Novel view synthesis is the task of generating new, photorealistic views of a scene from camera angles that were never actually photographed. Given a set of input images, an algorithm reconstructs the underlying 3D structure and appearance so it can render the scene from any viewpoint.
It is the core capability behind neural radiance fields and Gaussian splatting, which both learn a continuous scene representation from photos and then render unseen angles — producing free, interactive camera movement around a captured real-world subject.
Neural radiance fields (NeRFs) — A neural approach to novel view synthesis.
Gaussian splatting — A fast, point-based method for novel view synthesis.
Photogrammetry — An earlier, geometry-based way to reconstruct scenes from photos.
Streaming video reshaped media consumption; we're doing the same for 3D. Join a small team solving tough spatial streaming, content delivery, and developer experience challenges.
Technical deep-dives on adaptive spatial streaming, the infrastructure challenges of 3D at scale, and what we're learning as we build. Written by the team doing the work.
