How AI can turn your home video into a Hollywood blockbuster

Would you like to star in an animated film as an anthropomorphic animal version of yourself? Runway's AI video creation platform has a new AI tool that allows you to do just that. The new Act One feature can eliminate the need for motion capture suits and manual computer animation to match live action.

Act-One streamlines the usually lengthy process of facial animation. All you need is a video camera pointed at an actor and able to capture their face during their performance.

The AI that powers Act-One reworks the facial movements and expressions from the input video to match an animated character. Runway claims even the most nuanced emotions are visible through microexpressions, eyeliner, and other facets of the performance. Act-One can even produce dialogue scenes with multiple characters, which Runway says is difficult for most generative AI video models.

To produce one, a single actor takes on multiple roles, and the AI animates the different performances assigned to different characters in a scene as if they were talking to each other.

This is a far cry from the tedious traditional animation requirements and makes animation far more accessible to developers with limited budgets or limited technical experience. Not that it's always equal to the skills of talented animator teams with big film budgets, but the relatively low barrier to entry could give amateurs and people with limited resources the chance to play with character designs that are still realistic in portraying emotions. All without breaking the bank or missing deadlines. You can see some demonstrations below.

Animated runway

Act-One is in some ways an extension of Runway's video-to-video capability within its Gen 3 Alpha model. But while that tool uses a video and a text prompt to customize the setting, actors, or other elements, Act-One goes straight to mapping human expressions onto animated characters. It also fits with how Runway has released more features and options for its platform, such as the Gen 3 Alpha Turbo version of its model, which sacrifices some features for speed.

As with its other AI video tools, Runway also imposes some restrictions on Act-One to prevent users from misusing it or violating its terms and conditions. For example, you cannot create content featuring public figures and techniques are used to ensure that everyone whose voice is used in the final video has given their permission. The model is continuously monitored to detect any attempts to violate these or other rules.

“We are excited to see what forms of creative storytelling Act-One brings to animation and character representation. Act-One is another step forward in our goal of making previously sophisticated techniques accessible to a broader range of creators and artists,” Runway wrote in its announcement. “We look forward to seeing how artists and storytellers will use Act-One to bring their visions to life in new and exciting ways.”

Act-One may be somewhat unique among AI video generators, although Adobe Firefly and Metas MovieGen have some similar efforts in their portfolio. Runway's Act-One appears to be much easier to use than Firefly's equivalent and more available than the limited MovieGen model.

Still, there is more and more AI video competition as OpenAI's Sora model begins to spread and Stability AI, Pika, Luma Labs' Dream Machine and others release a steady stream of AI video production features. If you want to try Act-One, Runway's paid plans start at $12 per month.

You may also like...

Recent Posts

How AI can turn your home video into a Hollywood blockbuster

You may also like...

Sarah Boone's murder trial continues

“Trust the One We Call”

Will it rain on Halloween in the Bay Area? What awaits trick-or-treaters.

Recent Posts