A Unity ID allows you to buy and/or subscribe to Unity products and services, shop in the Asset Store and participate
in the Unity community.
New Unity Live Help updates. Check them out here!
Discussion in 'General Discussion' started by Deleted User, Jun 7, 2016.
It already no longer exists.
Hmmm, errr *looks back and forth* I think I found the issue with Unity terrain, it looks like Jello..
Don't be silly. The terrain can't have that much reflection.
A souffle, on the other hand... #UnityFood
Actually that's not a bad idea. Make an RTS were you control ants, put some heavy DOF on the camera an you've got a pretty unique aesthetic.
We've been saying that for years. The day still has yet to come.
Maybe we'll finally get that Unity Sandwich logo that @Tomnnn was always talking about.
I don't. Not really. I don't care if a game was made by Joe Bloggs in his Mum's basement or if it was made by Blizzard in their lofty towers. What matters is the game is fun, and that the game is worth its price.
I am sure some people have no differences in expectations but not so many I think. Many people say they have no difference in expectations but is it really true? Really? If you were given two links for games to check out. One link said "Blizzard's Newest Game!" and the other link said "A New Indie Game!" would there truly be no difference in what you expected to find on the other side of those links as far as presentation quality, scope of the game and so on?
Maybe you really don't have any different expectations. I think most folks do though.
Both those who use the engine and those who make the engine have been saying that for years.
If your actual game is open world and terrain based than jump on UE4 train, if not stick with Unity.
One of the problems with that viewpoint is that you're effectively excusing Unity for having a worthless terrain system.
I'd expect the Blizzard game to be something highly polished, but not targeted at my personal taste for games, and the indie game could be anything on a scale from total garbage to truely amazing. "Different" expectations doesn't necessarily mean higher or lower, imho.
I would expect the Blizzard game to be bigger and better. And also cost significantly more. If the games were the same cost in the same space I would play the blizzard game. Why bother playing something sub standard?
Nobody gets any points just for being indie.
Indies get points when they are better then a traditional AAA games, like Cities: Skylines. Or when they do something different, like Kerball or Factorio. Or even just doing something different to an old genre like Creeper World.
If an indie is going to publish an action FPS shooter I will compare it to Halo. And considering I can pick up last seasons Halo for $30, the indie game better be up to that standard.
For what it's worth, the first step in the grand Terrain plan was GPU Instancing support, which is in 5.4.
Terrain is not the priority number one. This is the user to be smarter and choose the right tools.
I've seen many games (and screenshots around here) made with Unity that have some nice terrains. Do they all roll their own solutions? I was thinking most of these folks were actually using the built-in terrain thingy.
Terrain is a major feature of modern engines. The fact that Unity's is practically worthless is a big problem.
By the way... neither of these automatically implies the other.
Talking about terrain, I've just tried a large terrain in UE4. Edit: Was complaining about performance, but restarted my machine and it's running @ 120FPS.!!! Odd.! Well @neginfinity is definatley right about "quirks"..
What I do like is, you get all that nice HM data from the RAW's without additional NM's they are only 256 resolution tiles..
Old Unity one for reference..
If I was going for a AAA look or a high res style, I would care. One thing I have noticed about Unity is that they come through, eventually. Maybe they won't match the exact, specific look of Unreal but there will come a point where their terrain systems are good enough.
What's super odd is... I see people bitching about having to pay more for Unity, and then bitching about how they want more features... what the F***, so you want more for less and you want it now yesterday? I love the human race.
We currently are not getting much (features come in very slowly, issues from 5.0 beta are still around, core engine features are still in an "eh" state*) and we get to pay more -> bitching makes sense to me.
*I have enlighten in mind for this and "eh" is very generous.
For the love of sweet baby Jesus... people make it sound like Unity just straight up doesn't work and they've been waiting for a critical bug fix since the fall of the Soviet Union. I just can't.
I am sorry, are you replying to me? Am I "people"? What is your counterpoint? I am confused.
