A friend of mine just received his D800 and offered me to test it against my 5DmkII.
I'm very interested in this Nikon camera and although I prefer Canon glass (I used Nikon for many years, in film days - F5 - and digital -D3) I also use Zeiss glass which is available in both mounts so I could keep the same fingerprint on some lenses.

For the test I used my 5D2 with the ZE 50 f/2 Makro Planar, since my dealer kindly lent me a used 50 Makro-Planar in ZF.2 mount so I could compare both cameras with identical glass… well pretty much at least, identical would have meant using the ZF 50 on my 5D2, but I don't have any Nikon to Eos adapter at the moment so this is the best I could do to mitigate the "glass factor". Incidentally that sample of ZF.2 Makro-Planar seemed ever so slightly softer than my ZE version, at least in the corners.

I have to say two things about the D800 focussing: the standard screen isn't very precise to judge focus, I had much more keepers with the EG-s screen on my 5D2 (and I'd say even with the standard screen), but it seems the D800 does indeed have interchangeable focussing screens, although there's no mention in the manual.
Live View is also poor implemented, at least compared to Canon's. Now you can shoot with the mirror-locked up but the vision is very fuzzy at high magnification, contrary to Canon smooth image and it is very noisy in the dark while my 5D2 is relatively clean even in at night. I shot two images at an exposure of 10 sec f/4 ISO 100 and I somewhat missed the focus with LV on the D800 because I couldn't see well the branches I intended to focussing on. No problem with the 5D2 instead.
And incidentally the 5D2 rendered a cleaner final image at that (relatively) long exposure, which is odd to me giving the overall cleaner shadows of the D800 at any ISO.
[correction: the D800 long exposure image is slightly misfocussed but as far as cleanness it is better than the 5D2].

A system though is not only about sensor, but also (if not mostly) about glass.
So how does a so-so lens on a great sensor compare with an excellent lens on a "lesser" sensor?

For this "lens or Mp" test I compared the Nikkor 28-300 ED VR on the D800 vs the 90 TS-E on the 5D2. Both lenses set at f/8 for maximum sharpness before diffraction limited D800 performance (although I discovered later that even at f/16 images from the Zeiss 50 are quite good with only a minor sharpness&contrast reduction and for DOF reasons I wouldn't think twice to shoot at that aperture on the D800, diffraction seems indeed an overrated issue).
Keep in mind that the 28-300 at 90mm has f/4 max aperture so at f/8 is already stopped down 2 stops and it doesn't sharpen more past that.
I threw in also a shot with a 90 Macro-Elmarit on my Leica M9, just for the sake of comparison.
As you can see a sharp 18-21 Mp is better than a so-so 36 Mp.
Nothing new here but I thought it helps to keep things in perspective.

Please forgive the poor uninteresting subjects, I hoped to shoot some sceneries of the Garda lake but sadly the weather didn't help during the weekend (I had the camera for only one day and a half), mostly raining and some wind, so you are stuck with my garden and kitchen. My wife was sick so I had also to attend my 5 months little girl, which didn't make for a relaxing test.

All images shot on tripod with Live View and self timer (which on both cameras in LV means mirror lock-up, even Nikon now finally), manual focus and no VR/IS.
Raws converted via ACR with standard settings (only set WB and sync between NEF and CR2), then saved in Tiff 16 bit ProPhoto RGB, opened in PS6 beta to do the screenshots on my calibrated MacBook Pro 17". I know it is not the best way to show images, sorry but I don't have the time to crop, save in jpeg and present the images for comparison so I used the screenshot method and let the Raws available at request.

I just noticed that I mistakenly set my 5D2 date to 2013… but I swear I didn't shoot the images next year ;)
My bad also about differences in exposures… some of the outdoors images have different exposure, +0.5 to +0.7 stops for the D800 so keep this in mind when judging noise. The interior shot though was shot at 50 ISO for the D800 and 100 ISO for the 5D2. The histogram was the same with identical apertures and shutter speeds so maybe the D800 is much more sensitive or the ISO label for Canon is different than Nikon's… either way Nikon images were always more open in the shadows and with a higher "luminosity" appereance.


RESULTS-COMMENTS

The D800 is simply amazing. Resolution and DR are excellent. Similar noise at high ISO (but better color retention) compared to the Canon, but once you print or look at the same size on monitor the D800 is cleaner (and sharper).
Anyway the old good 5D2 is still a great camera and the M9, the lower MP one, is actually difficult to beat because Leica lenses are usually fantastic and the lack of AA filter + CCD sensor that allows to push shadows quite a bit at lower ISOs, all make for an amazing IQ, slightly better than Canon and a match for Nikon as far as DR (up to about 800 ISO that is). Sadly I never liked rangefinder systems for lack of precise focussing and framing (I know it's me, but I also tried to use the M9 exclusively for some months to see if I can master it, no luck… the shooting experience is very poor for my taste). Even for this test I had to shoot many frames with a slight focus shift and pick the better shot later on computer…
I know I should sell it but the portability and image quality (especially wide open) are absolutely unique. Maybe I'll sell the M9 and keep the lenses waiting and hoping for some sort of live view in the next M10.

But to me the D800 is not enough to push me back into Nikon camp.
I prefer Canon lenses offering (and IQ in their latest lenses is amazing) and 5D ergonomics (both personal things of course).
Live View is quite a bit cleaner on the Canon (although to be fair I didn't misfocus a single shot [save the night shot] on the D800 because of the poorer LV image).
The easier manual focussing with my EG-s screen is also something to value more than 20% linear resolution and cleaner shadows. I prefer to have a higher keeper ratio at 21 Mp than much less at 36Mp.

In the end I really wish I had the D800 sensor in my 5DmkII, but since it's only a wild dream I'll enjoy my Canon system while waiting for the next higher Mp camera.
I hope it won't be only a match in pixel count like my Mp is bigger than your Mp… but mainly a step-up in sensor quality at lower ISOs (DR in primis).


______________


***** 1.6 Gb of Raw files are availble at the link below until May 3d *****

RAW FILES
______________


PLEASE AFTER CLICKING ON EACH THUMBNAIL SELECT "ORIGINAL" AT THE BOTTOM TO SEE THE IMAGES AT ACTUAL SIZE