Our migration from Unity to Unreal - the Good the Bad and the Ugly

@GarBenjamin, I'm following along your hunt for the "Goldilocks" engine, and I've typed up and decided to not post about it several times now. I just can't word it properly, I guess.

My main concern for you is that I would hate for you to find out you're wasting your time by using an engine with less "visual" toolsets. During normal gamedev workflow, it takes only a few seconds to ignore existing visual tools. On the other hand, it takes years to build and maintain equivalent tools. For example, just earlier this week, I decided I didn't want to use Unity's Sprite Editor. It would take too long to manually open up each spritesheet and slice them. So, I just wrote an AssetPostProcessor to slice up my sheets, and import the textures with the appropriate settings, automatically. And now, I don't have to look at the Sprite Editor for a majority of my dev cycle. Another example is, what happens when you need to run performance optimizations? What if you need a profiler? Would you build one? It's already been mentioned on the forums that it's possible to build a game in Unity with just code. If you ignore the Unity Editor, you're simply left with your code IDE. Isn't that the ideal working environment?

I'm saying that, all these editor tools are there. You don't need to use them, but what they provide in underlying infrastructure via the API can save so much more of your time than having to build it from scratch. I've spent too many years building my own tools for games that were just never close to completion, because I eventually stumbled upon a part of the toolset that I just didn't want to do, usually because I would realize how much more of my time it would take to complete.

That all said, you already know what it takes to complete a game to your level of satisfaction, so if it's really the engine that's holding you back from game completion, or if you're seeking the "journey" of gamedev and no particular goal, then who am I to say? I personally have another side project in Qt using straight OpenGL, so, if it's an itch that needs to be scratched, it's understandable. I just feel that Unity/UE4 provide so much firepower which directly translates to time savings once you learn what you can do *in code*, particularly when the goal is to actually finish games.

 

Well when I checked out Lumberyard it was way worse than Unity as far as the editor goes. Unity is highly streamlined in comparison. I guess it is just that for me game dev is fairly straight forward. I don't need something like a NASA command station to do this stuff. Just some basic lean & mean tools.

Like all I really need is a solid 3D framework that loads in models and textures and allows me to position, rotate and display them. Basically just a rendering engine. If I have that then I can build the other things I need. But anyway, so basically a lean 3D framework (with animation support built-in would be even better but not required) and then some tools to complement it for building. And I see these as standalone utilities. Like when I want to work on the game world building then I launch the 3D Map World Editor, load my previously saved map and continue building. It saves out a simple scene graph that I can use with the rendering engine (probably need to make a little in-between utility for that) as well as saving out any additional information I want (probably need to put that in the little in-between utility too).

For me personally things like the Lumberyard editor don't help me they just get in my way. I'd prefer a dozen or more little separate utilities. It's like 3D modeling and 2D art. I suppose there are people who think that stuff should be included inside these editors too. I much prefer them to be separate. Want to model go into Blender or whatever you like to use. Want to do 2D art launch the paint app / sprite editor of your choice and get to work. Well, I see all of this other stuff that same way and cannot understand why it is all crammed inside one huge editor.

Just a matter of preferences I guess.

 

I'm confused as to what you're looking to achieve out of all of this? You looking to be a games developer or technical guru? I get you do it for fun / hobby, but what's the end goal?

I point out flaws and shortcomings of all engines, just like if I was doing a car review I'd weigh the positive / negatives of it in relation to other cars. But the end of the day, they all get you from A to B...

So just because an edtior is a bit clunky I.E Unreal / Lumberyard, it ultimatley doesn't mean much.

 

Ultimately... to simplify. And I can see the trend is going in the opposite direction. One day I expect all work from laying out a scene/world to modeling, drawing the required assets and so forth will likely be done right inside the next gen game dev monolithic editors.

