Technology

Adobe Shows Experimental Tools That Turn Photos Into 3D Scenes, Characters and Sound

Martin HollowayPublished 4d ago3 min readBased on 1 source
Reading level
Adobe Shows Experimental Tools That Turn Photos Into 3D Scenes, Characters and Sound
Image by fancycrave1 from Pixabay

Adobe has shown a set of experimental tools at its Max Sneaks event. They can turn flat images into 3D scenes, let users build AI characters through conversation, and make photo editing work like copy and paste. The Verge reported the lineup on 11 November 2026. Sneaks are Adobe's showcase for experiments, and every project below is a prototype.

The most visual of the group is Project Sight Span. It takes photographs and other images and builds a 3D environment that users can move through and rearrange. According to the report, it works with any flat image, including photos, paintings and illustrations. It can also keep and copy the artistic style of the original.

The tool can create objects and backgrounds that sit outside the edges of the original picture. Image quality is restored once the user picks an angle. That detail matters to anyone who has tried to crop a small photo into something larger. The tool seems to treat the original frame as a starting point, not a hard limit.

Project Screen Test is aimed at a narrower problem: making AI-generated characters easier to create and use in videos. The process starts by describing basic personality traits and choosing a voice. The voice can come from a preset, a text description or an uploaded audio sample.

Next, the user gets to know the character by chatting with it in a FaceTime-style video call. They can outline its backstory, temperament, goals, flaws, fears and the way it moves. The interface works more like an audition than a text box. The user directs the character through dialogue, not only through written instructions. The tool can then generate animated screen tests of the character in a scene the user describes.

Project Sound Bender is an AI audio-mixing feature for customizing sound effects. The user uploads a video clip and describes the sound they want in a text prompt. The tool generates the sound effects and lines them up with the video automatically.

The showcase also includes a tool that turns photo editing into a copy-and-paste process.

The broader context here is a change in how AI image and video tools are guided. Most mainstream generators rely on a single text prompt. If the result misses, the user tries again. Screen Test treats a character as something that gathers traits over a session. That is closer to casting and rehearsal than to writing a prompt. Sight Span also moves from one fixed output toward a scene that can be reframed after it is made.

For the people who build these tools, the important questions sit under the demos. How does the system store a character's details? How does it stay consistent over a long conversation? How does it fill in parts of a 3D scene that a single photo cannot show?

In my view, the Sneaks format is a useful sign of where Adobe's research is heading, but it is not a promise of a product. Demos like these often differ a lot from what reaches the public, especially when they rely on carefully chosen inputs. Creative teams should treat them as early previews until the tools are widely available and can be tested on their own work.

Even so, the direction is clear enough to watch. If the 3D scene tool works as shown, one photograph could become the starting point for camera moves and layouts that once needed a 3D artist. If the character tool works, short videos could be planned with characters that have been rehearsed, not just generated. Sound Bender suggests that sound effects work, which can take an editor hours, could be drafted from a clip and a sentence of description.

Each of these would shorten a slow step in making content. Each would still need a person to check that the result matches what was intended.