Download

Whats Wrong with Epic? Dynamic Lighting, GI and Shadows!!

actually in the eula there’s a tiny little detail…

if you’re a studio and you get funding this means you already owe Epic money.

so the only way the engine is free is if you use it for free and release your games for free :rolleyes:

Ark and Battlegrounds are open world and are the most profitable indie projects for Epic so whether open world is the new MMO startup or not that’s no valid reason to not support them. Some of the problems these projects are suffering from are what is mentioned in the OP. Saying Epic already cared too much about open world, is just not a valid statement. They did with 4.8, introducing a set of tools that didn’t receive updates or bug fixes post 4.8 and to this date there are many open world related bugs and feature requests that remain backlogged because they’re not a priority for Epic -> open world is not a priority for Epic -> Helping the most profitable open world projects (Ark and Battlegrounds) is not a priority for Epic, which is where the weirdness begins and the more you look at things such as useless level editing in VR the more you realize a lot of Epic’s recent efforts are fruitless.

You think it’s weird because the focus is to expand beyond game development.
It’s not just about good looking games anymore, VR is increasingly becoming important for the markets of simulations built “around” the tech created for games and CGI animation; the stuff they’ve showcased at last GDC: movie rendering, vehicle design, VR training, etc… there’s a lot of revenue on those avenues and they clearly aim to capture that as it’s obviously cheaper to explore those markets instead of making more and more games. Beyond Paragon and Fortnite I don’t even see Epic focusing on games lately.
All of these companies involved need highly accurate static lighting, GI doesn’t even look good for what they’re doing with UE4.

Not sure what your point is, you pick two games out on an army that support your argument and thats meant to make it fact? I think if anything those games support the fact that Epic has provided the tools to make a successful open world survival game regardless of dynamic GI (that doesnt make it easy). I understand there are requirements outside of graphics that might be limiting but as I said its a design problem and maybe even management, if you dont have the resources to make an open world game then you should be looking to do something else.

As far as I can see Epic has give Ark and Conan: Exiles plenty of help, especially in marketing and support, not everyone gets lucky and becomes the next PUBG which is simply what Im saying. For every successful MMO theres 100s if not more of failed attempts and I dont see the point of supporting individuals but the industry at large, which is why I agree with you on the VR editor front because the numbers just dont support it being a success… yet.

Why should Epic support any part of the industry thats unproven? They give the engine away for free so I think the “do it yourself” answer is extremely adequate and I hate the entitled attitude. If this were an Ark or PUBG developer saying this then my tone would be different but no its Mr 1 post of rage… My answer would be the same though, they clearly have the money to do it themselves with their level of financial success.

I would just much prefer the time being spent making it easier to expand on UE4 ourselves, to manage these systems in a way they can be used for various types of games instead of being limited because someone screamed and jumped up and down forcing Epic to implement something half-arsed. Im very much about doing things right the first time even if it takes some extra planning and I feel rushing into LPV for eg wasnt the best idea or the limited level streaming options or lodding or asset management, the list goes on of features which could be more optimized for other types of games that dont require it all to be loaded and streamed in. For games that want to add more content in additional packages down the line instead of pushing out 20GB patches… Ive had lists which grow and shrink over years of using the engine, not just acouple but since the beta and there are still bugs that exist from back then, still issues that have only just been looked at, Im in a constant holding pattern myself because Epic simply dont have the resources to keep me happy but you dont see me on here raging out :cool:

Lighting isn’t particularly what this is about. It’s about maintaining balance between between both worlds instead of going all out on one side only. Epic isn’t the only company implementing VR features one after another, there’s nothing wrong with going the VR route, but Epic is the only company leaving most of the base broken while they’re working on futuristic VR features hence while people don’t necessarily hate VR, they do see it being unnecessary if the basic needs of a game engine are still missing. I still can’t digest that Subsurface Scattering is intentionally left broken, along with many other issues because VR the the future. The things they’ve been doing lately, as you mentioned in your post, is collectively a form of declaring the end of support for game development.

