The code change has landed, the build environment needs to be updated.
Christophe brought up a point that DGX servers might have difficulties with getting newer drivers. We’ll keep OptiX 7.3 as ifdef-ed support for the time beng.
ReSTIR is currently not being worked on actively. There are research papers coming out that seem to address issues that Weizhen run into.
Practical Info
This is a weekly video chat meeting for planning and discussion of Blender rendering development. Any contributor (developer, UI/UX designer, writer, …) working on rendering in Blender is welcome to join and add proposed items to the agenda.
For users and other interested parties, we ask to read the meeting notes instead so that the meeting can remain focused.
Hello, small question, by having this new metallic node in any of its flavors, does it mean that the texture baking is going to have metallic option too to bake?
The Metallic BSDF node PR is not adding a Metallic bake option. That will have to be handled seperately.
When you ask for a Metallic bake option, what are you asking for? The ability to bake metallic direct, indirect, and color? Because that’s already handled with the Glossy direct, indirect, and color.
Or are you asking for a Metallic map baking (E.g. Extract and bake whatever was plugged into the metallic input on the Principled BSDF)?
Exactly that, for the glossy workflow blender has nodes(glossy bsdf) but are not the default, the default its the Principled BSDF and that is metallic-roughness material pipeline, something that is really good because most game engine uses it, however believe it or not a lot of people uses blender with all its flaws to bake texture maps, the alternative its substance and adobe with its questionable licenses agreements, especially in Linux, there are not many more alternatives. So what its usually done is to connect whatever goes into the metallic socked of the principled to an emit and bake that.
There are addons that does versions of this connect/disconnect maneuvers but having that as an addon it has its own issues.
I am a volunteer I do not work at the Blender foundation. Although I do see merit in adding this option, I do not know how complex it is to add as I have not worked in this area of Cycles before.
So there is no confirmation that this feature will be added to Cycles. And if there is a desire from the Cycles developers to add it, there’s no clear timeline on when it will be added by.
This paper is about replicating the predictions of experts in cases where those experts disagree with each other, even if the experts are wrong.
In the case of Cycles, it is always about converging to the actual true values.
I’ve just downloaded the latest 4.3 build and am testing the Metallic BSDF node. I wonder if it would be smart to add N-K labels for industry consistency. Right now in Physical Conductor mode, the labels are IOR and Extinction. While the tool tip tells the user that IOR = N and Extinction = K, would it not make sense to just label them as IOR (n) and Extinction (k) so it’s immediately obvious?
One more question about the Metallic BSDF node. This is the only time I’ve seen three channels for N-K data. Sources like refractiveindex.info only list single N-K values. Is there a source that spells out the individual channels in a physically correct way for these N-K values, or do we simply populate each with the same N-K value?
Most sources provide the n and k values for specific wave lengths of light. The three channels in the Metallic BSDF node corrispond to Cycles R, G, B channels, and thus to get values for it, you need to get the three n, k values, one for a red wavelength, a green wavelength, and a blue wavelength (ideally with wavelength values that match Cycles as closely as possible. And sorry I do not know which wavelength values match Cycles the closed.)