At its Max Sneaks event, Adobe demonstrates how it wants to integrate AI into its software suite. Several demos illustrate what is possible.
During its annual Max Sneaks event, Adobe has once again presented a series of experimental AI tools. The projects show how artificial intelligence can accelerate, simplify and enrich creative workflows, from image editing to sound and 3D design.
Max Sneaks is the platform where Adobe shares experimental technologies with the creative community. The projects are in an early stage, but give an impression of functions that may later appear in commercial products.
For Image Editing
Most of the proposed AI-driven innovations have to do with image editing. Adobe wants to help users make quick adjustments in 2D and 3D.
- Project Surface Swap allows users to replace materials on surfaces in a photo. AI automatically recognizes the texture and adjusts new materials to the incidence of light and perspective. Think of virtually replacing a wooden floor with carpet, directly in an image.
- Project Light Touch gives control over the lighting in photos after the shot. Users can move or adjust light sources, change the atmosphere, or turn a daytime shot into a nighttime shot.
- Project Turn Style offers the ability to edit 2D objects in an image as if they were 3D. Objects can be rotated or moved, while maintaining their natural texture and lighting.
- Project Trace Erase removes unwanted objects from photos. The tool also detects and erases shadows, reflections and environmental distortions, largely eliminating the need for manual post-processing.
- Project New Depths enables editing of 3D photos. Users can adjust color, shape, and composition in radiance fields, three-dimensional image formats containing depth information.
- Project Scene It allows creators to build 3D scenes with image recognition. Objects in the scene can be linked to reference images, making it possible to maintain their visual style while they are freely positioned in 3D.
For Video and Animation
Adobe is also unleashing the AI functions on videos. Here too, speed and simplicity are central.
- Project Frame Forward applies edits to entire video clips via a single marked frame and a text prompt. This significantly speeds up the editing process without sacrificing visual quality.
- Project Motion Map analyzes vector images and animates them automatically. Based on the content, the AI determines how motion should look natural and expressive, without the need for keyframes or manual animation.
For Sound and Storytelling
Finally, Adobe shows some innovations related to sound.
- Project Sound Stager generates sound designs based on the visual content and emotional tone of a video. The user can then refine the result via text commands in collaboration with what Adobe calls a virtual AI sound engineer.
- Project Clean Take automates the post-processing of spoken audio. The tool corrects pronunciation errors, removes background noise and optimizes recording quality, which should be useful for podcasters or video makers.
All proposed functions are experimental. Whether and when they will effectively roll out to Photoshop and other products from the Adobe suite is unclear at this stage. However, the demos illustrate how Adobe thinks AI-driven functionality can be relevant in its professional software package.