What is 3D Gaussian Splatting?
A newer alternative to NeRF for 3D scene representation. Uses millions of 3D Gaussian primitives for faster rendering and higher quality.
How it works
3D Gaussian Splatting represents scenes as millions of 3D Gaussian primitives (small, oriented, colored blobs) that can be rendered extremely quickly. It offers several advantages over NeRF: significantly faster rendering (real-time vs seconds per frame), better handling of fine details, and more efficient training. In the context of AI video, Gaussian Splatting is used for creating 3D scene representations from photos or video that can then be rendered from new viewpoints, combined with AI generation for hybrid workflows, or used as conditioning input for AI video generation. Some studios use Gaussian Splatting to create virtual camera moves through real environments.
Tools that use 3d gaussian splatting
Related terms
Frequently asked questions
What does 3d gaussian splatting mean in AI video?▾
Need AI video produced by professionals, not generated by yourself?
Apostle is an AI-native video production studio. We use every tool on this page in real client work.
Get in touch