Thoughtful software for difficult data.

0:00 · Separate captures of renovation stages

This edit joins separate Splat Viewer recordings of apartment captures at different renovation stages. Each stage was loaded and explored separately.

0:02 · The older kitchen and adjoining spaces

The first capture moves from a room with parquet flooring into the older tiled kitchen. The camera looks towards the balcony and into a small adjoining space containing a ladder.

0:14 · Stripped surfaces and kitchen installation

Separate views show the kitchen with wall finishes removed, exposed masonry at an opening, visible service routes, and renovation materials.

At approximately 0:26, a later kitchen capture shows new cabinetry and fitted appliances, with materials still present during installation.

0:31 · The older bathroom

The edit returns to the older bathroom. Blue wall and floor tiles surround a bathtub, basin, toilet, and bidet. Dark staining is visible around the ceiling edges.

0:40 · The stripped bathroom and exposed pipework

Subsequent captures show the bathroom stripped to masonry. The old fixtures have been removed. A later view shows exposed pipe runs across the walls and floor.

0:50 · New tiling and shower enclosure

The final capture shows the newly tiled bathroom and a glass shower enclosure with black fittings. Some service openings remain visible. The camera looks around the shower before the video fades out.

0:00 · Introducing Splat Viewer

A Hypercliq title introduces Splat Viewer for Gaussian splats, LiDAR scans, building geometry, laser measurements, and splat training.

0:04 · Exploring the Ricoh E57 capture

The viewer displays the Ricoh capture as a coloured point cloud. The camera looks around rooms containing desks, chairs, windows, and doorways. A scan overview indicates the camera’s position and viewing direction.

An explanatory card identifies E57 as an ASTM exchange format. It says every scan in the file is loaded in colour with its recorded pose applied, and reports 6.4 million points opening in under two seconds.

0:11 · Processing the scan

Navigation continues through the captured rooms. A card explains that E57 uses the same processing pipeline as LAS and LAZ, without prior conversion. It names normals and oriented surface elements derived using local PCA, octree level-of-detail processing, and caching.

0:18 · Identifying building surfaces

Detected walls become orange, floors blue, and ceilings cyan. Furniture remains visible among the classified surfaces. The explanatory card says detected planes are used to colour points by their surface class.

0:24 · Showing the building layout

The view changes to an outside overview of the building. Translucent wall faces outline connected spaces, with door frames in yellow and window frames in cyan. The card describes each wall as a single quadrilateral and the result as a 3D floor plan.

At approximately 0:31, the geometry overlays disappear, leaving the captured building visible from outside. Another card says LAS, LAZ, E57, and PLY files can be opened through the File menu or command line in the same viewer.

0:37 · Closing feature summary

A closing card identifies Splat Viewer 0.6 as a Windows application. It lists support for Gaussian splats in PLY and LiDAR scans in LAS, LAZ, and E57.

The card describes GPU visibility culling and depth sorting, and octree streaming of scans exceeding 200 million points on an 8 GB GPU. It also lists building geometry with DXF floor-plan and IFC model export; distance and area measurements with wall-corner snapping and exported measurement annotations; Brush splat training from scans and photographs; typed or dictated notes pinned to viewpoints; recorded camera paths and COLMAP reference photographs; and panorama and labelled multi-view capture.

The closing text names Rust and Vulkan and states that no runtime dependencies are needed beyond a GPU driver.