I’ve been studying up on lighting this week and I can find no practical difference between Environment Lighting and Ambient Occlusion. What I mean by ‘practical’ is…
I took a scene with a few objects and rendered two versions, one with EL and the other with AO, each set to 0.5. No other lights are present in the scenes. To my eye, they look identical. So, from a practical POV, there doesn’t seem to be a difference.
I found an old thread on here that explained: “AO has to do with objects / surfaces close to gather, there by reducing lights reaching them. You get dark corners. Environment lighting is like fill lighting. It brighten up shadows.”
But, I really don’t see any difference in the final render. Is there another factor that needs to be added in (another light, maybe?) before the difference can be seen?
I’ll continue my studying/experimenting, but if anyone knows of situations where one is better than the other, please post.
EL is a fill light coming from all directions, faking environment and indirect, AO is the shadowing of indirect light. So technically speaking, AO is supposed to be multiplied by the environment lighting pass, and the result added to the direct light render. Turning it on in the world panel in “add” mode has a slightly different affect, where it’s just added to the direct lighting to fake ambient light. (resulting in the fully shadowed areas not changing, and the brightly lit areas getting brighter).
(ELxAO)+beauty will allow you to have colored EL while still getting contact shadowing.
Another facet of EL is that if a World is set up with a photographic image like an HDR, it will provide light sources based on sampling of that image. This is beneficial for scenes that should have a distinct color structure to the entire environment, like a forest, a deep sunset, etc., where the lighting is determined not by a single direct source like the sun but instead by less direct light and “bounce” from the environment. You can do the same with flat RGB World coloration but it ends up flat indeed, like tinted GI.
J, Chipmasque, I see some tutorials that need to be written here. (Hint, hint …)
One thing that I would add to the discussion is that you need to get to know the Histogram and the Color-Scope tools … the things which allow you to objectivelymeasure the color and lighting within the shot. Your “eyeball opinion” only goes so far: your eyes become tired, and they also become conditioned, and the qualities of your (perhaps un-calibrated) monitor also have their effect. But the graphs provided by these tools are based on actual data. You should rely upon them more than your eye.
Another useful trick is to take two versions of a shot, put them into Gimp or Photoshop as two “layers,” then subtract one from the other to produce an image that illustrates the differences. (Try it both ways; both are informative.) Once again, you’re using the digital computer to do what it does best.
Objective tools help you to quantify what you are seeing … and to illustrate what’s wrong (or “not [quite] right”) about a particular version … doing so in a way that points to a solution. (Your eye is generally better at recognizing “this isn’t right” than at knowing what to do about it. After all, if you’re ancient-man, it’s a lot more important to sense that there’s a tiger (or “something” …) in the grass, given that your objective is “to not be eaten by whatever-it-is” … than it is to correctly identify it.)
@sundialsvc4: I have to respectfully disagree with using data to judge whether or not your lighting looks good. People who spend all of their time looking at technical aspects of Blender end up never making anything, because whatever brilliant knowledge they think they have, it usually isn’t exactly true when it comes time to rendering(not speaking about you, just in general of what I have seen around here). Anyone can look at data, or what something technically is supposed to do, but every scene is different, and what something is “supposed” to do, is usually wrong, depending on the scene of course. Lighting depends on materials and textures, objects, the scene, etc., and the opposite of each of those things is true as well. Same goes for using AO or Env. This is kind of why a lot of people’s Cycles renders don’t look any good, because they are stuck staring at 20 nodes for one material instead of actually looking at the overall picture, where every single thing usually depends on every other single thing. Scenes change too much to just rely on data alone.
As for what chipmasque is saying, you just add an HDR to your world, change the environment settings to “sky texture”, adjust falloff and samples, and you’re basically done. It usually ends up quite noisy, so more samples are normally needed in the end, and nothing is going to tell you that more than looking at it and going “holy crap this is noisy as hell! More samples!”
Simple answer: AO is not a light. It’s meant to be ‘multiplied’ to darken the corners and places where there is no light - it is the occlusion of ambient light. (The ‘Add’ mode is a way to fake Enviroment lighting and should be considered evil)
Enviroment lighting is just that - the light that comes from the environment. It is a light source and thus brightens (adds, not multiplies) the image. It can also use an image (not only HDRs) to define what colour and brightness comes from which direction - AO is not a light and thus has no colour.
@rontarrant: Not for me it isn’t. Not everything needs GI, 3-point lighting, OR specularity. “Are there other approaches?” YES, many! sundialsvc4’s general idea is not bad, you do need to pay attention to some of the things he mentioned, just not necessarily in that order. I’d get “3-point lighting” out of your head unless it really applies.
