Adobe is joining several other players in the generative AI (GAI) space by rolling out its own model. The Firefly Video Model is powering a number of features across the company’s wide array of apps. At Adobe MAX, the company announced that some of those are available in beta today.
Generative Extend is a Premiere Pro feature that Adobe previewed earlier this year. It enables editors to add generated footage and audio to the start or end of a clip. This can help cover a transition or edit if the required shot isn’t available. All editors need to do is drag out the beginning or end of a clip. Adobe says the tool can also correct eyelines and actions that change unexpectedly in the middle of a shot.
Adobe said it worked with professional video editors over the last year to better understand how GAI could help resolve some issues in their workflow. While Generative Extend might give them the footage they need, other creatives may be less enthused. It may mean that reshoots are no longer required, taking days of work (and income) away from the cast and crew.
That said, Generative Extend is only intended for small tweaks. The feature is also limited to a maximum resolution of 1080p for now, so it’s not exactly cinema quality.
In September, Adobe previewed its text-to-video (similar to OpenAI’s Sora and Meta’s Movie Gen) and image-to-video features. Those are now available in the Firefly web app in beta, thought you may have to join a waitlist.
Adobe says that, like with other Firefly generative models, both the Firefly Video Model and the features it powers are designed to be safe for commercial use. Its Content Credentials watermarks are applied to whatever the video model outputs.
Elsewhere, Photoshop is gaining some features Adobe showed off earlier this year, with Generative Fill and Generative Expand now running on the latest Firefly Image Model (the company says this can generate images four times faster than previous versions). The Generate Similar tool is fairly self-explanatory — it can generate variants of an object in the image until you find one you prefer. Adobe released those features today.
In addition, Adobe is adding a neat feature to the Remove tool, which lets you delete people and objects from an image with ease, like Google’s Magic Eraser. With Distraction Removal, you can remove certain common elements with a single click. For instance, it can scrub unwanted wires and cables, and remove tourists from your travel photos.
Meanwhile, Frame.io V4, said to be the biggest update to the collaborative photo and video production platform since it debuted nine years ago, is available to all users. Adobe has entirely redesigned it to, for instance, improve workflows and upgrade the video player. It also emerged that Canon, Nikon and Leica will support its Camera to Cloud (C2C) feature, which allows for direct uploads of photos and videos to Frame.io. Adobe says that means most major manufacturers are now supporting C2C.
This article originally appeared on Engadget at https://www.engadget.com/ai/adobe-starts-rolling-out-generative-ai-video-tools-in-beta-163703601.html?src=rss