Texture Density Checking and Art Asset QA in Game Development: A Case Study and Practical Solutions

In game development, the accuracy and completeness of art assets are critical, yet accidents involving assets frequently occur, leading to severe performance problems such as client crashes and OOM errors.

In April, a major scene update caused a surge of external crashes because many small objects used overly large textures, resulting in excessively high texture density in certain areas of the scene.

What is texture density? It is the distribution of texels on a 3D model surface, usually measured as pixels per square unit. High density provides fine detail but can waste resources and cause performance issues, especially in less visible areas.

The incident highlighted that high texture density was a major cause of player‑side OOM crashes, prompting the need for systematic prevention.

01. Background

The client’s performance test, which randomly traversed the scene, failed to detect the density problem because it did not cover all regions.

02. Requirement Analysis

Stakeholder needs were gathered:

Programmers want an automated solution.

Designers and QA need intuitive visualizations of density distribution.

Artists require timely alerts on abnormal points for quick fixes.

Thus, two main requirements emerged: timely checks and visual heat‑map reports .

03. Solution Design

Historical Check : Basic asset checks (file naming, size limits, etc.) performed during resource production.

Texture‑Density Check :

During the resource‑production stage, artists can trigger checks manually or enforce them before export.

In the commit stage, pre‑commit hooks enforce rule‑based checks, while post‑commit hooks run more extensive analyses via the CDP platform.

During packaging, final checks such as CRC validation and package size verification are performed.

The core of the new pipeline uses Blender as a middleware. After replicating the scene in Blender, a Python script calculates texture density for each model and generates a heat‑map displayed on a web dashboard.

The solution also integrates with Unity’s texture‑density analysis tool for projects that use Unity.

04. Pitfalls and Lessons Learned

1. Delayed feedback : Running checks only at night left artists unaware of issues for days. The improved approach combines immediate post‑commit checks (triggered only when a commit message contains “texture density”) with scheduled nightly scans.

2. Insufficient information : Early reports only gave coordinates, which were useless for artists. Adding GUIDs and model names to the detailed txt report allowed direct identification of problematic assets.

3. Resource parsing challenges : Direct binary parsing was hard to maintain. Switching to a flatbuffer‑based definition generated Python/Lua bindings, making the data more readable and extensible.

05. Summary

The full workflow—from incident discovery, requirement gathering, pipeline design, to implementation and continuous improvement—demonstrates how systematic texture‑density checking can prevent crashes, improve asset quality, and provide actionable feedback to all stakeholders.

Key takeaways include establishing early production standards, proactively analyzing resource‑related issues, understanding multi‑party needs to enhance user‑friendliness, and building checks on a solid understanding of asset structures.