Dude,  I read the entire OP and I didn't get that until you posted a few posts down that you wanted peoples opinion on which they liked more rather then looked better.   I'd consider editing your OP.

I am just flabbergasted at the responses in this thread! I'm going to put it simply: Nvidia proved they have NO idea what the hell the Uncanny Valley means, while David Cage cleverly proves he does. One's a robot, the other is a human. If you are impressed with the textures and lighting in the Nvidia head you don't understand what the Uncanny means either.

If the one on the right looks more real to you then I'd definitely not let you set me up for a blind date. If you replaced his ears with pockets, his face would be a snooker table ... a point an eye:

watch this: http://www.youtube.com/watch?v=bLiX5d3rC6o

Of course you can compare both but you have to give the intention of the comparison. The PS4-demo shows especially eyes and eyebrow-movement where the nvidia-demo shows much more muscle-flexing around the mouth and the surrounding muscles which is much harder to calculate. In fact, this was the intention of nvidias demo - not to show a head which itself is mostly realistic but the muscle-flexments in realtime.

This video will explain to you what the Uncanny means: http://www.youtube.com/watch?v=OzxBpz7Xjl0

1. What exactly was your point here?  I say it looks creepy  you say it looks creepy.    Did you come up with this discovery by yourself? Congratz.

Its obvious that Nvidia showed creepy Dawn to highlight how much better the "hairless dude" is. Did I say something else? No so try not to jump to conclusions when you dont understand or please start to read the whole thing so you see the context.

And when you are at it also quote the full paragraph not the part that  helps your argument.

This is what I said:
"I just jumped on the creepy/uncanny-train ken started and said that they should use mocap like in LA Noire and refine that method. I mean the DAWN part of the Nvidia demonstration was totally cr(a/e)ppy because some person tried to immitate facial expressions and it didnt work."
(I highlighted the part you ignored)

All my comments here were made to show that mocaping faces should be standard.
(of course they need improvements etc. but its always better than 1 guy or team moving around the mouse cursor/s to position lips etc so it somewhat looks like a smile. They can do mouse cursoring for characters like Sonic etc. Not for realistic humans please.)

And to help this I said that dawn looked creepy BECAUSE some person tried to immitate emotions/facial movement and OF COURSE it didnt work because of that.  I never said  "hahahaha look at how shitty dawn looks like  nvidia n00bZ"

2.
You have some problems with  understanding what you read right?
I said " without changing emotion" where did I say there was no emotion at all?

They simply didnt really change the "old guy's" emotion because they just moved the eyebrows.   Just for your interest   your "lips"/"nose" (muscles around them) move ALOT when you change your emotions  that certainly didn't happen with the "old guy".  I see what the old guy is supposed to show but it fails at being a change of emotion.

You know its like people that ACT as if they smile and forget to move their eyes thats also not an EMOTION.

Real                       vs                      Fake


So when I see how I try to make a comment that at least has some plausible information than its more you that seems to be "incredibly bad at reading emotion. Or making a worthless comment. "

Thanks for reading.

To me it sounded like you weren't going to critisize anyone for picking the wrong one and were curious more about what people THINK is impressive graphically vs what actually is.

Though perhaps it's just I have too many conversations with Trash who goes on about the distinction a lot talking about how games like God of War 3 and Unchararted 2 often are said to have the best "Graphics" when in reality while they look great, graphically from a technical standpoint they are actually pretty unimpressive due to closed off set piece enviroments... and it's better to just say they look great.  Since there are games on the systsem that are better graphically from a pure hardware point of view.  (the above 2 being about design.)

Well what do you know he actually showed a demo of his work at GDC, keep on mind this is a shader demo more than anything. clearly elements like modling the inside of his mouth etc wasn't a priority even tho it does detract from it being fully realistic, and there are some other minor glitches especially around the eyes as well. 

180FPS on a GTX680

They need a more complete model (and maybe some more subtle acting) for it to really shine, but Jorge Jimenez' shader work is really fantastic tho it's a bit too shiny still IMO. I have been following his work for a while now, and he has really done some great stuff in rendering technology his work on MLAA and later SMAA plus his work on skin shading with Screen-Space Subsurface Scattering has already been intergrated into most major engines including Cryengine 3 and Unreal Engine 3. And his work at Activision is already looking impressive, he is mostly working on developing shaders for rendering eyes at the moment. 

Base model 

and with shaders

http://www.iryoku.com/blog