n and k curve data for refractive indices (IOR)

[DinoMuhic]

We need a more sophisticated way to apply IOR data (refractive and reflective indices), best would be as n and k data or curves for each RGB channel.

At the moment its really tiresome to build youself the nodes for each channel and apply gradient and fresnel nodes to each of them and then you still have to convert the n and k data somehow into curves before doing that.

[DinoMuhic]

To add: This addition could be incoorporated into the falloff node, using splines for each channel.

[bepeg4d]

Hi DinoMuhic,
when OSL will be integrated, we will surelly have more options for the Index parameter and much more.
In the meantime, please, look at this lua scripted graph for glossy material with n and k:
viewtopic.php?f=73&t=43973&start=10#p270402
You can save the metal material in your LocalDB and use it with any plugin.
ciao beppe

[DinoMuhic]

Hi DinoMuhic,
when OSL will be integrated, we will surelly have more options for the Index parameter and much more.
In the meantime, please, look at this lua scripted graph for glossy material with n and k:
viewtopic.php?f=73&t=43973&start=10#p270402
You can save the metal material in your LocalDB and use it with any plugin.
ciao beppe

Hey bepeg4d,

thanks for the link, I'm not sure though if it will work with C4D too but I'll try.
However this is such a fundamental thing it should be coded directly in the core app and not rely on lua, OSL or anything else, since that stuff (as far as I know) can never be calculated as fast as directly written code. Correct me if I'm wrong.
I have some experience with OSL in cycles4d and while it offers a LOT of possibilities, it can sometimes be very slow.

[Terryvfx]

You said fundamental but if you think that a realistic Fresnel is what's keeping you from realism you are sadly mistaken, I rather have the BRDF take into account the roughness value to decrease the Fresnel effect of the material (glossy fresnel in Vray) but that's just me.

Now there are a few resources out that you may or may not know to tackle this problem to some extent:

http://www.entagma.com/c4d-custom-condu ... c-fresnel/

https://vimeo.com/102082368

[Goldorak]

Hi DinoMuhic,
when OSL will be integrated, we will surelly have more options for the Index parameter and much more.
In the meantime, please, look at this lua scripted graph for glossy material with n and k:
viewtopic.php?f=73&t=43973&start=10#p270402
You can save the metal material in your LocalDB and use it with any plugin.
ciao beppe

Hey bepeg4d,

thanks for the link, I'm not sure though if it will work with C4D too but I'll try.
However this is such a fundamental thing it should be coded directly in the core app and not rely on lua, OSL or anything else, since that stuff (as far as I know) can never be calculated as fast as directly written code. Correct me if I'm wrong.
I have some experience with OSL in cycles4d and while it offers a LOT of possibilities, it can sometimes be very slow.

OSL on CPU maybe is the issue there? In Octane 3.1, OSL on GPU is custom compiled by us for max perf. so that new OSL shader code works as fast as the previously black boxed 3.0 shaders.

Re: Lua Script nodes, like the one mentioned above, which just provide procedural values to the material node are very fast, and would not be any faster if we baked it in - that may be moot in 3.1 when OSL can do all this and more. Even so, in cases where lua is not enough, we will soon have a native C API and native module system, for high speed procedural geometry and image operations that can be shared across plug-ins through the node graph.