GLTFExporter: Add KHR_materials_clearcoat support.

[Mugen87]

Related issue: Fixed #22751.

Description

Adds support of KHR_materials_clearcoat to GLTFExporter.

@mrdoob The PR also fixes an issue in context of KHR_materials_volume ( #22214). Right now, the exporter always includes this extensions when MeshPhysicalMaterial is used because the default of MeshPhysicalMaterial.thickness is 0.01.

According to the KHR_materials_volume spec, the extension always needs to be combined with another one like KHR_materials_transmission. The early-out condition of GLTFMaterialsVolumeExtension now checks for transmission usage instead.

Besides, any reasons why the default of MeshPhysicalMaterial.thickness is 0.01? Changing it to 0 seems more glTF conform.

[mrdoob]

Besides, any reasons why the default of MeshPhysicalMaterial.thickness is 0.01? Changing it to 0 seems more glTF conform.

Changing it to 0 sounds like a better solution to me. Would you like to do that in a different PR?

[mrdoob]

Thanks!

[Mugen87]

Would you like to do that in a different PR?

Yes, I do!

[optimus007]

Side note, when a glb with transmission is exported,

@donmccurdy 's gltf validator throws
a "value not in range" error with message "value 0 is out of range"

And it's pointing to "KHR_materials_volume/attunationDistance"

Maybe that also could be looked into while updating the thickness @Mugen87

[Mugen87]

It seems the three.js default 0 is no valid value for attenuationDistance according to the glTF spec.

[Mugen87]

@donmccurdy What do you recommend to fix this warning of glTF validator? Should we target a different default value for MeshPhysicalMaterial.attenuationDistance or enhance the exporter with a new condition instead?

[donmccurdy]

@Mugen87 The specification for KHR_materials_volume is not as clear as I'd like on this, so I've opened KhronosGroup/glTF#2090 and there may be some other comments there.

The default attenuation distance for glTF is +Inf, implied when the attenuationDistance property is omitted. Is that what we're using 0 to achieve? I think that's OK if so, assuming we're avoiding the divide-by-zero situation. We can just include a condition in the exporter:

if ( material.attenuationDistance > 0 ) {

  extensionDef.attenuationDistance = material.attenuationDistance;
  extensionDef.attenuationColor = material.attenuationColor.toArray();

}

[donmccurdy]

^KhronosGroup/glTF#2090 is resolved.

[mrdoob]

Is there anything we have to do on our side?