The Great Engine War 2014: Unity vs. Unreal vs. RPG Maker 2k3 vs GameSalad vs. [ETC]

Blueprint macros can help a bit on this front. Making good use of them is essentially no different from dividing your C++ code into multiple functions.

Essentially the same for me.

About the same as using any scripting language. Blueprint is merely a scripting language that exists in visual form rather than text.

 

Well, except that it is visual, which is rather a huge difference. It makes sense for modelling a relatively simple domain specific language (a la ShaderForge, the UE4 material editor, World Machine...) - as long as the problem you're solving in that language is limited. But even in the case of something like World Machine, just a slight move beyond simple graphs (and that includes making use of WM's macros) makes everything a rather huge, unmaintainable mess. This in spite of WM not even being close to attempting any real flow constructs - it's a DSL, not a full-fledged scripting language. Unlike Blueprint.

No need to even try it out (which I have) in UE to have that point come across. Just look at the UE4 tools demonstration video and listen to the presenter himself basically admitting how "fairly expansive" just a simple AI script for a butterfly becomes in Blueprint. In code, you could display the entire script with a generous amount of whitespace and indentation on... hmm... around half a screen - and read through and understand it all (I've actually tested that on a few people who have a couple of weeks of superficial programming experience) in less than a minute or two...

Blueprint is really nicely designed in terms of visuals - but it's still a fairly bad idea for anything slightly more complex than "if distance between character and door is less than 4 feet, open door".

 

Well I wouldn't go that far, I've used it for all the GUI stuff and it was simple. All the character controller stuff, also very simple I'm coding the AI but apart from that I've found Blueprints very simple and handy.

 

I still use Blitz3D to knock out little projects and especially for prototypes.
I can take the same project and knock it out in Blitz at least 3 times faster than I can with Unity.
The Unity architecture, although very flexible and powerful, makes many things more difficult.
Just simple things like changing a font and presenting text. In things like Allegro and Blitz this is simple and easy.
In Unity we get GUIStyles and GUISkins. I get that it makes for a more flexible and powerful system overall but at the same time it just adds yet another thing, another layer, that has to be done before you can actually accomplish anything. It just all seems a bit over-engineered to me.

That all being said I am still working with it. I think once I get some decent wrappers around all of this stuff it will become an easy to use game development platform.

 

I actually wound up writing some wrappers that I wind up copying into all of my projects (The Asvarduil GUI System.) It really helps me build GUIs much more quickly (usually about 2 hours per GUI, and most of that is dominated by creating the artwork.)

 

Good thing guistyle and guiskin is a thing of the past, right?

 

Well, that is just an example of what I see as "bloat" and over-engineering.

Basically, I just want to be able to get in there, use the minimum functionality necessary, and get things done.
So, for example, I made a little prototype platform game in Blitz a couple weeks ago. I am converting that to Unity. To display the relevant scoring information I decided to just do the kind of thing I have done in the past... old school style.
Launch Paint Shop Pro. Draw the font characters. Import into unity and just use sprites to display that stuff on top of the game playfield. The kind of thing we did with bitmapped fonts many moons ago.

 

In before the lock? Technology is like religion, and someone is going to want to die on them there swords.
Gigi

 

@visual scripting and blueprint: Personally, as a developer, visual scripting is harder for me than scripting in C#. It's a completely different experience than linecoding. It might be easier to learn if you're starting with nothing, but nothing is easier to learn than what you already know how to do. So for me, Unity is still the right choice, at least until I hit the limitations of the Free version (which I don't expect to happen anytime soon, personally)

 

, you're a funny guy.

 

Actually this topic is intended to be a black hole for such behavior. And, so far it seems to be working, but we'll see when we get about 10 pages in.

Sadly, Trolls can only be contained; their regenerative powers (and whopping 3000 HP) make it hard to truly defeat them. However, they are weak to hammers...

 

I would like to know if Unity 5 will be able to achieve these kinda results? (without having to use the asset store)

I know the old argument, that it is up to the artist, but here is one of the best thing I have seen (of similar nature) in Unity

Though they are both very talented artists, the ones made with unreal just have that extra something, which I assume may have to do with the engine used.

