PDA

View Full Version : Diamond Render



jimiclaybrooks
07-21-2017, 12:19 PM
Hey guys, I've been online looking for any tuts on how to render a diamond using the dielectric node. Since I'm new to the node editor, I need some help with this. There's a blender tutorial that talks about the prism effect that diamonds have, and their refractive quality, which is
what I'm after in Lightwave. As you can see here, my render looks too much like glass, and I'm not understanding why. 137430 Can anyone point me in the right direction? thanks.

Iaian7
07-22-2017, 12:32 PM
Unfortunately the Dialectric node doesn't include a dispersion option. You can use the Reflection/Refraction nodes to build your own setup, however, and enable dispersion that way (it's a percentage setting just below blur).

jimiclaybrooks
07-22-2017, 12:56 PM
thanks for that insight, much appreciated. I'll certainly give it a try.

Iaian7
07-22-2017, 01:21 PM
Well, I was totally wrong (I checked before replying, and still completely missed it). The Dielectric node DOES have dispersion, and to use the setting you need to add some amount of surface roughness. Any amount, even 0.1% works, then you can add as much dispersion as you want.

In the Basic tab:
137444

In the Advanced tab:
137443

prometheus
07-22-2017, 01:58 PM
Well, I was totally wrong (I checked before replying, and still completely missed it). The Dielectric node DOES have dispersion, and to use the setting you need to add some amount of surface roughness. Any amount, even 0.1% works, then you can add as much dispersion as you want.

In the Basic tab:
137444



In the Advanced tab:
137443


Thatīs what I thought :)
Also..do not have roughness set to zero, it needs a bit of roughness, the higher roughness the more softer blurred refractions, also try set a high amount of refractions blur to see it kick in properly, then tune down.

Iaian7
07-22-2017, 02:27 PM
It's interesting to note that the dielectric node renders prismatic effects very differently: the dispersion colours are sharp instead of blurred, it does not affect GI, and of course internal reflections are included using the partial internal reflections option. With a simple setup using the Refraction node, however, the dispersion colours are blurred (smeary) and light bounced through refraction in the GI takes on the dispersed colouration (easily seen in the third frame below). Either method can be overdriven (using very high percentage values), but the Dielectric always seems to render sharp faceted colours. Depending on the look you want, choose accordingly.

137447

137446

The third example combines a Reflection and Refraction node using a Mixer node controlled by a Fresnel node, which is then piped into the Specular buffer...it's the easiest way to bypass all the unnecessary parts of the surface while using the Diffuse, Specular, Reflection, and Refraction buffers for different layers of a render. If the dispersion setting is dropped from 500% to 50%, it renders exactly the same as the second example.

jimiclaybrooks
07-22-2017, 02:29 PM
Thanks guys, I appreciate your help with this. At present I'm bogged down with a model that I have to finish, and then I'll be able to get back to the gemstones. Again, I appreciate your help. I much prefer the middle sample, but the third most colorful sample might be the one I'll end up using if I can get this right. But all three are really nice.

prometheus
07-22-2017, 02:34 PM
It's interesting to note that the dielectric node renders prismatic effects very differently: the dispersion colours are sharp instead of blurred, it does not affect GI, and of course internal reflections are included using the partial internal reflections option. With a simple setup using the Refraction node, however, the dispersion colours are blurred (smeary) and light bounced through refraction in the GI takes on the dispersed colouration (easily seen in the third frame below). Either method can be overdriven (using very high percentage values), but the Dielectric always seems to render sharp faceted colours. Depending on the look you want, choose accordingly.

137447


137446

(also, that the third option simply combines a Reflection and Refraction node using a Mixer node controlled by a Fresnel node, which is then piped into the Specular buffer...it's the easiest way to bypass all the unnecessary parts of the surface while using the Diffuse, Specular, Reflection, and Refraction surface settings for different layers of a render)


Do I misunderstand you here? I get blurred dispersion colors if I raise the roughness in the dialectric material, but that may not be what you ment?
Great samples by the way.

And you use no air polys, one singlel surface, and not double sided?

Iaian7
07-22-2017, 03:12 PM
Do I misunderstand you here? I get blurred dispersion colors if I raise the roughness in the dialectric material, but that may not be what you ment?
Great samples by the way.

And you use no air polys, one singlel surface, and not double sided?

Thanks! If you raise the roughness, yes; but that also blurs the refractions which may not be desirable. If the goal is a sharp refraction with exaggerated dispersion, that's where the differences really start to show. With both Dielectric and Refraction nodes set to identical values (0.1% roughness or blur, 250% dispersion), we get the following results...

Dielectric material (note the sharply defined colour shapes):
137448

Refraction node (note the blurry colour patches, while maintaining the sharp faceted refraction):
137449

However if you try to blur the dispersion of a Dielectric material using increased roughness (2.5% in the example below), you get the following results (note the globally blurred effect with no sharp facets visible, almost like the interior is frosted):
137450

So the Dielectric node allows you to create very sharp facets of prismatic colour (top example), while the Refraction Node renders a much softer dispersion without blurring the refraction (middle example). The bottom example doesn't seem so helpful to me when it comes to gemstones or prisms...it lacks the clarity and sharpness that might be desired (depending on art direction!).

And no, there are no air polys, it's just a single surface, and it's not double sided...thus the lack of internal reflections on the non-material setup right now. Feel free to download the scene files above if you want to see how its setup.

Iaian7
07-22-2017, 03:38 PM
For completeness of testing, this is what it looks like when the Refraction node's blur setting is increased. You have to use settings 4x-10x stronger than the Dielectric material to match a similar level of refraction blur, and even then, the results may have more dimensionality (the refraction retains more linearity and feel less broadly diffused to me).

Refraction node, 2.5% blur (minimal change from 0.1%, but noticeable on a few edges):
137451

Refraction node, 10.0% blur (just about matches the Dielectric material's 2.5% roughness, but with more angularity):
137454

Refraction node, 25.0% blur (finally matches the lack of any sharp edges in the 2.5% roughness Dielectric example, though far blurrier in other aspects):
137455

jimiclaybrooks
07-22-2017, 08:54 PM
Thank you for the sample scene, it'll be a great help, along with this thread.

prometheus
07-23-2017, 03:32 AM
Thank you for the sample scene, it'll be a great help, along with this thread.

Iīll chime in and want to express thanks to ian for the scenefiles, In cases like this...we would surely like to have the next lightwave version out, that should be able to give the same results in VPR as in final renders, cause playing around with glass and diamonds with vpr makes it really not even worth using that much.

I also need to take a look at octane for diamond renders as well, and take notes on what it does better or not.

madno
07-24-2017, 03:46 PM
Hi,
not the OP but thanks for this thread. Parital Internal Reflections can really make a difference.

Ialan7 custom 2 surface:
137465

Dielectriec - no part refls
137464

Dielectric - with part refls
137463

A render with Dielectrict (no disperson color in GI caustics - but anyway, caustics are not prominent in the render)
137466

gar26lw
09-27-2017, 06:14 AM
Fantastic