As expected, biases are easier to believe than evidence.
That's one reason why I like these kinds of comparisons and sample images. I can judge for myself and use my own criteria.
At the same time, it's amusing to watch some people twist and turn as they try to rationalize their biases and even more entertaining when they couch it in pseudo-scientific lingo and pretend to be the ultimate authorities on sensor noise.
My own, undoubtedly biased opinion:
The differences between full frame and APS-C sensors used to be small. Now they are approaching tiny.
The differences between sensors of different brands but of roughly the same size used to be tiny, now they are approaching miniscule.
The major difference, at least at higher ISOs, seems to be the number of pixels, regardless of brand. The new Nikon D750 seems to be about equal to the 5DIII and 6D, while to my eyes, the D810 is absolutely horrible at higher ISOs.
I was disappointed in the Fuji X-Pro 1. (The only mirrorless camera I would ever consider) Given its lower resolution, I expected it to do better. But then it's sensor is getting a little long in the tooth by APS-C standards.
Given the tiny differences and the fact that Canon is clearly holding its own in the sensor quality realm, I think it is completely logical and admirable for Canon to place more emphasis on other areas such as improved autofocus.
If your goal in life is to lift shadows by five stops and shoot dark rooms that have the exposure set for an open window, then another brand might be your better option.
As for my own personal situation, as a 5DIII owner, I'm intrigued by the 7D, but a better investment of my resources would be in either the new Tamron or Sigma 600 zooms.