I have discovered over the years that the best way to get super clean shadows with the 5d2 is to overexpose the image. I have not really tried it much with portraiture but in industrial, landscapes and architecture and interiors it works very well. I am quite surprised how much detail there are still left in what appears to be 'overexposed' highlights and because the shadows are 'overexposed' they are clean when I 'pull' them in post processing. More images and some thoughts why I think my 5d2 is still good enough for most applications here at..http://www.ivanmuller.co.za/blog-item/expose-right
It certainly can be good enough for a lot of things, and some technique can extend that for sure.
BUt, I got rid of mine tho, as I'd purchased it for landscape when it came out and it was the most disappointing camera I've owned.
Just did some extra test shots before I sold it recently; flat shade target, shot at 1 EV intervals from -5 to +3 as metered.
Histogram peaks line up perfectly on the grid in DPP at those levels.
Then pushed them a little in post using ACR or DPP. Don't even need to go +2 stops before I saw pattern noise on not only the -5 EV shot, but the -4 and the -3 EV shot would show FPN with as little as +1 EV push!
That, to me, is a camera with severely limited dynamic range if the appearance of any FPN is the criteria.
That was a deal-breaker for me, should have got rid of it long ago.
I did get some great shots with it, but it just didn't work the way I needed/wanted/expected it to.
I'd like to know if other 5d2s fare as poorly with the same kind of test or if mine was just a complete lemon.
6D seems considerably better, not sure if enough to appease me yet.