I am still every bit as enthused about the D3x after another week of shooting as I was when I wrote the initial post Leping quotes. One interesting fact I notice is that most of the most positive remarks I see about the D3x are from actual D3x users, while a lot of the negative comments are from folks who haven't actually used one. To be fair to the competition, I haven't shot the 1Ds mkIII extensively, nor the Alpha 900 (a few shots on each at PhotoPlus in lousy conditions at high ISO), so I cannot fairly compare my D3x to those two cameras (having seen large sample prints from the A900, it seems to have a bit more shadow noise than the D3x, but that could be the particular prints). I do have 10,000+ images on the 1Ds mkII (I switched to Nikon, rather than moving to the mkIII, in large part due to handling considerations - I have one hand, and Nikons are easier for me to use, so when Nikon made a high-resolution body, I switched), and can say with confidence that the D3x is in a completely different league with nearly twice (by eye) the total resolved detail of the mkII. 1.5x, I would expect from the increase in resolution alone, but the additional increase is from something different in the design, perhaps the low-pass filter, which Nikon has been boasting about. The only number I have on resolution is that Imaging Resource has published their test chart (although they haven't put a review out), and I can read about 3200 lines of strong detail (a little better than the Alpha and 1Ds III, which are around 2900, and a lot better than the 1Ds mkII at 2200 - which corroborates my eyeball estimate of twice the detail on paper - 3/2 the resolution in each direction would be 2.25x the total detail on a print, and 3200/2200 is a tiny bit under 3/2). I haven't yet seen any formal dynamic range numbers for the D3x, but I am seeing on the monitor and in print over a stop above what the 1Ds mkII could do in similar conditions - this would put it at least in the Alpha 900/D3 class of the best DSLRs made, possibly even half a stop better than that.
I've managed two more landscape shoots in the past week, in which the camera performed very well, as I expected. Perhaps the most exciting test I've performed was to borrow a 24x36 inch printer (HP Z3100) and make a print from a D3x image to that size. It held up beautifully to that extreme degree of enlargement, and my next major investment will be to move my maximum print size from 16x24 on a 17 inch printer up to 24x36. My figure of "roughly twice the print detail of a 1Ds mkII" comes from comparing the 24x36 inch D3x print at close range to a 16x24 1Ds mkII print and finding that the detail per square inch on the page is about the same (the D3x is actually a tiny bit better, which may come from superior tonality and DR), when the D3x is printing twice as large. Anyway, I'm as comfortable with 24x36 on the D3x as I was with 16x24 from the 1Ds mkII.
In addition to the 1Ds mkII and various lower resolution DSLRs, the other meaningful comparison I can make is to different sizes of scanned film. The D3x hugely eclipses 35mm Velvia scanned with any scanner I've ever used, but this is hardly a surprise - 6-10mp DSLRs routinely returned images that were superior to 35mm film scans, and the 1Ds mkII is certainly far superior to any 35mm scan I've ever seen. I am also quite clear that the D3x surpasses even medium-format scanned film (645 and 6x6). In pure detail, 6x9 cm Velvia scanned on a Super Coolscan 9000 is a very close comparison, but the grainless nature and superb tonality of the D3x image at ISO 100 would lead me to choose the D3x image as the better overall in almost every case. 4x5 film still has a visible advantage over the D3x, which is unsurprising given the huge difference in area.
The one caveat I have about the D3x is that, to get its full performance, you need to treat it as a 6x9 cm camera (since that is its film equivalent). Would you shoot 6x9 cm (or print 24x36 inches, no matter what the image originated on) without a tripod? The D3x sans tripod will perform at least as well as any lesser DSLR without a tripod (24.5 mp blurred by handholding is still better than 12 or 16 mp also blurred by handholding), but its full performance is achieved by handling it as a 6x9 camera - tripod-mounted, released with a cable or the self-timer, very carefully focused. Its depth of field is narrow (because the sharpest parts of the frame are so sharp, small differences in other areas are noticeable), and I will probably purchase a PC-E lens or two eventually to give myself another tool to increase DOF (the superb LCD and flexible live view will serve as a near equivalent to a ground glass).
-Dan