Hi!
Just a question, I saw Unity5 will be using Enlighten to provide real-time GI - why isn’t UE4 using it as well? I guess the cost ($) isn’t a problem, so I would like to know if the UE Team plan to implement something better or if Enlighten isn’t great enough to be used as main system of UE4.
I hope you guys understand the question.
Thank you!
Actually money is everything if we are talking about Enlighten.
It is most expensive middle ware in industry. More expensive than licensing entire game engine.
I don’t have single doubt, that if ARM (who owns Geomerics), would make reasonable offer to Epic, I’m 100% they would integrate it into engine.
Enlighten as middleware covers pretty much everything in terms of global illumination (indirect lighting). Light maps, dynamic lighting.
It’s honestly hard to believe for me that they made deal with Unity. But that hopefully make more possible to make deal with Epic. After all, UE4 is also expanding to mobile platforms.
The other issue might be level of source access in UE4. I don’t know if Enligthen can be only integrated as binary, but reading one vague PDF presentation, I don’t think it should be a problem.
Enlighten also has a lot of issues, there’s still a lot of preprocessing that needs to go on, and there’s many situations that it doesn’t work for. Basically, it only is useful if you intend to do a dynamic time of day and want nice lighting.
Enlighten already has a plugin for UE4. The difference is that Epic does not have the same deal as Unity to include it with the standard engine license. If you’re really interested in their solution you can contact Geomerics directly and they can give you a quote on the licensing cost. My guess is it would be pretty steep though as I don’t think their UE4 plugin is geared towards Indie devs.
http://www.geomerics.com/enlighten/
In my opinion though, I don’t think what Enlighten offers is that much better than the LPVs that are already built in to UE4. The main difference is that Enlighten is production ready now while LPVs need quite a bit of work. I would wait and see what develops though. Now that the marketplace is open, you can bet there will be those trying to sell various DGI solutions. These will probably not be as multipurpose as what Epic offers, but may offer much better results.
Thank you for the reply!
Actually, you can see Enlighten in the list of features here - unity3d.com/5 - so looks like they have a deal, what I think weird is that if Unity can handle this licensing costs, Epic has high chances unless the licensing terms are different for both companies.
Or maybe Epic team want to build a better tool? I never used Enlighten before, but looks pretty good (maybe someone used it can give futher feedback?).
Thanks for the reply @Spade and @darthviper107.
I understand, actually afford the license to Enlighten isn’t that easy because it really isn’t indie friendly at all. My option is wait to see what UE4 will implement on that, but I was just wondering why not: costs or technical reasons.
Maybe a Epic’s worker reply here?!
Kind Regards!
My guess is that Epic wants a better solution, their original dynamic GI system gave much better results without the problems that Enlighten has, only issue was performance outside of high-end systems.
One big issue with Enlighten is that it doesn’t work with dynamic objects, so things like characters don’t get any benefit, and if you want to do some really dynamic stuff in a level then it doesn’t help you. It also has issues with small objects, and it’s limited in the types of lights that it supports.
My guess the reason that it’s in Unity is that most people don’t actually want Enlighten if they know what it actually is like. This way they get it to be more popular. But it’s a better solution than what Unity can do, since it can be used for static lighting as well and adds a form of dynamic GI if the situation works for you.
The lighting system in UE4 is way better than Unity, so it wouldn’t get as much of a benefit.
Again, thanks for the reply!
I’m with you, I also think that Epic plan to build something better to UE4, not saying that UE light isn’t really amazing.
The guys that work with the rendering system visit the forums? I think they can answer my question easily hehehe
I somewhat disagree. Preparing UV maps for lightmaps, is not easier than than preparing cage meshes for your geometry for Enligthen. As matter of fact you could have those cages already done in form of collision geometry or lower LOD. At worst it takes the same amount of time.
Enligthen precomputation is faster, since you don’t really need to bake individual lightmaps, which store directional information about lighting condition on level. You just bake light transport information. Hell, you could prolly bake light transport info, when you import assets, as it is not dependent on how your level or lighting is setup.
In return you get instant relighting and dynamic lighting support for infinite amount of lights and bounces + as bonus emissive lighting of arbitrary shape (where are emissive lights in Lightmass ?).
