I processed both sets of files in Adobe Camera RAW using identical settings: Sharpening at 70 / 0.5 / 36 / 10, Luminance and Colour Noise Reduction both set to zero, and the Process to 2012 with the Adobe Standard profile. These settings were chosen to reveal the differences in sensor quality and isolate them from in-camera processing. The high degree of sharpening with a small radius enhances the finest details without causing undesirable artefacts, while the zero noise reduction unveils what's really going on behind the scenes - as such the visible noise levels at higher ISOs will be much greater than you're used to seeing in many of my comparisons, but again it's an approach that's designed to show the actual detail that's being recorded before you start work on processing and cleaning it up if desired.

The difference between this page and the previous one is I have downsized the D800's processed RAW files from 36 to 20 Megapixels to better match the resolution of the 5D Mark III. I've discussed the pros and cons of presenting scaled versus unscaled image crops on my Nikon D800 noise down-sampled results page (for JPEGs) so I won't go over it all again here. Just suffice it to say some people like to see one type of result and some people like the other, so in the spirit if fairness and completeness, I'm presenting both in this review! So on this page it's the turn of the down-sampled RAW files, and as before I used Photoshop to perform the process using the Bicubic Sharper (Best for reduction) option.

On the previous page with 100% unscaled crops you saw how the D800 and 5D Mark III actually share very similar noise characteristics. Indeed when viewed at 1:1 with noise reduction completely turned off, the noise patterns looked almost identical. The big question here then is whether down-sampling the D800's RAW images can actually deliver a result with less noise than the Canon 5D Mark III?

Pixel peepers could stare at the images below all day and maybe persuade themselves there's a minor advantage to the Nikon, but to be honest I can hardly see any difference myself. To all intents and purposes, downsampling the RAW files without noise reduction hasn't given the D800 any perceptible advantage over the Mark III.

Of course this is with noise reduction completely turned off. If you applied the same amount of noise reduction to both original images and then downsampled the D800 image to the same resolution as the Mark III, then you may enjoy the benefit of smaller and less noticeable noise artefacts. Feel free to try it yourself with the crops from the previous page and tell me what you think in the Cameralabs forums or my social media channels, but this after six pages of results comparing the D800 against the 5D Mark III, this is as far as I'm going to go myself.

The bottom line is when shooting in JPEG with the default settings, the D800 enjoys a detail advantage at the lower ISOs, while the Mark III enjoys less visible noise artefacts at higher sensitivities - a draw if you like. But switch to RAW and turn off the noise reduction for a level playing field and both cameras deliver essentially the same amount of noise at 100%. Since the D800 does so while also capturing more detail, it's the winner at any ISO in RAW. So overall I'd say the D800 has the potential to deliver superior image quality at any ISO value, so long as you're shooting in RAW, but in JPEG there's a definite point where one model does better than the other. Ultimately this means the D800 enjoys the edge over the 5D Mark III in terms of still photo quality. For video and speed the Canon may be better, but for still photos, the Nikon wins.

That said, I hope these results pages show that both cameras are capable of delivering superb image quality, especially when shooting in RAW. For me it also reconfirms the excellent output from the latest version of Adobe Camera RAW, using the recipe described. Try it for yourself on your own RAW files: Sharpening at 70 / 0.5 / 36 / 10, Luminance and Colour Noise Reduction both set to zero, and the Process to 2012 with the Adobe Standard profile, or if you prefer, Camera Standard. Let me know what you think.

In the meantime, that's the end of my Nikon D800 vs Canon EOS 5D Mark III tests, but this section isn’t quite finished yet. The D800 so impresses in terms of resolution, it begins to tread on the toes of medium format cameras, but how close are they really in practice? To find out, I got hold of a 60 Megapixel Phase One IQ160 and compared them head-to-head - see my Nikon D800 vs medium format quality page!

Alternatively if you've already seen enough, head on over to my Nikon D800 sample images, or skip straight to my Nikon D800 verdict.

Canon EOS 5D Mark III (RAW using Adobe Camera RAW) Using EF 24-105mm f4L IS USM at 40mm f8		Nikon D800 (RAW down-sampled to 20MP) Using AF-S 24-70mm f2.8 at 40mm f8
L (50 ISO)		L (50 ISO)
100 ISO		100 ISO
200 ISO		200 ISO
400 ISO		400 ISO
800 ISO		800 ISO
1600 ISO		1600 ISO
3200 ISO		3200 ISO
6400 ISO		6400 ISO
12800 ISO		H1 (12800 ISO)
25600 ISO		H2 (25600 ISO)
H1 (51200 ISO)		51200 ISO not available
H2 (102400 ISO)		102400 ISO not available