What I'd like for my own dev environment is:
App to make 2D art
App to make 2D animation (just auto-building the code for defining the anim tables)
App to make 3D models and texture them.
App to rig and animate the models.
The World Map Editor to build the world environment, lay down collision zones, position enemies, items and so forth. Saves a couple different representations of that data. Basically one set of data is a render file optimized for rendering. The other set of data is a query file optimized for querying the environment in the world.
API to easily load, position, animate and render the objects created previously.
Sound app.
Music app.
API to easily load and play the audio.
API to easily accept input.

That is basically it. Audacity and other available apps already take care of some of these things.

So once these foundational items are in place it is just a matter of programming to bring it all together.

And of course the other end goal is for it all to have been a very enjoyable experience.

 

I'm not positive I would consider this a simplification.

You could easily have Unity handle this task. You'd simply have to write a custom export script for those files.

 

Except this would be a terrible design, exactly like having a sports car with a trailer on the back and built in microwave oven. It would just get in the way, and suck. The innovations in content creation take vastly different paths and just as much money as the innovations in game engines. The two cross over, but they can't travel together.

The reason terrains are frequently built into engines is because terrains are the worst possible performance killers - if you want them to be, and thus need to be as close to the engine as possible, where possible. This isn't the same for 3D assets in general.

IMO far better to give the developer compatibility and the ability to add such tools if desired, and you could probably build blender into Unity just fine with enough patience with the source. Maybe someone should.

But having all the tools in one place doesn't really make much sense. How can you then innovate?

 

Doesn't blender technically have a game engine built-in?

 

I think it would be terrible too. However, that seems to be the way things are heading and I bet there are many people right on these forums who wish the art creation process were new menu options inside Unity. Do all of their 2D art and the 3D modeling etc right in the Editor.

 

How does not having all the tools in one place make any sense? In fact from a level design perspective it's been common place for the last twenty years in AAA engines. Lumberyard / CE has crydesigner, Quake's ID had a level editing tool, Valve's source has hammer, Unreal has BSP which is soon to be replaced by geometry 2.0 (more full featured). Snow drop has en mass art tools, including a procedural city generator (bit niche of course) but y'know it has them.

Y'know why? Because it's simple and easy, if were talking innovation? Exactly what's so innovative about art tools for games? What do we have new really in the last decade? Sculpting? Being able to paint textures (which I was using vertex / texture paint / layer weight tools in engines who knows how many years ago). Engines have even been built atop of Maya..

There is only ONE reason not to have these tools in engine, because you could have an art UI view. Programmers may never need know it's there...

That reason being, most engines struggle to be optimised / efficient / simple and bug free as it is.. Until they get to the point where they're bored, there's too much going on already..

@GarBenjamin

Sounds like you want a slim version of Unreal, it'd take you much more time to build that through a framework than just get used to something. Unless you want to do it for the challenge?

But it didn't really answer the question, you want to be a games developer or a technical architect / llc programmer?

 

They can call it that, but I wouldn't call it that.

 

I prefer specialized software like Zbrush so i can concentrate on creation instead of gameplay.Perhaps Blender is the way you want ? A modeler ,sculpter, animation tool that have a game engine ? Also Stingray should suits you a lot as the workflow is straight forward between Maya and the editor ?

There is no perfect tools or engine, and lot of work around some times, you should start your game instead of waiting for the perfect software that will never happen.

 

Nowadays pretty much all largest asset development softwares are crossing all sorts of areas. Years back it was usually the rule that best softwares are those that are specialized on certain areas, as time was spent better from software devs on improving the quallity of features. Today you have everything in it, retopo,sculpt, uving, hard surface, but the main difference is how well was it actually implemented for use, its not all about functionality, especially not when it comes to design. 3d coat is still from that view probably the most interesting and inspiring in terms of how many features was actually implemented in it covering wast areas, and still everything in it is working as good as you would want it too, not to mention that at least speaking for myself i had only 1 crash of that software so far, which is impressive considering that it was and is being made by tiny team. As for Zbrush being specialized not really sure if you can call it that. It has tons of cross features, which cover nearly everything on basic side with exception of animations.

