Texture Mapping: First Voices, Then Faces

Let me explain by way of example.

Imagine it's 2015. The latest craze is the new Original Star Trek TV series on TV. That's right, the new original series. Meaning: Kirk, Spock, Bones, Scotty, the whole original crew, of the original Enterprise, back as if for Year 4 of the original series.

Why not? Think about what you'd need: obviously a set of really good scripts. Ok, let's say you have that. Technically you would need to be able to create digital characters that looked, acted, moved, and spoke just like their counterpart actors from the 1960s show.

Imagine software that does texture mapping --- not just of any old textures onto 3D shapes, tools we have plenty of now ---- but faces and bodies and clothes, digitized right out of the original series episodes, and wrapped around 3D representations of William Shatner, Leonard Nimoy, etc. Then, take the voice recordings of these people, and apply their vocal patterns --- the characteristics of the way they speak and their vocal waveforms, and wrap them around a generic vocal reading the lines of these newfangled 2015-era Star Trek shows.

If you're still following me: imagine some teenager geek somewhere writes, produces, directs, and acts the entire series, through a powerful set of these tools. I think this is coming. And it won't just be Star Trek. It'll be The Beatles. Clint Eastwood. Mike Myers. You name it. Any popcult phenomenon. As long as we have original footage and original audio of the person, why not create a digital counterpart that can act, speak, walk, whatever, through whatever situations, stories, etc you can imagine.

So anyway, it's something I think is coming.

And then I saw this story in today's New York Times.

Seems like I might have to revise the 2015 date downward...

Posted by brian at November 24, 2003 11:41 PM