"modding down" is done by the users. A mod now just hid those posts :PSo he posts incorrect information, yet I'm the one modded down. Perhaps it's too much to expect mods to do basic fact checking before exercising their itchy trigger fingers
Global illumination techniques
Chris_F, Prune -- consider this an unofficial warning to BOTH of you to cut the personal attacks and general nasty rhetoric (aka flame bait). I've removed most of the posts that have no contribution to the thread.
I don't really get what the fuss is about anyway because specular reflections are one the easiest things to do. Much more impressive is the computation of caustics, many GI techniques can't handle those. It's obvious that their technique handles both from reading the paper and looking at the images.
I'm not sure why you think specular is easy to do - they're OK (with lots of inaccuracies) to do for static scenes via image based lighting, but for dynamic scenes to calculate them without visual artifacts you need to both gather & process indirect lighting information at very high frequencies, which is very hard computationally. Personally, I'd even go as far and say that we wont find a good solution for this (at least for the general case, i.e. materials of all kinds of glossiness) until we end up using ray tracing anyways.
I was talking about offline ray-based GI, which I should have clarified. I also agree that ray tracing will eventually overtake traditional real-time rendering before it can achieve equal quality.
I was talking about offline ray-based GI, which I should have clarified. I also agree that ray tracing will eventually overtake traditional real-time rendering before it can achieve equal quality.
Oh yes, in that case I agree.
Chris_F, Prune -- consider this an unofficial warning to BOTH of you to cut the personal attacks and general nasty rhetoric (aka flame bait). I've removed most of the posts that have no contribution to the thread.
Apologies; my bad.
Also, this is relevant to the thread: http://graphics.cs.williams.edu/papers/PhotonI3D13/
This is pretty much follow up to the 2009 stuff I linked to above. Note that there are multiple shots of glossy here. I find the results very impressive, especially since most of their test scenes have little direct lighting. It's near-realtime on a GTX 70, so it should be usable for real-time on the 880 coming out in June. I would place my money on photon mapping overtaking propagation volumes in real-time applications over the next few years.
Thanks agleed , for sharing your thesis and code . It was so helpful.
I read about a technique called "Rasterized Voxel-Based Global Illumination " by Hawar Doghramachi . I read about it in GPU PRO 4 . It generates good results in respect to the comutation time . Why isn't anyone talking about that ?
Thanks agleed , for sharing your thesis and code . It was so helpful.
I read about a technique called "Rasterized Voxel-Based Global Illumination " by Hawar Doghramachi . I read about it in GPU PRO 4 . It generates good results in respect to the comutation time . Why isn't anyone talking about that ?
GPU Pro is a pay series, and there's more than enough free material for any one game or engine to implement that paying for more isn't the highest priority. At least that's my impression. Speaking of voxel based GI "Layered Reflective Shadow Maps for Voxel-based Indirect Illumination" sounds interesting. http://www.highperformancegraphics.org/2014/program/ So hopefully the paper will be up in a few weeks.
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement