Allow IORs below 1? about openpbr HOT 6 CLOSED

Just to provide an external reference point, the MaterialX graph for Autodesk Standard Surface uses 1.0 as the uisoftmin for specular IOR:

https://github.com/AcademySoftwareFoundation/MaterialX/blob/main/libraries/bxdf/standard_surface.mtlx#L29

from openpbr.

@kikou wrote:

• IOR below 1 is physically possible. People usually object you cannot have a velocity higher than the speed of light, but since this is the phase velocity, it does not carry information and thus can exceed the speed of light. See https://en.wikipedia.org/wiki/Refractive_index

• IOR below 1 is useful in production when the default medium is not air, like for underwater scenes. It can be more practical to use the inverse IOR for bubbles for example.

from openpbr.

IOR below 1 is physically possible.

Physically possible, but exceedingly exotic, as in nothing you encounter in everyday life has such an IOR (you could say, physically implausible).

IOR below 1 is useful in production when the default medium is not air, like for underwater scenes.

Yes I'm aware, though I think what is happening there is that people are exploiting the fact that shaders treat the value as the ratio between the IORs of the interior and exterior, which can be <1 if the outside is denser (e.g. as in light exiting an air bubble in water, ratio = 1/1.3). Really the system should determine that ratio from the properties of the media (as is done in nested dielectrics), which will all have IOR >= 1 (e.g. air=1, water=1.3) unless they are exotic.

Though in practice systems may not have implemented nested dielectrics etc., so I know it's a useful hack. We could at least make it a soft-min at 1.

from openpbr.

@peterkutz wrote:

Making it a soft-min at 1 seems reasonable.

from openpbr.

In #56 the range of IOR is set to $(0, \infty)$ while the recommanded/UI/soft range is set to $[1, 3]$.

from openpbr.

Was addressed in #56.

from openpbr.