I didn’t pick two random games, I picked the two most profitable games which in fact, those games are built on the tools provided by Epic but are looking so poor, and run so poor, not because of design problems, but because of bugs not addressed by Epic. I don’t even know how you come into the conclusion of “Since there are many people that can’t create proper open world games, Epic shouldn’t provide proper open world support”. The point is not to have 100 successful open world projects out of 100 open world projects. Even 1 successful open world project out of 100 is enough for generating millions of dollars for Epic, which is a proven fact. Which makes it a fact that open world tools deserve more attention than they are receiving atm.

You fail to see every open world project as a potential successful open world project which makes you think that “do it yourself” is the adequate reaction unless a project has been successful already and have made tons of cash for Epic only at that point they can come in and make request/voice concerns, but in fact, the success rate goes up if they have Epic’s support since the beginning and that’s not supporting people individually, that’s supporting the tools they put out and putting out more useful tools that help all those individuals into bringing about a successful project. I think you understand that if 1 person complains about the lack of proper open world support it doesn’t mean there’s 1 person interested/benefiting from peoper open world support.

Epic has the resources to fix most of the broken tools and bugs. Again, proper SSS, Proper Grass tool, Proper Tessellation, Physical sunlight, spline decals, Fixing ShadowDepths problems, etc. etc. are basic, basic needs of an AAA game engine in 2017 and if you want to believe all the basic needs are ignored due to shortage of resources then you might need to think about where those resources are dedicated, they’re dedicated to VR and generally features that happen to have less than 5% audience which makes the existence of such threads created by the OP a valid act. It’s not raging for being unhappy, it’s raging at a ladder that doesn’t have steps on the bottom half.

VR is not the future it’s just another tool a Wii controller type, when VR first came out, everyone jumped on the bandwagon thinking it was going to get them gold out of the box, we were some of the few who saw it coming, it hit the wall hard, sales numbers were far from expectations, a Wow factor that lasted 2 minutes. Articles and reviewers suddenly stopped talking about it when every day prior it was all about VR everywhere, and now so many companies with invested interest are still trying to beat this half dead horse. VR is useful in some places and i can see where it can be, such as a couple looking to buy a house and the clients of a prestigious agency wants to “impress” by having them put on a VR headset so that a guy can convince his wife that bed A can look good in that corner vs the other. Or more effectively military use or pilot training etc…

Does this justify a VR editor, in my humble opinion a big no, in fact i challenge anyone to tell me how this feature would be used effectively in an actual production pipeline other than just being a showcase. I don’t mind it but the way it is advertised as being “the future tool for editing” just raises some eyebrows as to the intentions and direction the engine is taking.

I too feel that it would be great for Epic to focus on things that are broken or matter in the engine, such as a functional solid full body IK system, or proper SSS, or faster integration of the new cloth pipeline with all the features, or better landscape editing. I understand these things take time but prioritising them over other areas will help cover the foundations better before moving on to other things, but its Epic’s decision at the end and we just have to live with it if we can’t persuade it to change.

Finally while I understand people trying to make open world games, I too would argue that it is a hugely ambitious task for a small or even a medium sized team working with the current tools, and that they would be better off polishing a more practically doable game by the team. But it is a free world.

Also in terms of GI which is related to Lighting, keep in mind that while it would be great to have it, it is not really the do all and be all solution at all, just look at games like Horizon or the Witcher in some instances, they have good lighting in an open world without GI, lighting is an art on its own and no amount of GI will add magic to your frame and most of the screenshots i’ve seen with GI in Cryengine have that very flat look because they didn’t really pay attention to the light design and the bounce lights have created a flat ambient all over.

I get the impression youre just replying and trying to start something without actually understanding what was said. Did I not say in my post I would prefer if Epic did stuff properly? I was explaining in more depth why I thought Epic supported open world more because they answered the calls of people jumping up and down in the past and it lead us to where we are now.

I used to think this was true but after being around the UE4 community for so long and seeing how slow things happen at Epic I dont believe it. I think most of their resources are tied up making games or supporting people using their engine not in advancing their tool set.

