If I add a Principled Shader node, it has an Index of Refraction of 1.45 and a Specular Reflection of 0.5. The physically correct specularity of an object can be worked out from its index of refraction as:

$$specular=\left(\frac{ior-1}{ior+1}\right)^2 \div 0.08$$

Putting 1.45 into that formula yields a specularity of $2025\div4802\approx.422$.

So why when I add a Principled BSDF node, does it have an Index of Refraction of 1.45 and a Specular of 0.5, instead of something like an Index of Refraction of 1.5 and a Specular of 0.5?

  • 1
    $\begingroup$ As far as I know refraction is only used for materials with transparency, and 1.45 is close to the index of refraction of various commonly used forms of glass, which is a frequent recurrence for artists in every day scenes. Other than that, "Why?" is a question only the developers who coded it could possibly answer $\endgroup$ Aug 1, 2022 at 8:50

1 Answer 1


You are mixing things up there. As the Blender manual states:

"Amount of dielectric specular reflection. Specifies facing (along normal) reflectivity in the most common 0 - 8% range." The index of refraction doesn't affect this in a way defined by this formula. Since you are talking about defaults, the Principled BSDF has a default value of 0 for Transmission - which means there is not even any refraction that could be affected by the IOR. By the way, the formula that's given there which you have in your question uses 0.08 i.e. 8%, whereas the manual also speaks of the "most common 0 - 8% range." So by using only 6.75% you'll have 0.5 for specularity.

And to quote the Wikipedia article on Specularity:

"In computer graphics, it means the quantity used in three-dimensional (3D) rendering which represents the amount of reflectivity a surface has. It is a key component in determining the brightness of specular highlights, along with shininess to determine the size of the highlights."

The difference in Blender is, instead of shininess going from 0 (not shiny) to 1 (completely sharp shiny) it uses Roughness the other way round, from 0 (not rough at all) to 1 (completely rough).

The specular value of 0.5 is a good average value which usually results in realistic looking reflections for dielectrics. If you think it is too shiny or not shiny enough, you can still change this value. Sometimes the IOR is referred to as index of reflection for non-transmissive materials, and there are lists of IOR values on the internet to find certain materials - refractive/transmissive materials as well as non-refractive/non-transmissive materials. But 1.45 is a good average value for common opaque dielectrics here, if you don't need perfect accuracy (and to be honest, slight differences can be hardly visible on non-transmissive materials). Now, why are there IORs for non-transmissive materials? I quote another website here because I am too lazy to write it down myself:

"Anyway, you take a look at IOR lists, see glass and diamond on there, and probably water and milk among other fluids and solid refractive materials.

But then you see asphalt and aluminium and start to wonder if you have a grasp on what IOR is because last I checked, you can’t see through asphalt or aluminum foil, and there really shouldn’t be an IOR (Refraction) value to be set, right?

Well here’s the thing:

The Redshift Material already hints at it. The default material setting couples the IOR of Refraction and Reflection. If you set an IOR of 1.5 in Reflection, this is copied into Refraction as well.

So even though you can’t see through asphalt, it still has an IOR value."

  • $\begingroup$ Sometime after 17:54 in this video, Andrew said that the majority of dielectric materials have an Index of Refraction of 1.5. So, why does not the Principled BSDF shader use an Index of Refraction of 1.45, instead of 1.5 by default? $\endgroup$ Aug 15, 2022 at 7:20
  • 1
    $\begingroup$ @BlenderUser9000 I have no idea how much Andrew knows about dielectrics, I only know that he often jokes about him not having much knowledge about mathematics and physics and doing everything by guesswork. I read the value 1.45 some years ago on some page which I don't remember dealing with the physics behind IOR in CG and so on. I'm sure the value 1.45 or 1.5 isn't actually what the majority of dieletrics have, they will only average around that. And anyway, Andrew's 1.5 might just as well be rounded up from 1.45 and you don't see a big difference between them anyway - so why be so nitpicky? $\endgroup$ Aug 15, 2022 at 8:35
  • 1
    $\begingroup$ @BlenderUser9000 Okay, I didn't find the page I read that a few years ago, but Google brought up many results. First of all, dielectrics vary in IOR as it was to be expected. Some sources say between 1.2 and 1.8 (which averages to 1.5, but without knowing the distribution inbetween those values the average could be anywhere), another site said dielectrics have IORs around 1.5 to 1.6, a book on C4D said1.44 for dielectrics, somewhere else is stated 1.5 would refer to plastic. So the 1.45 is as good as everything else and as I said, it's hard to recognize anyway. $\endgroup$ Aug 15, 2022 at 8:49

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .