There are two recent pieces that give a somewhat strange verdict on the noise levels between Sony cameras.
The first is the dreaded DxO and their comparison of the NEX-5n, NEX-7 and a77 cameras. They have a much better result for the NEX-7 as compared to the a77. This according to Sony is the same sensor, so how can it have a "worse" reading? This is skewed by the high ISO results. The comparison between those two cameras and the NEX-5n is here:- http://www.dxomark.com/index.php/Cameras/Compare-Camera-Sensors/Compare-cameras-side-by-side/%28appareil1%29/734|0/%28brand%29/Sony/%28appareil2%29/736|0/%28brand2%29/Sony/%28appareil3%29/737|0/%28brand3%29/Sony
The other is a piece by Michael Reichmann at Luminous Landscape on the NEX-7 compared to the NEX-5n in terms of noise. http://www.luminous-landscape.com/reviews/cameras/sony_nex_7_rolling_review.shtml#28
Despite providing evidence to the contrary this article asserts:- "But, and it's a very big but, now go to the top of the samples where
the NEX-7 files have been downsampled to the same size as the NEX-5n.
What you'll see, or at least what I and several other photographers to
whom I've shown these agree on, is that the differences are reduced to a
quibble, even at 100% on screen. In prints or any non-pixel-peeping use
of the images, there is really little to choose between them, except
maybe at ISO 12,800 where the 5n has a slight advantage on screen, if
not in real-world prints. Like I said – a quibble."
I don't have a NEX-7 to compare but I do have an a77 so I decided to do my own tests to see what was going on here.
The following is a comparison between the a77 and NEX-5n using the same lens, a Sony a-mount 30mm f/2.8 macro. Since I attached the lens to the NEX-5n using the translucent mirror adapter, this should take the effect of the mirror out of the equation. Many are currently speculating that the light loss from the mirror has affected the DxO results. We however have no confirmation of this since DxO are.... well.... DxO!!!
By clicking on this file through to flickr then Actions > View all sizes > Original you can see a large high-res version of this file. I also downsized the a77 files to the same size as the NEX-5n files, as Michael Reichmann did.
N.B. The a77 images were all taken using the very latest firmware v1.04.
N.B. The a77 images were all taken using the very latest firmware v1.04.
As ever I processed these files from raw using the default setting in Adobe Camera Raw in Photoshop. Processing for each file was identical.
The first thing that becomes immediately obvious is that the a77 has a "cooler" colour balance. This isn't particularly significant as it can be very easily altered in processing. In terms of noise the a77 files are worse than the 5n equivalents, and by the time it gets to ISO 3200 and above the a77 images start to look quite unpleasant. Incidentally the jpg files from both these cameras are a lot better, but I thought a raw comparison would be more useful to see how they compare.
I followed this up by doing a comparison between a77 and NEX-5 files when the NEX-5n images are upsized to the same size as the a77 files. These are full-size files and are pretty huge. You can access them in the same way as before. Click on them > flickr > Actions > View all sizes > Original. Because of the size, I haven't done an all in one comparison but uploaded 4 separate files at particular ISO points.
ISO 100
ISO 800
ISO 3200
ISO 12800
Again the a77 files are noisier. They are sharper, but that tends to disappear when you apply some noise reduction to the files.
So what does this prove? Well as ever I'm going to say what I think on the results from the cameras I have at my disposal. The NEX-5n I have produces "cleaner" files from raw than the a77 I have, when the files from the a77 are downsized to the same size as the 5n I have, and also when the 5n files are upsized to the same size as the a77. I obviously cannot make any comments on how this will apply to a NEX-7, though I might speculate a bit!!
I'll post a jpg comparison soon.