Reminder to everyone about to throw a fit because something is better/worse/exactly what they thought it would be:
- Look at the total measurements for the whole range of the camera, not just the overall score or the 'best case scenario' scores.
- Colour depth really does not matter.
- ISO sensitivity matters a lot, yet is hardly ever brought up by anyone.
- The processor makes almost as much difference as the sensor, so don't talk about just the sensor as if it's the only thing which matters.
- 99% of the time absolutely any camera made in the last ~10 years will end up giving you essentially the same results, with only a few specific, specialised purposes really demanding more from any one aspect of the body.
- If you really care about getting the cleanest images and the most dynamic range, you'd probably already be used to compensating with image averaging, HDR, and other similar techniques.
With that in mind, the only part of the 6D2's measurements which is at all surprising is how low the true sensitivity is. ISO 100 is actually 63 and ISO 200 is actually 127. That's a record-low for Canon, and I believe a record-low for any SLR. That means that using like-for-like settings, the 6D2 underexposes by nearly two thirds of a stop. [Is this a fact? Can you prove it?] To compensate and get the same exposure you'll need to either use a faster lens/open up (potentially dropping in sharpness and increasing vignetting), lower your shutter speed (yay motion blur) or raise the ISO (more noise and less exposure latitude to play with).
That's a huge problem. Yes, the 80D also severely overrates its ISO, but that wasn't quite this bad. Yes, Fuji do it too, but again, not by this much. Nikon and Sony haven't, as far as I'm aware, ever fudged their ISO sensitivity by any more than half a stop. (Though even half a stop is too much to get away with, in my view.)
This means that, if we assume you don't want to change your aperture or shutter settings and only raise the ISO to compensate for the 6D2's low sensitivity, you won't really get any cleaner images than you would with a 5D mark 2 at a 1:1 pixel level. At 1:1, the 6D2 is only 1/3rd of a stop cleaner... but it shoots 1/3rd of a stop darker. So if you have to raise the ISO by a third of a stop anyway, you end up at the exact same place.
Now, the 6D2 does of course have slightly higher resolution, which helps it along. It has about 2/3rd of a stop lead when the 5D2 and 6D2 files are scaled equally, which after compensation, would still leave it 1/3rd of a stop cleaner. But 1/3rd of a stop after 9 years is pretty awful.
Not only that but if you throw in a 5D3, which is also 1/3rd of a stop cleaner once the images have been equalised, you're matching the 6D2. Now, that 5D3/6D2 image quality is far more than good enough for most uses (so is the 5D2's, really), but that's still very poor considering the time between releases and also the price of a second hand 5D2 or 3 vs a brand new 6D2.
Dynamic range is the exact same deal, although the 5D2 actually leads very slightly at the lower ISO settings. The 6D2 matches them in range but not in exposure, so once you've increased sensitivity of the 6D2 to match the 5D2/3, it's fallen behind them. And again, the higher pixel count helps when you compare the equalised measurements, but not enough to actually put it ahead. The only variance is at ISO 3200 and above, where the 5D2 drops off sharply. (The 5D3 remains very slightly ahead of the 6D2 once you compensate for the 6D2's lower-than-advertised sensitivity.)
For people who want long exposures matching the lower ISO settings of other cameras which Canon usually lacks (e.g. Nikon usually has a ISO 64 setting which isn't just ISO 100 pulled down, like it is with Canon's extended range ISO) settings, you could see this as a benefit; buy a 6D2 and set it to ISO 100 and you're essentially at 64. But if you're not one of those few people trying to get the absolute longest shutter you possibly can, this is a big problem.
TL;DR version: What you're looking at with the 6D2's fundamental image quality is a 5D2 with more potential for cropping, assuming you want your final images to also be of the same brightness. The 6D2 is only better than the 5D2 if you open up the lens or shutter by half a stop or so and do not crop (or any other action which would throw away pixels) so the higher resolution can work in the 6D2's favour once you've scaled the image down or printed it at its final viewing size.
Again, if any of that bothers you, re-read the very first part of my post. Also consider that the 6D2 has far better functionality over the older 5D2 and 3. Considering how most photos these days end up only being seen printed at relatively small sizes or scaled down to fit Instagram, I'd say the 6D2's technical image quality is more than good enough for most people and the functionality is far more important. The few people who do genuinely need better IQ (for huge printing/viewing, very heavy cropping, industry catalogue work, legitimate fine art, etc) likely already have better. (That kind of person is unlikely to have even considered a 6D2 in the first place, as they're more likely to be shooting with medium format, already have a 5DSR, etc). I highly doubt a single person who buys a 6D2 today would be bothered by the image quality in terms of the final viewed image.
The 6D line is a 'prosumer' line meant to satisfy the richer hobbiests as a primary all-round camera, or fill in nicely enough as a back-up for professionals. In that regard, the 6D2 seems to do fine. It's horrible that Canon is flat-out lying about the ISO settings, and they should be taken to task for that, but in terms of the end results it can be compensated for easily enough, and the other features of the camera make it more than worthwhile for the kinds of user it is aimed at.