 

That's exactly what I said about the terrain. These tools do not replace 3D package functionality. They expose workflows which are mandatory for the engine to function well, ie terrain, bsp. These are special cases. The other cases you mention (procedural city) are very, very game specific, and such things also exist in Unity - via asset store.

Oh yeah - the bsp tools... really suck.

Also, epic and Unity combined can't make Maya*. They'd have to buy it, and half their users don't want to work with Maya*.

* Replace with tool of your choice.

 

Meanwhile we only have very basic primitives within Unity.

 

That is exactly what I am getting at. The concentration part you mentioned.... focus. As you said when you are working on asset creation you want to be inside a dedicated asset creation app... Zbrush in your case.

While I might try out Blender for game dev just for the novelty of it that is exactly the opposite of what I am looking for.

Basically, I think all of us are at widely varying views on this thing. Some would like every tool stuffed inside one massive editor. Wouldn't matter if it required dozens of menus with dozens of sub menus with each of those having a hundred menu options... those people would be praising such beasts.

Others like focused tools free of distraction and just free of anything else period. Focused solely on the task at hand. Clarity of focus I guess. It seems like I am at an extreme at this opposite side because (for example) @Ryiah said she didn't know if she would consider my list of dedicated tools as simplification. To me it absolutely is. Because each tool is only there when I explicitly need it and launch it. When I am inside that tool the focus is clear because the tool only handles certain things. When I am not into that area of work I don't even need to see it. Not even a menu item for it (although I could certainly handle a master app that had a single menu item for every major piece... 2D Art, 3D Modeling, etc that launched the dedicated tool.. maybe 4 to 6 options total in that master app interface).

I guess I am very task oriented. When I am working on something I like to concentrate solely on that thing. And also I believe that tools dedicated to doing one thing and only one thing are generally much better than those found inside all-in-one solutions.

Again, this is just my view of it. This is true for me. It may not be true for other people.

Anyway, just call me a weirdo and we can all go forward. LOL

I need to get back to working on this thing.

 

I guess you guys completely ignore the entire modelling tools and systems on asset store. Why would native modelling be good for you? There is even a super robust BSP modeller on the store.

I could see sense in it for something like a live link to Blender and other packages, so you edit the geo live in say, blender and it updates instantly in Unity, or even an integration of blender somehow, but you can forget keeping the UI uncluttered with the latter case.

And if only simple tools with no clutter is what you ask for, then asset store has this covered.

 

Yup BSP's do suck, but in all fairness the system is a bit long in the tooth now.. How long has BSP been around? It was definately around in early versions of Unreal 3, maybe Unreal 2? The whole point of them is, you set up your scene then convert BSP's to static geometry anyway. Then export out of engine into said package to add "detail"..

But they have nothing on more modern design tools (like hammer / crydesigner).

Why couldn't Epic and Unity combined not make Maya? I dare say making an engine is harder, softimage was pretty good and their team was tiny..

 

What focus issues? What impact would it have to just add one drop down item saying (Modelling view)? Do your level design and flip back by one keyboard shortcut?

It doesn't have to be cluttered, neither does it have to be that difficult if it's well thought out. End of they day, you can only work on one thing at a time so why are you "un-focused" on the task at hand?

If you have sub-menu's, clutter with little explanation of what said icons do and everything stuffed together you have to search for. Y'know what that's called? Crap UI design.

It does sound a bit like an excuse, like when I'm faffing around with lighting / shaders when I should be doing something else.. I'm complaining the lighting is hard to get right, whilst secretly I enjoy doing it ...

