From what I know both Radiosity and Ambient Occlusion are methods for simulating global illumination in a scene with ambient occlusion being the least accurate.
I would like to ask in which cases is it better to use ambient occlusion and when radiosity and why to choose so??
From what i know ambient occlusion is better suited for use in outdoor scenes as a general skylight whereas radiosity is better for indoor scenes(correct me if wrong)
Please post the exact difference between them
radiosity is the old system in Blender which can be used to bake color and shadow information into vertex colors.
it is good for stills in the old time and good for level design. but not good for animation
or models with a lot of details
it is probably replaced with the photon mapping which is more precise, but maybe slower.
ambient occlusion is a way to darken or brighten areas of geometry depending on how
far they are apart from each other. this does not take color into consideration.
very good if approximate ac because it is blazing fast for animation. not very realistic
but for motion picture pretty sweet.
I have found many times that I generally prefer the AO look to the GI look…it seriously depends on the style though.
Well looks like the do almost the same thing but radiosity takes also colour into consideration and AO is just an approximation algorithm for GI. Am I right ??
If you were to give standar definition for each one what would that be ?? I am searching the whole web but I can’t find any !!! I need two definitions that show their difference !!!
mh no not at all
AO is a fake GI because it darkens cracks or the area where surfaces get close to each other.
this is in some systems also called as dirt shader.
GI is different - GI deals with indirect illumination based on light tracing or photon mapping.
this creates the effects of color bleeding, diffuse and indirect illumiantion, and also caustics.
GI is also local meaning it depends on the light source.
AO is global wide, you can only tell a material to be more or less affected.
Ive always wondered if there was a “cheap” way to do directional AO with colour bleeding. Crytek has managed to pull it off with ther light propogation volumes malarky before, even claiming to have realtime GI.
nor really
you can fake it with extra lamps with specular turned off and using layers to hit not the main object itself.
When is GI going to come to blender’s internal renderer? I can’t install any external ones, so I can’t get any photo-realism just yet I guess.
With the right setup in the Blender compositor you can actually squeeze some additional realism out of AO renders. Better, but fully intergrated true GI will erase that need.
it will take some time till the code is ready
Why?? .Python versions and OSX.
You can install external renderers.
I’ve got Lux, Yafaray, and Sunflow working just fine here.
I was outlining the reasons for my problem. Not the problem itself directly… regular blender needs a certain version installed of python, all the external renderers need an older one (for macs), and you can’t have a different version installed as well.
why can’t you have different versions installed? OSX is a Unix. Learn to use its powerful command-line and environment variables. You can then have your pythons installed anywhere…
Im still confused as to the difference.
AO no light source, lights up everything, in that, you dont need another light in the scene.
radiosity, Im aware VUE uses this to create realistic light. its a light source but is meant to be used as a sky dome to represent global light?
hemispheric light, another way to create in-direct lighting, its meant to represent the light from the horizon, not the direct sunlight.
GI- whatever the hell global illumination is. sounds like more of the same.
So pardon my naivety, I just dont “get” it.
answers in English please.
these are all techs to do parts of the same thing, I’d say…
Please correct me if I’m wrong, but that’s how I understand it:
AO only lights according to the sky
hemispheric light more or less equals AO but probably does a different approach to reach the same goal
Both neither take into account light sources nor mesh emitters.
Radiosity does NOT take the sky-light into account (except if you create a hemispheric mesh, textured with a sky and set to emit light) - nor does it take into account lightsources. It only works with Meshemitters (or in theory area-lights)
GI basically is the best of both worlds + the world of lightsources. It calculates any bouncelight, no matter, wether it’s from a sky, a mesh emitter or a lightsource.
Though, even if all of them would work for all situations, their results still would look kinda different, it seems.
GI produces rather smooth lighting, most often while AO generates very sharp shadows. It happens sometimes that both techniques get combined to give the actual desired result.
That’s how I understood it 
Of the four techniques, GI seems to be the most realistic and the most expensive…
ao basically works like that:
at every location of a pixel in the scene shoot randomly directed rays to the sky. if the ray reaches the sky -> lit, if it doesn’t -> not lit. the average of all results gets used as a lighting value for the pixel.
it’s a very crude approximation and doesn’t really work for indoor scenes.
real global illumination is a lot more complicated.
They are different techniques to simulate indirect light contribution to a scene.
AO is totally fake: it uses no light or color information, it only relies on how close objects are to each other and then “occludes” the space between them with fake shadows, regardless if there’s light directly hitting that point or not.
Radiosity if mostly a mesh-only technique that shades (high) polygon model vertex with color and diffuse light contribution from close objects, starting from an origin light source. It can simulate only diffuse indirect light contributions, not caustics.
GI techniques handle it all, including caustics and subsurface scattering. Photon mapping is one such technique and it more complete than radiosity and faster than what comes next.
And what comes next are physically-based, unbiased techniques, like path tracing and bidirectional path tracing, which ditch the several passes approach of separate and approximate direct, indirect, diffuse, reflective and other parameters developed over the years in favor of physical correctness. They employ Maxwell electromagnetic equations and integrate them over the scene by stocastically tracing light paths through the scene to sample it in countless iterations.