On a slightly different node, this fan made pokemon on cryengine, is looking pretty impressive too (the render quality anyway).

 

I don't see why Unity 5 shouldn't be able to. Most of the looks in the UE4 demo come down to the assets, and Physically Based Shading (Unity 5 will have this as well), since that makes it easier to get materials to look "right". That and good post-processing.

 

W

What you are describing is more level and environment design. Making a cool world (level, corner) to run through is nice, but that's not a finished game.

 

The UE4 video is a staircase with reflections and rain. The Unity video is architectural visualization. I could most likely replicate the UE video in Unity, but I won't bother.

The staircase is not something finished, it's just a mood demo of a few square feet. The Unity video is a finished and most likely paid for project that can be used to actually visualize a space you can walk through. I understand it might seem more impressive, but the reality is that the projects and their goals are completely different, not to mention that the UE4 one isn't even a project.

Now the CE one clearly has some problems:

- First is the fact that it's just CE assets and some pokemon models (that isn't a game, it's more like a quick prototype).
- Second Nintendo isn't going to give you permission to work with the Pokemon license, this has been proven a bunch of times for other Nintendo properties.
- Third everything they have added is sub-par...


It's cool as a fan project, but if they think it will actually lead to anything, well... good luck.

Visually you can say that the other engines are superior, but the only project with value is the Unity one. I can't even compare them: a staircase, a house and some models in Sandbox environment.

 

http://zerolight.com/#/virtual-showrooms/4

http://nvyve.com/our-work/interactive-walkthroughs/

http://survivetheforest.com/

 

I went to meet these guys last week, in Helsinki. Only 2 Artists and its all done in Unity 4:

http://mindfieldgames.com/

Oh, and other than a few obvious moments, thats in-engine and part of a full-game.

 

Oh, and here is a Making-Of video, from ADN, who scanned someone, and imported them into Unity 5 to show off it's rendering capabilities.

 

I think the best way to compare the visuals would be for an artist to recreate the same scene in both UE and U5. Not just use the same assets, but to try to achieve the best image quality. Until then we can't really compare anything.

 

Well skipping past the fact that's obvious beyond belief, the level design / artwork / animation portion is still a massive part of the process and one where the most amount of money get's drained in larger 3D games.

Also were talking about engines not licences, it's obvious Nintendo won't license it's prime IP out.

Fact of the matter is in terms of OMG GFX, there all based on a DX11 / GL renderers. The key component is how much work YOU have to do to make it look good and keep a game performant.

I'm not impressed with any of the showings, the arch viz demos from UE4 are a bunch of cubes and a showcase for lightmass / maybe the particle system for the rain, the arch viz from Unity suffers from poor post and lighting. The pollen trailer is clever, a cool showcase for two artists but as soon as the Bee moves out of scope it starts looking all "Unity again".

The forest would of looked much better in CryEngine!. Not to say it's priority number 1 as Skyrim wasn't exactly a "gorgeous" looking game but sold very well, then again it's three years old nearly.

 

Could you post some visual examples from different engines, that impress you? You seem very hard to please!

 

Witcher 3: Capable of rendering far distance geometry, great terrain / foliage shading / particles / lighting. It is the whole package:

Crysis 3, more of the same. If you use CryEngine and some decent artwork it's pretty much how it comes out the box..

Unreal 4 more of the same, although I've not seen much come out of UE4 yet. Just shiny examples as opposed to games:

 

I had a really big post, but the fact is I decide I'm not going to go down this road.

The gist of it is this: are we comparing what can be made by a modest team, with a modest budget in these engines or what corporations that are from almost 100 to almost 1000 can throw together in the span of 3-4 years (with source access to their own tools and the teams dedicated to their development being in house)?

What you are expecting is out of your price range, even if you can get Unreal for $20 and CE for a tenner. I have seen your project in the past and for an indie team it was good, but what you see in the Witcher 3, is not something the engine will solve for you, it helps, but it won't solve it.

 

You don't know our "price range" and nothing has been shown bar a few concepts to acknowledge the project exists, what gives hope is not what AAA's are doing but invdividuals in the WIP section. There are plenty of examples of small teams and one man bands doing good looking games in UE4, I fully believe you can create games like Arkham Knight / Witcher 3 in UE4. Maybe not to same scale / detail or complexity. But it comes with all the tools necessary to create whatever you want.

