FOG interaction with Portal On - Off comparison

[sikotik13]

May I ask why, after reading the thread linked previously, you still seem to expect it to look different? The portal, by apparent design, looks to be intended solely to make the path-tracing more efficient. That would just make rendering to the same point faster/cleaner, not change the image. Am I missing something implied, or are you doing these tests to see if the image would change?

According to the post you linked above, if it does, it's a bug, and not supposed to happen.

Just curious, as I am trying to learn more about all the lighting stuff myself.

[linvanchene]

edited and removed by user

[sikotik13]

Interesting, I am apparently getting different results than other listed here.

No Portal

Portal Off

The only difference I see in both No Portal and Portal Off is that the scene is far noisier at the same resolution, indicating to me that the portal is functioning exactly as it's supposed to. They also both appear visibly darker to me, and the light seems more concentrated around the window, less on the floor and walls themselves. I suppose given some other variable, it could appear lighter, but light that is unfiltered tends to be closer to white, hence why white spots appear until resolved enough to smooth them out on textures. They look pretty much identical to the portal on render when it was just over 1k s/pix.

Portal On

Portal On looks drastically cleaner to me, and the light appears to be more spread out, as it should be. Guess ymmv or something. *shrug*

-Note: since the file names got messed up a little, all were rendered to 4,000 s/pix.

[DrHemulen]

Which kernel are you using? I think Directlighting is cheating a lot, there's light in the room, even when it's fully closed (when the plan isn't a portal but just fills the "window").

[sikotik13]

Ah, I didn't realize it didn't show the original names at all. It's all default PMC kernel, since I gathered that portals are entirely based on path tracing, thus direct lighting wouldn't use them anyway, as it's biased.

[linvanchene]

edited and removed by user

[sikotik13]

Best guess from looking at images: you are confusing the hell out of the program with normal inversions stacked with normal inversions. Would explain the odd square with the portal off, the normals are still reversed of the reversed sphere. Portal on would probably just send it into spastic fits since there are now normals in every direction it's supposed to be tracing. I'd give up too, lol.

Maybe someone more hands on with the innards of Octane could tell you why the crazy effects, that was my best guess (and I'll probably stroke out if I'm right, lol). They do look neat though!