I think this is right here demonstrates exactly why DXO has something to offer. Your opinion here is simply that...your opinion. If you actually held a poll about those two photographs, I would be willing to bet that you would NOT get a 90%/10% ratio, where most people could tell just by looking at those photographs which was which. I would bet such a poll would end up closer to a 60%/40% ratio.
It's interesting how different people see the same set of images differently. It's funny that you bring up a "poll," because I did just that after these shots were taken. I sent these two images out to a half-dozen colleagues of mine to see if they were able to determine which image came from each camera. Every single one of them correctly identified which shot came from which body, and unanimously agreed that the image shot with the 5D looked substantially better. Keep in mind theses are professional automotive photographers, each with decades of experience, that routinely scrutinize images like this. So you're right. The ratio wasn't 90/10. It was more like 100% of respondents that agreed with my assessment
In fairness, viewing shrunk down images on a message board do hide some of the obvious differences. If you're really bored, I'd be happy to email you both jpegs, but I ain't sending the damn raws
On a somewhat related note, clients generally have no idea what kind of equipment their contributors shoot with. All they know is the quality of the images you submit. That said, my editor immediately complained about how the images shot with the 7D lacked contrast and shadow detail with blown-out highlights. This despite my best efforts to address these issues in post.
That wouldn't be entirely because how each of us sees is subjective, but also due to the differences in computer screens, computer screen calibrations, etc. To me, those photos look relatively similar, however I have a calibrated screen tuned for post-processing photographs for final print. Because I print and judge my print qualities from how things look on-screen, the blacks in those photos look pretty even-keel. I'd be willing to bet, however, that one of them would indeed stand out as having "better" blacks if I viewed it with the screens I have at work, as they are calibrated for an entirely different purpose, and are a bit lower contrast (which would enhance shadow details.)
The difference is obvious on a $h!tty monitor as well
There is also the simple point that we don't know for sure how dark the deepest shadows are in the leaves of the trees of the 5DC shot. They may look "better" simply because they are not nearly as deep as the ones the 7D had to work with. That may be the case with all the 5DC shadows, where as the 7D may have had to deal with deeper shadows everywhere. You can't really make an objective comparison with two entirely different shots like that...you don't know for sure exactly how the shadows of each shot compare. You need a consistent, calibrated photographic source to properly measure the differences (even if they are "useless differences"), and that would be an area where DXO excels.
You make a good point. This is by no means a scientific test, and it would never stand up in a lab. However, it was never meant to be a scientific test. It just so turns out I had a a car to shoot, my 5D took a dump, so I busted out the 7D as a backup, using it in the same manner with the same technique in which I always shoot. You can question the difference in background lighting in the foliage between the two shots, but you weren't there
All I can tell you is that, in terms of the backgrounds, the image captured with the 5D looks MUCH more like what I saw through the viewfinder that the image captured with the 7D. There were all kinds of beautifully backlit green pine needles in the 7D's viewfinder, but none of that showed up in the captured image.
DXO mark publishes low-level measurements run through a standard set of mathematical formulas. While their numbers may seem odd, I find them valuable at times if for no other purpose than to demonstrate that physical hardware specifications make a picture not. The best example are DXO's MF camera ratings, which generally appear rather crummy compared to the latest and greatest from Sony, Nikon, and Canon. Empirically, modern-day digital MF sucks (regardless of niche.) Practically, they are still the best money can buy (by a long shot) for the niches they service.
I'm not saying this to be a smart@ss, but have you shot with multiple bodies at length in order to asses how DxOMark's ratings stand up to your own personal observations? I find that sometimes their rankings seem legit, while at others they're completely off. For instance, I shot with a 20D for a long time before moving up to a 5D. The IQ of the 7D reminded me a lot of the 20D, and sure enough, both bodies rank similarly on DxOMark ratings. I'd say their rankings of the 5D, 1DII, 1DsII, and 1DsIII seem somewhat useful when compared to my personal experiences with those bodies as well. That said, according to DxO the 20D and 7D aren't that far off IQ wise compared to the 5D, but I'd beg to differ. Your results may vary