A new feature of Adobe’s generative AI video tools, which creates video clips from still photos, has been teased. The software behemoth’s April demonstration of the in-development Firefly video model, which will support AI video and audio editing features across Adobe’s Creative Cloud apps, is expanded upon in this most recent peek.
The newly released promotional trailer features a film created with Adobe’s Firefly text-to-video capabilities, which were disclosed earlier this year but not shown in action. Using a range of “camera controls” that mimic camera angles, motion, and shooting distance, users can create video clips with the tool and modify the output.
Also Read: Adobe Launched Adobe Express Mobile App for Content Creation for Free
For the Firefly video model, an image-to-video functionality that can produce clips using particular reference photos was also presented. According to Adobe, this might help fill in any gaps in the production schedule or create more B-roll films.
The resulting video quality appears comparable to what we’ve seen thus far from OpenAI’s Sora model, which Adobe is also “exploring” as a third-party integration for its Premiere Pro video suite if the example footage is any indicator of the ultimate release.
However, Alexandru Costin, VP of generative AI at Adobe, revealed to The Verge that the text-to-video and image-to-video tools can only create videos that are up to five seconds long.
The claim that Firefly is “commercially safe” since it was trained on publicly available, openly licensed, and Adobe Stock content gives Adobe’s model an advantage over Sora and may allay some worries about copyright infringement.
Also Read: What is Adobe’s AI Premiere Pro for Video and Sound Editing?
Later this year, the text-to-video and image-to-video functionalities will both be made available in beta as stand-alone Firefly applications. Over time, Adobe promises to include the new Firefly video paradigm into its Creative Cloud, Experience Cloud, and Adobe Express software.
Additionally, the firm unveiled a few more previews of the impending “Generative Extend” feature for Premiere Pro, which functions similarly to Photoshop’s Generative Expand tool for picture backgrounds in that it may expand the duration of pre-existing video material. According to Adobe, this will also be available “later this year” at an unannounced date.
Refer to the official blog for more information.
Also Read: What is Adobe Firefly AI Integration with Substance 3D for Creation and Product Visualization?
