Originally posted by boardhead If so, try scaling the K-5 image to the same size as the K10D image and compare them again. This is the fair way to compare the two cameras.
I know, and I took that into consideration, mostly looking at shots scaled to my screen (1680x1050). I didn't really have an image that's 100% comparable (I didn't find anything with quite as difficult light as yours, I think), but I tried to look at a few ISO 1250 shots taken at comparable EV values, and while the K10D can have some nasty VPN if you look in the shadows, I think at least half-dark skin tones from the K10D at ISO1250 look slightly better. It's hard to tell, though, because I don't really have many above ISO 800 shots taken at such high EV values.
For lower light than in your shot, I suspect the K-5 may be even better than ISO 800 from the K10D.