 

I don't work that way. It is not better to lay down your ideas on paper or drawing in screen concepts before starting the game ? Than you can start creating the first models you need, than you move on gameplay.
I don't think mixing 3D art creation with programming in a same session at same time makes sense.
If i code gameplay i do that only , if i notice models details that need change i take note , but i stay focused on gameplay. Next session i work on new models or i improve the existing ones.
Otherwise i would loose focus on polishing each task (gameplay coding or 3D art work).
I mean when the game is about open world , with terrain, characters and buildings , there is not BSP work. This is different workflow, this is not the same way of working like you could have with Source Engine.

 

I thought we were talking about stock Unity, without any tool assets installed... I would definitely not want my 2D/3D/audio content creation editing to be done inside Unity. But the merits of stock Unity are already several years' worth of effort.

 

I think overall, you'd hate Blender. Its original design philosophy was to have every UI control available to the user all at once. And now they're in a pickle, deciding to tuck things into tabs, and design leadership is still against modal windows. But if you can get past that, it's just as usable as Maya. Maya has the opposite problem, in that nothing is exposed to the user, and you have to dig around to find what you need, but once you are experienced in it, the interface is clean and focused, because the UI is not cluttered to begin with.

 

The issue with these arguments (or debates) always come down to a misunderstanding of workflow. In 3D games, it's common to have level editing features for many reasons like lighting / materials / navmesh / import pipeline differences / smoothing groups.

It's just far easier to get these bits right in editor and it's a common practice spanning decades back in AAA games. These tools are generally focussed on one purpose and one purpose only, that's level design and most AAA game engines expand on this, like Unreal's spline tools and CryEngines river / water toolsets..

When we start talking about "art" in general, well that's a completely different topic.

 

Ha ha. Well I don't know man. I am trying to describe my views of it. It's odd really. Generally, I have no problem blocking out everything else and focusing on the task at hand. I get so caught up in what I am doing I am oblivious to the fact there are loads of people just watching and so forth. But for some reason I just cannot click with these all-in-one editors. So, am trying to think through it as to what exactly it is.

And I think it is simply that they just seem so overkill and basically unnecessary to me.

I guess maybe that is it in the end. If want to do some modeling I will open up Blender or even Anim8or. If I want to do uv unwrapping I use Ultimate Unwrap 3D same thing I was using years ago and was happy recently to see that it is still around. If I want to write some text notes I open Notepad++.

In nothing else do I feel the need to have a multi-tool editor and I see no need to have one for game dev. Like I said, it is basically just a personal preference. I guess in the end it is the only way to explain it. That and maybe I am one of the "less is more" kind of folks. I'd rather have just some base functionality that is very well done. Very solid and then I can build onto that anything I want without bumping into issues with canned systems, "their" (whoever) design ambitions and so forth.

Right now I am thinking I could work with something like this:


With each of those buttons taking me to the appropriate standalone tool.

It's not really needed though other than a way to wrap everything up within a single interface.

Anyway, time for some D3 and then back to digging into OpenGL.

 

This is what i like about Epic, because they make games they bring amazing tools for the editor to help you make things easier and faster
www.youtube.com/watch?v=4gBAOB7b5Mg
Should Unity try to make a big AAA game ?

 

Well, umm yes and no I guess.. Like I'll never need 2D tools, I never need options for mobile but somebody else does.. So do we stick them in another room because I don't need said tools?

Personally it only becomes an issue when things get in the way or is extremely convoluted. Take Modo for example, to get a color map baked out, you select the master shading node (little circle thingy) set it up as a diffuse co-efficient whatever that is, change the image size from within the master node in the shader tree. Then search through some more menu's to bake something render to something outputs.. That type of workflow annoys the living **** out of me..

Another example, setting up the UI.. To turn off trackball rotation, it's not under mouse ohhh no.. It's under preferences and then OpenGL.. Again annoys me beyond belief.

I don't know why I use it.

@zenGarden

Yes, yes they really should..

 

Well there's 2 precedents for all this:

Blender's game engine and stingray. But Stingray seems like the practical approach. Just get some form of backend live communication going between major 3D apps and your game engine. This is practically same thing for none of the cost.

