Unity 5 and Unreal Engine 4 are now free of charge!

[toneddu2000]

Great news for those who looked at these 2 engines with interest but in doubt for their prices. Epic and Unity decided to release as 100% free their engines!
Unreal engine will be 100% free as every update, developers will only pay 5% royalties quarterly (of gross revenue) when game is shipped after first 3000$ of income. Unity will be free and royalty free for personal license (team will be anyway obliged to buy professional license) with ALL the features of professional license: royalties will be applied only for games made on mobile platforms if developer exceeds 100.000$ revenue/funding limit in previous fiscal year.
Wow!

Personal note to gb: If you were in doubt to use Unity for SJ, now it could be the right time!

[jitspoe]

It's crazy how the competition has driven engine prices down to nothing!

On a sad note, this means fewer people will be looking at GPL'd idTech engines.

[motorsep]

Why should they look at idTech engines?

If starting fresh, there is no reason one should even look at any GPL engines. The reasons used to be for such move were knowledge of the tech / lack in finances / love for tinkering without having any pressure (name yours). Now there is almost no reasons going with idTech engines (or any GPL engines for that matter).

I personally is still with my Doom 3 BFG derived engine, but I began learning UE4 just to be on the safe side (loss of programmer; video drivers mess; etc.) I don't quite grasp noodles of Blueprints, and C++ code is too complex in UE4 for me. So it would be even harder for me to dive into the code side of things switching to UE4. Art assets I have can be ported somewhat easy into UE4.

So, time will tell

[toneddu2000]

The reason I sticked to FTEQW is quakec. I used both Unity and UDK, and quakec remains simpler than both scripting languages (UnrealScript and C#). I don't even think of C++ UE4 codebase, that would be impossible to me to use it and I don't like BluePrint logic. Excluding language, the list of advantages is endless:
- radiant is a joke compared to modern level editors
- complex tools for cinematic scenes
- skeletal animation tools for complex animations
- Unity Asset Store (and, maybe in the future, Unreal MarketPlace) that practically makes you build a prototype from scratch of every game type (FPS, RTS, gaming, point&click adventure, walkthrough, etc..)
- particle effect tool for stunning explosions,etc
- rendering pipeline with modern features for modern platforms (mobile/console/desktop)

The only doubt I have is network. I found too difficult to write network code in Unity. It seems (but maybe I'm wrong) that in FTE everything is all set up. You just only need to put in there

[motorsep]

I don't like Blueprint myself either. I haven't been able to grasp the concept yet. Normal code seems to have a cleaner representation and flow than Blueprint noodles.

C++ isn't that bad if it's clean. I find my way around idTech 4 code (game code at least). I haven't seen UE4's code yet, but there are plenty of C++ tutorials from Epic. When I go for UE4, I am hoping to either learn C++ or have a programmer to write code in such a way that Blueprint is used as decls in Doom 3 and for level design purposes, but not to visually "code" entire gameplay logic.

[goldenboy]

QuakeC is pretty easy, but FTE has too many bugs, LordHavoc is no longer fully maintaining Darkplaces and Quake engines in general lack features and rely on outdated, unmaintained tools (q3map2 anyone?)

Quake engines can probably be used for relatively simple oldschool shooter games (or sidescrollers etc) up to a tech level that roughly corresponds to the year 2004. That's more than ten years ago.

Vis is a problem because it was made for Quake 3-type game levels. There is no proper way of culling or LOD for large areas that contain a mixture of indoor/outdoor stuff. Realtime lights are not efficient enough - they don't get culled properly in FTE last time I checked. In a large level with a lot of torches or whatnot this will strongly impact performance. Realistic realtime sunlight is not possible. Lightmaps don't get enough contrast.

Proper lightmapping of large environment models is impossible because the whole lightmap technology was made for planar faces, ie brushes.
Collision of large assets seems possible but is not tested much.

Physics support is bad. Support for fast arrays and vectors is limited to FTE. Heightmap terrain is in its infancy compared to other engines and doesn't get enough testing and tweaking.

Fast brush based blockout is no longer an argument because 3D packages such as Blender can do similar stuff, and blockout with planes isn't really any more difficult.

If any of the advanced Quake engines got a dose of stringent bugfixing, rugged testing and good motivation, it could lead somewhere. But that's not currently happening.

-> Unity looks like a winner.

[motorsep]

Fast brush based blockout is no longer an argument because 3D packages such as Blender can do similar stuff, and blockout with planes isn't really any more difficult.

That's where you not quite right, Sir BSP / CSG requires you to have limited art skills and no knowledge of any of 3D app. All there is just blocking out level and testing it, tweaking and nailing gameplay. Without CSG, you have to know 3D app pretty well on the modeling / UV mapping side. And then there is exporting. With BSP / CSG there is none of that. It's fast, streamlined workflow. If you mess up on mesh, you have go back, adjust it, re-export it, etc. Not with BSP / CSG. And if you don't have BSP / CSG as in Radiant, and you are not an artist, you need an extra body on the team, even if that was not an immediate need.

