3 replies to this topic

[T_Ice]

Hello,

 

I have been reading up on Light Indexed Deferred Rendering (LIDR) but I couldn't find an answer on how to handle ambient occlusion.

 

What options do I have here? Is Screen-space ambient occlusion (SSAO) still a feasible option (albeit via a detour)?

 

 

Thanks in advance for your time.

 

[Lightness1024]

hmm, could you explain your feeling on why you would have to make a detour ?

from my feeling, lightprepass, full defered, tiled defered, index defered, heck even forward rendering can all do SSAO. you just need the depth image and a post process pass.

other than that, you can also do much better ambient occlusions using true geometry aware algorithms like voxel cone tracing, or nvidia point (rather discs) based evaluator...

http://http.developer.nvidia.com/GPUGems3/gpugems3_ch12.html

http://http.developer.nvidia.com/GPUGems2/gpugems2_chapter14.html

...

[T_Ice]

With detour I meant filling up a G-Buffer that contains the depth or position information of the entire screen.

 

Thanks for providing the links, I have seen them before but never full read them.

[Hodgman]

LIDR already requires a depth-pre-pass, which you can use to generate SSAO for the scene before your lighting pass.