← All resources

Spatial hashing for raytraced ambient occlusion

Auto-discovered

Blog Advanced

Kostas Anagnostou

Visit resource ↗

Description

Spatial hashing for raytraced ambient occlusion is a November 2025 Interplay of Light post by Kostas Anagnostou on caching RTAO in a sparse spatial hash instead of dense 3D textures.

It covers nested pcg hashing of quantized position and normal, xxhash32 checksums, linear probing, packed hit and sample counts, distance-adaptive cell size, age-based eviction, early stop after enough samples, and optional tangent-plane jitter, with static-scene timing versus noisy RTAO.

It is for advanced real-time GI programmers comfortable with compute buffers and hemisphere sampling. Dynamic objects are left for later work.

The article is free WordPress HTML with diagrams and code. No separate download is required.

It is useful when RTAO is stable on static geometry but too noisy or expensive, and you want a practical hash-grid cache with LOD-aware cells.