Atomos has added a technology preview of PlanckOne FoleyAI to its ATOMOSphere cloud platform. The system analyses video stored in a workspace and generates synchronised Foley elements such as footsteps, fabric movement and impacts.
The notable design choice is that the model works from picture alone. It does not need a production mix, dialogue stem or temporary sound track as a guide, and the generated elements are delivered in separate layers that an editor or sound professional can adjust rather than as a single flattened result.
For fast-turnaround production, the integration could provide a structured first pass before a human sound edit. It may also help teams create guide tracks earlier, while keeping editorial control over timing, selection and final mixing.
The feature remains a beta technology preview. No public accuracy benchmark, rights framework for generated sounds, release date or pricing has been supplied. It should therefore be treated as an assistive production tool whose output still needs creative and technical review.
