Friday, February 8, 2013

More Nikon D5200 Samples: DPReview... err... Review



DPReview has their sample images up for the Nikon D5200. These are probably the best so far with this camera, as DPReview is usually the most thorough and standardized out of all of the testers. Too bad they are also one of the slowest. To see the images, pull up any full camera review and go to the image comparison pages. I used the D600 review page, referencing the D7000, the Sony NEX-7 and the Nikon D3200 and the D600.

So this is what I'm seeing with regards to the Nikon D5200 RAW samples:

  • One stop behind the D600 noise-wise. In other words, the D5200 sensor is technologically on the same generational level as the D600, with the largest difference in image noise due to the size of the sensor and implied light gathering area. Actually, to achieve this, the D5200 would have to be slightly more advanced than the D600, as the more closely packed pixels on the DX format would lead to more read and dark noise and dark current. (And in fact, it is, as Sony, until now, does not use copper fabs like the Toshiba sensor in the D5200)
  • Better at all ISO's than the Sony NEX-7 sensor. Given the amount of time between these two cameras, it would be a disappointment if there was no difference.
  • Equal resolution to the D3100 throughout the ISO range, but holds on to colour saturation better as ISO rises.
  • Better rendition of fine detail than the D7000. This didn't come through on many of the other sample images by other sites, and goes to show how important setup is when trying to create comparable test images. The take home message is that if you are not exercising good shot discipline, then the D5200 will make images that are the same as the D5100/D7000.... only bigger.
  • Comparable dynamic range to the D7000 in real world terms. DxOMark actually ranked the D5200 higher than the D7000, but DxO's numbers are hard to translate into meaningful real-world differences. I think that side-by-side, you wouldn't notice the difference in most circumstances.
  • RAW samples show much less blotchy reds than the D7000 and D3200. The noise in the D5200 reds has a similar grain to that in the blues and greens. The D7000 red noise has a larger grain.  In other words, my guess is that the D7000's noise suppression system has to "work harder" with the reds when rendering out of camera JPEG's/
  •  Both are bested by the D600. This intrigues me the most about the D5200 sensor, can't wait to see more real life example of what if can do with reds, and if it makes for good skin tones as well.
 
This is the best set of sample images yet. Imaging-Resource come second, but their comparometer only shows differences in JPEG output, and from what I can tell of their D5200 samples, their traditional still-life smorgasbord is not quite optimally focused. If you only had the Imaging-Resource samples to look at, you'd be hard pressed to see any meaningful advantage the D5200 would have over the D5100/D7000. This is not the case with the DPReview samples; the advantages of the D5200 do come through, though at times they can be subtle. 

No comments:

Post a Comment