Unity largely solves it with its support for max, maya and blender: just save in the 3D package, tab to Unity and it's changed and updated (native, not the FBX, although it does convert behind the scenes).

 

Well , there is also one on Unreal forums "Moving to Unity"
https://forums.unrealengine.com/showthread.php?99972-Moving-to-Unity!https://forums.unrealengine.com/showthread.php?99972-Moving-to-Unity!
No tool is perfect, just take what fits the best to your game.

 

Only if it was a proper "bridge", I'm talking a two way synchro between engine / DCC. In which you can update freely from boths sides. Hit an update in Unity and it imports the level / assets in which you can then modify in the DCC and click sync again to update the scene / then playtest.

I could see no downside to a solution like that and it would be majorly awesome.

Probably would be best of both worlds.. I used Maya and the send to Unreal / Unity option just dumps the .FBX in the selected folder.

 

No, I haven't ignored them. I'm just wary of buying expensive solutions without any source access. I'll still end up purchasing one or two of them at some point though simply because Unity won't get around to making their own any time soon.

Right now my current asset of interest is SabreCSG but even without source ProCore is very tempting.

 

Unity does not make games , so they have no need to make great tools or have today terrain graphics for example.
I understand people that don't want to bother with Asset Store for basic features that are missing ( vertex painter, cinematic tool, spline tool , better shadows, visual coding ...).

 

This was added in just to keep xamarin away; very good reason imo.

 

Because those tools on the asset store are *very* expensive and Prototype/basic probuilder/whatever it is called now doesn't have everything you'd want from level design tool.

Why did Unity spent time on UI? There are dedicated tools on Asset Store for that.
Why did Unity spent time on 2D tools? There are dedicated tools on Asset Store for that.
Why did Unity spent time on UNET? There are dedicated tools on Asset Store for that.
Why did Unity spent time on PBR? There are dedicated tools on Asset Store for that.
Why did Unity spent time on Enlighten support? There are dedicated tools on Asset Store for that.

See my point now?

 

Unity don't have necessity for quality features and production tools , because they don't make games. They release features for everyone , but you must pay for quality and advanced tools (cinematic motion blurr and dof, cinematics tool, spline tool, vertex painter ...) . This is the way Unity works.
I think you can't ask Unity to become similar to UE4 with many things out of the box, while you can't ask UE4 to become very easy with C# for example

 

That depends if there are unity devs reading this thread. Now, I'm not sure what is the financial situation of Unity as a company, but if it's good, I think buying company behind Procore/SabreCSG (though procore would be IMO better) and then integrating it as standard engine feature would be a good move both from technical standpoint (you just integrate existing code instead of making new one, heck you could even put Procore as part of Standard Assets) and from PR one (now Unity has nice level editing tools! YAY!).

 

Most of your examples are very sound but this one is kind of out there.

 

No ,because there will be no income for Unity throught the Asset Store, if the tool becomes free and integrated on the editor.

 

Not sure what point there is bolting an entire application onto Unity is, tbh. Your point means Unity should also bolt an entire 2D package like Photoshop on as well. Let's not forget Logic Pro and FL Studio too.

Probably won't hurt to throw an operating system in as well.

 

AFAIK there isn't much of income in Asset Store by itself anyways + there are other items people buy there.

 

Your analogy/hyperbole fails. Photoshop/FL Studio are general-use programs. Procore/Sabre/Whatever Unreal uses now are tools specifically made for level design. No sane person would use them for making other stuff than level's geometry.

 

And that's how unity technologies got a wonderful idea to create their own video game console.

 

Unity main income is coming through the assetstore. They wont include features when they can get $$$ due the assetstore. Examples like UI, yes they integrated / updated it. Why? Because the market was dry. The sales of nGui went down impressive, therefore the cow was melked. I bet we soon see other features implemented as soon as their sales going down on assetstore since everybody purchased them.

