Atomos (7.B19) has announced the integration of PlanckOne FoleyAI with its ATOMOSphere cloud platform, giving filmmakers and sound professionals a faster way to create synchronised Foley audio directly from video stored in their ATOMOSphere workspaces.
Developed by Mocha Chai Laboratories, PlanckOne FoleyAI can analyse anything from a single shot or scene to a complete edit. It identifies the actions, materials and environments shown on screen, then generates synchronised diegetic sounds such as footsteps, fabric movement and impacts.
ATOMOSphere is a cloud-based production ecosystem designed for video professionals, photographers and content creators. It enables users to upload, store, edit, share, review and collaborate on media from anywhere in the world.
Through the new integration, users can send videos from an ATOMOSphere workspace directly to PlanckOne FoleyAI for analysis. The process requires no audio stems or reference sound; PlanckOne FoleyAI requires only the visual information contained in the footage.
PlanckOne FoleyAI can deliver a synchronised starting point within hours, reducing a process that might otherwise require weeks of preparation and scheduled Foley stage time. The generated audio is separated into independent, fully editable layers, including footsteps, fabric and impacts.
“Foley is a vital part of creating a believable and immersive soundtrack, but it can also be one of the most time-consuming stages of post-production,” said Peter Barber, CEO of Atomos. “Integrating PlanckOne FoleyAI with ATOMOSphere gives filmmakers and sound designers an intelligent starting point much earlier in the process, while keeping creative decisions firmly in the hands of the people shaping the final mix.”
“Foley has always been a craft built on time, patience and a room full of props, and nothing about this partnership changes that,” said Chai Yeewei, CEO of Mocha Chai Laboratories. “Launching PlanckOne FoleyAI inside ATOMOSphere means the tool arrives where the footage already lives, so existing ATOMOSphere users get an early look at the launch version on their own projects – without new software to install, new files to move or any change to the way they already work. Their feedback will shape where we take it next.”
Sound designers remain in control throughout the creative process. The cadence, weight and texture of individual layers can be adjusted without affecting the rest of the mix, allowing the generated material to be refined to suit the action, characters and creative direction of each production.
PlanckOne FoleyAI is built using open-source large language models and trained using Mocha Chai Laboratories’ extensive in-house sound library and real-world post-production data. It is designed to integrate directly into established editorial pipelines, helping production teams adopt the technology without disrupting their existing workflows.
“PlanckOne FoleyAI allows projects to add Foley at the offline stage with little fuss, bringing texture to what is usually a plain-sounding offline cut,” added Lee Junhan, Technical Lead at Mocha Chai Laboratories.
The integration expands ATOMOSphere beyond media storage, review and collaboration, creating a direct connection between cloud-based video assets and AI-assisted sound post-production. Editors and sound teams can move more quickly from picture to a structured Foley session, with editable audio ready for further creative development.
PlanckOne FoleyAI Integration: Key Features At-a-Glance
- Analyse video directly from an ATOMOSphere workspace
- Generate synchronised footsteps, fabric and impact sounds
- Requires no audio stems or reference sound
- Create an editable Foley starting point within hours
- Adjust individual audio layers independently
- Integrates into existing editorial pipelines
Availability
PlanckOne FoleyAI is currently in beta and the ATOMOSphere integration shown on the Atomos booth is a technology preview. A localised native version and subscription service are scheduled to launch soon.
Visit https://www.atomos.com
Visit https://mochachailab.com
