An Unreal Decision

I'd like to see that cube placed in the typical developer's project... like this one:

 

Alright this demo is aimed at mobile developers and pretty basic all things being said, the material on the box is so basic it's ridonculous. A constant vector3 tied to the base colour, two constants applied for roughness and metal. Also there is a comparison here between lightmass and LPV, the difference to me is huge. See how much lighting can have an effect?

 

@ShadowK, what hardware are you running that on? What's the frame rate?

When we tried it on the Fire TV it was dizzying to some degree.

 

These silly console analogies would make sense if the PS3 had a similar cost to the Xbone in today's market. But it doesn't so there goes that little bit of silliness.

It's also absolutely ridiculous to say that you can't purchase Unity 4 anymore, you can. That's what you get if you give UT money for a license today.

 

Unity 5 isn't even in Beta, it doesn't have a release date. I don't personally consider that 'on the horizon'.

 

Also, waiting for a new version to come out just to make a "fair" comparison is not realistic or practical. Unity 4 is just as capable of producing the same quality in the screenshot as UE4, just not out of the box compared to what UE4 provides.

 

By your logic, we should be comparing Unity 6 vs UE5. Neither is available today, but technology moves very fast.

Seriously though, we cannot compare unreleased alphas with products that are already available. Unity 5 is currently an unreleased alpha. I cannot download it or install it. I have absolutely no idea when it will be available. Unity 5 might ship in 2014, or it will ship in 2015. And when Unity 5.0 ships, I have no idea how many version 5.x features will be included in the 5.0 release and how many will be held for later point releases. I am excited about Unity 5, but there are a lot of unknowns surrounding it right now.

The versions I can currently compare are UE 4.4 and Unity 4.5. Since I am not a Unity private beta tester, I cannot even compare using Unity 4.6.

 

Well this is running on a GTX 780m with an I7 4930 / Evo SSD and I get pretty much a solid 60FPS from it as it should do. That being said, I'll agree the performance as a whole is pretty bad I can easily tank this GPU when you start putting together an actual game. Same goes for my desktop with a 780TI in it..

I sent the cave demo to a friend and tanked his 7970, I think it was looking around 30 - 20 FPS. Luckily UE4 does deal with lower frame rates very well.. Below 20FPS and it's lost, just constant stutter. 30+ FPS is generally ok though..

This could be one of Unity's redeeming features, performance for larger games and how simple it is to tweak until performance is right.

You can have "shinies" to the hilt, it's no good if it'll run like a dog on any thing but the latest most powerful cards.

@jp122

I disagree with you completely and don't get me wrong, I'm not an Unreal fanboy. I'd even go as far to say I'm not a huge fan of it. Circumstances caused changes, nothing else.

 

The extremes of C++ or BP visual scripting leaves a big hole in the middle. While I grew up on C++, I find it painfully low-level these days. Visual scripting is horribly bloated and unwieldy for someone used to text coding (though great for non-coders). There are C#, JS, and LUA bindings in development though (Unity: publish your source if you want to survive).

The screenshots show "this is where we got with Unity, this is how far we've already gotten with UE4". It's ironic that people here are just looking at the screenshots and then screaming "UNFAIR!" when it is only they themselves that are being unfair by not reading the article.

Personally, I found the article talked about many advantages of UE4 that are no benefit to me (eg. BP, easy for non-coders, easy for larger teams), and disadvantages to Unity that do not apply to me (eg. 64 bit editor, though I agree it's pathetic in this day and age), Mac OSX issues. There is no "best engine" that applies to everyone.

 

Since we've already seen some nice Unreal 4 cubes, here's a Unity one...



Unity 4 Pro, 100% vanilla with default standard shaders. No PBR, no asset store shaders, only what's included in the box.

 

This whole cube thing has turned this into my new all-time favorite Unity Forums thread. Please... please continue comparing cubes until we see who can come up with the most impressive one. This is hilarious!

 

And how much does Unity Pro cost compared to UE4?

I like Unity. I downloaded and tried Unity Pro. It made my programmer art look decent enough to my eyes. Of course not magically. You had to put in some work. Then I had to return to Unity free. I was all

I spent the last year on and off, writing my own little framework. Waiting for a price drop so that maybe I can finally get Unity Pro. I'm disappointed by the blog post, but I still want to believe! I'm waiting patiently for Unite and with fingers crossed, hope they blow us away with something not yet announced.
Its good to dream.

