Adobe Unveils 4 AI Projects for 3D Scenes, Characters and Photo Edits

3 articles · Updated · The Verge · Nov 11

Adobe used its Max conference to preview four experimental AI tools: Project Sight Span for turning flat images into navigable 3D scenes, Screen Test for building characters through conversation, Sound Bender for prompt-based audio mixing, and Rinse Repeat for reusable photo edits.
Project Screen Test lets creators define traits, voices and backstories by talking directly with AI characters, then generate animated screen tests, while Sight Span expands a single image beyond its frame and restores detail after users choose a new camera angle.
Sound Bender generates synced effects from a video clip and text prompt, then lets users blend sounds on a square control canvas; Rinse Repeat copies Photoshop edits such as blemish removal or product shadows onto other subjects while keeping them editable.
These 'sneaks' are tech previews rather than shipping products, but Adobe said they reflect a push to make complex photo, video and audio workflows more intuitive as generative AI becomes central to its Creative Cloud and Firefly strategy.