(Credit: Adobe)
Adobe is bringing three new generative AI tools to Premiere Pro, all of which are focused on meeting the needs of video editors in their everyday work rather than on their buzz factor.
The idea is to “solve real-world video-editing tasks” and to reduce the hours needed to perform them, says Adobe’s director of video editing products, Paul Saccone.
Three new generative AI features are expected by the end of the year: Object Addition, Object Removal, and Generative Extend. Adobe will tap its own Firefly AI as well as technology from Pika Labs, OpenAI (with Sora), and Runway (Gen-2).
Object Addition is just what it sounds like: Add a picture on the wall of a scene, for example. The introduced object tracks with the existing footage so it doesn’t look fake. The image above shows the tool adding a generated pile of diamonds to a case based on a text prompt.
Object Removal is also fairly self-explanatory: The example here is removing anachronistic items accidentally captured on film in a period piece (like Starbucks on Game of Thrones).
Generative Extend may be one of the more useful AI tools coming to Premiere Pro. This generates frames in your video for things like getting timing exactly right, having enough content for a transition between clips, or matching the video to a soundtrack. Below, you can see the AI-generated additions on the timeline.
At a briefing, Saccone also mentioned that AI-powered audio workflows will arrive in May 2024. These include automatic tagging and some new interface visuals.
As with all of its products of late and especially those involving AI, Adobe is making a commitment to responsible innovation, says Saccone. This includes Commercial Safety, Content Credentials, and Content Authenticity initiatives. "For us it's about solving hard workflow problems, less about flash and more about functionality," he says.