Teaser
What is Metacast?
Unity Metacast is a platform that uses volumetric technology, which encompasses the process of capturing, viewing, and interacting with the real-world, from moving people to static objects, in 3D. This content can then be viewed from any angle, at any moment in time, giving audiences the ability to see every bead of sweat, blow, takedown and submission, as if they were going toe-to-toe on the famous Octagon canvas themselves.
Metacast doesn’t get into the actual work of hardware or capture systems. It works with partners such as Microsoft, Canon, and other volumetric capture providers who provide the raw volumetric data. Unity then takes over from there by providing an entire pipeline – from ingestion to Unity, content development, delivery, rendering, to interactive experience consumption on various devices.
How does the pipeline work?
Metacast: ML modules
As a Senior Machine Learning Engineer, I spearheaded the end-to-end development of two core ML components integrated to Metacast’s 3D scene capture and athlete analytics pipeline. Leading a cross-functional team of four engineers and researchers, I architected scalable solutions leveraging ML reconstruction to optimize 3D data acquisition costs by 40% and implemented advanced human pose estimation frameworks to enable granular behavioral analysis.
The project lifecycle encompassed multimodal data curation, iterative model prototyping, rigorous validation (achieving 98% pose estimation accuracy), and seamless deployment to production environments. Despite technical complexities, we successfully deployed both modules within an aggressive timeline, driving measurable efficiency gains for stakeholders. The links below detail technical milestones and demonstrations.
Press Coverage
- "Unity unveils Metacast real-time 3D platform for sports broadcasts" VentureBeat - October 05, 2021
- "Unity demos real-time volumetric 3D for UFC" Broadcast Now - March 23, 2022
- "Sport like you've never seen it before" CNN Sports - November 21, 2022