Skip to content

Long Volumetric Video

Rebuild long volumetric video from multi-view RGB footage

Visit official site

What Long Volumetric Video is

The technique is described as a way to reconstruct long volumetric video from multi-view RGB footage, using a Temporal Gaussian Hierarchy to model it compactly.

What you can do with it

  • Model long volumetric sequences in 4D
  • Handle regions that change at different speeds
  • Render thousands of frames in real time

Getting started

  • Open the online demo page
  • Read the paper through the arXiv link
  • Download the renderer and framework code

Pros & cons

✓ What we like

  • Different speed regions are modelled efficiently
  • View-dependent effects are kept with compact storage

! What to watch out for

  • Input is multi-view RGB video footage
  • Setup needs the renderer and framework downloads

FAQ

What representation is used?

The material names a Temporal Gaussian Hierarchy as a 4D representation.

Which datasets are named?

SelfCap, DNA-Rendering and Sports are listed.

Last reviewed: 2026-11-15

More AI video generation tools

View all →

How we review