DXOmark lists these cameras as having better DR than the Canon 5d:
D80
alpha 100
canon g12
40d
s95
d40x
k200d
km
d60
50d
d200
500d/t1i rebel
60d
k10d
gh1
d300
d90
I've had personal experience with some of these, and there's no way that they even get close to 5d DR in real life. Some - in particular the canon g12 powershot and s95, and also the older dslr models like the d80, k10d, d200, 40d, d60, d40x & alpha 100 are just laughable...
It shows them as having higher DR.... wait for it... at iso 100. Even at ISO 200 the pocket cameras already fell behind, and by 800 the rest (save the d90) trail behind as well. This is possible and believable with newer sensors. I shoot with a 5D, trust me. I love the sensor and wouldn't trade it for any of the above cameras purely because of how flat the DR curve is as the ISO goes higher. I still trust their measurements. I had a 40D before the 5D, and I found it to compare to the 5D exactly how their curve displays. similar at iso 100, less at 800 and 1600. This went from "I've experienced the inaccuracy in real life" to "there is no way these inferior cameras are as good as mine". That's the bias that makes me distrust the other review sites, and trust DXOmark's measurements via machine, cold, calculating, uncaring, brand agnostic, machines.
DR is defined by two limits - saturation (one of the color channels at a value of 255 in an 8-bit file) at the upper end and a certain - excessive - S/N value at the lower end (this is where digital noise overrides any available shadow detail). So, DR is essentially is dependent on the definition of the acceptable S/N ratio at the lower end. This also explains the various and contradicting DR values one can read in different reviews.
Apparently you don't mind noise in the shadows - this makes me think you have little experience with digital cameras: Digital noise looks very much different from analog grain (different distribution in location, size and structure), and once you'll be confronted with it, I'm sure you won't like it. I usually add simulated film grain (noise) in Photoshop whenever I have noisy shadows. It goes without saying that this goes at the expense of shadow detail and thus DR.
I've only shot with digital, but I've been spoiled by having a 5D. Lifting shadows is not a problem. Pixel peeping might make the noise look ugly, but in print it's fine. For example I have printed
this shot (iso 1600) at a size of 90x120 cm on canvas and the noise is inoffensive to me, and I did a fair bit of shadow lifting and highlight adjustment for maximum dynamic range to fit into the readable image.
There is no use mentioning 255 values unless speaking of jpegs. Raw files have far, far more than 255 levels per channel, closer to 64,000 values if I remember correctly. Useless difference till you start adjusting curves and exposure masks, dodging and burning, etc. But since I do lot of that, I care about DR in the raw files.
All I know is that the 5D scores well on DXOmark, and my personal experience using the files at all ISO's lines up with their score. And when comparing the files shot from other, more poorly scoring cameras, I notice their lack of DR, particularly at higher ISOs.
This is neither here nor there, as we don't know what sensor is in the X100, and even if we did, we don't know if their particular implementation is better or worse than Nikon's with the d90. I will have the camera in hand far before DXOmark gets around to testing it, but at that point I'll look at how my experience lines up with their scores. I expect it to line up quite well.