I already said that I was perhaps cynical about open world games, more power to the people that do it but Ive yet to play a good one in all my 30yrs of gaming :cool: (Disclaimer: I have not played The Witcher 3)

Oh and P.S. Apparently PUBG is not an indie game https://www.bluehole.net/en/game/pc

I think it’s fair to feel the whole VR thing has gotten a little bit too much of the spotlight compared to absolutely ZIP where it concerns dynamic GI.
Yeah apparently there is a market for it cause otherwise it would make absolutely zero sense.
Fact is that market is completely incompatible with anything regarding dynamic lighting, hence the sour grapes (or well, at least a slight bit of salt).

Last thing we heard regarding dynamic GI was RyanB getting told to bin his efforts cause of different priorities. Word on SVOTI being very difficult to implement and maintain is way older even.
I get that things happen. Some things you like (Volumetric fog in my case), some things you might feel are absolutely retarded to spend 2 minutes of dev time on (VR editor in my case).
There’s probably tons of people feeling exactly the opposite. No problem.

All I’m saying is that it would be swell if some dev could actually comment on any plans regarding dynamic GI.
Is it being discussed? Has it been recently? What do they think? Just a heads-up. Perhaps that’s naive though.
I’d like to think it’s not since Epic generally is quite open as to what’s cookin. Maybe that’s also an answer though :frowning:

[MENTION=202133]William K[/MENTION], Don’t shoot me, I didn’t intend to imply VR is the future. :rolleyes:

Dynamic GI is 1 of the 3 points OP has mentioned. Now while dynamic GI and Epic are worlds away, the other two points which are complaints towards overall dynamic lighting performance are completely valid. Is dynamic lighting one of the most expensive features in every game engine? yes. Is dynamic lighting cost in UE4 reasonable when compared to another engine? no. So in my opinion it’s far better to get the dynamic lighting cost down/fix every backlogged bug report that’s tied to dynamic lighting and more specifically ShadowDepths before implementing any sort of dynamic GI. It’s not far from imagination that SVOGI was removed not only because of the cost SVOGI has, but because how poorly dynamic lighting performs even without dynamic GI.

What calls have they answered?
What open world tool they’ve added that you find unnecessary to have?
How can adding a few basic open world tools in 4.8 be the reason for the issues today? (Considering today’s issues are not positive).

Compare the UE4 release notes with every other game engine release notes. Compare the frequency of updates. Epic is almost one of the fastest in both adding new features and rolling out major updates. The only one issue is, resources are mostly tied up advancing their tool set for car manufacturers, VR partners and generally “partners” which happen to be not in the business of game development. What was shown in last GDC proves it.

Agreed, one of the things I would like to see in dynamic lighting is some kind of improvement of shadow cascades to make the transition between baked and dynamic light more seamless, this is especially true on characters shadows and especially true when you are using longer lenses in the game which would force you to calibrate the cascade distances at a more precise and sensitive numbers, but when you move said camera then the whole thing needs to be re-calibrated again (non 3rd person cameras more 3d platformers or fixed camera workflow) and current;y there is no way to adjust the cascades through blueprints or other more accessible ways).

Other annoying issues revolve around the SSS behaviour, and lack of SSS back scattering for skin, proper IK and so on… (this is just me repeating myself).

In any case this is just one of those threads unfortunately.

I still believe Lightmass combined with LPV gives pretty neat results if worked properly; the problem is, it’s a LOT of work to setup manually all that stuff required to make LPV work, it’s just too much manual work.
UE4 has Enlighten support, but that sht cost upwards 6 digits.

Road tool, LPV, the Kite Demo for eg and Epics games are also hardly indoors, things like outdoor lodding used in Paragon. There are alot of tools that are simple not needed for smaller games but Open World is a popular buzz term so it makes sense Epic went along with it for marketing but the realities of the genre are far less optimistic. The Witcher 3 and Breath of the Wild being exceptions.

My point is they rushed to support people making open world games because everyone was jumping up and down (to repeat myself for the 3rd time btw) and implemented the tools in a fashion that was not largely beneficial for everyone and actually detrimental in the damage it caused to the stability of the codebase.

