On 2017-04-03 08:41 AM (-4), clipka wrote:
> For genuinely convincing glass material, I think it is essential...
>
> - to control any desired tint via distance-based attenuation (`fade_*`
> in interior block) rather than pigment (i.e. use `rgbt 1.0` as pigment)
>
> - to control the balance between reflection and transmission via the
> fresnel mechanism (realistic `ior` in interior block; `reflection { 1.0
> fresnel on } conserve_energy` in the finish block)
>
> It may also help to control the brightness of highlights via physical
> parameters rather than tweaking (i.e. use `specular albedo 1.0` and
> `fresnel on` directly in the finish block).

The question I have is how to best simulate interface between the glass 
and a liquid without incurring either a coincident surface or spurious 
IOR changes due to non-coincident surfaces.

Post a reply to this message