Even in Unity you have people like Manufactura 4K doing great level asset packs, so it's not un-reasonable. I just doubt we'll drop 20 - 50 million in marketing.

It's apparent what you can do by just using it yourself, there's no guesswork or opinions to change this. It's still early days and projects take years, we'll see..

 

I think what is really needed is a standard scene, complete with textures, which can then be taken into each engine (by whoever has the time) and tweaked to get the best possible look. A visual benchmark so to speak.

The scene could be something that shows off different aspects of the engine capabilities, so it includes:
- Interiors
- Exteriors
- Dynamic Lighting
- Dynamic Reflections
- Etc

 

I know this is like opening the door to a fireball storm... but why do some of you put so much emphasis on the visuals? I appreciate the talents of a good artist as much as anyone but at the end of the day this is what I think has become a big problem with games.

Too much focus on the "gee whiz look at the shiny objects and the shadows" and too little focus on the game-play. If a person wants to experience the ultimate in real world visuals then open the door and walk outside. Awesome frame-rate, very realistic lighting (one would hope anyway) and subtle shadows. All are there waiting to be discovered. Or put in a movie. I just don't understand placing so much focus on how things "look". Unless at least as much focus is being put on how it plays. And that is usually not the case.

 

Because the customers do, but I don't believe it needs to be at the bleeding edge of graphical technology , just consistent. The issue stemming from what's your competition going to look like in 2 - 4 years time when you release your game?

So you try and make it the best you can now so at the point of release it doesn't look 5 - 10 years out of date before it's lifted off the ground.

 

The problem with a standard scene is that even if you match it in every engine it's not representative of anything really. ShadowK has used Manufaktura40K's packs and still had issues. Who knows how big he wanted his environment to be for example.

Basically what he needs - I assume - might be a huge streaming world. That's great and it's nice that other engines can do that out of the box, but the vast majority of games don't need that feature. The example scene won't cover that, so matching the visual fidelity is a none issue, since you have to match other features first.

I understand his frustration, my game has passed Greenlight and it's Unity based. Getting performance to be optimal on a dense terrain is difficult and a pain, but I wouldn't switch engines just for that. For one, last year, Unreal was asking for 25% of revenue, their tools were crap compared to Unity and you had to conform to their workflow to get things working.
Cryengine was a joke with people waiting for licenses since 2012 and losing projects to forum admins...

As we have asserted in multiple threads in the past engines are tools, use the right tool for the job you want to do. The problem here is that sometimes you have to do a lot of stuff to understand which tool is better for the job you want to do. Even ShadowK says it looks promising based on forum threads, I guess when he done what needs to be done we will know if it was for the best and most likely will be, because Unity's main problem area is huge open worlds. (this is mostly due to a terrain tool that hasn't been battle tested and I think was the result of a ninja camp or external project brought through, it's quite possible you can do better if you invest the time).

 

They might not absolutely need it, but they sure can use it to eliminate loading times or at least keep them very short.

 

My main interest is the visual right now, I want to make photo realistic scenes. Other features of game engines can have their own tests/benchmarks done too for the interested parties.

The only way to really compare all the engines, is to compare each one of their features one at a time and then have an overall summary. Based on that summary, people can better understand which engine suits their needs.

I don't think there is any problem in having a benchmark type scene, it measures one aspect (in this case rendering), in a standard manner across all engines. How you load those scenes and any issues around that is another matter.

 

Other than just visual, have you consider workflow/issues when doing bigger projects? I guess most of you have not done bigger projects, or you will be frustrated at Unity.

 

Switch engines just for that? Where are you getting this info from?

I'll agree, CryEngine was always one to run away from. Unity's problem (in the past) was just big games in general not huge streaming worlds there are TONS of developers on this forum who spoke up, 32-bit limitations caused headaches, you couldn't export to beast or it would fall over. Umbra was hit and miss at the best of times, there were issues with Physx admittedly that wasn't Unity's fault but another frustration. We were always having to scale back to the point where what we set out to do was trashed because of the engine.