It is like that for a lot of people. If you haven't stumbled upon issues (that you reported years ago, and still aren't fixed) then great, good for you, but you don't get to tell other people if they should be frustrated or not with the state of Unity 5.x
What's super odd is... I see people bitching about other people bitching for having to pay more for Unity, when they themselves are going to use the free version, are not doing game dev professionally and are probably never going to give Unity a single dollar... what the F***, it has nothing to do with you, yet somehow you think your opinion is more valid? I love the human race.
I'm not exactly sure who you're aiming that at? I was a pro user that had several licenses.. So if I'm not qualified to "bitch" about it then who is exactly? I'm also one of the fair few where Unity charging "more" wouldn't mean a thing to me, I'd even border on encouraging it IF they could at least keep somewhere within the last half decade of their competition.
P.S I'm not sure why you're "caring" or not either? It's not about "high res style" because that's the easy bit.. Getting it to perform well is another matter and I'm still not convinced that UE is the best fit in some circumstances like openworld, although I am actively looking into reasons why.! I'll share a conclusion when I have one..
For indoor scenes, UE runs like lightning and with baked lighting looks awesome.!
CryEngine is the best and fastest with SVOGI activated for open world games.
Right tools for the job, always. Unity can do everything, it doesn't mean it is suitable to everything.
The only problem with that is I'd have to use CryEngine..
CryEngine has a monolithic archaic workflow system (sorry if I'm insulting some CE dev here). If you can live with that then good for you. But yes their renderer is (out of the box) the fastest/prettiest one and it's also made for dense/large/heavy populated maps.
If they would be able to totally change the workflow and make it as smooth as Unity or UE4 , it would become suddenly lot more attractive to lot of indies. But this is not the direction and priority they are taking unfortunately
I believe it would require them to rewrite their editor, and possibly parts of the engine. But their priority has always been graphics-heavy applications, not indie developers. It's nice and all that they opened their tools to everyone, but it is not for everyone.
It's like a racing car versus a sports car (CE vs Unity): the racing car can't carry passengers, can't handle rough roads, it's so raw that you can get yourself killed while maneuvering the supermarket. But once you manage to sit in it (which could take a while), and you are in a race track, it goes way faster than any similar road car.
UE4? Maybe one of those "street legal race cars", like the Porsche 911 GT3 RS: hard racing seats, metal bars on the luggage area, rock-hard suspension, but still walks the street, and still flies on the track.
None can beat a proper road car outside the track, though. That's why I use Unity everyday
I really think you seriously loose perspective at some point, you just keep pushing and pushing.. I whipped open Witcher 3 for some "context" on landscapes and the first thought that popped into my head was, that's not a very nice looking game..
It's because I've been staring at my game so long I've become slightly "wall eye'd".. A couple of days away and I'll think Witcher 3 is awesome looking again.
Seriously, time to reign it back a 100 miles mr. Shadow..!
(Before someone says report this on the UE forums, I have this is a comparison..
GRRRR!>>> So please let me know if I'm being a moron here, I thought the tris amount could be invariably causing issues with framerate.. So I did stress testing, I ramped up the terrain to max settings and zoomed out to view the entire terrain, I have 46 MILLION Tris in view and I'm getting 90 FPS and 11MS render time. No matter if it's full screen / windowed..
I add some dense foliage, within 4000 units (which is a smaller draw distance than W3 for example) and now I get 40 FPS (if I'm lucky) in full screen view.. Although I get 90 if I'm windowed mode, obviously the amount of pixels being drawn becomes higher the bigger the resolution of the screeen, therefore it's clear as day to me that's a deferred rendering issue right?
The defacto answer is "aggressive LOD's" but what's that going to do? I'd never have more than 20 million tris in view even if I made a dense forest and said screw it to the culling distances.
I tried it in Unity / CE and I get around 75 - 90 FPS with tons of foliage and grass at 1440P.!!
So I said I'd explain what I thought was the issue, I think that's pretty much what it is.. Although feel free to correct me if I'm wrong.!
Well if lods are too aggressive, you dramatically raise draw calls so lods generally get slower the more you have. Trick for me was to combine into larger lumps further out.
