Adobe has unveiled some of its upcoming AI-generated video tools, including a new feature that can create video clips from still images. This latest preview is based on the in-development Firefly video model that the software giant demoed back in April, and is set to power AI video and audio editing features across Adobe's Creative Cloud applications.
The new teaser ad shows an image created using Firefly's text-to-video feature, which Adobe announced (but hasn't shown) earlier this year: the tool lets users create video clips with text descriptions, then adjust the results using a range of "camera controls" that simulate the appearance of camera angles, movements and shooting distances. Adobe also demonstrated the image-to-video capabilities of Firefly's video model, which can generate clips using specific reference frames, which Adobe suggests could be useful for creating additional B-roll footage or filling gaps in production schedules. If the sample footage is any indication of the final version, the quality of the generated videos is likely to be on par with what we've seen so far from OpenAI's Sora model.
Adobe is also "considering" the model as a third-party integration for Premiere, its professional video software. However, as Alexandru Costin, Adobe's vice president of generative AI, told , playback time is limited, with videos created by the text-to-video and image-to-video features lasting a maximum of five seconds. One advantage of the Adobe model over Sora is the assurance that Firefly is "commercially safe" because it's trained on publicly licensed and Adobe Stock content, which may alleviate concerns about copyright infringement. The text-to-video and image-to-video features will be available in beta as a standalone Firefly app later this year.
Adobe says the new Firefly video templates will eventually be integrated into Creative Cloud, Experience Cloud and Adobe Express apps. The company also showed off a few more clips of Premiere Pro's upcoming "Generative Extend" feature, which lets you extend the length of existing video footage, similar to Photoshop's generative expand tool for background images. Adobe says this will also happen at an unspecified date "later this year."