radiosity raytracer???

[STRAT]

whats ppl's thoughts on this?

 

http://www.cgtalk.com/showthread.php?s=&threadid=61157

[John Dollus]

Most popular solutions today use a blended approach to achieving the final image.

[Paul Griger]

Here's a copy of the novel that I posted over at CG Talk

 

I have always understood raytracing and radiosity to be two completely different processes. Radiosity and photon-mapping are forms of Global illumination. Raytracing is not a form of GI.

 

I can only define them best by quoting Jeremy Birn’s excellent book, “Digital Lighting and Rendering”;

 

Raytracing =

“simulates the natural reflection, shadowing, and refraction by 3D surfaces. … the rays for Raytracing begin at the camera…”

Again, Jeremy defines GI simply as =

“illumination that takes into account light transmitted from other objects.”

So radiosity is a form of Global Illumination because Radiosity =

“an approach to rendering indirect light, in which light is transmitted between surfaces by diffuse reflection of their surface color.”

So by definition Raytracing isn’t a form of Global Illumination because it doesn’t calculate the level of light in a scene. It only calculates the surfaces that can be visible or not visible via transparency, reflection or hidden in shadows. Raytracing doesn’t light a scene AFAIK

 

Yes, I guess one could say that GI’s indirect light solution is “Raytraced” because the indirect rays of light are being “traced” from one object to the next and that information is then fed back to the scene camera. But the two terms shouldn’t be synonymous.

 

Now RE: C4D, AFAIK, C4D calculates GI via Photons, not true radiosity.

 

Radiosity is =

“calculated by storing shading information for each vertex of the polygonal objects in a scene. … the resolution of your geometry is linked with the resolution of your GI.”

Where as lighting w/Photons a =

“GI solution that depends upon a number of photons… emitted from a light…bouncing between surfaces.” This creates a photon map that stores the GI solution. .”

To me that sounds more how C4D does GI, but I could be wrong.

 

It is interesting that Jeremy notes this:

 

“the name game: Some people define radiosity broadly to mean “a rendering principle that calculates diffuse reflection of light and color among surfaces in a scene.” Based on this functional definition, photon mapping could be said to be a type of radiosity. However, many computer scientists reserve the word radiosity only for specific algorithms, and refer to photon mapping as “an alternative to radiosity.” You are safer using the blanket term global illumination when discussing all these indirect lighting processes in general.”

for me, I don’t have much of a problem if people confuse different forms of GI, whether it’s radiosity or photon mapping. AFAIK, Lightscape is one of the few programs that probably uses true radiosity algorithms.

 

What gets me is when people interchange the words radiosity and raytracing.

 

Here is an example.

 

http://groups.yahoo.com/group/Chief-Architect-ArtLantis/

 

This group uses Art*Lantis. I previously used [and own] A*L and it is a nice program. It even has a setting within the program that lets you set the “GI”. But the program doesn’t really use GI or radiosity, it uses local illumination to calculate the light in a scene, sort of like our Environment Object and setting the color and strength of such along with direct lights. But this is not GI or radiosity! Yet, the users at the forum say this:

 

“Post your Opt files for others to view and learn from, lighting techniques, radiosity settings and anything else that would benefit others.”

Who can blame the users? It’s obviously a confusing issue for many and is propagated by software companies that use these terms arbitrarily.

 

It just would be nice to see the terms Raytracing, Radiosity, and Global Illumination used with consistency. Otherwise you get people with knock off programs thinking that they can do photo-real scenes lit with GI, when that really isn’t the case.

 

BTW, the users at the above forum are comparing A*L w/Chief Architect 9.0 which now has POV-ray integrated to do “radiosity”. Now that it has been released though, users are complaining about being frustrated because it is so hard to get any descent renderings from CA 9’s “radiosity” engine. Plus “radiosity” rendering times for CA 9 are running from half a day to days and counting for CA 9. What a surprise. Thank you for C4D!

 

 

So as the future brings GI type tech to less expensive programs, more people will be using programs that are claiming to have capabilities similar to more professional programs. And these users will be expecting similar results as the professional rendering programs, only to get frustrated and give up.

 

 

On second thought, maybe that’s not such a bad idea so the real rendering is left to the pros

 

 

Best,

 

paul