Currently unreal doesn’t have a solution for dynamic GI. Light Propagation Volumes are a pain to work with, not at all optimized, and don’t even look that great. If I remember correctly DFGI was a thing but I haven’t heard anything about it in forever.This part of the engine feels unfinished and hasn’t been touched in a very long time, which is strange considering how important it is for some games and architectural visualization.
I know there are 3rd party solutions like Enlighten but I don’t know if I have enough organs I could sell on the black market to afford it.
The problem is no good general purpose dynamic GI solution exists. There’s a few that work well for exteriors and landscapes, but fall apart for interiors. Generally they are better than LPV, but still have similar issues.
Enlighten still has the cons of baking like seams, artifacts, long bake times etc, like lightmass, but are even worse. Also Unity already has it so if you really need Enlighten, there’s an option. And Epic probably wants a solution that doesn’t have the same cons of lightmass.
Maybe we’ll see something cool from GDC this year.
what do you think guys, is this thread going to be 15 pages long again in a few days? : )
I’ve found working in Unity’s lghtmap baking and propagation volumes much better and with far better results in general. It’s interesting what those guys are doing, I think when Octane comes out of lightmap baking Beta, Unity will arguably hold the best render engine for baking among the public engines out there.
This or that new technique usually doesn’t matter. The’re like 2 viable solutions you can do or combine and all the others have some huge disadvantage that’s not obvious. None of which matters because unless you’re going to do it yourself you’re not going to get it. For reference it’s some variation of signed distance field raytracing, which UE4 half supports but doesn’t seem to be advancing at all (in fact previous features are being disabled). And realtime environment map/light probe updating, which works fine in Call of Duty and is basically what The Witcher 3 and The Division do variations of. But Epic seems to have no plans on supporting.
Which is too bad. Ironically the 3 biggest UE4 games are all large world games (Fortnight Battle Royale, PUBG, ARK) in an engine still not designed to run such that well. No, gotta get photorealistic baked lighting that takes way too much effort to actually produce. Not the best strategy in hindsight, but then that’s hindsight, can’t do anything about it now.
here we go again, but I still hop on the train
I don’t really need a perfect dynamic GI solution. Something precomputed like Enlighten in Unity would be enough for me.
But it is up to Epic to decide. We will see…
I would be happy if they just get LPV to a somewhat usable level. Every time there is even a small engine update I have to go back and change the config and it sucks. I’m creating a stylized game with very simple textures and you wont believe the difference a little bit of bounce lighting makes.
You would think that with so many games out there using ue4 that they would be making a ton of money off the engine and that in itself would be reason enough to add these big features that the engine needs. Maybe I’m wrong and they aren’t making that much off the 5% royalty fee.
This is indeed super frustrating. The idea that the development of major engine features is dictated only by the two or three games Epic is working on, rather than the needs of millions of customers/developers which Epic provides a service hence the engine for. That is not to say that it should be either or, but a healthy balance between the two. In other words just because a feature is not needed for one of their internal projects, shouldn’t exclude the possibily of it’s development. Which is the case here.
I feel it’s mostly the case of lack of competition in general coupled with internal politics and decision making. Companies can tend to take it easy when there’s no real competition. Cry engine is gone with the wind, and unity’s visual side needs an overhaul. that leaves UE4 the go to choice for such projects. I only wish Valve kept at Source and introduced some real third party competition into this market, it certainly had all the potential with mods being developed. Valve had no excuse, those guys struck gold and oil at the same time.
only a few more posts to go and we’ll be on page two and i can claim that 20 : ).
wtf are you talking? Cryengine has a Dynamic Global Illumination solution and I’m pretty sure it’s a rock solid engine right now. I didn’t check what they do last year, but if they focused on making a game engine instead of having an AR/VR/MOVIE/“Whatever Epic has done last year” focus, I can imagine they are pretty close or even better than Unreal in certain topics. But sadly I don’t see any Cryengine experts here with Unreal knowledge, only people who think they know what’s good, judging by what the majority is using. You also make it sound like Unity is no competition, only because it needs a visual overhaul. All in all, I really don’t think it’s because of lack of competition. They have a small core engine team that can’t focus on the needs of everyone. Even if the majority requests it, if they don’t see the potential to make serious cash with that, they won’t bother.
He said big features. Big features for a small group. Even if the demand looks high, how many projects will actually benefit from that? Look at all the projects that use Unreal Engine. Now look at the projects that need dyn. global illumination? And now take a look which of these projects will make some serious cash? Answer: Not so much. They will still look into feature requests and have to decide on each case how much time it takes for them depending on their staff.
Crytek wasnt paying their employees. Star Citizen, the biggest game using their engine has moved to lumberyard, and now crytek is suing the developers. They have had a rough couple of years. https://www.polygon.com/2016/12/10/1…oyees-not-paid
Its a big feature that any game using a dynamic skylight should be able to take advantage of. Thats a lot of games. Besides only 1 big game would need to make use of the feature to make it worth while. Isn’t that epics current logic anyway?
You still didn’t mention 1 game that has the potential to get enough money to pay 1 developer to implement this feature and why in this game cheaper solutions are not an option.
I will be the first to point out that there are priorities that need attention in the engine before a GI solution, and the methods that i’ve seen so far of realtime GI are still somewhat questionable for me.
But I would ask what have many other features currently in the engine that have had time and resources spent on contributed back to the engine in terms of financial profit to justify hiring a developer to work on them?
Ex. Has live action background integration features benefited UE financially? like that car example they where showing countless times over in an event.
Also like Krunku said, I don’t think Cryengine is competition to UE because it’s in trouble, also the forums and docs have been dead for the past few years now. No ones going to risk it with that kind of lack of support or company future unknown. Epic I believe knows this, which is why I mentioned it.
Unity I support best, It’s a personal bias for good reasons for me and my team, and i know that i’m more impressed by the creativity and art of the games published on it vs the usual 3d games i’ve seen in Epic’s reel. But facts on the ground are that it still needs some good push to fix up not only the visual side (in terms of larger scene handling and node based shaders, lighting methods) but also some of it’s animation and dynamics features. Which is why UE still hold that extra ground even in a flawed state in some areas (ehem cloth).
Cryengine has a Dynamic Global Illumination solution and I’m pretty sure it’s a rock solid engine right now. I didn’t check what they do last year, but if they focused on making a game engine instead of having an AR/VR/MOVIE/“Whatever Epic has done last year” focus, I can imagine they are pretty close or even better than Unreal in certain topics. But sadly I don’t see any Cryengine experts here with Unreal knowledge, only people who think they know what’s good, judging by what the majority is using. You also make it sound like Unity is no competition, only because it needs a visual overhaul. All in all, I really don’t think it’s because of lack of competition. They have a small core engine team that can’t focus on the needs of everyone. Even if the majority requests it, if they don’t see the potential to make serious cash with that, they won’t bother.