Prism Effect - Light Dispersion

[archidork17]

Can anyone suggest how to simulate a prism light dispersion in V-Ray for Rhino? I have a tutorial from Maxwell, but want to create the same effect in rhino with a small animation where the source light changes position around a prism.

Any help would be very appreciated!

[dalomar_cg]

Re: Prism Effect - Light Dispersion

To fake dispersion you will need the v4 beta. Make a normal refractive/glass type material, and then add two more refraction layers (for a total of 3). Make sure that the IOR values of each layer is slightly different than the previous ones. Now you will need to set a transparency color for each one. Instead of using gray scale values as you normally would for transparency, you will need to set the color of the layers to red (255,0,0), green (0,255,0) and blue (0,0,255). It doesn't matter which color you use for each layer as long as each color is used once. Lastly, but most important, in the material options you will need to check Disable Volume Fog. That should do it, hope that helps.

[Travis Serio_cg]

Re: Prism Effect - Light Dispersion

i still cant make that work. in brazil it works perfect, but in vfr my glass materials still turn black with multi colored facets.

[dalomar_cg]

Re: Prism Effect - Light Dispersion

I've noticed that it depends allot on viewing angle and the amount of internal reflections. Below is a simple example. Nothing changing except the angle. The first result is rather predictable, while the second is very dark. The third actually produces a very odd effect. The way dispersion is done with vray for max is very different(a blend material I believe), so I'll play with that tonight and see if I notice any major differences. It almost looks like a glitch in how we are handling the fresnel falloff. I'll talk to Joe about it tommorrow.


Predictable


Dark result


odd result

[Travis Serio_cg]

Re: Prism Effect - Light Dispersion

yeah, i get result 2 when trying to do it with diamond shaped geometries. However I could make it work with a simple sphere.

[archidork17]

Re: Prism Effect - Light Dispersion

Yes, I'm getting the same results with a prism as Dalomar. I'm using caustics because I wanted to get RGB split behind the object but it doesn't seem to work. I'll try with a sphere next.

[archidork17]

Re: Prism Effect - Light Dispersion

Here are some tests with different positions of the light source around the objects.


Light behind the objects.


Light in front of objects.


Light to the right of objects.


Light to the left of objects.

[dalomar_cg]

Re: Prism Effect - Light Dispersion

So in thinking about this a little bit I think I may have an idea about why this might be happening (or I'm completely wrong). See when changing the transparency to a color rather than a grayscale value, this is effectively filtering the effect of that layer. The slightly different IOR values cause the difference in color thus causing the dispersion effect. However this is essentially a negative process because it discards the information that is filtered. This would lead to certain areas of refraction that get completely discarded, thus leaving it black. In vfmax, dispersion is created through an additive process with the vrayblend material (which we don't have). The max way of doing dispersion is very similar, but because the effect is additive and not negative this prevents the black refractions from appearing.

I'll run this by Joe, and we'll see what comes of it. Dispersion is something I've been looking to get working for a while now, and although its close its still not what it should be.

Here's the CG thread if you wanted to check it out http://www.chaosgroup.com/forum/phpB...asc&highlight=

[joe_b]

Re: Prism Effect - Light Dispersion

Vlado told me that creating a dispersion effect requires you mimic the different wavelengths of light by making multiple layers with varying IORs. Perhaps if someone did a little research on the actual differences between the colors in nature they would get a better result? Maybe simply approximating to +- 0.1 IOR for each layer is too much of a leap and is what is leading to these odd results. I am not sure.

But the dispersion material I was told was to create 3 refractive layers each with slightly increasing IOR. Make the transparency color for each layer to be the white complement of red then green then blue. The white complement means 255-x for each component. Also you need to check the option that disables volume fog in our options page. Otherwise the shader would just average the 3 IORs instead of keeping 3 separate ones.

Maybe I can get Vlado to weigh in on this thread to help clear up any mistake. Its possible our shader isn't handling something correctly.