Welcome to Fractal Forums

Fractal Software => Tutorials => Topic started by: Minataz on January 16, 2017, 09:54:59 PM




Title: What is MC rendering and how and when should I use it?
Post by: Minataz on January 16, 2017, 09:54:59 PM
Quite new to the program and I have come across the MC rendering tab. Could someone answer the questions in the title?


Title: Re: What is MC rendering and how and when should I use it?
Post by: Sabine on January 17, 2017, 12:06:16 PM
Since I have only a vague idea  88) about the theory behind MC, a paste from wikipedia:
Quote
Path tracing is a computer graphics Monte Carlo method of rendering images of three-dimensional scenes such that the global illumination is faithful to reality. Fundamentally, the algorithm is integrating over all the illuminance arriving to a single point on the surface of an object. This illuminance is then reduced by a surface reflectance function (BRDF) to determine how much of it will go towards the viewpoint camera. This integration procedure is repeated for every pixel in the output image. When combined with physically accurate models of surfaces, accurate models of real light sources (light bulbs), and optically-correct cameras, path tracing can produce still images that are indistinguishable from photographs.

How to use it? Get your scene ready in the main  MB3D-window. Click the Import Parameter-button in the MC-window and then Start Rendering. Stop rendering when you are happy and save the image.

When to use it? That is really a matter of how much time you want to invest for a often (but not always) better lighting result. In mb3D I use it when the fractal renders Really quickly (MC is slow!) and when I am not really happy with the standard diffuse light mb3d provides. I have two images here http://sabine62.deviantart.com/journal/Monte-Carlo-vs-MB3D-515961258 which will show how much a MonteCarlo-render may differ from the standard mb3d render (all light-and colour-settings are imported 'as is' into MC).

Hope this helps!


Title: Re: What is MC rendering and how and when should I use it?
Post by: kram1032 on January 17, 2017, 02:17:06 PM
Monte Carlo Raytracing is a method that will, if implemented correctly, guarantee (with caveats) to converge towards a physically accurate result. It essentially simulates how light would act in the real world. Naive implementations of it are slow but these days some versions exist that are just crazy fast, capable of producing neat, lowish-noise images of fairly complex scales in seconds.
Until the initially inevitable noise has gone, however, it could take hours or even days: It basically depends on how far you want to go.
For fractal with their infinite detail, meanwhile, these methods to my knowledge, are pretty hard to get to work right: You'll often run into situations where you'd think there should be light, but it quickly becomes somewhat unlikely that light just happens to be bounced back your way and it rather tends to "get stuck". That's also where the limitations mentioned above come in. For instance, because fractals have infinite detail, there would inevitably be surface interference effects which may yield weird properties like ultra deeply saturated colors (https://en.wikipedia.org/wiki/Pollia_condensata) or iridescence  (https://en.wikipedia.org/wiki/Iridescence) - but with standard Monte Carlo Raytracing, such effects normally aren't considered save for specialized shaders. It would simply be too computing intensive for a technique that already is known to be rather slow.

Though the method is nevertheless quite effective as you can see in the image Sabine linked - at least if you have the patience for it! (note the render times and also that the transparent surfaces in her image aren't yet noise-free. It looks gorgeous but such artifact will take almost indefinitely to go away. The algorithm is only guaranteed to yield "the correct" result "in the limit" of infinite rendering time)

Basically, if you want very realistic lighting and have plenty of time on your hand, MC methods probably are the way to go.
It also works for quick previews: Images can be sufficiently noise-free after just a few seconds or minutes, giving you clues what to change or whether you want to go for full length.