 

No, it's not their main income. If it were their primary source they would have quickly caved on situations where people threatened to quit using Unity if they don't have access to the source code, access to the dark skin, the ability to remove the splash screen, etc.

 

I think @hippocoder was getting at the general line of thinking.... people use a 2D art program to make art for their 2D and 3D games. Looking at it in a more dedicated way replace Photoshop with Anime Studio (vector art) and Pyxel Edit.

The main point is when does it stop? How much stuff do people want crammed inside these editors? People need music and sounds too. Should things like Acid Music, Audacity and Bfxr be included in the editors too? How about providing extensive support for input controllers? Might as well stick all of In Control functionality in there as default. Okay, actually I think that functionality should have been in there all along.

 

It stops when it's no relevant use within it's context, level design tools have been about since 3D was created. It was adopted by AAA and been a thing ever since.. It's primary purpose is prototyping, it's quicker to prototype in engine than it is to do it in a DCC, so that's why it was so widely adopted:

Either @hippocoder is being a joker, or doesn't "get it"..

 

There are already a lot of things in the asset store that do that, same with animations. I think I prefer that method rather than it being built-in because you can keep Unity smaller.

 

This is nonsense. It's still pro licenses.


Regarding the bolt on stuff: The thread context so far was we should build in a complete modelling tool. My response was: where does it stop? Why modelling tool? why not 2D tool? why not sound tool? It never stops.

As for prototyping with simple geo, I am coping extremely well blocking out with primitives tbh. I have probuilder and a bunch of other CSG asset store offerings (which are actually a lot better than UE4's built in tools) and I don't use them as I'm happy to prototype with primitives.

For serious prototyping for level design, something like probuilder is useful if your game absolutely requires doing the block out in editor, and asset store solves this. If you want it built in at UE4 quality, you better realise that there's better available on asset store, since a lot more care is taken over that than basic built in support (which is useful, I agree - but far from the suggestion of an entire modelling package...)

TLDR: DO agree basic modelling support is helpful, with emphasis on basic. I'm sure Unity is considering it - they have added basic 2D prototyping elements. Discussion derailed to prototyping when REALLY the original suggestion was for a COMPLETE 3D package built in among other tools.

Anyway get this (free): https://www.assetstore.unity3d.com/en/#!/content/11919

 

For the purposes of discussing the underlying explanations for @GarBenjamin's needs, I'd rather stick with the tools that Unity already has in its stock package. For example, the tools I mentioned, the Sprite Editor, the Profiler, and what @hippocoder mentioned with terrain. And other systems like Mecanim or even just Unity's animation state system, the lighting tools (which I frequently read that are broken), the asset processors for drag-dropping resources, the real-time hierarchy and scene visualizers for debugging, and the many others that I am missing.

What of these tools and systems feel like they should be separate programs, or would otherwise gain in either quality or production efficiency by replacing what Unity provides?

 

Many UE4 features and new ones are available throught Blueprints so UE4 editor don't have grown.

 

I just tried to play Ark and this may very well be the first UE4 game that I actually played. Can some of the people having experience in playing UE4 games please tell me if Ark is an outlier or if all UE4 games have borderline unplayable performance? I thought with dayz I had already seen pretty much the worst performing game ever, but Ark seems to run even slower on my PC. When I adjust the settings so that I get ~30fps at the starting beach (didn't get further than that) it looks way worse than any decent Unity game I've played (e.g. Verdun, Sattelite Reign, or even 7 Days to Die).
@Tomnnn I think you played this too, right? How was your experience?

 

IMO, ARK is slow because the development team went overboard with visual effects.

You can check the list of UE4 games here:
https://en.wikipedia.org/wiki/List_of_Unreal_Engine_games#Unreal_Engine_4
Most of them will be TBA, because the engine is faily new, plus it changes fairly quickly.

On of the released games were Kholat, Hatred and The Park. I don't exactly remember a lot complaints about performance with those games.