In case you missed it, Rosebud.AI released Tokkingheads a few days ago (https://tokkingheads.com/). TL;DR you can use their GANs to animate headshots- including lip movements and eye blinks. I am fascinated by the potential for generative models as a tool for storytelling and have been playing with Tokkingheads for the past week. In a recent talk, Lisha Li, CEO of Rosebud.AI described the technology Rosebud is building as: "storytelling at the speed of thought." When I heard this it sent a chill down my spine. I've been dreaming of this future and here it is!

1 Hour Challange

Ok... I've been talking up GANs and generative models more broadly, but I haven't actually told a story with them. I decided to timebox 1 hour to give it a shot.

Steps

  1. make a quick storyboard in cavana
  2. use Artbreeder to generate my two characters, Ygritte and Ivan.
  3. Record myself saying all the lines
  4. upload the characters to TokkingHeads and video of my (poor) voice acting.
  5. animate the TokkingHeads based on my videos
  6. Stitch the animated videos together.
  7. ?
  8. $$$

1. Storyboard

Tool: https://www.canva.com/

https://s3-us-west-2.amazonaws.com/secure.notion-static.com/30dd2aeb-ee72-40ca-af3a-bf9c407761fd/Untitled.png