And as final bonus, you can bake what you see on scene, to lightmaps, almost instantly, because most of needed information is already precomputed. Will those lightmaps be as high quality as those generated by Lightmass or other offline renderer ? Probably not. But they will be generated in seconds instead of hours.
I would love to have this option by default. I wouldn’t ever use Lightmass again.
Edit.
It doesn’t work with dynamic objects. Well, yeah Lightmass doesn’t work with dynamic objects either ;).
But in most ofcases, you need dynamic lighting, while most objects are static. Fully dynamic GI solutions are not viable. And probably still won’t be with the current generation of hardware.
There are fully dynamic GI solutions though, that’s what was originally in UE4, and for the people that have the hardware power then it’s fine, and probably in a year or two then mid-range hardware will easily be capable of doing that.
I would say it’s much easier to setup lightmap UV’s then to remodel everything. You might even have them already if you did it the same way for your textures. Doing UV’s is more like a puzzle, where you can’t really go wrong, just make sure to put seams where they should go logically and it will flatten correctly. Modeling is a different matter.
And from what I understand, Enlighten still has a very slow pre-processing that has to be done as well.
LAst time I checked cages for Enligthen were, very, very simple. 10-20 faces. I wouldn’t call creating something like that hard or complicated ;).
I have worked with the UE4 integration of Enlighten, and I have to admit it works pretty solid and you dont need to build cages or so since they have implemented a very good autounwrap algorithm.
However, for us…it sadly fell apart because at the time we tried it, it was not ready for levels of our size and because of that, we had the most issues.
But in smaller environments it was really and precomputing times were acceptable 
Looks like that for big worlds nothing better than real-time GI, because it doesn’t uses precomputed textures maps, correct? So, Enlighten isn’t the best option for open worlds too…
Thank you for the comment!
It wasnt (which was almost a year ago)! Maybe it is now 
Regarding what was written above…a lot of it is actually not true! Enlighten supported all light types, they even had a skylight cubemap that was shadowed in realtime and characters and movable objects work as well because Enlighten uses a Probe system for dynamic objects So dont judge too hasty if you dont know all the details^^
Actually I would love to try it out again and see how it has improved because the workflow is still way better than Lightmass and you dont need a time of day to take advantage of it. Think about shooting off lights, turning lights in rooms on and off, using a flashlight, having GI from explosions and fire…really neat stuff! Also all the cubemap captures from Unreal work dynamically with Enlighten, which means indirect specular updates in realtime depending on the light situation in your scene.
Actually Enlighten has only one major downside (if they managed to improve on the large level issues we had in the past) and that is the price^^ Its just **** expensive.
UE4 already uses a probe system for dynamic objects, but there’s many cases where that’s not enough, like for doors. And for lots of dynamic objects there’s no solution there, like say vehicles which need more than just a general probe augmentation. Or if you want to do really dynamic environments.
Enlighten really isn’t that much better than LPV, and it would be a mighty challenge integrating it into Unreal. Just because a Plug-in exists doesn’t mean it’s a simple drag-and-drop installation. Also, were it ever integrated, Geomerics would have to give out their source code. That’ll never happen.
I would far prefer Epic to work on their own solution in their own time, or better yet, work with nVidia to come up with a great solution. Good things come to those who wait
You’re joking right? Enlighten is almost like lightmass…but real-time.
Video 2:
Neither is creating lightmaps
I have to admit, you must not have used it… I can’t really say anything about its Unity integration due to NDA rules. I suppose I can say, I’m glad UE4 didn’t go with it…
@ Joshua, it’s still an extra step in the chain. The amount of processes in artwork you have to go through is bad enough as is and the amount of threads I see on lightmass issues alone shows it’s not quite “there” yet. Personally I don’t think you can beat a good dynamic GI solution even though they all seem to have major drawbacks, but static lighting and big terrains generally cause headaches. Hopefully VXGI might add something (wink wink). Nobody says CE games are ugly and in all fairness the performance is pretty great…
It’s very cool that you guys are working / improving on automatic LM generation though and in some cases the in-built dynamic lighting is good enough… Skylight + dynamic lighting (none GI) seems to do the trick.
You have the point. But you can’t beat that. Baking scene for Enlighten in Unity 5 takes, few seconds ;p.