Global Illumination alternatives

[alekmych]

+1 - there is no other way. Every new honorable title just have it. GTA 5, The Witcher 3, Fallout 4, The Division. Hardware also getting there. Games just become bigger, more like sandboxes and backing is not the way it should be done with lighting. I'm not saying about effects - volumetric stuff, weather system, physics. Look at the ARK - its so heavy on resources and looking so 2010-ish.

[darthviper107]

Each of those games have a solution that really only works for that game. Until hardware gets better that will allow most people to use VXGI/SVOGI/etc. then the best you can hope for is a low-quality solution that will work with only a limited number of game types.

[MrRabbit]

I haven't read someone mentions 'Image-Based Photon Mapping' (ISPM):

https://www.youtube.com/watch?v=GckOkpeJ3BY
and more info here: http://graphics.cs.williams.edu/papers/PhotonHPG09/ (with source code)

As far as I understood:
-no need for voxelization, uses the actual polygonal mesh
-not view-dependent
-gives Global Illumination, Ambient Occlusion and Caustics with one solution

It is from 2010 but I think with todays hardware, dx12 and optimizations (doing the path tracing part on the gpu rather on the cpu as the paper proposes) this technique could be even more performant.

[TriNityGER]

I haven't read someone mentions 'Image-Based Photon Mapping' (ISPM):

https://www.youtube.com/watch?v=GckOkpeJ3BY
and more info here: http://graphics.cs.williams.edu/papers/PhotonHPG09/ (with source code)

As far as I understood:
-no need for voxelization, uses the actual polygonal mesh
-not view-dependent
-gives Global Illumination, Ambient Occlusion and Caustics with one solution

It is from 2010 but I think with todays hardware, dx12 and optimizations (doing the path tracing part on the gpu rather on the cpu as the paper proposes) this technique could be even more performant.

Quite interesting, would love to see this in UE4.

[Brejlounek]

Hi, sorry to bring up an old thread.

I see that screenspace deep gbuffer method seems to work really well and fast, its only problem is being screen space. Would rendering a scene from one or two more views in a very low quality really impact performance that much? They wouldn't be full quality passes, just very basic LOD meshes (only very close ones or far big ones being rendered) and low textures of solid colors, no tessalation/skinning/complex materials etc. and really lowres render target, GI is very low frequency effect.

NVidia's current VXGI works amazingly well, but it's really slow in interior scenes for current generation of GPUs (half of the framerate for 8 cones is really bad). Would the screen space method be really that much slower in comparison to VXGI when rendering additional views in a very crude detail and low resolution?

[iniside]

http://www.aduprat.com/portfolio/?page=articles/PBGI

Something i stumbled upong today. Technique has very high quality albeit in presented implementation is very slow (200ms). It works similiar to Geomerics, but it's fully realtime (do not need to precompute geometry visibility).
Might be interesting to consider it.

[AE_3DFX]

https://www.youtube.com/watch?v=mECv52eSjBo any use to EPIC? discovered that link from Blenderartists.

[darthviper107]

That's not all that different then the voxel solution, it just uses less points

[iniside]

That's not all that different then the voxel solution, it just uses less points

Technique posted above have nothing to do with voxels. It's much higher quality, works mainly on static objects and requires hefty amount of precomputation, but does support fully dynamic lighting and dynamic occluders.

Also it's very ast at runtime is very memory efficient. Looks like nice compromise for semi static big open world games.

[darthviper107]

It's not voxels, but it's doing a lot of the same type of thing, some open world games use a similar but much simpler technique.

[iniside]

It's not voxels, but it's doing a lot of the same type of thing, some open world games use a similar but much simpler technique.

That's stretching. You just as well say that Enlighten does the same thing as VXGI. Yes they do real time GI. But that's where similarities end

Presented technique is completely different from Voxel based techniques. It's far more similar to techniques used in The Division/ACU and other auto placement probes which hold precomputed light transport.
Novelty here comes from it's sparse nature, dramatically lowering storage/memory foot print.

[poke1103]

That's stretching. You just as well say that Enlighten does the same thing as VXGI. Yes they do real time GI. But that's where similarities end

Presented technique is completely different from Voxel based techniques. It's far more similar to techniques used in The Division/ACU and other auto placement probes which hold precomputed light transport.
Novelty here comes from it's sparse nature, dramatically lowering storage/memory foot print.

I looked at VXGI and it looks like it kills people's GPU and framerate. It's quite nice, but I have only seen videos of it on a small map, not sure how it would look on larger landscapes, not to mention the fact that it's an NVIDIA product so by default its going to run worse on AMD GPUs.

[II_ADN_II]

Any new on this topic?

[ZacD]

Wait until DXR features start rolling out in early 2019 or version 4.22.

[Frenetic Pony]

Wait until DXR features start rolling out in early 2019 or version 4.22.

Hopefully those features just speed up the SDF tracing they were doing anyway, so you get backwards compatibility on PC for next generation games. That stuff was blazing fast and had a good flexibility, would much rather see it finally completed and totally useable than going for something shiny that gives as big a performance hit as reflections do on BFV.