Principled BSDF: Change minimum IOR to 1.0 #111158

Alaska · 2023-08-16T04:11:20+02:00

Alaska commented

2023-08-16 04:11:20 +02:00

Set a soft limit of 1.0 for the minimum IOR on the Principled BSDF.

❤️ 1

Alaska added 1 commit 2023-08-16 04:11:31 +02:00

9f2425ffdc Principled BSDF: Change min IOR to 1.0

Set a soft limit of 1.0 for the minimum IOR on the principled BSDF.

Alaska added the

Module

Render & Cycles

label 2023-08-16 04:12:06 +02:00

Alaska requested review from Brecht Van Lommel 2023-08-16 04:12:16 +02:00

Alaska requested review from Weizhen Huang 2023-08-16 04:12:26 +02:00

Alaska requested review from Lukas Stockner 2023-08-16 04:12:32 +02:00

Alaska commented

2023-08-16 04:18:16 +02:00

After recent changse, the IOR now partially controls the specularity of opaque surfaces in Cycles. And to make it easier for people to play around with this effect, I changed the soft limit for min IOR to 1.0 (since most materials are in the 1.0+ range). This is so if you drag the IOR slider down, it will stop at 1.0.

You can still set the IOR below 1.0 by manually typing the value in, and I personally think this fits better with the workflow for when IORs below 1.0 are typically used. IORs below 1.0 are used for nested dielectrics, and to get the right result you need to do some math to figure out the correct IOR value for the two interacting materials.

After recent changse, the IOR now partially controls the specularity of opaque surfaces in Cycles. And to make it easier for people to play around with this effect, I changed the soft limit for min IOR to 1.0 (since most materials are in the 1.0+ range). This is so if you drag the IOR slider down, it will stop at 1.0. You can still set the IOR below 1.0 by manually typing the value in, and I personally think this fits better with the workflow for when IORs below 1.0 are typically used. IORs below 1.0 are used for nested dielectrics, and to get the right result you need to do some math to figure out the correct IOR value for the two interacting materials.

👍 2

Alaska requested review from Sergey Sharybin 2023-08-16 04:18:50 +02:00

Alaska changed title from ~~Principled BSDF: Change min IOR to 1.0~~ to Principled BSDF: Change minimum IOR to 1.0

2023-08-16 04:19:17 +02:00

Brecht Van Lommel approved these changes 2023-08-16 10:14:54 +02:00

Brecht Van Lommel merged commit 007b7d0353 into main

2023-08-16 10:18:05 +02:00

Brecht Van Lommel referenced this issue from a commit

2023-08-16 10:18:05 +02:00

Shaders: Change IOR soft minimum to 1.0 in Principled BSDF

Brecht Van Lommel deleted branch change-min-ior

2023-08-16 10:18:06 +02:00

baoyu commented

2023-08-21 15:32:36 +02:00

First-time contributor

IORs below 1.0 are used for nested dielectrics, and to get the right result you need to do some math to figure out the correct IOR value for the two interacting materials.

Can you share a simple demo blend file of this use case?

For example, water inside a glass cup, till now, I just let the water mesh slightly intersect with the glass mesh, and assign the glass with 1.5 ior, the water with 1.33 ior.

What should we do after this PR?

> IORs below 1.0 are used for nested dielectrics, and to get the right result you need to do some math to figure out the correct IOR value for the two interacting materials. Can you share a simple demo blend file of this use case? For example, water inside a glass cup, till now, I just let the water mesh slightly intersect with the glass mesh, and assign the glass with 1.5 ior, the water with 1.33 ior. What should we do after this PR?

Alaska commented

2023-08-22 01:17:38 +02:00

@baoyu for as long as I've been using Blender (roughly 6 years), to render a nested dielectric accurately you typically needed to set one of the IORs to below 1.0. This pull request doesn't change that.

But just for reference, I've attached a .blend file to show you how you might accurately create a glass cup with water in it. (Note: When working with nested dielectric, you have to be careful to make sure your normals are facing the right direction, and the IOR values are correct).

Make sure you look at the different materials, where they're assigned, how the geometry is laid out, and the geometry normals in the file I've attached.

Note: Nested Dielectrics (presumably automatically, or semi-automatically figuring out what IOR values to use when materials intersect) has been listed as one of the "further goals" of #99447. So this feature may come to Cycles in the near future.

