Tried UE4, I'm Back

It always seemed to me like Unity had the edge in user friendliness and Unreal in high end features. And slowly for each version they both move towards each other. Unreal improving user friendliness and Unity the graphical capabilities.

I think Unity 5.x will be comparable to Unreal in terms of graphics. Right now there are a lot of components that show potential but are not exactly "there" yet. I mean in 5.2 it's the first time I've played with Enlighten and the results made me think "oh cool, I can work with that". And even then it's still kinda buggy and it still has ways to go until it reaches its full potential.

 

Ok, so I goofed - Linux support is currently "Community supported". Which is why it is not on the main site.

Sorry

 

I think more content / workflow examples would be cool for learning. Like Blacksmith demo, Epic just released Infiltrator to study which is awesome.. It's like the sort of game I'm looking to make so, of course over the moon. It'll make life much easier to reverse engineer the example.

Biggest thing I learnt from BS demo is, avoid terrain system .. Nah J/K skin shaders were cool, atmospheric scattering examples were nice and to say it was done by a small team seemed very professionally done.

But ultimately I agree, it'll be interesting how things come together.

 

I thought that was a given with Unity.

 

We're seeing similar trends in gaming with AAA vs indie. Rocket League might be the best I've played this year!

Sounds worth a few minutes looking into. This weekend I looked into azul3d, pandajs, wave engine, mink engine, and oxford commas. I found a few of them interesting / useful.

Oh... I guess this engine isn't for me. It definitely looks nice, but as a hobbyist I only shell out ~$100 or more for dodgy VR equipment that promises to deliver at least 2 years after they promise to deliver.

I've noticed this as well. The showcases and tutorials really show they're trying, but it appears to be a struggle. Compare the blacksmith demo to samaritan. Compare the years on those release dates. Unreal visuals are simply unreal. But as for user friendliness, I think I have officially chosen unity for VR... at least initially. I want to get something running and performing reasonably well with little to no graphics in as little time as possible. The choice was easy

Maybe at some point in the future one of the two giants will consume the other, and we'll have the best of both. As much as I love UT, I hope it's unreal because their revenue model scares me the least. "Use it for free, but give us a small portion of your success". Does anyone else do this? That's the kind of software I can invest my time in and not worry about!

--edit

Checked out paradox. That "free for 1.x" certainly is suspicious, although the "open source coming soon" hopefully means 2.0 will continue to be free. Atomic's comparison area made me giggle. "choose one" for the modules on the free version.

I think I'm gonna go back to text based games and write them with GO. I spent the last week and a half looking for some kind of simple engine that made use of GO, and I couldn't find one. So text based console game it is!

 

Make a Multi User dungeon. Does not need to be massive just awesome

 

I'm not crazy about internet enabled games, but I do enjoy LAN. I'll write up something singleplayer & small in GO to learn the basics, then try out some network code. I know very little about GO, I've written only 3 programs in it.

 

Go is really awesome, I don't know anyone who has written nontrivial amount of code in it that doesn't love it. Enjoy

 

A Multi User Dungeon doesn't strictly have to require the Internet. You could create something similar to NWN2.

 

@HeadClot88 Atomic is very far from Unity or Unreal. Did a quick test. Undo in the scene view doesn't even work, lack of import formats, bare bones documentation, and many other things in a short time. The problem I see with smaller engines popping up are they miss things like scalability for high end systems/complex render paths, years of optimizations for driver/platform issues, OS integrations ...

 

Serious question: what should I use instead of terrains? I've noticed adding a simple terrain with 1 texture adds about 5ms to Camera.Render in the profiler. Should I use the terrain collider only and render it with tiled meshes instead?

 

I feel @ShadowK would likely know a better answer to this. Being a hobbyist I don't worry about performance too much as most of my work never makes it online. There was an asset on the store I was considering trying out, but my finances tend to be pretty tight. Saving up for a new graphics card currently has priority and will take a few months.

 

I probably found my answer here:
http://forum.unity3d.com/threads/terrain-reduces-mobile-performance-by-35fps.312728/#post-2031773

@hippocoder wrote:

So I'll try that and see if it helps. If anyone has anything to add to that topic I'm still interested!

 

Interesting way to approach the free version: You chose one option from the 3D module, Android export, or iOS export. So you can have anything, but not everything. That might have suited Unity, but that ship has sailed.

