Gradient glass and caustics

Started by polo, April 14, 2022, 06:11:38 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

polo

Hello,

I would like to render a piece of glass with a gradient in it and to see the caustic.
But I encounter two problems:
-first I can't create a "glass solid" material with a gradient color in it, I managed to sort of bypass the problem by creating a "dielectric" material but I feel like it doesn't look like glass exactly...

-Second problem is I can't see the caustics, I obviously toggled the caustics in the Lighting mode but nothing significant is changing...
I would like to create an effect similar to the attached picture:


Thanks,

Paul

INNEO_MWo


Anindo Ghosh

For the dielectric not looking like glass:
- If the IOR matches that of the glass(~ 1.47 to 1.65), and other parameters are at the dielectric defaults, dielectric looks pretty much indistinguishable from the built-in Glass (Solid) material, to my eyes.
- In the context of the tumblers in your reference image, for the ring-like light and shade, a gradient on the transmission channel won't yield that. Instead, a bit of Displace geometry using Marble textures with some subtle noise works for me, see my sample render and BIP file.
- Adding a ground plane didn't help much, my attached render is without any ground plane. Perhaps I'm misunderstanding the purpose the ground plane should be fulfilling.
- I am on KS 10.2, so couldn't open your BIP. However, here's mine, attached, if it helps.

One more thing: When using a color gradient for the Transmission channel of the Dielectric material, using 100% saturation results in a drastic difference in appearance, compared to using 99% or lower, for any of the color points of the gradient. Please see the material graphs for the two tumblers in my BIP file.

Will Gibbons

Here's what I came up with... not a perfect match. This did expose something though... I found that texturing the transparency distance didn't yield any results. Not sure if it's a bug or not. Regarding the caustics...

1. Use bump textures (displacement isn't needed in my experience) to build up the textures you want. I used the radial brushed as well as some noise textures.
2. For lighting, I used a spot light with almost pure black HDRI. Add a ground plane as either diffuse or plastic instead of using the built-in ground plane.

Those should help get closer to what you're after.

polo

thanks for your answers,
-yes I added a ground plane, and the IOR of the glass is around 1.50
-I've also put a bump texture (displacement or bumb didn't change, but bumb was much faster to process)
-I added a spot light and I see a beginning of caustics, but I mostly see the dark shadow and the stronger I put the spot, the harsher the shadow is.

Will gibbons--> even by adding this effect I didn't manage to get those subtle lines
+ it seems that you both don't have too much harsh shadows around, is there something I'm missing? a box to tick somewhere maybe...

INNEO_MWo

first try - change falloff and radius from the light material to zero and render again. Do you see harsh shadows?

Will Gibbons

I suspect that something is off...

1. Not sure if you're in photographic mode, by try using linear response curve... looks like your image is very washed out. This should make your highlights brighter
2. The smaller the diameter is, the sharper the details in the caustics will be
3. The bump height of each individual texture node may be too low to see a result
4. Is your model to scale? Your spot light is super bright... that's making me think your scaling is off
5. Is your model just a surface, or is there real thickness? It doesn't look like any refraction is happening in there

polo

#7
Hello
Sorry for the late answer,
1:it's was normal mode
2: ok, it seems to be better with a smaller diameter but the brightness has to be set to 200000, which is ok in this scene, but If I want to have caustics appearing into an interior scene and I have to set one pin at this value it is going to burn the entire image...
3:I tried different bump map and displacement with different height, and part of the answer was here and and it had to be between 0.01 and 0.001 to work.
4:the model is to scale, the glass part is 150mm diameter by 150mm height
5: there is a thickness of 3mm

Another mistake I made is that in the environment the ground shadows where ON, I turned them OFF
Regarding the material, all the roughness has to be 0

so here is the result, it still needs some work but It's much better and closer to the effect of the reference picture already

Will Gibbons

This looks much better than the first time around!