I know a decent vita developer going to great lengths to change shadows / post / performance upgrades, I know MMO devs that have been really up against it. What forced the switch is we couldn't wait forever for Unity 5.0, GRFX never came into the equation. It was more about getting the game off the ground...

Yes, Unity 5.X changes this.. But UE4 is overall still a lot less work, it has everything you need and I'm not biased here. I'd probably recommend against using UE4 for mobile ventures and my my UE4 likes to throw a hissy fit every now and again.

It appears to me that mobile has always been the goal of Unity and nobody want's the switch engines, it's a grueling task that can set you back months. But some times it has to be done..

 

The problem with Unity 5 vs UE4 debates is this simple thing: UE4 already has a lot of the things Unity 5 brings to the table. And they've proven they can pull them off from what I hear. Unity on the other hand has a track record of putting in half baked features and then taking years to finish them :-/

UE4's dev speed and the fact that you don't need to shell out a ton of cash any time they add a new major feature is a big draw for me. I'm definitely checking them out for my next project. Learning curve be damned really, there are just so many benefits that I don't think any reasonable developer should stop themselves from giving it a shot.

And yes, the little Unreal logo does make a difference to a lot of gamers.

I hope to be proven wrong by Unity 5 but honestly, a lot of the things important to me have been pushed out to 5.x and that just means its going to be years before I ever see things like an usable terrain engine, better support for big worlds, prefab improvements, etc.

We'll see when I'm done what things look like. Unity has done well for themselves the past few years no doubt, but now that UE4 is here shaking up the market, I'm afraid the success may have gone to their heads given their recent statements about price and such. I don't know what their financials look like but I bet they're losing the majority of the hobbyist market with UE4 soon (aside from the people making slender clones with Unity free and mobile app clone games, teens without expendable income will probably go Unity or Gamemaker).

I'm sure they have a lot of other demographics to please. So who knows, maybe they're just doing what's good financially for them and they just don't care to be the choice for serious hobbyists.

 

All your problems stem from world size, correct?

 

Incorrect or partially, but it's of no consequence anymore.

 

Right now I just enjoy using Unity Free as a hobby and I'm enjoying the new GUI. I like UE4 better right now because of the price and all the features it has out of the box. Unity mecanim is better for animation in my opinion than the tools in UE4 But UE4 is stronger in many other areas. I like them both but i'm a little mad at Unity over the Pro price but I'm sure I will get over it one day because the 4.6 beta is so cool UT don't delete this. I gave credit to both engines LOL!!!

 

The Blueprint system Unreal is beautiful, That cannot be denied. However once you start developing a game, you realize that the number of blueprint nodes that you use is ridiculously high if you want to make anything but simple features for your game. If you don't believe me check out the blueprint samples in some of the sample programs that you can download from their marketplace. For example there is a real-time strategy example that somebody is working on and displaying in the forums. When you look at the blueprints for simply drawing a square around characters to select them you are overwhelmed by the number of nodes in the blueprint, and you start to realize why coding is important and where visual programming has its limits. As for these videos that are displayed, I honestly think that both engines can do them fine, but if you want a completely accurate 100% comparison you would have to have the same team do the exact same video and then analyze the methods/time/etc for how each team accomplished it and go from there. So perhaps for some users here, it might be a good idea to use unity free version and buy Unreal for a month or so because it is so cheap anyway. I love playing around in unreal because it is just so damn beautiful, but when I make something I have to use unity because it takes out so much of the complication, especially since I don't want to touch C++.

 

http://www.twitch.tv/unrealengine/b/562806326

Very nice:
- substance plugin for UE4 is free
- can load and customize substances in UE4 editor (what is adjustable is selected on creating substance)
- the game is shipped with substance materials, so the maps are generated on the client computer (up to 95% download size reduction)
- substance materials can be changed at runtime, e.g. in Blueprint

 

Identical to Unity's support. I do love the demonstration of Substance Painter ten minutes into the video.

 

Wonder how long it will be before this thread gets locked or deleted?

Anyway my favourite engine is Unigine.

 

My favorite engine in terms of look and performance is CryEngine.
My favorite engine in terms of workflow and intuitivity is Unity.
My favorite engine in terms of developer support is UE4.
F this world!