Autodesk folded an actor-founded VFX startup into its suite

The shot

Wonder Dynamics was a cloud-based visual-effects startup, co-founded in 2016 by actor-producer Tye Sheridan and VFX artist-filmmaker Nikola Todorovic, whose product, Wonder Studio, used artificial intelligence to map a filmed actor's performance onto a 3D character without traditional motion-capture hardware. According to a trade-press report that cites Autodesk's own statement, Autodesk, a major 3D design and entertainment-software company, acquired Wonder Dynamics, with the deal reported 22 May 2024 and its financial terms not disclosed. The acquired product now operates as Autodesk Flow Studio, formerly Wonder Studio, folding an AI-assisted performance-capture pipeline into an established 3D software ecosystem.

What the documents show

The report quotes Autodesk's own framing of the deal's purpose: the acquisition will give more creators the ability to add 3D animated characters to their projects while minimizing the learning curve and automating complex, time-consuming processes, and it describes Wonder Studio as helping artists animate, light, and compose 3D characters within live-action scenes at lower cost than competing tools, Autodesk's own characterizations, relayed by the outlet, not independently benchmarked claims. What the acquisition changed in practice is documented separately in Autodesk's own blog account of visual-effects studio Boxel Studio's work on Superman and Lois: a named, credited production, using markerless motion capture technology, powered by machine learning, that captures performances from a single camera angle. That is verified production use, distinct from the acquisition's promotional language.

The workflow

The rebranded product's own site describes an artist filming a performance with a single camera, then exporting the motion capture, a character pass, an alpha mask, and a clean plate as separate elements the artist can bring into whatever 3D software is already in use on a production. The site states the automation covers what it calls the objective portion of the visual-effects work, the mechanical mapping of performance to a character rig, leaving what it describes as the subjective portion, creative choices about performance, lighting, and integration, to the artist.

What the tool does not change

None of the three documents claim the tool eliminates an artist's compositing or animation review pass; the product's own site explicitly frames its automation as covering only the objective, mechanical portion of a VFX pipeline and hands the subjective portion back to a human artist. The Boxel Studio account describes the tool solving a specific production constraint, capturing performance on a fast-moving set without traditional mocap hardware, not replacing an animator's or compositor's judgment about how that performance should look once composited into a finished shot.

• Which parts of a given VFX shot fall into the objective automation this tool documents, and which stay subjective?
• Is a single-camera capture adequate for this production's specific performance and blocking needs?
• Has this now-Autodesk product's licensing or pricing changed since the acquisition folded it into a larger suite?

An acquisition announcement's promotional language and a documented production's actual account of using the tool are different kinds of evidence. The Boxel Studio case is the more concrete one: a named show, a named use, and a specific technical description of what the software did and did not automate.