Adobe has previewed the new Adobe Firefly Video Model, which aims to enhance the capabilities of film editors and video professionals.
The advancements are powered by Adobe Firefly - a generative AI that will assist in filling gaps in timelines and adding elements to existing footage.
This Video Model will extend the Firefly family of generative AI models, which currently includes an Image Model, Vector Model, and Design Model.
The new Firefly-powered Text to Video and Image to Video capabilities, set to be available later this year on Firefly.Adobe.com, will also be directly integrated into Premiere Pro. Premiere Pro is Adobe's industry-leading video editing software.
These advancements aim to bolster the workflows that video editors are familiar with.
Text to Video capabilities include features such as generating video from text prompts, offering various camera controls such as angle, motion, and zoom to finetune videos. Additionally, users will be able to reference images to generate B-Roll that can seamlessly fill gaps in the timeline. The Image to Video capabilities will allow users to transform still shots or illustrations into live-action clips.
Ashley Still, Senior Vice President of Creative Product Group at Adobe, said, "Building upon our foundational Firefly models for imaging, design and vector creation, our Firefly foundation video model is designed to help the professional video community unlock new possibilities, streamline workflows and support their creative ideation. We are excited to bring new levels of creative control and efficiency to video editing with Firefly-powered Generative Extend in Premiere Pro."
The Firefly Video Model is designed to support a broad spectrum of use cases. For instance, it could be used to leverage fine-grained controls like camera angle and motion to produce compelling animations and effects. It could also assist in generating inserts into key video shots with camera motion and pair video with professionally captured footage for more efficient storytelling.
Adobe emphasises that, like all Adobe Firefly generative AI models - including the Image 3 Model, Vector Model, and Design Model—the Firefly Video Model is designed to be commercially safe.
It is trained solely on content that Adobe has permission to use and does not utilise customer content.
To date, Adobe Firefly has generated over 12 billion images globally. These figures underscore the potential impact and widespread adoption of Firefly's generative AI capabilities. The forthcoming advancements aim to further extend this impact into the realm of video production.
