WestLangley

Right. This is the same issue as #13288. This is a three.js problem caused by skinning in world space. The change to skinning in world space began in #4812: >Previously,...

I have seen users use `MeshLambertMaterial` for the performance, others for the illumination model. // I am trying to make the shader code more maintainable. Part of that is removing...

>Why not directly using MeshPhongMaterial with zero specular? Because it is not the same, but `MeshPhongMaterial` could be modified so it is the same.

>Why not directly using MeshPhongMaterial with zero specular? @Mugen87 Did you mean zero-shininess? Because even if shininess is zero, the fresnel goes to 1 at grazing angles. So specular reflections...

@donmccurdy >I think your (E) and my (A) are the same, right? Yes, if your (A) is referring to the new per-pixel Lambert with the normalMap support. >I guess what...

>Okay, I'm convinced^^, too. Yay!

>I guess we no longer need this file? We do need it. Gouraud replicates legacy Lambert.

Current limitations: 1. Equirectangular environment maps do not work for this custom material. There is apparently a conflict when the renderer auto-converts the equirectangular map to a cube map. WebGL:...

>The shader materials in [WebGLBackground](https://github.com/mrdoob/three.js/blob/68f8ba5f0b3b7eb1a74024d7fcc0ba8689583006/src/renderers/webgl/WebGLBackground.js#L89-L98) do it like that. Oh, see #15187. FWIW, I think that pattern is no longer needed, because the code injection (`getTexelDecodingFunction()`) was removed.

>Good news! However, if you want automatic cube map conversion for MeshGouraudMaterial, you still need the envMap property on material level. Right. It's main purpose was to enable automatic uniform...