I cant honestly do this because the engines I know with the better feature sets for open world games are all in house, Snowdrop, Frostbite 3… Decima and there are more but Epic is actually an oddity in the industry not the rule. The issue for me is largely that bleeding edge graphics do not represent the games industry at large and those that need that level tend to have inhouse engine teams (even if they are using UE4) so I would much prefer to see the focus on features which help to make better games with the engine because its clear thats where the competitive advantage is atm not graphics :cool:

I just think the world could do with a few less clones and a for more large ambitious games that arnt simply a string of buzzwords tied together! Game projects die every day even for the biggest badarsed studios on the planet.

Road tool? you mean landscape spline? I’m aware people ask for spline decals to have proper roads: Unreal Engine Issues and Bug Tracker (UE-13142) and no one is answering the call though.

As for Kite demo, from my perspective (and many others) it was a form of marketing to pull more audience, because you know, majority of the communities are all about open worlds and an AAA engine without any open world support is to be laughed at so they implemented a set of tools to pull people in and say “we have it”. I think the fact that open world support ended right after GDC showcase proves the point. If Epic ever intended to have any proper support for open world then kite demo should’ve been polished to the point it’d run on all PCs of mid range hardware with ~100 frames.

But hey, Epic answered every performance complaints by “It’s a tech demo, not supposed to run well”. I was among the people who also happened to say the same like to complainers, but after a while, let me remind you epic, this is not a tech demo community, this is a game dev community with intention to create games that scale well on vast majority of hardware. Yeah let’s not blame the texture sizes and tri counts we both know that’s not the real issues, the real issues are all those accumulated backlogged bug reports that tend to not align with Epic’s goals to be addressed.

I think the matter of fact is they rushed to have the open world tools for the showcase stage. I don’t see any point in blaming the community for asking the tools they need. If it was for people then the support should’ve been continuous until the tools worked flawless before switching to support VR partners and car manufacturers.

As for graphics, I agree with you graphics isn’t the first factor, but I also believe we’re not discussing graphics in this thread, we’re discussing existing tools and supporting them before adding in new features and changing agendas.

You assume that open world game should have the same content density as corridor shooter (loosely speaking).
I disagree. Sheer world size in Open World game is feature and selling point. It doesn’t need lots of hand placed content or story or things like that. It needs nice looking, somewhat immersive world, which is dynamic in its nature (it’s actually simulated, not scripted with capture towers…).

To that end, yes Dynamic GI is imporant. It gives you better looking world, which is actually quite important for open world games, where world is playable content on it’s own.

You clearly just don’t like open world games. I do like, I get mad when open world games are trying hard to be like small story driven games, because it’s not where their strenght is, and that’s the reason why there is so much survival sandbox games. Because there is mark for it, and consumers of those games like them not for the story, but for the world and multiplayer.

I think the point is well taken that EPIC better solve the problem soon because there are other players to market that are already showing off better solutions. Some don’t have the same user base, some don’t have the same capital but the wild-west of VR is changing all this. It’s being felt more in Los Angeles, San Fran and in the Vancouver where you have the largest contingent of developers across all spectrums of next-gen platform (not consoles) and lighting is where investors are putting their capital. There is a HUGE engine we know who is having a potential to unseat EPIC in several venues from it’s prominence simply because of the way they handle lighting. If you’re behind closed doors of several different second and third round of capital gathering for companies who you’re going to see leading the way in title and entertainment markets, they ask the same questions as MASSIVE did, at the beginning of this thread, but they answer it themselves and say if EPIC cannot do it there is nothing stopping us from talking to those “other” groups and they are. EPIC publically seems to be just fine in how they’re being used. Privately, they’re is a MONSTER PUSH (E.L.E. size PUSH) for EPIC to solve the dynamic lighting issue now. EPIC has to solve the issue within the next 9-12 months or you will see some, as of yet public players, that will be stepping up with better solutions at least that will WOW on the lighting end and they already have a number of huge studios and directors offering 8 and 9k figures at these solutions right now… that are not EPIC because they look better, not because they can solve all the problems but because their lighting holds the frame-rate, is fully dynamic, there is no baking and they have enough of the other issues solved for a reliable platform to give them the solution EPIC can not. So hopefully we’ll see something very soon, or… we’ll see something different very soon :slight_smile: I, for one, need EPIC to get their **** together when it comes to solving the dynamic lighting and GI situation. These are not things that you find in POLYGON or 3DWORLD, you might find them in the “Reporter” or “marketwatch” but most of this is only information inside the large Chinese & west-coast investment firms planting funds in entertainment.

