Please add procedural animated caustics

[Elvissuperstar007]

says the photon engine is fast?)) he wants to render 2000 samples with low rendering resolution in almost 2 hours and this is on 4090

[nuno1980]

says the photon engine is fast?)) he wants to render 2000 samples with low rendering resolution in almost 2 hours and this is on 4090

Yeah. But what are kernel settings?

Can you share your one stone (eg: diamond) but should export to OBJ or Blender?

[Elvissuperstar007]

it’s not the stone, the stone is configured according to the physics of diamond, there are a lot of them scattered in the scene, and the scene itself has 19 gigabytes of video memory

[nuno1980]

it’s not the stone, the stone is configured according to the physics of diamond, there are a lot of them scattered in the scene, and the scene itself has 19 gigabytes of video memory

Thank you. But your ior of diamond is incorrect but yes ior should be 2.384 because you don't forget to use "dispersion coefficient". Or if you want to change "dispersion mode" from "disper. coeff." to abbe number with 65.6 and then you stay ior 2.417.

But I will test to render your FBX...

[skientia]

Was splitting the "scene" in two rendering kernel ever considered?
This is common in VFX and animation productions. Splitting "scene" elements for them to be rendered separately for more efficiency and optimization purposes. Here, the caustics of the "diamonds" could be rendered with all other elements as matte, then composited to the "normal" render that uses Path Tracing kernel and is the inverse procedure: diamonds are matte and everything else is "normal". This is an over simplified explanation of the workflow.

There other factors to consider. Scale is one, but also lighting.
Here, the lighting is soft similarly to your other post. Soft lighting makes caustics barely visible. This is a physical phenomenon, not a "renderer flaw". The more sun-like the light is (sharp on a cloud-free sunny day), the easier caustics are calculated and sharper as well (blurry lines).

Another way to think of caustics is the correlation it has with shadows when it comes to lighting:

[Elvissuperstar007]

By the way, I divided the phonon part, the Path Tracing part,

I can archive the entire scene, but this is 3D max, and you can test it yourself, caustics do not appear as quickly as it seems, but it’s not even a matter of caustics, in photons the rendering of the scene occurs more slowly. I set the dispersion as I saw in a wild diamond, I am a jewelry maker and on a clean stone I set it to 0.02, but for a wild one 0.01 is better.
This doesn't change the question, why not add procedural caustics? I am sure that all people associated with cinema and animation will only say yes! Yes, we can use texture sequetion from other programs, but it may not be stable, any animation rendering that uses sequetion in parallel can break and crash. And most importantly, it takes up precious video memory in difficult scenes.
Most of the procedural maps added to the octane library are meaningless, no one uses them, but this one will

[Elvissuperstar007]

viewtopic.php?f=27&t=84470
The caustics work depending on the size of the source, I know this, it’s not a matter of caustics, but a fake effect for animations, At one time we promoted realistic metals and hair, now my goal is for developers to create caustics like this

[nuno1980]

Sounds... your ultra-low perfornance and too high VRAM size used are abnormal.

I got ~33 MS/s and only 3 GB VRAM used on my RTX 4090:

diamond stones test.jpg

lol

[Elvissuperstar007]

I didn't write anywhere that there is no caustics, I wrote that heavy scenes with caustics are very long, and I need fake ways to get caustics that are hidden behind the wall, why should I create a whole ocean for the sake of caustics behind the wall? If it is hidden, and I only need the effect on the ceiling from this ocean, I don't understand why you are laughing! You haven't even seen my scene in your eyes, and you're testing five stones that I have separately also render quickly. you are laughing at the one who helped make the direct engine more realistic, if I write what is necessary, it means it is very necessary.
just look at the library of octane shaders, who made them, and then smile, it was all done long before octane could do anything, everything was done fake by creating test scenes and attempts to repeat the original. , these were distant 2009-2013

But it should not just be in the procedural maps, but it should be a function in the light sources of variations of procedural light maps, one of them is caustics with animation and dispersion function, it is like absorption parameter only fake. Why should it be in the sources? People rarely go to the procedural map library and don't know and IES maps should change shape in real time without loading these maps, this should all be the base of the light source!

[nuno1980]

Your scene is heavy. How many polygons does this have?

Am I laughing? No, "lol" doesn't mean laugh. But yes "lol" means the head horizontally and slowly quaking (= "no") + "happy".