Micro-Meshes

For over 16 years, Simplygon has been at the forefront of advancing automated 3D content optimization. In line with this tradition, we can now present one of our latest innovations. In collaboration with NVIDIA, we now enable you to leverage our SDK to prepare your meshes for Micro-Mesh based rendering pipelines.

Micro-Meshes offer an incredibly efficient method for encoding displacement data, which is applied onto a base mesh. Although designed with NVIDIA GeForce RTX 40 Series GPUs with hardware acceleration, this technology is not exclusive to that hardware as it is also open-source and versatile across platforms. While it's primarily engineered to enhance meshes for real-time ray tracing, it also offers advantages in traditional rasterization pipelines.

This feature can either be integrated directly into rendering pipelines or used as a storage solution on disk for highly detailed meshes. When loaded into a game, these meshes can be tessellated and rendered through conventional rendering pipelines, with adjustable tessellation levels. This adaptability allows the same micro-mesh data to be used across various Level of Detail (LOD) settings, relying on the low-poly base mesh for distant LODs.

One of the key advantages of micro-meshes is their storage efficiency, requiring only about 1 to 6 bits per triangle, depending on the asset and tessellation detail. Because micro-meshes store only displacement data, they achieve superior compression compared to methods that store coordinates, UVs, and normals for every triangle.

Micro-mesh creation relies on tessellated attributes, which store data more frequently than per-vertex storage. Using a displacement caster, Simplygon transfers data from the original mesh to the optimized mesh during material baking, making it micro-mesh ready. Key applications include:

• Remeshing with Micro-Meshes: Using tessellated attributes configuration, the Remesher can cast displacement data onto an optimized low-poly base mesh, enhancing it with micro-mesh data for improved visual quality.
• Base Mesh Creation with Triangle Reduction: The Triangle Reducer can create a base mesh that closely mirrors the original model's topology, enabling seamless micro-mesh integration.

This feature is a result of a close partnership with NVIDIA, reflecting our commitment to helping game developers fully leverage cutting-edge hardware capabilities.

Interested in efficiently storing large models while conserving memory? Contact us for an evaluation.