 

In terms of community issues we had a lot of the same until changes to the forum structure. Remember when collab wasn't locked up nice and tight? Hmm? Remember those days? "hai, we maaek fpsmmotps city builder, join team nao!"

It unfortunate that there isn't as much forum support. They do seem to shovel out updates regularly though. I suppose for the more independent types that's okay. I've come to appreciate the engagement we get from the devs on the forums and answers a lot more than when I first started.

A group of friends tried to use UE4 to make something recently, even had coders that preferred CPP!... Lasted about 3 weeks. Didn't produce anything to show, unfortunately. Could just be a lack of experience to blame but coincidentally they did stop teasing me about my progressive Unity projects after that.

I agree with your concerns about base features, but I think the crutch of the Asset Store is both good and bad in that regard. On one hand its there for basically anything you need and the good products basically solve your issues without any trouble so you can just grab them and move on with your life as if they were native...

On the other hand IMO having those secondary tools there has allowed UT put off updating some critical features like Input and UI so I think its a double edged sword. The solution was there (NGUI etc, ReWired/cInput/InControl) so it was acceptably delayed to prioritize other features and fixes that we like so much. I don't think it really matters at the bottom line, if you need a feature/tool just buy it and expect returns when your product is published. The amount you'll spend on tools from the store is really insignificant - if it isn't then you may be doing something wrong.

 

Meshes, generate them in world machine / sculpt in something like Z-brush and re-topo via Z-Remesher, I also use UV Master in Z-brush for UV's of course. Then use something like Mari / Substance painter to texture it, looks sweet..

There's foliage tools I believe in BS demo? So you could mimic that setup.

@LaneFox

The asset store covers around 90% of features for me personally, there are some middleware integrations like APEX which would require Unity to sort out. Had mixed success with the rendering pipeline, seems there's always a way to do things but a lot of it's "workarounds" in places.

A lot of stuff we added via plugins, so it wasn't too bad. From a developers perspective, you of course can get around a lot of problems.. It's mainly a time thing, reliance on the asset store can be a mixed bag.

If I had a couple years of "downtime", I'd try and develop a lot of solutions / integrations and annoy Unity every 5 minutes asking why X doesn't work and how we can improve it. We'd get there eventually so there is of course always a way..

 

I'll probably develop a minion / follower system. But if it's all text I'll have to be careful with how it's handled. If a user has 50 minions they might get tired of waiting for their turn lol.

There's a few oddities but I do like it so far. Go won't let me use a goto to skip a variable declaration I'm not using a goto because I care about good design or practices!

Do they have to know what they're talking about? Because if not, I have an idea. How about you fake a terrain with your own heightmap system and for the visuals you simply flash an image in a separate camera and don't refresh it

 

Yeah thats probably the only other thing I would add to my post above. Unity will stand behind their product and support/maintain its features while 3rd party assets are not guaranteed and that can be a risk for long term project lifecycles.

 

Use what works best for you and your team.

Both have their ups-and-downs, but you can't make a wrong decision with either tool, as they are both fantastic in their own ways. The barrier to entry is lower than ever, so give them both a fair shake.

I wrote a detailed post about both engines operate under the hood when it comes to things like asm.js in the browser and IL2CPP, for those who want to go deeper in comparison.

 

UE4's editor was atrocious (I think build with QT by non-UI programmers) but based on what Ryiah said I'll try it again this winter.

 

The Go community is very big on writing idiomatic Go code, heck you can see it just in the fact all Go programs have to be formatted the same way. I think it is a good thing, it produces very readable code that is easy to share. I've had a hole in my heart since I stopped using C and Go has filled it quite nicely I think it is a really strong language for writing network servers in, including game servers. I've even considered rewriting CircleMUD in it, funny enough.

 

I'm actually a little deflated at the moment, one of the main reasons I stuck out UE was because of the flagship Infiltrator demo.. Being straight up honest I've done better in Unity, that's not a brag either as I always want something to look up to. Something I can reach to the stars to try and achieve.

You look at video demo's of Arch Viz and Tech demo's and they look like some sort of mystical un-achievable piece of awesome. Then you try them in editor or a build and I'm like meh...

Somethings not quite right here.. It's a lot of pain to go through to take up UE, don't get me wrong tools sets in UE are amazing and they do a great job keeping current.

