Proposed 2.0 materials common

[lasalvavida]

This is my perspective on how to update the materials common extension.

• Drop lights out (it should be a separate KHR_lights extension as in glTF 2.0: New KHR_lights extension #945)
• Use the existing pbr and pbrSpecularGlossiness properties to compute the specular lighting parameters in the context of the lighting models
• Very simple extension, just define a technique, and use the existing material to do the rest
• enum for techniques to be more in line with the rest of the spec

{
    "materials" : [{
        ...
        "extensions" : {
            "KHR_materials_common" : {
                "technique": 0 // Blinn
            }
        }
    }]
}

Feedback appreciated

[pjcozzi]

Thanks @lasalvavida!

I think @McNopper would be a good person to review and collaborate with here.

[McNopper]

Please let us optimze/finalize #947, which will become the offical KHR_materials_common extension sooner or later. This will be a "real" Phong, Blinn etc. material model, with the well known parameters which already do exist.

As far as I know, @bghgary is working on some math, where you can make out of Phong etc. a PBR material. As far as I understood your suggestion, you are doing the opposite way: Making out of a PBR material a Phong etc. material.

I think both directions do make sense in a manner of compatibility e.g. an engine does not support a specific renderer.
But a quality loss (PBR -> Phong) is to be expected and the impact should also be discussed in the glTF working group.

If we follow this path, I suggest we define an extension e.g. KHR_materials_conversion, which allows to convert between materials.

[pjcozzi]

Also see #824 for discussion on spotlight falloff angle.

[pjcozzi]

@McNopper OK to close as duplicate with #947 and #945?