to see what a year's difference makes between sensors of similar megapixel count.


ISO 800


There is no meaningful difference between the D5200 and NEX-7. The out of camera Sony JPEG output looks slightly more saturated (but not by a lot) and a bit more sharpened, but you would expect that between the two manufacturers. Curiously, between the two Sony cameras, there isn't a lot of difference in noise at the pixel level either... you would expect that the NEX-7 would have more shadow noise than the NEX-5N, but again, it's not much of a practical difference. At the pixel level, the D3200 already shows slight degradation from noise, and is easily identifiable as the noisiest of the four cameras, though not by a lot.

ISO 1600


At ISO 1600, there is a visible difference in shadow noise between the Sony cameras.... not much, and it would probably go away if you downsize the 24mp file to 16mp size. When viewed at 100%, there is again not much difference between the D5200 and the NEX-7. However, once you go down to the pixel level, the texture of the noise of on the D5200 is better than on the NEX-7... the grain is finer and shows less noise reduction artifacts. In fact, the quality of the noise is refreshingly "Nikon-like"... it's fine, and not splotchy at all. Default colour saturation remains a bit less than the NEX-7. The D3200 is again a tad bit fuzzier than either of the other 24mp cameras. Flat patches of colour show more noise and less saturation than the other cameras, but edges are still rendered reasonably sharply. By comparison, the NEX-5 has less noise but the same noise reduction artifacts as the NEX-7, so my guess is that had there been equivalent test shots, the Nikon D7000 would be better than the Sony cameras, possibly even with the D7000.

ISO 3200


At ISO 3200, the NEX-7 still looks good at 100%, but edges start looking fuzzy when you view at the pixel level. The NEX-5N still renders lines and hard edges sharply, proving once again that the 16mp sensor used on this and the D7000 makes for a comfortable carefree shooting experience through most of the ISO range. Noise reduction artifacts are again more prominent on the NEX-7, with blues suffering the most. On the D5200, edge detail is not as crisp as at ISO 1600, but it's still better than the NEX-7 at ISO 3200. Noise reduction again shows a lighter touch. The pattern of the grain is coarser than at ISO 1600, but still has a very film-like texture. It's visually better than the D3200, where the noise pattern is courser, and while not showing the artifacts of the Sony, does show more chroma noise than the D5200.

ISO 6400


This is as high as I would go with any of these cameras in practical situations. At full size, they are still usable, but noise is very apparent at 100% viewing. At this level, the NEX-7 uses a steeper tone curve than the NEX-5 to darken up the shadow regions in order to mask the noise. However, the same trends at ISO 3200 apply here: at the pixel level, the D5200 looks better, and preserves edge detail better. (In case you are wondering, Photography Blog's bookshelf test target doesn't have a lot of fine detail to judge these cameras on). Again, the D3200 lags behind the other three.

Conclusion


If I'm honest, none of these are terrible cameras. Even though the D3200 is the noisiest of these four at high ISO, it's still better than if you are coming from an older camera like the D80 or the D90. Though it's early, I think it's reasonable to say that the new Toshiba-sourced sensor in the Nikon D5200 is competitive. I am comfortable in saying that it produces visually more pleasing output than the Sony NEX-7, which was state of the art for APS-C sensors in 2012.

So if preliminary output is anything, the D5200 is more than just an obligatory update. (Ken Rockwell: "I wouldn't pay $800 for a D5200 when I can get the pretty much identical D5100 for half the price. I don't see anything significant to make it worthwhile to throw more money at the newer D5200.) Here's how to sum up: the D5200 has the D7000's sophisticated AF capabilities, the added flexibility of the flip-out screen, and Nikon's best APS-C image quality (assuming that dynamic range makes the same strides as did resolution). I would say that given these qualities, the D5200 has had a very understated rollout. More enticingly, if this is sensor is also used in the D7000 and D300s replacements (D7100? D7200? D9000?), I think advanced users will be satisfied that they will be getting more resolution without much sacrifice in image quality.


2 comments:

  1. Great review, I'm currently deciding to go with the D5200 or the D7000 and noise from the iso plays a big part in my photography as I particularly focus on night time photography. A nice read!

    ReplyDelete
  2. It's really hard to go wrong with either, but I'm finding that the colours with the newer D5200/D7100 are a little truer to life / less plastic in the higher ISO range. The difference, is subtle though.

    ReplyDelete