3 stories · covering 11/2 6:30 – 11/5 6:30 (JST)

A hands-on benchmark of web 3DGS renderers and version 2.0 of a Unity viewer, plus a paper on creating movable 3D assets from a single image.

  1. 01 Six Web 3DGS Renderers Benchmarked, With Procedure and Code Released Under MIT Tools 3dgs-renderer-bench.vercel.app
  2. 02 Unity 3DGS Viewer Reaches 2.0, Adding Direct PLY/SPZ Loading and SPZ 4 Support Tools github.com
  3. 04 PhysAg3D, a Method for Creating Movable 3D Assets from a Single Image, Published on arXiv Research arxiv.org

Issue 2026.11.05 Tools 2 stories・Research 1 stories / Next issue: 11/8(Sun)7:00

In this issue

01Tools

Six Web 3DGS Renderers Benchmarked, With Procedure and Code Released Under MIT

A benchmark comparing six renderers (nine backends), including Spark, PlayCanvas, SuperSplat Viewer and Radix, has been published. It uses PLY files of up to roughly 14 million splats. A reproduction harness and a browser-based lab for testing your own models are included.

Why it mattersThe benchmark states explicitly that sort-key precision and the lag of asynchronous sorting can affect visual appearance. It notes that when nearby splats share the same key, the alpha-blending result can change. Image quality needs to be checked in addition to FPS.

Source: 3dgs-renderer-bench.vercel.app +2 moreRead more

02Tools

Unity 3DGS Viewer Reaches 2.0, Adding Direct PLY/SPZ Loading and SPZ 4 Support

aras-p's open-source 3DGS renderer for Unity is now at v2.0.0. It drops asset creation, loads PLY/SPZ files directly, and adds a stochastic rendering option that needs no sorting.

Why it mattersFor production teams working with 3DGS in Unity, being able to display PLY/SPZ files simply by placing them in the project makes it easier to set up preview viewers and prototypes. It should also suit workflows where scan results are edited in other tools before being brought into Unity.

Source: github.comRead more

04Research

PhysAg3D, a Method for Creating Movable 3D Assets from a Single Image, Published on arXiv

A paper has been published on PhysAg3D, a method that combines a 3D generative model with a vision-language agent to create 3D assets with joints and deformation from a single image. It says the method handles both rigid-body motion and soft deformation.

Why it mattersIn game and XR production, it could reduce the effort of modelers setting up movable objects such as doors and drawers one by one. However, practical accuracy and how it would fit into workflows cannot be confirmed from the source material.

Source: arxiv.orgRead more

Get it by email

We send the latest issue every 3 days at 7:00 JST (in Japanese).

The newsletter is sent in Japanese only. You can subscribe on the Japanese page.