Maybe I'm wrong, but still a little bit disappointing.

 

This sounds interesting, can you elaborate a bit more?

 

Edit: MMMMnn never mind, talking out my bum >.

 

UE4's UI is a custom solution they made called Slate. They only included Qt/GTK for system dialogs on the Linux port. With 4.9 though this changed and now Slate handles those aspects.

 

The problem with UnrealEngine4 is, that all the shiny demos they've been shown so far are either done by a army of industry veterans and/or are unusable with common gaming rigs. It gives small indie devs a skewed picture, what they can really achieve.

Of course, there are some impressive one-man stuff like that PT clone. But that is something Unity can achieve too, given you know what you do and buy a couple of third-party plug-ins.
Just a couple of days ago I saw Layers of Fear and thought it is powered by UnrealEngine because of the FX and lighting. But it turned out to be Unity.
If you aim for a actual game and not just some showcase demos. Unity is fine enough. But there is some stuff Unity needs to fix, or get. Like a node-based shader- and scripting editor, better standard shader vertex painting etc.

 

The mods in the UT forum are also incredibly unbiased!

I suppose that makes sense. Still, sometimes a guy just wants to jump over code. I wonder if goto would be able to jump variable declarations if said declarations were inside a closure... hmmm...

I'm looking forward to writing some console based network programs. Maybe this venture will lead to a great solution for something unity related in the future.

You seem to know something about GO. Know of any window / graphic libraries?

 

Definitely not it's strong point(yet) but I have been poking around a few lately. If you are talking about stuff related to games your best bet is probably either go-sdl2 for desktop or the brand new bleeding edge golang.org/x/mobile package for android/ios. Both seem to have their issues and I'm not sure I'd recommend them at this point, but they are out there and you can get some sprites moving around the screen with either.

There are a number of go game engines however, such as https://azul3d.org/, don't ask me if they are any good never checked them out.

Just write your server in Go and use Unity for the client- win win.

 

I don't really think its fair to say that these are things that we need to 'fix'. Not having these features built-in to the engine, and out-of-the-box, hasn't ship-stopped the hundreds and hundreds of thousands of games that have been released with Unity; but they are a nice addition to have to the toolset we currently offer. But currently, these things are available by either learning to write shaders and code (Via our and/or community's tutorials), teaming up with someone who has these skills (Via the collaboration forum or similar forums), building these tools yourself or having a look at the offerings on the Asset Store. Of course, looking at our roadmap, we do have plans in the future for Visual Scripting and we have spoken in a few Unite roadmap talks about things related to a shader-editor..

 

Unusable with common gaming rigs? Or unusable with normal computers? There's a very big difference between Intel HD and an actual graphics card. Intel may have come a long ways in the recent years but they're still several generations of capabilities behind AMD and NVIDIA.

Most of my computer components have finally hit the five year mark with some of them being even older. They weren't very expensive at the time - I think I paid all of $80 for my processor - yet I have no trouble running UE4's editor and the editor is much more demanding.

 

That's what I figured, at least for now. I saw gocos2d but it didn't look promising. azul seems to be too early at this time. Definitely excited for the future though.

 

It's more along the lines of someone else has done it for you, we have ShaderForge and before that there was Strumpy. Plenty of VS tools out there, the fact of the matter is Unity's community and diverse skillset is what's awesome. Pretty much anything you want that doesn't require direct access to source is available for a small fee (or sometimes free).

There's never anything stopping you, plenty of people were building their own engines and / or using XNA and releasing games long before Unity became a dominant market player.

Without the community supporting you, I'm not sure how things would of turned out? I could hazzard a guess you wouldn't be in the same position. Unity is simple to use which attracted a huge user base, which then more advanced users flocked to it as well because they could earn a living out of the asset store.

In the day and age of ever increasing games and competition, the only factor now is time. In a commercial sense, you'd never build your own engine or framework anymore as an indie. Only reason I mess around with my own engine once in a blue moon is to keep skills sharp, I wish I had time to implement modern tools and middleware into Unity I honestly do. Sometimes I believe it may be a better idea to sell the shovels..

 

Yes I meant usual gaming rigs based on Steam's hardware survey. Because most of them couldn't run something like the Kite demo, or the Samaritan demo back in the days. It's simply showcase stuff.

