Friday, September 9, 2016

Computer Generated Imagery Is Crossing the Uncanny Valley

Here's an example of the uncanny valley we've been using as an example since 2014:

Seinfeld, Virtual Reality and Mild Revulsion
The Uncanny Valley, Interior-Design Edition

Greg Miller
The "uncanny valley" usually applies to human aesthetics. It describes that vague sense of revulsion you get when you see a fabricated person—a robot, usually—who looks aaaaalmost human … but not quite....
You know what the image is supposed to represent and you know it is not that place.
On the other hand...

From The Next Web:
CGI is moving fast, very fast. Not just for added effects in movies, but also for creating completely lifelike 3D models.

The biggest problem facing the animation community is an effect called the ‘uncanny valley’. We’re confronted with this when we see something that looks nearly human, but just not quite.

Humans are so good at recognizing other humans, that’s it’s extremely hard to create a thing that looks like a normal, healthy person and not a zombie or a corpse.

It’s explained in the graph below:

We’ve become pretty good at making CGI look like it’s almost real, but the real challenge is making it seem like its normal.

Japanese artists Teruyuki Ishikawa & Yuka Ishikawa — otherwise known as Telyuka — started a project in 2015 to create an extremely realistic computer-generated schoolgirl. Her name is Saya, and she has been improved on since then.

This is the 2016 version (click to enlarge):
And these are some pictures of the 2015 version (again, click to enlarge):....MORE
But beyond even the computer generated imagery is this guy:

This ain't your dad's animatronic... dad
At Atlas Obscura.