Rendering complex scenes in real time remains challenging due to strict performance constraints. Geometric level of detail (LoD) is widely used to reduce cost by replacing high-frequency geometry with prefiltered representations. ReSTIR significantly improves real-time rendering quality by reusing samples across space and time; however, its effectiveness degrades in the presence of complex geometry, where high-frequency detail reduces reuse efficiency. Moreover, prior ReSTIR methods require the same mesh topology across frames, causing sample reuse to break when switching LoD. In this paper, we introduce a surface point mapping that enables sample reuse across frames containing meshes with different topologies. Building on our method, we enable LoD-aware ReSTIR by maintaining valid spatiotemporal reuse under geometry LoD changes. Our approach restores reuse efficiency in LoD scenes and significantly improves rendering quality compared to prior ReSTIR methods.">

Rendering complex scenes in real time remains challenging due to strict performance constraints. Geometric level of detail (LoD) is widely used to reduce cost by replacing high-frequency geometry with prefiltered representations. ReSTIR significantly improves real-time rendering quality by reusing samples across space and time; however, its effectiveness degrades in the presence of complex geometry, where high-frequency detail reduces reuse efficiency. Moreover, prior ReSTIR methods require the same mesh topology across frames, causing sample reuse to break when switching LoD. In this paper, we introduce a surface point mapping that enables sample reuse across frames containing meshes with different topologies. Building on our method, we enable LoD-aware ReSTIR by maintaining valid spatiotemporal reuse under geometry LoD changes. Our approach restores reuse efficiency in LoD scenes and significantly improves rendering quality compared to prior ReSTIR methods.">

Real-Time Level-of-Detail Rendering with ReSTIR | NVIDIA Real-Time Graphics Research

Search

Real-Time Level-of-Detail Rendering with ReSTIR

Yu-Chen Wang,<br>Markus Kettunen,<br>Daqi Lin,<br>Chris Wyman,<br>Shuang Zhao

July 2026

Cite

doi

PDF

Video

Abstract

Rendering complex scenes in real time remains challenging due to strict performance constraints. Geometric level of detail (LoD) is widely used to reduce cost by replacing high-frequency geometry with prefiltered representations. ReSTIR significantly improves real-time rendering quality by reusing samples across space and time; however, its effectiveness degrades in the presence of complex geometry, where high-frequency detail reduces reuse efficiency. Moreover, prior ReSTIR methods require the same mesh topology across frames, causing sample reuse to break when switching LoD.<br>In this paper, we introduce a surface point mapping that enables sample reuse across frames containing meshes with different topologies. Building on our method, we enable LoD-aware ReSTIR by maintaining valid spatiotemporal reuse under geometry LoD changes. Our approach restores reuse efficiency in LoD scenes and significantly improves rendering quality compared to prior ReSTIR methods.

Type

Paper

Publication

SIGGRAPH (Conference Track), 2026

ReSTIR

Level of detail

Rendering

Markus Kettunen

Daqi Lin

Chris Wyman

Related

Compatibility-Guided Neighbor Selection for ReSTIR

Multi-Layer Reservoir Splatting for Temporal Reuse under Disocclusion

Gradient-Domain ReSTIR Path Tracing

A Gentle Introduction to ReSTIR: Path Reuse in Real-time

Amortizing Samples in Physics-Based Inverse Rendering using ReSTIR

Cite

&times;

Copy

Download