That's like saying "Oh look that dude did swim seven Km with just one leg and arm". It is possible. But it's much comfortable to do so with all limbs. It's 2015 and such things like node-based editors are the standard in game-engines, not the icing of the cake.
And referring to the AssetStore for Unity's shortcomings never does you a favor. I barely buy anything there not because it cost's money, but I never can be sure how well it plays with other third-party tools, and how well it plays with Unity updates, let alone how good the support is, mostly from lone developers (DaikonForge anyone?!) So I prefer certain features already build in, with official support.

 

I've become doubtful of whether or not Steam's survey is worth paying any real attention to. At this point I feel like Steam has become the desktop equivalent of Android's Play Store with the seeming ease it accepts games through Greenlight. Practically anything seems to be acceptable now.

Many of these games aren't requiring much in the way of hardware because they want to target as many people as possible and thus are drawing in people who are running systems that are very much borderline on graphics capability. Unfortunately unless I missed something the survey won't show hardware usage for a given game or genre. At least not publicly.

 

That might be true.
But I regardless of that, I doubt there are a ton of people who have extreme high end machines either.
I didn't say you can't make games with Unreal Engine for common gaming machines. But Epic tends to create demos that are some years ahead. The Kite demo for example ran on a Titan X with 30fps.
And it's done by veterans who pushed the limits. They even were flying around the world to scan real world locations and props. Nothing a indie has resources for.
And that was my whole point. Epic targets Indies, but shows stuff that is not easily reachable by a small team. Unity seems more down to earth in that regard. It's not as fancy but usually fine enough for a real game.

Still, UT needs to step up their game and put more artist friendly tool like node-based editors, vertex painting, cut-scene editor etc. into Unity. It's cool that I can write my own shaders and code, but I'm a artist and Unreal is way more artist friendly.
I simply don't want to bother with writing my own extensions, or buy them from the Store, for the reasons I wrote above.

 

This boggles the mind. I don't even know where to begin with how silly this statement is.

Here goes: Do you honestly think that using Unity is comparable to performing a physical task with a major physical disability?

 

Before or after Unity 5 Free gave us the entire suite of engine features?

 

Yes.

 

If you don't know how to program but you know how to use visual editors, yes. So maybe the metaphor would have been stated better to say that programmers who can't program are handicapped [like missing a limb] and visual editors are like prosthetic limbs?

 

We are many lurking into UE4 because of it's great toolset.
Vertex painting is really very old and simple features, i just find incredible Unity not bringing that to users with some ready to use multi texture shaders example
About the shader editor, yes this is great for non shader coders as we can tweak, try things and make interesting new effects, unity should buy ShaderForge and improve it like they have done with the new GUI.

 

I think Unity needs to solve more important problems than native Visual Editors. Stuff like new terrain, more optimization, etc. Right now on the asset store you can get quite few visual scripting/shader editors even for free with source code.

 

I didnt know CE was user friendly about anything! lol Just their texture pipeline makes me want to scream.

 

I agree the terrain needs some work, not even complicated things, like being able to import .r16 files, that would be great! (I hate having to rename an 8x8 tiled set to .raw). But having dealt with it for the last few days, it still has alot of really nice capabilities if you use some of the asset store extensions, mainly for me its been Terrain Composer and RTP shaders.

 

to be fair with CE, you have several control options to export dds files, it's not hard at all, geometry is a bliss as well, the editor is really simple and painting vegetation is a blast (i'm very frustrated with unity tough), things become really tedious with characters and animations.

 

What about C++ , documentation, lacking tutorials , AI configuration files ? What about outdated scaleform tools and support ? It easy easy for some intermediate C# coder ?
(I don't think )

 

Very much agreed!

I would not mind it if they updated the .net runtime as well. I know that it is very unlikely to happen. But It needs to happen eventually Unity cannot run on .net 3.5 forever. Just saying.

 

That's going to happen after IL2CPP gets out everywhere (as far as I'm aware), aka a long time from now.

 

Yeah, I doubt that IL2CPP is going to be a big hit especially with mobile devs. Which is why I am suggesting the .Net Upgrade. Unity Tech should just bite the bullet and just do it. It will save them allot of pain and the devs that use unity allot of pain.

I wont go into specifics here. I will leave it at that.

 

IL2CPP is required for iOS, so I don't know what you mean with that.

They are also bringing it to consoles.