What I've denied is your claims that Canon's AF system in the 5D III cannot do certain things. I tested those things. My 5D III performed fine doing AF with the 16-35mm f/2.8 with close, moving subjects. It's not 100% perfectly accurate, but I'm sorry, I don't believe for a moment that the Nikon system is 100% perfectly accurate either.
But that's not what I claimed. I said that, especially in comparison to Nikon's 3D focus tracking
, it's ultimately so unreliable as to be practically useless for my shallow DOF wedding photography using fast 24/1.4 and 35/1.4 primes. Because I could not trust it. It gets confused too easily, and this'd come as no surprise if you performed my little thought experiment on how the algorithm actually works. And I'm *not* alone in my opinion. Sure, rewind 10 years, and the 5D III's subject tracking would've been the best thing since sliced bread. It's a very cute, clever algorithm. And I think it still works quite well for sports. It just hasn't kept up with what the metering sensors in the 1D X and Nikon's can do, or what the imaging sensors can do in mirrorless cameras now (although without phase detection, the latter are still lagging IMHO). Arguing against this is literally arguing against the benefits subject recognition via a sensor bring. And, again, you'd know this if you just picked up a D810 & put it in 3D tracking and waved the camera around a bit. It sticks to subjects almost as well as the Sony AF method you praise.
As for the lens stuff, there's too much to respond to, but generally my entire point of those examples was this: I pulled a few of the lenses I own to show you that your statement that generally
Canon lenses are better clearly requires more proof
. Old lens design, new lens design - I picked both, you're still going to find something to complain about. But my point stands - you can't just go and say that generally a 5D3+Canon lenses will outresolve a Nikon D800E+Nikon lenses. That statement needs proof, b/c I easily pulled up 4 examples proving otherwise. And it's ridiculous to suggest anyway - that a sensor with literally 50% more resolution is somehow going to perform worse when paired with similar, high-end lenses.
Also, I wasn't referring to the overall DxO score; I was referring to the P-MPix score, which takes into account lens MTF & sensor resolution and characteristics.
In case you haven't sensed a them here yet over the years - I'm fighting your general
claims. Like Canon high ISO is better (glad we've dispensed with that finally!), Canon lenses are better, Canon sensors have more DR at higher ISO, Canon lenses + a 22MP sensor somehow still generally outresolve Nikon lenses with a 36MP sensor, etc. etc.
These claims are misleading.
Also, I'm not a Nikon fan. I'm a fan of some of the technologies
Nikon uses, as well as some of the technologies
in the Canon system (dual cross-type, wider baseline, wireless flash, DPAF). I'm a technology
proponent. Therefore, I cringe at false generalizations that ignore the realities of better, and worse, technologies