1 views collected around this technical thread.
This article provides a comprehensive guide to SAR Creator, a TypeScript‑based interactive solution used for building 2D/3D game scenes, covering workflow, asset creation, canvas adaptation, bundle splitting, texture compression, script lifecycle, dynamic loading, and configuration for ByteDance's Spring Festival activities.
The article presents a Filament‑based PBR material editor that unifies PC and mobile workflows by providing real‑time ImGui previews, JSON‑driven persistence, model optimization, texture compression, and export/import of mobile‑ready material packages, thereby lowering authoring barriers, accelerating iteration, and ensuring visual consistency across platforms.
Texture compression dramatically reduces memory and CPU decode costs while allowing GPUs to read data directly, making formats such as ETC1/2, ASTC, PVRTC and S3TC essential for WebGL and mobile game graphics, though developers must manage lossy quality, size constraints, and platform support.
This article presents a texture compression method that eliminates hollow (alpha‑zero) regions by reorganizing textures into macro‑blocks, storing index maps for reconstruction, and demonstrates significant memory savings and unchanged visual quality on mobile devices, even after video encoding.
This article presents a macro‑block based texture compression technique that removes hollow (alpha‑zero) regions from textures, dramatically reduces memory usage on mobile devices, and restores the original visual quality during rendering, even after video encoding.
This article shares practical techniques for improving Cocos Creator game performance by reducing black‑screen time, minimizing CPU draw‑calls, and managing memory through resource compression, atlasing, cache modes, node pooling, and texture compression, all illustrated with step‑by‑step examples and screenshots.
During Alipay's massive Spring Red Envelope campaign, the frontend team introduced WebGL 3D scenes and particle effects, tackling memory, transmission size, and rendering performance through model reduction, texture compression, GLB packaging, progressive loading, and GPU‑based particle optimizations to achieve stable, low‑latency user experiences.