This is Signal vs. Noise, a weblog by 37signals about design, business, experience, simplicity, the web, culture, and more. Established 1999 in Chicago. Visit the Product Blog for more information on our products.

So close … and yet sooo far.

What’s more amazing than this example is how much better this technology will have to get before it actually can fool a person.

I suspect we will get there pretty soon, which is scary.

Spooky.

The novelty here is the mimic part, of course—i.e. taking/analyzing something observable by video.

Good find, for all that.

And the uncanny valley strikes again…

This totally reminds me of the Johnny Cab from Total Recall.

http://www.youtube.com/watch?v=0H5k—n7sFI

Next stop: Mars.

Needs to have a mouth that can do more than whisper.

do these scientists happen to work for Cyberdyne.

Pretty impressive. They should be careful to make jokes about the robot takeovers etc though, because there are a lot of uneducated people who will assume this somehow brings us close to that without understanding the technology behind it. You can guarantee there will be some sort of backlash to this kind of research (evangelical christians?) as it makes progress.

It’s funny that they pan back at the last minute as if to illicit the response “Wait? That’s not a real human?!”

That was completely and utterly unconvincing.

The Japanese do a much better job and it’s still not even close.

Jules: A blind robot that creeps you the eff out.

Try again, nerds.

Looks like the kid from magnolia

What a grump, someone needs to take that mounted head to the park to feed the ducks. I ain’t hanging out with no rebot if it’s going to be in a bad mood all the time.

It doesn’t blink…

I always thought our robot overlords would be far more metallic when they took over.

Uncanny valley, indeed.

interesting how it’s overall so lifelike, and it’s facial expressions are very realistic, yet it’s mouth moves like a sock puppet.

Kinda disappointing in the realistic movements, given what Disney started almost 50 years ago with audio animatronics.

Is that a copy of “Getting Real” on the bookshelf?

did you know that about most of the approximately 100 billion neurons in the human brain are for computing sensation and sensor data? I guess that machine looks so creepy because it’s mostly doing something, and not receiving anything.

Uhm one thing, when I read this post there was a shutterstock ad in the right upper corner, and it matches with the content in a funny way … :-) ... http://www.travelsheep.com/wp-content/uploads/2008/11/picture-5.png

This is pretty impressive, but one of the most common mistakes when trying to mimic speech, for example in animation, is to not move the mouth wide enough.

If you stand in front of a mirror and try it – you will see that you can make your mouth as wide as your hand.

There really is a lot of squash and stretch that the bottom half of your head and jaw can do – more than most people think if they’re just working from memory or away from video reference.

Very cool stuff though.