Please point out to me how you think Epic is pandering to open world developers. In my opinion they have done nothing to fix open world game development. What did we get from Unreal? The Kite Demo! That thing was about as friendly to real time as my cat is to baths. It was all “Hey, look at us, we’re Epic and our engine can do large open worlds” NOT!!! It was a marketing ploy and no more.

GI is an issue whether you want to use it or not. As said earlier, it should be an option at least. If you need it, YAY it’s there. If you don’t, leave it off.

I would also like to point out that a large chunk of the best selling games are open world. Horizon Zero Dawn, The Witcher, Ghost Recon, Skyrim (all TES games), The Legend of Zelda: Breath of Wild, Fallout, Assassin’s Creed, Red Dead Redemption, Batman: Arkham City. Grand Theft Auto V, 90% of the MMO’s on the market, and the list goes on and on. So if you haven’t played a good open world game in 30 years, you obviously haven’t played many games.

I don’t want to make an open world game because it’s trendy, I want to make an open world game because that’s what I enjoy playing!

Both CryEngine and Lumberyard have implemented SVOGI, and it works! Already being used in games that work! In addition, Lumberyard is also using OIT which is order independent transparency. Guess what! It works! The tech is out there, just read the f**king papers and stop acting like nobody is using it and/or implementing it into games. As someone said earlier, Unreal really isn’t free, so we should be able to expect it to work, not just for mobile, not just for VR, but also for open world games. These other teams of game engines have been working to implement the newest tech into their engines, why can’t epic? Oh, btw, Tessellation has been broken for years. That needs to be fixed also. And shut up about “why do you need to use such and such feature, other teams have found work arounds”. Those teams have a large budget and more times than not, just build their own engine.

Regarding work around’s:
From Horizon Zero Dawn, "World lighting is interesting as well. Horizon makes use of pre-calculated global illumination but to make this work with its dynamic time of day system, the engine blends between six different lighting bakes as the day progresses. This is combined with other lighting passes to effectively give the impression of real indirect lighting with moving sunlight. Horizon also uses a multitude of lighting passes in building each scene, starting with emissive or glow lights such as the sky followed by local direct lights such as these torches. This is then followed by the static indirect lighting and sun bounce passes before the sun and sky are rendered. Next are reflections of light across the scenery just before the final texture layer is presented.
Yeah, it’s a work around, but I’m not really sure how that is more cost effective or time effective than SVOGI. I guess if money grows on trees for you and you have a large team it’s doable.

Look, the bottom line is, If I have to fix a game engine to do what it should do, then why the hell would I pay them royalties for something that didn’t work? Hmmmmm, good point!

Oh, and one last thing. To those of you saying or thinking “well, if you don’t like it, don’t use it”. After a couple of years in development, I can’t afford to make the switch now. It would be wasting 100’s of man hours of time and money, so as unf**king happy as I am with Unreal Engine, I can’t afford to make the switch or I would have already.

TOTALLY Agree. Epic does nothing to handle open world for years.

Bump Bump This Up

  1. There’s very few CryEngine games, if it was really the perfect open world game engine, we’d be seeing a lot more use out of it.

  2. Most of the open world games you listed do not have dynamic GI, it’s static and faked. One of them even uses Unreal Engine. They almost all have different solutions because they have different needs and performance concerns. Plenty of large open world games exist in UE4 and look great, Fortnite, Battlegrounds, Astroneer…
    ​​​​