Documentation

Support

Asset Transformer SDK


User Manual

Python API

C# API

Changelog

Discussions

Asset Transformer SDK

Decimate meshes

Optimize mesh density by decimating polygons to balance visual quality and polygon count, with strategies for quality-driven or target count decimation.
Read time 1 minuteLast updated 21 hours ago

A huge part of data preparation consists in optimizing mesh density. Finding the perfect balance between visual quality and the polygon count is essential. The most common way to reduce the polygon count is to decimate meshes. You can aim an acceptable mesh quality or at a precise polygon count. Each strategy has benefits and disadvantages. Your choice depends on your needs.

Decimation strategy

Quality

Quality-driven decimation is designed to preserve visual quality. You should choose this strategy first because you can drastically decrease the polygon count with most models. This strategy is particularly suitable for models that have hard edges, such as industrial CAD models. You can preserve important topological singularities that help recognizing these model types. Examples of such singularities are holes, fillets, and chamfers.

Target polygon count or decimation ratio

This strategy is designed for situations that require a specific polygon count. With this strategy, you can drastically reduce the polygon count, but visual quality is affected. Decimation may distort normals, texture coordinates, and mesh borders. You can add constraints to limit the loss of visual quality. For example, you can preserve texture coordinates from breaking. This strategy is particularly efficient on polygonal and organic meshes, such as game models. You can also use this strategy for CAD models that have hard surfaces.