The easiest way to kill framerate is with fillrate and transparent polygons. Foliage is exactly that.
Modern GPU should have no problem handling millions of triangles, as long as the data is fed into it efficiently (see vulkan demo). Riva TNT 2 could render 2 million tris, for example. That gpu is from 1999.
To get more precise answer, you'd need to profile the whole thing and see what the render is actually doing.
Aye, you're not telling me anything I don't already know.. I checked for DC's etc. and 190 isn't exactly going to bring it to the floor.. It's overdraw etc. that's causing it, the question really is why is UE so heavy compared to the rest? Then the follow up is going to be, what am I going to do about it?
Only reason I mentioned tris is because Epic bang on constantly about LOD's / poly count etc.
The other issue is scalability, if you drop it down to let's say medium it looks horrible.!
How exactly are you measuring the FPS in each scenario? I found out that the results I get between stats window in the Unity Editor, Profiler in Editor, Profiler in Build and Fraps' FPS counter, vary wildly and don't make sense to me at all. I've had big differences between what the stats window in the Unity Editor told me and what Fraps measured on the game window. Like 90 vs 60 FPS (from my memory, might be slightly different). I don't have the other engines to compare to. I'm just confused what counts as an accurate measurement in Unity.
If you are measuring in the Unity editor, Fraps may be reading fps based on screen refresh, while Unity is reading from internal rendering speed. I believe you should use Unity's number in the editor.
@ShadowK Sounds like for all of its other issues Unity is ultimately more optimized than UE. And I am not saying that to start an engine war discussion. Just saying maybe it is a possibility considering Unity seems to cater more to lower end devices and particularly mobile. It is possible, right?
Maybe you would be better off just switching back to Unity, simply do the best you can for the graphics quality side and get it done? Nobody can expect more than that. I always get the impression that UE is almost made for people targeting high-end machines. So one benefit with Unity may be a larger audience even though the game may not look quite as pretty. But maybe that is okay? If it means that more than 1 out of 1,000 gamers can play the game at a decent speed.
No, it is definitely the other way around. Unity chokes on scenes interior/modular which UE4 handles without much trouble.
Terrain component may be slower, but not the engine.
Fair enough. I have no idea personally about UE. I've just read a lot of comments from different people talking about how massive and resource heavy it is both for the Editor and in-game. Granted in-game is probably from folks trying to max out graphics.
It IS resource heavy, requires more memory, etc. However, once it boots properly, performance is better than in unity.
At least that was my experience when working on modular castles with melee combat.
You'd need visually equivalent testcases to make a fair comparison. I'd hazard a guess that his UE4 scene looked better.
If anything I'd see him switch to the new engine he's currently actively contributing to. Switching back and forth between Unity and UE4 makes no sense imho.
@Martin_H I agree. No idea how performance is for the new engine. And actually I think @ShadowK himself is working on rendering so it can make things look as shiny as he wants. But I figured he doesn't have time to focus on building an engine to use for the game he is currently building. If he does then I agree that sounds like the ideal solution.
And yes I am sure the UE scene looked better on max. But he said that performance was low. To match CE and Unity performance he had to decrease graphics quality and it looked "terrible". No idea if that means it looks like the best he can achieve in Unity, better or worse. I was just thinking if he can deliver a better performing game that also (for that level of performance) looks better in Unity then it might make sense to just use Unity.
Last time I tried "ARK"(UE4) it looked rather mediocre @~40 fps on my system and "The Forest"(U5) looks better @~60fps. So, you've got a point. But I'm pretty sure there's a ton more to the choice between U5 and UE4. And my example is not a sensible comparison either, since both games are pretty different. It's anecdotal at best.
We had a thread about graphic quality. It got shut down, if I remember correctly.
I'm sure any slowness in games like that has to do with all the world-crafting. Real-time occlusion vs. pre-built is probably going to make it slow no matter the engine. I don't think any current engine is really optimised for it, including MineCraft's
Thanks for the feedback! Comments like that motivate me when I'm burnt out to keep going to finish it!