Yeah, I know all that. And you can't compare the JPEG's generated by the cameras, because most people who want quality already know to shoot RAW and have their computer comvert to JPEG. Comparing in-camera generated JPEG's is not even close to comparing apples to apples if you want to gauge how the new sensor does at differen't ISO's, as the new Digic 5+ processor is some 30X more powerful than the one in the 5DII, and thus has time to do good conversions of the RAW data.
Sorry, I should have been more positive. You are totally right that RAW is the way to make real comparisons.
The truth that I am trying to communicate is that the sharpness of "RAW" images when posted as JPEG images really depends on the processing done to them. The ultimate maximum level of sharpness depends on the sensor and AA filter of course, but I have seen scores of messed-up comparisons where a "better/sharper" camera's RAW images were just over-sharpened, and then compared to under-sharpened images from another camera.
The classic example of dishonesty with sharpening RAW images is to "prove" to newbies that they should "always shoot RAW." They give super sharp screen images produced by Lightroom from RAW images that are sharpened for viewing on a monitor, side-by-side with unsharpened (or even blurred) JPEG images from the camera at the lowest sharpening level. And then they crow about their conclusion that shooting RAW is the only pure way to take pictures, when it was all cooked.
And the very first word of my post was that noise isn't about sharpness.
I'm not sure what point you are trying to get across, anyway. I am defending the 5D3 based on accurate testing done by DP Review, and you are bringing up an irrelevant controversy about whether experienced shooters know how to shoot RAW vs JPEG and the power of the new Digic 5+ processor.