I presume Mount & Blade counts as “open world?” And it was not a big-budget game when it was developed.
Anyway, you can do an open world in Unreal Engine today, with some combination of level streaming and origin shifting.
It requires a significant amount of elbow grease for the programmers on the team, but compared to the amount of work required by the artists to make sure the open world is actually interesting and has value across the size of it, that’s nothing.
You can make a 2x2km world today as a single level (perhaps with sub-levels loading when you get close) quite easily. And building useful content that is not repetitive, to cover a 2x2km area, is a LOT of work! The technology is not the bottleneck.
Yes I’m not talking about programming;
I am talking about filling a whole world full of 3D assets… How one is able to do that without expending millions on 3D Assets.
Yeah… Look at how much people “loved” that game, for example, No Man’s Sky.
(I even had to google it because I had forgotten the name of the game already)
This is where the problem lies these days. “Open world” with zero polish and content.
Horizon and Witcher (big fan of both) did an incredible job and even with those insane budgets and productions they are not perfect. But with a small or medium sized indie team we wouldn’t even attempt something quarter that size, but somehow wants to make open world games these days, maybe it’s a fetish : ). Our general rule of thumb is: if you can’t polish it, can’t afford it, can’t finish it, then don’t make it.
What good is an open world feature when you can’t populate it with decent content/gameplay/story and if you are big enough to afford all those three then you wont be using Unreal Engine, or if you are, you would afford an army to build upon the engine and not ask Epic to do it for you. So both ways UE staff are better off investing their time in the features that actually would make a difference in our day to day tasks.
I believe many people underestimate the monumental task of building a medium to large scale good game let alone an open world one.
There might be some areas where the engine needs more, but I honestly disagree a bit.
Unreal is constantly being loaded with new features every PR. It’s getting annoying. Features and systems that were working have been broken and backlogged, and things that have been broken for a while remain broken. While some of these can be fixed yourself with some time, or searching out pull requests on github, it’s just frustrating. Some features get added and never fully developed, sorta left in that state (gives options for users to expand on it though)
Since 4.12 it’s been a case where moving to the next engine version seems like a toss up between a new feature or optimization and a system that was working, that will be broken in the next version… and you can already see it is backlogged for a fix x.x
The final engine version withou bug, another of the chimeras that have the beginners.
But that has easy cure, start creating games for clients or publishers that impose hard deadlines or you do not collect money.
Quickly learn to stay in a version throughout the project, take care of the bugs that accompanied you during the development, even make some of them a feature of the game
and of course the wonderful world of workarounds in which you will soon become a master.
There is no illusion here of a “final version” I stick to one version of the engine, though will -have- to upgrade due to a serious problem that Epic do not know the commits to fix are. Work arounds are fine but after several versions of the engines, with things being broken and new niche features (I’m not discounting some of the fixes that have come though) I don’t think I’m the only one who is wishing for a stability release like 4.10
Conversations about changing engine versions usually go along the lines of “I need to upgrade” “This version breaks X” “But this version fixes X… But it breaks Y” “The version after that backlogged X and tried to fix Y but it’s still broken”
I bet there is fair reasons why this is happening, but I think posting more threads requesting more features for Epic to maintain is going to hurt in the end unless they hire 2 dozen more engineers to cope with the workload.
Make the game about wandering around in a vast open desert of dunes while experiencing various forms of delusions due to the fact that you’re slowing dying of dehydration.
I’m just curious, I’ve heard this demand many times from some here. How does a feature like this help you with your tasks in Unreal vs just taking the proper time in a 3d app which would be able to model/unwrap faster and more efficiently and later just import it. ( judging from the features shown, it looks like a very simplistic version of a box modeling tool for static meshes where you would be able to do simple poles, walls, tables etc.) modeling these and importing them from any 3d app should be a breeze already, and make sure all the assets no matter how small in the game are properly tracked and saved outside the engine like they usually should be specially when working with a team or in a production environment. So i’m just more curious to figure out the usage of such a feature in engine from a practical point of view.
Well this come from the old school engines idea, there some games keep using this way, right now the tools that Epic added have a pretty slow workflow, the idea is:
You get auto UV, auto lightmap uv and auto colisions in realtime each time you edit the meshes, that is a secondary part the use of this is get a total control of the sizes in the maps and preview and edit in editor without need to reimport or replace something in the map, allow you do edit any wall at any time and edit it and control sizes and collisions in a perfect way, really useful in games where you need control of sizes and space, that is the feature that is made for: Place a cube in the word drag without select the faces and make a wall, press space copy that wall and you can make a new wall and all get auto collision, uv and uv lightmap
And define the whole world layout, later you place the meshes is like some kind of system where you can drag in 3D to design the map and later you can use the base for walls and grounds and that kind of things.
