How to accurately setup a Sapphire Crystal (wrist watch crystal) material

[zachraven]

Hi everyone,

I'm trying to setup a material to accurately render the sapphire crystal used in watch crystals. I've gotten close with trial and error, but i'm looking for some help to get as close to real life as possible. My model is a piece of sapphire with an opaque coating on the inside surface, like paint.

I've attached a screenshot of the material i've created. I've used the dielectric material and using real world values i've looked up (refractive index and Abbe number) i'm close.
I've also attached another screenshot showing how i've modeled the part. Blue representing the sapphire crystal, red representing a separate part with the paint material.

Some Questions:

1. Which material type should i use? Glass, Gem, Dielectric, etc.
2. If i should be using Dielectric like in my example, how do the settings look? I'm confused about transparency distance, and if i should be using the refractive value for both inside and outside.
3. My model is a piece of sapphire with an opaque coating on the inside surface, like paint. How should that be modeled? I approached this by splitting the interior surface of the crystal to apply the paint material, shown in one of the images.

I'm trying to get as close to real life so i can judge the refraction and reflection of the part before i have one made, as they are very expensive, so i'm turning to you! Let me know if you need any other info that i didn't attach.

[hesamzibamim]

send your real sapphire picture that you want i will made it for you.

[richardfunnell]

Hi Zach,

A few answers that may help:
1. Dielectric is definitely the material to use here; Solid Glass, Gem, and Liquid are (from what I know) somewhat simplified or optimized version of the same base material/shader type.
2. Your parameters look good, except for the "Refractive Index Outside" value which should be set to 1 (air). Inside/outside is much more relevant when using layered glass/liquid layers. I'm assuming you are referencing an existing 1.768 IoR value?
If your material is clear, then the transparency distance shouldn't be a factor (unless you have a non-white transmission color, which is affected by transparency distance). I wouldn't suggest a non-white transmission value if you are using any dispersion (like you are).
It will add a bit to render time, but if you introduce roughness to a Dielectric material I would suggest bumping your samples up to 32 (both general and Dispersion Samples).
3. Splitting the surfaces is exactly the workflow I would recommend in this scenario, no need to complicate it further.

It would help to share an image of what you're trying to achieve, but from a technique standpoint you should be in good shape! Good luck!