@baoyu for as long as I've been using Blender (roughly 6 years), to render a nested dielectric accurately you typically needed to set one of the IORs to below 1.0. This pull request doesn't change that. But just for reference, I've attached a .blend file to show you how you might accurately create a glass cup with water in it. (Note: When working with nested dielectric, you have to be careful to make sure your normals are facing the right direction, and the IOR values are correct). Make sure you look at the different materials, where they're assigned, how the geometry is laid out, and the geometry normals in the file I've attached. Note: Nested Dielectrics (presumably automatically, or semi-automatically figuring out what IOR values to use when materials intersect) has been listed as one of the "further goals" of #99447. So this feature may come to Cycles in the near future.

Glass example.blend

1013 KiB

❤️ 1

baoyu commented

2023-08-22 06:22:54 +02:00

First-time contributor

Thank you @Alaska for the demo!
I did a comparison between your method and mine.

The difference is there, while I do see your method is a bit tidous to model and shade.
I learned my method said above several years ago from arnold render's manual(in which I now find your method talked also).

Aside from that in arnold one can and should set the priority of the material. However when I get to use blender cycles, the 'overlapping' trick seems work quite well without the feature of having nested dielectrics priority.

I don't know why I don't have right to upload files here. This is the blend file of side by side comparison.

Thank you @Alaska for the demo! I did a comparison between your method and mine. ![image](/attachments/6e4e87e9-31cb-4ff6-a053-bdb176afcf27) The difference is there, while I do see your method is a bit tidous to model and shade. I learned my method said above several years ago from [arnold render's manual(in which I now find your method talked also)](https://help.autodesk.com/view/ARNOL/ENU/?guid=arnold_user_guide_ac_transmission_ac_nested_dielectrics_html). Aside from that in arnold one can and should set the priority of the material. However when I get to use blender cycles, the 'overlapping' trick seems work quite well without the feature of having nested dielectrics priority. I don't know why I don't have right to upload files here. [This is the blend file of side by side comparison.](https://drive.google.com/file/d/1tKZbDi82pBwzoIAZTZsOC-fj0_i_jxFN/view?usp=sharing)

image.png

2.0 MiB

Alaska commented

2023-08-22 07:37:12 +02:00

@baoyu manually dividing up the scene into different parts and setting the correct IOR for each medium change is annoying and it is difficult to get right in some scenes. But it is more physically correct than overlapping the two meshes.

However due to the annoyingness of setting it up, and the difficulty of explaining it to people, most people will just use the method you use (overlapping water and cup), and if it looks alright to you and the many other people that use that technique, then it's fine. Most people are using Cycles to make art, not to make perfect physical simulations.

Anyway, that probably ends our discussion. If you want physical accuracy, use the annoying method I showed. If you want ease of use, use the method you use. And in the future, hopefully proper nested dielectric support is added to Cycles and you can get the ease of use of your method with the physical accuracy of my method.

Also, it seems I got some of my IOR notes and math backwards in the original file I sent you. I have updated the original file with the fixed notes and IOR values.

This just proves it's kind of annoying to setup because if you rarely do it, then you can forget the correct order to do the division.

@baoyu manually dividing up the scene into different parts and setting the correct IOR for each medium change is annoying and it is difficult to get right in some scenes. But it is more physically correct than overlapping the two meshes. However due to the annoyingness of setting it up, and the difficulty of explaining it to people, most people will just use the method you use (overlapping water and cup), and if it looks alright to you and the many other people that use that technique, then it's fine. Most people are using Cycles to make art, not to make perfect physical simulations. Anyway, that probably ends our discussion. If you want physical accuracy, use the annoying method I showed. If you want ease of use, use the method you use. And in the future, hopefully proper nested dielectric support is added to Cycles and you can get the ease of use of your method with the physical accuracy of my method. --- Also, it seems I got some of my IOR notes and math backwards in the original file I sent you. I have updated the original file with the fixed notes and IOR values. This just proves it's kind of annoying to setup because if you rarely do it, then you can forget the correct order to do the division.

👍 1

Alaska commented

2023-08-22 09:39:46 +02:00

@LukasStockner I've re-read over various notes in #99447.

It seems you want to support a specific case for IOR = 0 with the specular layer. This pull request/commit interferes with the ability for users to "naturally" set the IOR to 0. So feel free to revert the change if you want.

@LukasStockner I've re-read over various notes in #99447. It seems you want to support a specific case for IOR = 0 with the specular layer. This pull request/commit interferes with the ability for users to "naturally" set the IOR to 0. So feel free to revert the change if you want.

Sign in to join this conversation.

No reviewers

No Label

Download

What's New

Blender Studio

Manual

Developers Blog

Documentation

Benchmark

Blender Conference

Development Fund

One-time Donations

Principled BSDF: Change minimum IOR to 1.0 #111158