MLSLabsRenderer-Lite has been introduced—a specialized plugin for Unreal Engine 5 optimized for high-speed rendering of 3D Gaussian Splatting (3DGS) and dynamic volumetric video (4DGS). By utilizing a custom CUDA pipeline instead of the standard Niagara particle system, the tool achieves impressive performance when working with massive point clouds.
What Happened
Developers from MaLanShan Audio & Video Laboratory have released the MLSLabsRenderer-Lite plugin, which enables the rendering of over 5 million static Gaussians at 50+ FPS, as well as support for 100,000+ dynamic points at speeds exceeding 100 FPS. The plugin supports integration with UE Sequencer, allowing it to be used in animation and video production workflows.
Context
Traditional particle systems, such as Unreal Engine's Niagara, are not designed for the efficient rendering of high-density neural representations like 3DGS. Moving to specialized CUDA renderers is becoming a necessary condition for scaling neural field technologies (NeRF/3DGS) in real-time applications.
Why It Matters for the Industry
Integrating specialized CUDA renderers removes performance bottlenecks when working with high-density point clouds, transforming neural fields from research artifacts into applied tools for real-time graphics. This opens new possibilities for integrating photorealistic neural fields into game engines and the virtual cinema industry.
Why It Matters for Users
Users gain the ability to seamlessly integrate complex 3D/4D content obtained from object scanning into Unreal Engine 5 workflows without critical FPS drops. The tool allows for rapid prototyping of high-quality scenes using the standard Sequencer pipeline.
What Is Not Yet Known / Limitations
There is uncertainty regarding the product's distribution and licensing model due to discrepancies between commercial offerings on the official website and the availability of releases on GitHub, which requires further verification of functional limitations.
Sources
Author
Look at AI, Editorial Team