Plus some tools includes too a terrain system you can make tunnels etc or no plane walls with auto collisions what is way better than need todo all of that in a external editor without a direct control in sizes as you can get in realtime.
Triggers and collisions are build over this system.
I guess it’s a matter of what different people are used to.
In our case coming from a traditional animation background, we tend to do everything even a prototype in the 3d app and measure everything there in terms of scale, layout, etc… then move to stage 2 (in engine) rough lighting lookdev, later to assembly and so on.
I guess building even a wall in unreal will take us more time because we still have to compare that wall with our assets and make sure that those match up to scale etc… We would worry about UV and collision much later on.
Also it would be a big no for us to edit or touch up any imported asset on a model level in the engine as this would create conflict with the original asset shared outside the engine.
I guess I saw the presentation with someone editing this sort with VR and that just further added to the confusion, I’ve said it before it’s simply alien to us to have an artist using VR editing this way in a production environment, it’s not only tremendously slow/not intuitive, but will also burn our eyes 10 mins in, sitting in front of a monitor all day is bad enough : ).
A VR game I may understand but actual work is simply strange.
I wondered about this myself when first getting to know ue4. It seems like such an obvious thing to attract people interested in creating stuff themselves and maybe progressing to more advanced development.
Epic has already put out a bunch of content in the form of examples but why isn’t there a well organised full set of asset packs with a few different themes?
It should be reasonably easy to create a great environment out of the box with content packs and a ready made template. Sure you can buy packs but ignoring the cost its a mixed bag. Think about why modding became so popular. Gamers have a ready made game full of assets to play around with and change. Making an outdoor environment for example should be more like cryengine with a few good content packs and better tools and defaults. There’s a lot of messing about to get something good going right now.
At the end of the day i get that epic is mainly engine and dev focused rather than casual user but ue4 could be so much better and it wouldn’t take much.
If you mean content examples for learning purposes then maybe yes and no, because even the best content example can teach you only so much from the basics the rest is just a lot of self learning and hard work. I say it again I would rather have more documentation and tutorials on advanced topics than just packs. You would be surprised how much time you waste on simple pros and cons of what method to use where and how and why during a workflow for the basics when it comes to Unreal, we spent months figuring out some things the hard way when it would’ve taken a few hours/days if things were better documented in some of those areas.
If you are speaking of modding and packs for game usage purposes to populate your game or map with them, then I don’t see how this is any different than asset flipping. Why waste time when your game will look just like any other game that is going to use the same packages all over. From a game development point of view it is a waste of time as any success seeking or serious game development will require to have their own original assets made down to modifying grass and trees (considering grass and trees are the easiest asset flipped assets to use in an environment that goes least noticed) then again the amount of and uninteresting trees and vegetation i’ve seen in games these days is staggering, because they mostly use the same packages with almost Zero modifications or molding to them. Of course that also has to do with composition and lighting but that’s another story, if they are not bothering with phase one than they wont be bothering with phase two to begin with.
Epic’s Staff time is little, and in that little time i’d much rather see them focus on the things that truly matter in every day situations using the engine rather than making it a colorful package for all, and i’m speaking as a beginner in the engine (not a beginner in the industry), I don’t want Epic to waste their time giving me new shiny buttons or packs or other fancy useless tools such as (editor VR to build my landscapes) I want them to give me a monochromatic UI with bare bones ugly text but one that is less buggy and well documented with solid features that will help the game creation process from a small team point of view.
I was thinking of it mainly from a new user or casual creator point of view. We have a starter pack and various examples but these are basic and unorganized. Something more substantial like a realistic theme content pack which could be used to create a full landscape with lots of materials, foliage, building pieces, roads etc. Plus more fully featured templates with enough in there to get basic ai and interaction going could go a long way. UE4 is like half way there already.
In Unreal, you know that the character is just under 200 units, so a “realistic-ish” indoors will use 300 unit tall walls, and a “game-ish” indoors will use 400 unit tall walls.
It’s basically built to make the standard gameplay shortcuts that players understand the easy/default thing.
Regarding geometry tools in UE4, the old-school BSP tools have worked all the time, and I find them to not be as bad as some people say. (Not best in class, for sure, but totally usable.)
There are three markets for in-engine geometry tools:
Indies that don’t necessarily have the budget for an external 3D tool at all. (The might buy characters from the marketplace, but you can’t buy your levels)
Prototypers who are looking for gameplay information. How wide should a moat be for standing jump, running jump, or adrenaline boosted jump? How high should a shelf be for the pull-up climb? How thick do platforms need to be to reliably block all kinds of moving geometry?
Level designers who need non-visual or auxiliary assets. A trigger zone that roughly matches the irregular shape of a podium. A pile of crates seen on a dock in the distance that the player will never get to. Exact collision geometry for a ramp that launches cars across a canyon.
My guess is that most people who are really looking forward to high-quality in-editor geometry tools are in category 1).