So all in all, BSP / CSG is a must! Unity seems to have it with ProBuilder (not brushes per se, but afaik workflow is the same as with brushes). UE4 doesn't have such thing yet. It's really painful to work with CSG in UE4 at this time.

-> Unity looks like a winner.

I am torn between the two. Unity allows for many platforms, but there are so many little issue not visible until you dig into it. Plus now it has forced DRM and something new and intrusive with splash screen (I didn't mind old splash). Also they added forced stats collections from end-users PCs (those who play games made with Unity 5 free). Not to mention that Android / iOS require extra modules to take full advantage of the platforms, and consoles require Pro version.

So in those regards UE4 wins. It's just quite raw at this moment and I wish they has Blueprint in a script code version, not just in noodles.

[toneddu2000]

FTE has too many bugs

please,PLEASE,PLEASE, tell me which they are! I'm creating a commercial multiplayer game with fte and I need to know if I'm using a game engine which will give me problems FOR SURE and it will make impossible to release my game. In this case I'd switch to one of those 2 engines described earlier.

[motorsep]

please,PLEASE,PLEASE, tell me which they are! I'm creating a commercial multiplayer game with fte and I need to know if I'm using a game engine which will give me problems FOR SURE and it will make impossible to release my game. In this case I'd switch to one of those 2 engines described earlier.

Play Quake and expansions with FTE or DP, and see if there are any issues. If there are, and you are making something else than Quake-like, you will have a ton of issues. I guarantee you that.

If you know QuakeC well, there should be no issues learning C# or C++ pretty quickly.

[toneddu2000]

you are making something else than Quake-like

Please #define Quake-like. A mod that uses quake .dat and art assets (mdls,wads,quake1 maps, and so on) or a game that, in the "mood" (a fast paced fps, for example), reminds Quake?

[motorsep]

you are making something else than Quake-like

Please #define Quake-like. A mod that uses quake .dat and art assets (mdls,wads,quake1 maps, and so on) or a game that, in the "mood" (a fast paced fps, for example), reminds Quake?

I mean once you start using ODE, csqc/menuqc, skeletal anim stuff, you will unearth issues. So if you go with Quake-like non-skeletal models, q1bsp (maybe q3bsp is ok), Quake's qc code with some enhancements, you should be ok.

[toneddu2000]

I mean once you start using ODE, csqc/menuqc, skeletal anim stuff, you will unearth issues.

That's exactly what I'm using! I'd like to know Spike's opinion about that. Is really FTE so unreliable to make modern games?
In effect, today I downloaded r4843 and launching dedicated server, console said:

Code: Select all

ODE plugin failed: Engine too old.

But it's the more updated version!
So, maybe I'm wasting time using FTE for my game? Please Spike answer me!

[Spiney]

In related news; nVidia just open sourced PhysX
https://developer.nvidia.com/content/la ... ree-github

[toneddu2000]

physx github link is broken

[goldenboy]

You won't find too many bugs in vanilla quake with FTE, since I tried to report them all and Spike did fix a lot of them. There were some game breakers in the mission packs IIRC and those should be fixed now. You will find more issues in Quake 2 for instance, and I have never tried Quake 3 with FTE.

On the flip side, I tried to do a lot of things in my game that really took the engine to the edge of what it could do. I used most of its non-Quake features. All these things should have been doable, but if you break new ground then you are likely to run into bugs that simply no one else encountered before. Spike does fix bugs. but in the end there was a list of two dozen bugs or so and they just lingered. Some were highly visible, such as fullbright models if a patch mesh used the same texture, or solid world outside heightmap terrain. Bugs like the latter actually break the game. Please note it's entirely possible that they are fixed now.

On the whole I would be unwilling to depend on any of these engines for a commercial game, unless I had a really good programmer on my team who can fix bugs (I can program, but I am not an extended Quake engine specialist and I really don't have the capacity to become one.) I am under the impression that both DP and FTE are hobbyist projects and if you run into a lot of bugs after release, you might be caught between an angry crowd and an engine you can't fix yourself. This is one good reason to not use them unless you are sure you can cope with a post-release engine bug tracker. Basically all risks are yours and yours alone.

With Unity, you can at least point at them when something happens and you can be sure that the engine is tested by a shit ton of people, so it will be more stable no matter what. It's just reasonable to use the stable and well tested, well documented alternative with the huge community.

motorsep: OK, but if you make (3D) indie games, especially in modern engines, I would expect that you eventually have to learn a 3D modelling suite anyway or have someone on hand who can. You won't get far without it

And yeah, it's not that easy to decide between Unity and UE4. I think in the end, you can make a game with both. Question of taste.