Spatial Capture
Three ways to capture the real world in 3D: neural radiance fields and Gaussian splats for volumetric scenes, photogrammetry for textured meshes, and point clouds for raw spatial data.
Three Methods, Three Outputs
NeRF / Gaussian Splat
Volumetric Scene
A scene learned from photos or video that encodes how light behaves in the space (reflections, translucency, specular highlights) and renders new viewpoints in real time.
Walkthroughs, virtual tours, web-delivered 3D, film previs, archiving spaces as they feel.
Navigable volumetric scene (WebGL, splat file)
Photogrammetry
Textured Mesh
Reconstructs polygon geometry with UV-mapped photo textures from overlapping images. The mesh can be edited, rigged, animated, 3D printed, or dropped into any standard pipeline.
3D printing, game assets, VFX, product visualization, object scanning.
Textured 3D mesh (OBJ, FBX, GLB, USDZ)
Point Cloud
Raw Spatial Data
Millions of 3D points captured by LiDAR or depth sensors, each with color and position. No surfaces, no mesh: the rawest form of scanned geometry.
Surveying, architecture, construction documentation, measurement, forensic archival, raw input for mesh reconstruction.
XYZ coordinate data with color (PLY, LAS, E57)
How They Relate
A point cloud can seed a photogrammetric mesh, a scan can inform a NeRF training set, and the same walk-around footage can produce all three.
Quick Comparison
NeRF / Gaussian Splat Captures
Each capture is a live volumetric scene running in your browser. Click and drag to orbit, scroll to zoom.