 

Nice, like how beast kisses the floor but it's not quite lightmass. Looks better than UE4's LPV solution (I'd hope it does as it's baked). Not sure how we got to a cube off !..

You can make good looking games in Unity, but that's not what it's about. Lack of limitations is, 64-bit editor or lack there of (that was in UE3 / UE4), theres no advanced VFX system like UE4 (or even UE3), no cinematics tools, no out the box large world support and no physx origin set, no material editor, no in-built apex use cases, no source code (even though I'm not fussed about it).

List kind of just goes on. But that's not what really bothers me either.

TL;DR

What I've always really wanted to ask, is Unity bothered about the PC / Console market? Is it too much work for too little gain? I want Enlighten and 64-bit editor yesterday. But did this come about because mobile devices are getting more advanced and windows are dropping 32-bit OS's on their asses?

 

I'm curious what is causing the bleeding off of light that is distorting the edge of the cube.

Yes. Any schmuck can make a prettier looking scene, but cubes don't lie. CUBES ARE TRUTH!

 

Beast? There's no lightmapping here buddy, that's all real-time.

As far as the cube off goes, well I just wanted to see what I could come up with in a short time since Headclot88 claimed that a simple cube in Unreal would look better than what you could do in Unity in a month. Plus to see what could be done going pure vanilla, since many people on these forums claim you need to spend hundreds of dollars on the asset store just to make something look decent in Unity or to be even remotely comparable to UE3/UE4...

Well it depends on whether you factor in royalties or not...If you're making a full feature length game with high end visuals, chances are you're aiming big, and in that case UE4's royalties can potentially make it cost significantly more than Unity Pro. It depends on your situation of course.

It's mostly caused by the depth of field for some reason. It only happens when the object is in front of a blurry background; the blurrier the background, the more the bleed is visible (but only around the really bright areas oddly enough). If you crank up the DoF blur distance, it's much more visible.

 

Please make a game only with cubes. In bigger more demanding scenes, Unity gets very unstable. At least unreal has a long history of big titles shipped.

 

Of course.
We should all be so lucky that we actually get to pay those royalities.

 

Sad to report that when I tried switching my scenes to text serialization Unity crashes in the middle of it, out of memory.

So I guess my point sort of still stands until Unity 5 and potentially beyond?

 

If you want to run a successful full time business and you don't need to pay those royalties then chances are you're in trouble.$103k per year per developer is the break-even point, and in most westernized countries $100k per FTE per year is the rule of thumb my business training (years ago) used for target turnovers.

 

Even as you point out, Unity is losing $20 a month, $240 a year to Unreal. Unity has a great opportunity to make more money from free/hobbyist users by offering the Pro version, for non-commercial use, at some price similar to UE4.

 

Although I agree that Unreal is offering a much better value proposition (I have moved my project over to Unreal 4) and am also bringing Unreal online in the school in which I work (Unity's education licensing is expensive and far too restrictive) making pro free for non-commercial use would be suicide. They would see a massive drop in revenue as a result as I estimate that ratio of shipped products vs pro subscriptions is very very low.

Unity need to focus on filling the gap between their up front pro cost and the subscription based option. The subscription pricing is bizarre and represents poor value for money. Unity also need to address the duplicate pro licensing costs (especially desktop being a requirement for mobile).

Moving pricing to one side, I see UE4's pace of development the biggest threat to Unity. UE4 is already making huge advances with each point release. You can only imagine what it will be capable of when Unity 5 finally ships.

 

I compare generations of products that are the premium offering which is UE4 and Unity 5. What's the point of comparing old technology with new when I know I'll use Unity 5 when it's released or UE4? Unity 4 is not an option when considering a switch to UE4. What would you do, switch to UE4 and then switch back again when Unity 5 comes out?

If Unity 5 doesn't give UE4 a run for it's money then I'll be concerned. I already have an Unreal subscription and have loaded some demos & I like what I see.

Anyway, i have a game to work on, lets just agree to disagree.

 

I agree. I think this is what concerns me the most. Unreal have their S*** together so far.

 

This, so much.

