Hi Guys,
Playing around with some scene assets I came across something that I was unable to solve efficiently. The scene features a crystal chandelier, and only that. Black environment, and nine small lightbulbs. As far as know, everything is set up to real world values. The features I'm interested in here are the small bright colored highlights (pronounced here with some post flare) produced by multiple dispersed refraction of the tiny bulbs inside the lamp. Rays need to find those small lightsources through many many glossy refractions, and obviously, this needs a tremendous amount of samples to clear.
So, my question is: theoretically what kind of rendering algorithm would be the best to solve this scene? Not just what is accessible right now in VRay, but in general. I'm betting something like PPM, and initially I was guessing MLT as well, but a friend of mine who works mainly with MLT based solutions gave it a try and got even worse results than what I was able to pull off with VRay.
Simple curiosity.
Cheers,
A.
Playing around with some scene assets I came across something that I was unable to solve efficiently. The scene features a crystal chandelier, and only that. Black environment, and nine small lightbulbs. As far as know, everything is set up to real world values. The features I'm interested in here are the small bright colored highlights (pronounced here with some post flare) produced by multiple dispersed refraction of the tiny bulbs inside the lamp. Rays need to find those small lightsources through many many glossy refractions, and obviously, this needs a tremendous amount of samples to clear.
So, my question is: theoretically what kind of rendering algorithm would be the best to solve this scene? Not just what is accessible right now in VRay, but in general. I'm betting something like PPM, and initially I was guessing MLT as well, but a friend of mine who works mainly with MLT based solutions gave it a try and got even worse results than what I was able to pull off with VRay.
Simple curiosity.
Cheers,
A.