Closed iwae closed 1 year ago
Could be a solution for https://github.com/cocos/cocos-engine/issues/13998
There is a very good comparison in this article https://juejin.cn/post/6931954784018628621
So we should include EXT_meshopt_compression
as a candidate.
I modified the title to model compression integration, the output of this issue should be a detailed solution, executable tasks and plan.
There is a very good comparison in this article https://juejin.cn/post/6931954784018628621 So we should include
EXT_meshopt_compression
as a candidate. I modified the title to model compression integration, the output of this issue should be a detailed solution, executable tasks and plan.
I would suggest to use MeshQuan with 2 reasons, 1st the ASM of meshOpt performs unstable at some platforms, 2nd some platforms don't support WASM. So with the choice of MeshQuan, it would be a more completable solution, since it simply just compresses the float data without changing the data structure, which does not require special uncompress methods/libs with it ( and also it mays save several tens of KB)
Use Case
In other engine cases(https://static.ssl.jimingkeji.com.cn/popmart_show/?source=fma&sceneId=107003#/home), use Draco can significantly reduce the mesh/animation size.
Problem Description
Currently, the mesh stored with binary costs much more storage compared with Draco compressed file's size, which may could be improved with Draco compression.
Proposed Solution
No response
How it works
No response
Alternatives Considered
Moreover, by using Draco loader with wasm+worker, which increases its decoding speed as well, which has been widely used in Cesium for large projects(http://ecnuzlwang.top/draco/ )
Additional Information
No response