Can’t wait for those… hope they are delivered in July as planned…otherwise im gonna cry 
Well, my experience with Geomerics was slightly different. It was a relatively small company with small budget (definitely not AAA), and it was an early stage of project development. We’ve asked Geomerics about Enlighten evaluation and received a positive response.
Seems to be random because the project I was working on even reached it’s KS goal and was very successful but couldn’t get it.
Looks like roulette game… Anyway Enlighten still too expensive for the huge amount of UE4 developers, so any native DGI solution would be welcome. In my own (rendering) wish list it is situated on the 2nd place, right after volumetrics.
Should wait and see if the volumetrics will be widely usable or not. If their cost is high they’d be just useless.
Something like this Physically-based & Unified Volumetric Rendering in Frostbite - Frostbite would be good enough
It would be great if Unreal Engine will start support CLion IDE for C++ projects. Coding in Visual Studio is really slow and painful process… At least after I used Intellij Idea for years.
As an individual who migrated from using and testing Indigo Renderer, which is a Photorealstic, Physically based rendering engine. It often would take hours to render high res, high quality imagery even with high end hardware, often times a single image, meaning one frame, could take hours or even days to render.
Now, also take
… Accidentally hit post before I was finished -_-
Now also take into account that we are talking about real-time lighting computation here. Technology has come extremely far, but we aren’t yet at the point where you can have your cake and eat it, too. Even a stationary directional light on the production setting can easily take a 30-50fps hit. Concessions need to be made to hit certain performance thresholds. Especially since most developers want to support as much hardware as possible in order to reach a wider audience.
I don’t believe that a expecting a fully dynamic GI that will come without a significant cost of either performance or quality is fair. No matter what kind of coding magic is being used. Even moreso because of how adaptable the engine is for different use scenarios. Sure, there may be some real-time engines that can manage high quality real-time GI and yet still have decent performance. But all of those are almost certainly engines built from the ground up with that particular goal in mind and use in mind. Such engines are only used internally and for one maybe two types of games. If anyone has ever used CryEngine they’d understand how restrictive it is compared to UE4.
Something like fully dynamic GI isn’t something that can merely be tossed in like most other features. Unlike the newer shading models, there’s very few shortcuts that are feasible.
I understand it’s something people want a lot. But at the same time, they also need to remember that people at Epic are likely working on it. But it’s a very intense feature, such things take time to get right. And even if they’re not working on it, I can’t blame them. There are some things that just aren’t possible in a commercial engine yet.
It is correct that UE4 isn’t built for real time from the ground up, and then it’s hard to implement real time features init. But it’s not like they cannot make changes to it. In cryengine you can throw hundreds of dynamic lights in a level and don’t worry about anything. In UE4 put a few dynamic lights and suffer the consequences. If a directional light could take 30-50 FPS hit in some cases it’s not a reason to halt other advancements because this simply light is already taking so much of the resources. It’s the directional light and more generally, the lighting that needs to be fixed/improved and that is very important. The static game environments are coming to an end and yet when we talk about anything real-time we hear no no no it’s hard answers.
Friendly reminder that once upon a time, this was a thing: https://www.youtube.com/watch?v=MOvfn1p92_8
The engine was originally built for fully dynamic lighting, but priorities have shifted over time when they realized the hardware of the day couldn’t do what they wanted it to do. Yes, SVOGI had problems, but even ignoring that feature the dynamic lighting is running really well here. I am actually rather curious if this build or the current build would have faster dynamic lighting just in general, without any GI, because it was just a higher priority then.
cries on the floor
What a beauty…
Well, I know no history about EPIC and Unreal Engine. Gotta take what people say and go with that. haha.
OP updated. Keep the requests coming.
Not added the GSC
A movement component for root motion, proper replication, would be nice.
GI wise it would be nice to have high end rendering features when taking into consideration that UE4 is being used more and more for other things than just a video game. Could put all of the performance hungry stuff into a for rendering only mode as in say a UE4-Vis version.
Now it is worth noting that according to GeForce Articles, Guides, Gaming News, Featured Stories the original demo was running on a single GTX 680, not even with SLI. For the demo at the time, that was running at 30-40 FPS at 1080p (framerate did fluctuate a lot all the way up to 80 fps but that could just be from the editor’s long standing weirdness of running slower when you’re clicking in the viewport to move around), you can see the framerate info in the editor’s upper right corner since that was the default setting at the time. In the fullscreen demo though, which is what counts, it was running at 30 FPS or higher the entire time.
If that chart you listed is anything to go by, the GTX 1080 should be able to easily attain 1080p 60 FPS with that old demo. Even without any advances in speed that we’ve had, like the ones the wizards at Crytek apparently put into place in the CryENGINE V’s SVOGI, it’s not a completely unattainable goal anymore.
Of course. It’s a useful feature for more than just games. I highly doubt planar reflections are going to be used in every scene either, but having it as an option expands what the engine’s capable of for both games and non-game purposes. I would personally advocate for using a realtime GI solution even as a sort of “preview mode” for games that you’ll end up baking light in anyway to cut down on artists’ iteration time. It would be a great help to know exactly what it’ll look like before you hit the bake lighting button and wait an hour or two to see the results.
I also think that there is an opportunity here to get ahead of the quality curve of developing games for 5 years from now instead of sticking with what works today. If it’s marked as alpha only I’m OK with that as long as I don’t have to go digging around in some INI file. 
Add a “Make MMO” button.
There’s no doubt that Epic still needs to make major improvements when it comes to the lighting system as a whole. There are certainly inefficiencies and if I knew how I would be delving into code to find means of improving things myself. You’re right about CryEngine being very capable in dynamic lighting. However, with that said, if you were to attempt making a high quality scene with an entirely populated level that is essentially the same as in UE4. We’re looking at an overall performance difference of 20-30 fps between the two, but that’s primarily due to Cryengine not having instancing capabilities.
I mean, if they would add the great lighting functionality of Cryengine to UE4, it would easily be the best engine out there by miles.
But I don’t think that’s possible due how light is managed and utilized in each engine being fundamentally different.
I think that there absolutely needs to be improvement in terms of lighting as it relates to performance and flexibility, that’s no question. I just don’t think that expecting a a well-performing and high-quality fully dynamic GI from an engine like UE4 is fair or even reasonable, (at least not for the near future, anyway.)