3d-games_skill

Overview

This skill covers core 3D game development principles for rendering, shaders, physics, cameras, lighting, and LOD. It focuses on practical techniques to achieve good performance and visual quality in real-time engines. Use it to inform design decisions, optimize pipelines, and avoid common anti-patterns.

How this skill works

The skill breaks down the rendering pipeline into vertex processing, rasterization, fragment processing, and final output, and explains common optimizations like frustum and occlusion culling, LOD, and batching. It outlines shader roles (vertex, fragment, compute), collider choices and physics heuristics, camera types and behaviors, lighting strategies, and when to bake versus render in real time.

When to use it

• Designing or profiling a 3D rendering pipeline for performance
• Deciding whether to write custom shaders for effects or optimization
• Choosing collision shapes and physics approaches for gameplay entities
• Implementing camera systems for player feel and collision avoidance
• Planning lighting strategy and shadow performance for levels

Best practices

• Prefer simple colliders for gameplay and reserve mesh colliders for static terrain
• Use frustum/occlusion culling and batching to reduce draw calls
• Implement multiple LODs and billboards for distant objects
• Bake lighting and shadows where possible; use cascaded shadows for large worlds
• Profile shaders and simplify expensive passes on constrained hardware

Example use cases

• Create a third-person camera with smooth follow, look-ahead, and collision avoidance
• Design LODs for an open-world scene to cut triangle count at distance
• Replace mesh colliders with primitive shapes for physics-heavy scenes
• Author a custom fragment shader for water and optimize with cheap variants on mobile
• Bake static lighting for interior levels while keeping dynamic lights for players