[Global Network Technology Comprehensive Report] On September 12, according to technology media engadget, Adobe recently announced that it will launch its latest AI video tool-Firefly later this year. This innovative tool is designed to help movie editors easily create high-quali

[Global Network Technology Comprehensive Report] On September 12, according to technology media engadget, Adobe recently announced that it will launch its latest AI video tool-firefly later this year. This innovative tool is designed to help movie editors easily create high-quality AI clips from text, still images, and existing videos with powerful shortcuts. The

firefly series of tools include text-to-video, generative extend, image-to-video and other AI functions. They will serve as complementary technologies to adobe premiere pro, providing unprecedented convenience for video editors. With these tools, editors can easily complete tasks such as navigating between shots, removing unwanted objects from a scene, smooth jump transitions, and searching for the perfect b-roll, saving a lot of time and focusing on creative exploration.

Among them, the firefly text-to-video function allows users to create AI-generated videos based on text prompts, while providing control tools for camera angles, movements, and zoom. This feature can fill in gaps in the timeline and even convert static reference images into photorealistic AI videos. Adobe says its video model is particularly good at representing natural-world videos, helping to create establishing shots or b-rolls on the fly, reducing production costs.

adobe

In the promotional video, Adobe shows several impressive examples. With detailed text prompts, firefly can generate establishing shots of a volcano erupting, a dog relaxing among wildflowers, and mini woolly monsters dancing. These high-quality outputs demonstrate the tool's huge potential for use in television, film and commercial productions.

In addition, Adobe’s image to video conversion function has also attracted much attention. The feature starts with an uploaded image, converting it into a vivid video with text prompts. In the demo, a scene of a couple walking across a forest stream was perfectly filled in with AI-generated clips, demonstrating the tool’s superior ability to fill in gaps in footage.

It’s worth noting that Adobe emphasizes that its Firefly video model is safe by design and is only trained using content that the company has permission to use. The company says its generative AI models are trained exclusively on the Adobe Stock database, which contains 400 million carefully selected images, illustrations and videos and contains no intellectual property, trademarks or identifiable characters. At the same time, Adobe also made it clear that it has never used users' works for training.

It is reported that Adobe’s firefly video model will be launched in beta later this year.