Niantic Spatial released SPZ 4, an open-source format for 3D Gaussian splats, giving developers and businesses a standardized, accessible way to work with one of the most realistic emerging 3D capture and rendering techniques available today.
Gaussian splatting has quietly become one of the most exciting developments in 3D capture technology, producing remarkably photorealistic scans of real-world spaces and objects, and Niantic Spatial’s decision to release an open-source format for it, SPZ 4, is a genuinely significant moment for any business working with reality capture and 3D visualization. This blog explains what Gaussian splatting actually is in accessible terms, describing it as a 3D rendering technique that represents a scanned scene as millions of small, semi-transparent points of light rather than traditional polygon meshes, producing a level of photorealistic detail and lighting accuracy that older 3D scanning methods have struggled to match. It walks through why the open-source nature of this new format matters so much practically, since proprietary or closed formats for emerging 3D technology tend to fragment an ecosystem, locking businesses into specific tools and making it harder to move captured data between different software platforms, while an open standard lets developers, studios, and businesses build and share tools around a common foundation rather than reinventing compatibility from scratch for every project. The piece covers what this means concretely for industries relying on reality capture, including real estate virtual tours that could achieve dramatically more realistic property walkthroughs, manufacturing and construction digital twins that could represent physical facilities with far greater visual fidelity than mesh-based models, and heritage and cultural preservation projects that could capture historical sites and artifacts with a level of detail previously reserved for expensive, specialized scanning equipment.
A section will address the practical adoption question businesses evaluating 3D capture technology should now be asking, whether their current or prospective reality capture vendor supports Gaussian splatting and specifically the SPZ 4 format, since standardization at this level tends to accelerate how quickly this rendering technique moves from cutting-edge research into mainstream commercial tooling. The blog also touches on the performance side of this format specifically, since a well-designed open standard for Gaussian splats needs to balance the technique’s famous visual fidelity against the file size and rendering performance considerations that make the difference between a demo and something that actually works smoothly in a real business application like a WebXR virtual tour or an AR visualization tool. Gaussian splatting technology, open-source 3D capture formats, and photorealistic reality capture are the throughlines here, treating a technical format release as a meaningful step toward higher-fidelity 3D experiences becoming standard practice.
