^Failed attempt at joke.

I don't see why people think that this is outside of the realm of possibilities. It seems the next logical step. I just don't know what would be done with something that reads lips like that. Maybe Kinect 2 will read your lips to see what you say when you mess up and tell you that you're a potty mouth when you curse (even if it's under your breath).

I can't believe more people aren't outraged by this claim and calling BS on it.

I thought the last Kinect update was suppose to do that ... I'll believe it when I see it. The last update hasn't impressed me tbh. Apart from being able to stand sideways I can't see much improvement?

sounds true if you ask me, although it's still kinect...

What's to be outraged about? Even if the claims about its precision are overblown, neither you or me know what Microsoft have going in their R&D labs...

reminds me of this...

"Dave, although you took very thorough precautions in the pod against my hearing you, I could see your lips move"

Sounds interesting but I'm kind of skeptical on the lip reading, unless we're talking another "only works well with English" type thing. The increased bandwidth, etc. and higher resolution though all make sense.

I mean I'd like the lip reading to be true but that seems an awfully tough challenge from the worldwide perspective if you're talking reading lips for multiple languages and dialects.

Sounds a wee bit like the initial over promise/under deliver with Kinect with videos of scanning anything you wanted in and so on.

By the sound of it, I think the lip reading is merely an example to illustrate the precision it will have. It doesn't actually say they will be doing it.

Yeah but lip reading is technically something else.  if they just mean it can resolve lips the shouldn't use the term.  I'm nitpicking on language and context use.  Lip reading means decoding and understanding words from position of lips, not being able to see them.  Assuming they can see them it should be possible I'm just doubtful about many words, languagues, etc. they could easily support at launch.

TBH it would be better example if they're talking about resolution to focus on fingers given the origianl "can it, can't it" debate around Kinect resolving individual fingers.

Ah, that way around. Just to add though, while the unit itself would be unlikely to actually read lips, I don't think it would be that complex to do, to some degree at least, if the data is of high enough precision. If you could simply detect vowels to some degree, it could be used to support voice commands in getting higher precision. Or maybe just put one person in control of who gets to pause a movie, so you avoid trolls on your couch.