It's been quite a week for users of Adobe's video tools. First, Adobe unveiled some useful new features in Premiere Pro aimed at improving overall workflow efficiency; in particular, making color grading a cinch. And now the software giant has announced plans to integrate generative AI features into its video editing software.
The new text-to-video feature, powered by Adobe's Firefly Video Model, will enable you to generate video clips directly from text prompts. This feature will be complemented by a wide array of camera controls, enabling fine-tuning of elements such as angle, motion and zoom.
The idea is to enable editors to generate B-roll footage that seamlessly integrates into their timelines. Adobe suggests you could also use text-to-video to generate elements for pre-viz, such as atmospheric overlays, or inspiration for motion graphics.
A related feature is the image-to-video tool, which can turn still images or illustrations into dynamic video clips. This could prove useful for video editors looking to add motion to static elements, or create transitions between scenes.
Extend your clips
Another AI-powered feature coming to Premiere Pro in beta is called Generative Extend. Similar to Photoshop's Generative Expand tool for image backgrounds, this tool will enable editors to automagically increase the duration of existing video footage. This could be a handy backup where a shot needs to be held longer or when B-roll is scarce.
Of course, any use of AI raises worries about commercial safety and potential copyright issues, so Adobe is making a lot of noise about its model is trained only on content that Adobe has permission to use.
This is not surprising, as there are already other apps out there that can generate video from prompts, such as Runway and Luma Dream Machine, so this seems like a good way for Adobe to distinguish itself as the choice for video industry professionals.
That doesn't, of course, deal with the separate issue of using AI to generate video of real-life actors; one that was at the heart of the 2023 Actors strike which had devastating effects on the TV and movie business.
Does Generative Expand come under that category? That might seem mean spirited if we're just talking about a few seconds of AI generation. But what if it becomes a few minutes, or a few hours, further on down the line? Actors could see it as a "thin end of the wedge" issue and studios, as ever, will need to tread carefully around this stuff.
All of this is a bit moot right now though, as Adobe has not yet announced a specific release date for these features, stating only that they'll be available "later this year."
At this unspecified point, text-to-video and image-to-video capabilities will initially be accessible through Firefly.Adobe.com, while Generative Extend will be integrated directly into Premiere Pro.
In the meantime, the company has opened a waitlist for the beta release of the Firefly Video Model. Interested professionals can sign up to receive notifications and early access.
See our rundown on the best video-editing software currently available