While awesome, it's also risky. It would suck so hard if you're dependent on these bindings and the ones maintaining them would simply get bored and stop developing them. Relying your entire code on a 3rd party with no official support is not something I would do.

 

An estimate that you have no way of proving. While I would love a free non-commercial license for Unity Pro, I don't feel that is necessary to retain hobbyists. Most of us understand that developing an engine costs money and would be willing to pay a reasonable amount.

Paying $1,500 or $75/mo per platform though is a bit much when we have no plans to enter the market professionally.

 

It seemed like it maybe doesn't like how many scenes I have. I'll see if its crashing in one in particular and attach that one if possible. My project is too big to attach at this point so I can't send the whole thing.

Regarding cubes and such:

I don't think anyone's saying you can't make pretty games with Unity. I've certainly seen some impressive ones and we're working hard to create a nice looking game with it as well.

The biggest problem with Unity is that it seems to be tailor made for small games, not big ones. Big worlds in Unity, as someone building one can tell you, are a HUGE pain in the neck. The tools are all basically working against you on it, basically.

1) Outdated terrain system: I actually got better performance from removing all my terrain trees and creating them as game object 3d meshes. I'm talking about nearly 2x FPS boost here, no billboarding or anything.

2) Umbra culling just ins't great: No matter how much I've messed with it, how many hacks, visible or invisible obstacles, Umbra simply does not provide good enough performance for the hit it takes. I'd have to be culling billions of triangles with it for it to be worth the performance trade off of it doing the cull checks it does every frame. It's like 6ms or something ridiculous last I checked.

3) No support for open worlds: SECTR Stream was a godsend in this case for me. I found Unity to choke up with a moderately populated scene (like a 1024x1024 terrain with some trees, grass, maybe 10 - 20 buildings with items/enemies.An actual level, you know?). I had to learn to close out all my hierarchy objects, kill the scene view, turn off any objects I wasn't editing, then open scene view again, watch as the editor slows to a crawl since Scene view wants to render the whole scene and you can't change it's camera render distance. Even then the scene would crash on me randomly out of memory. I've learned to save my scenes every time I make a change now due to the frequency of this happening.

With SECT Stream I can cut up the scene into chunks and edit them individually which means Unity is working with scenes it seems it was made for, tiny little scenes with like a hundred or so GameObjects.

4) Procedural = Pain. Unity's workflow is super unfriendly to procedural levels. Pretty much none of the Unity Pro tools will help you there. Umbra? Nope, needs to be baked in editor. Beast? Nope, baked too. Navmesh? Care to guess? The entire Unity workflow is built around baking things. Baking baking baking and more baking. Good if you like to bake I guess, but procedural worlds can't be baked until after the level is generated at runtime and none of these tools allow that. At least static batching works though.

I'd really like to see Unity step up and fix all these issues. UE is a proven engine that has been used to build games big and small, you see no articles about how horrid it is to build X game with it. It can handle both small and big worlds. Unity seems to be stuck on their mobile niche but wants to command a bigger premium than UE4 simply because there's that captive audience.

But let's not forget that the captive audience may be captive atm, but they won't be forever. A lot of people are too far along to switch their projects. But I see more and more people switching or starting their projects in UE4 when Unity would have been the default. Is it a problem now? Maybe not. But I wouldn't be surprised if it is in the very near future and once someone ditches your engine, I think you'd be hard pressed to get them back. I've gone through an engine change before and it wasn't fun... you lose a lot of work. Any Unity to UE4 converts are probably there for a good few years or more, if not forever imho.

I know I come off as anti-Unity sometimes, mostly because it crashes so much on me that it makes me want to bash my head on the keyboard :-/ However I have an Unity project and several games I've already made with it. When it works with me, I do enjoy it. However it saddens me to say that *a great majority of the time, it feels like it works against me*.

I want Unity to get better and to succeed. I don't want to throw away my codebase and my workflow and the learning of the inner workings only to have to do it all over again. But if things remain the same, and I'm looking at building a big world game again, I don't think I'd use Unity for it. A game jam game? Sure, done plenty of those with Unity but I expect Unity doesn't make their money with game jams since most of those folks use Unity free.

 

Well if you're going to make a glaring mistake like comparing apples to oranges, what other inaccuracies and fallacies might be present?

 

Yep, it's the same old crap over and over. Locked.