2023 Adobe Sneaks Tease Generative AI Video Features and More Coming to Creative Cloud
The Adobe Max 2023 conference’s Sneaks will present a series of AI-powered features for Creative Cloud applications that are in the early stage of development and could become available for users in the future.
Based on Adobe Firefly’s family of generative AI models, the teased features include text-based video editing, improved object-aware image editing, AI 3D model generation, and more.
This announcement comes along just as Adobe unveiled the new versions of Adobe Firefly at the same event yesterday.
Let’s look at all the new projects Adobe is working on!
Project Fast Fill – Generative AI Video Editing
One of the most exciting applications of Adobe Firefly that to preview at the Adobe Max 2023 conference today, Fast Fill, is essentially “Generative Fill for video.”
Using Adobe Firefly’s generative AI capabilities –already available in Photoshop enabling users to add, modify, or remove objects or expand images– Fast Fill would let users perform video edits using text prompts as commands.
Additionally, other functions teased today were Dub Dub, which would leverage AI technology to automatically translate voice recordings or audio tracks to multiple languages, preserving the voice and syncing of the original; Scene Change, for unifying two videos with different angles into one single subject and scene; and Res Up that would be basically an AI video upscaler.
These features would be a part of Premiere Pro, and After Effects, the main Adobe video editing tools, and it’s a highly expected one as video is the latest frontier to be conquered by generative AI technology, from video generation to video editing.
See the full announcement from Adobe here.
Projects Poseable and Neo for AI 3D Modeling and Design
They also unveil a few features in the works for 3D modeling and graphic design involving 3D elements.
Project Poseable exploits Firefly’s state-of-the-art, 3D-capable image generator in a feature that can transform text prompts into 3D renders that consider depth, perspective, and alignment.
Project Neo is another interesting development, a function to integrate 3D elements into 2D graphic designs using AI easily and without requiring specific knowledge in 3D object manipulation.
These are not the only projects to be unveiled, as there are a few more potential features to automate lettering and turning doodles into proper sketches, among others.
Project Stardust: Improved AI Object Manipulation for Photos
By loading the video, you agree to YouTube's privacy policy.
Learn more
Last but certainly not least, there are improvements already in the works for Photoshop. Generative Fill and Generative Expand, the first Firefly-powered tools added to the flagship image editor, are hugely popular, and a sneak peek at what is to come would enhance the creative possibilities for Photoshop users.
Project Stardust is among the most interesting, implying improved, intelligent object control within photos. This feature would let editors select complex objects and move them, modify their elements, fill the space they originally were seamlessly, and more. For example, it would make it possible to select a person in a photo, change their position in the image, and even change the color of their clothes, all using AI.
Also relevant is a feature tentatively named See Through, which would use AI technology to remove glass reflections from photos, a tedious and not always possible task to do manually and one that significantly improves photographs.
Adobe Sneaks: A Preview Into the Future of Creative Cloud
Sneaks is a staple segment of the annual Adobe Max conference, where the company lets the public see what they’re working on and what sort of upgrades and additions they can expect to be added to their applications in the future.
This year, the above-listed are just some of a total of eleven inventions revealed that are in the early stages of development.
What do you think of these sneaks? Which feature are you more eager to try soon?
