What is Adobe working on? MAX Sneak Previews on experimental AI tools
[14:00 Fri,31.October 2025 by blip]
A recurring segment of Adobe MAX is the sneak preview of experimental AI tools currently under development. Some of these later become part of the official feature set (usually under a different name), offering interesting insights into what to expect in the future across photo, video, animation, and audio.
Project Frame Forward
Many of the sneak peeks initially concern photo editing, such as Project Surface Swap, an AI-powered texture recognition tool for seamlessly selecting and replacing materials, or Project Mood Light, a generative AI tool that allows users to reshape lighting moods after the shot was taken.
However, some are already being adapted for video editing—for instance, Project Trace Erase, based on Diffusion Transformer Models, removes objects along with their shadows, reflections, and environmental distortions, resulting in natural, context-aware photo edits. Project Frame Forward utilizes this functionality, offering similar capabilities for videos.
To remove an object from a video, it is sufficient to make it vanish without a trace in the first frame using Trace Erase. Based on this image, Project Frame Forward then transfers the change to all subsequent frames. This works even with complex, fast-moving clips—one of the preview examples shows an aerial shot where a race car, complete with smoke trails, is seamlessly removed while preserving the camera movement and all racetrack details.
Project Frame Forward can also be used to add new image content consistently across all frames via generative fill, ensuring that elements like reflections and shadows appear appropriately in the image.
Project Motion Map, also for moving images, brings illustrations to life. AI analyzes static vector graphics and then automatically animates them in a meaningful way, without the need for keyframes or manual rigging.
In the audio domain, Project Sound Stager aims to make creating sound design for one's own visuals incredibly easy. By analyzing the visual language, tempo, and emotional mood of a video, it automatically generates multi-layered soundscapes using professional sound design logic.
Editing dialogue is set to become simpler thanks to Project Clean Take. This uses AI to correct mispronunciations, isolate voices, remove background noise, and refine the speech output—all in a matter of seconds.
;)
