Radiosity and GI, what's the difference?

[Per-Anders]

radiosity is a form of gi, radiosity is generally a foward casting algorithm, i.e. rays are shot from lights and bounce around the scene, whilst most other gi solutions are backckwards casting or hybrids (photon), c4d's gi is from what i can tell a photon mapped gi. this subject has been done several times here so it's worth doing a search.

[WDABUILDERS]

Even beyond the direction of the flow...

radiosity is generally mesh based to create it's accuracy. The mesh recieving the light is divided for accuracy. Lightscape for example you controled the mesh density to give the quality (accuracy) of light being raytraced around the scene. Max's radisoity is similar, C4D uses somewhat different approach but fundementally it deals with light using a method to subdivide the surfaces for accuracy, but not totally mesh based. Either way a ray of light's effect is controled by divisions on the geometry surface.

Photon based lighting the accuracy is controled by the size of the photons associated with the rays being cast into the scene. Over simplified but shows the difference.

GI is both a catch all term in vernacular accepted understanding for real world lighting in CG and used specifically for Photon based lighting. It's hard to differentiate at times.

Wax

[Carina]

I would say the biggest difference between radiosity and other GI methods is that radiosity has (in it's purest form) nothing to do with raytracing.

Traditionally, in radiosity a solution is found by evaluating the geometrical relationship between every single pair of surfaces in the scene, then using these geometrical relationships to set up a system of linear equations representing the light transfer in the scene. Basically the radiosity of a surface i is the sum of the emission from that surface and the sum of the radiosity of each surface j in the scene weighted by the geometrical relationship between i and j. So essentially, you have a mathematical solution of the light transfer in a scene.

This results in a solution that is completely view independent. Obviously the above is a very very simplified version of what actually goes on..

While you can, in theory, use radiosity to calculate glossy reflections, it's very costly, which is why you'll really never find a radiosity system that supports more than diffuse light transfer.

As an alternative to radiosity there are path tracing methods (where rays are cast from the camera into the scene, or from lights into the scene, and are then allowed to bounce around to achieve realistic lighting) and photon mapping. All of these (radiosity, path tracing, photon mapping) fall under the GI category. They're just different methods of achieving realistic lighting, each with their own benefits and drawbacks. They are different from local illumination, as local illumination only takes the primary light received directly from a lightsource (i.e. not from other surfaces) into account.

[jeremybirn]

I don't get it. I use cinema 4d and they used to have radiosity and then recently they renamed it to Global ilumination. I thought it was the same thing just different names. But then i vue5 it has scenes with radiosity and scenes with GI. And Final render says it uses GI instead of radiosity as it's better.

I am completely confused.

The posts above are correct about Radiosity being one form of GI.

In much of the 1990's, radiosity was basically the only form of GI that was available in commercial software, so among many users "radiosity" became the common word for software that renders with indirect light. Cinema 4D tries to be easy to use, and to label functions with words that are familiar to its users, so its developers apparently called its GI implementation "radiosity" because that was the word users were asking for. As users have grown up, and learned about different GI implementations, they changed the label in the interface to the more accurate name.

-jeremy

[cpnichols]

Just a little side note. For whoever is interested, the term radiosity appeared before CG. I means to measure the radiation energy off of surfaces. As you my know, radiation is any type of ray, gamma, alpha, beta, x-rays, ultraviolet, or even visible rays. Since the methods that the people above show, leave a plot on the surface of the light values that that surface is emitting, they called the GI method radiosity.

[WDABUILDERS]

LOL..yea that's right! Lightscape was based on a Heating algorythm for HVAC (heat,Vent Air Conditioning) enginneering. Lightscape being the grand daddy, in many eyes, of Radiosity.

[cpnichols]

Yeap... that is what I started doing GI on, around 1993 (for me). I giggle when people look at things like photometric lighting as something new.

[Carina]

The CG "radiosity" method that was introduced in 1984 was directly based off heat transfer research, so it's logical the terminology is based on real life physics as well.. Lovely to see how beautifully skewed some terms have become though :)

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.