http://www.flickr.com/photos/28177041@N03/5152110098/sizes/z/in/photostream/ (For explanations and shooting/conversion details, please look here.)
(Click to Enlarge in a New Tab/Window.)
It can be noted that for the K-5 picture, whilst virtually has no shadow noise, it has more obvious purple fringing across the high contrast edge (see the left cheek of the boy's face). Do note that the same lens was used for both cameras and test shots and the shooting conditions were the same.
Besides, whilst the K-5 has little shadow noise, it just can't give more highlight headroom than the K-7. See the right face of the boy, it can be yet noted that highlights are clipped for both images.
With the low shadow noise of the K-5 and sensor, there is a trick to play, though, if the user just wants to have more highlight and details retained. That is, to shoot in RAW, intentionally underexpose the image and then apply +ve digital compensation to bright up the whole image. Whilst for doing this, shadow noise will be increased, highlights could be preserved better.
Last but not least, as for the skin tone colour response and textural feel, I think K-7 is still doing better for this shootout and samples. Overall, I found K-7's picture renders somehow more natural skin tones and has less plastic skin feel. Furthermore, the K-7 picture could look sharper, too, putting aside the annoying noise that is superimposed on the image.