Slashdot Mirror
Simple Open Source 3D Game Engines?
Zenitram asks: "I'm trying to find a good open source/free, 3D first/third-person game engines. I can write basic scripts and make basic programs in various programming languages, but when it comes to making 3D worlds I do not have the skill set. Most of the Open Source programs I've found are not for beginners. I've found some pretty good commercial programs, however the game I am making has no chance of ever making a profit. As such I don't really want to invest money on a personal project. Any advice?"
You can try: http://www.alice.org/
It might not be as simple as what you need, but Crystal Space 3D might work.
24 beers in a case, 24 hours in a day. Coincidence? I think not!
Ogre is a great 3d graphics library, and cross platform.. Unfortunately it is not (nor will be) a complete gaming library. What I recommend is that you learn the c++ skills (which will serve you well) and use Ogre. Then you can make great games without having to learn the math involved (which is a lot) and the graphics rendering (which even with a masters in math I still think is a nightmare.)
Try Blender. It is good if you want to start learning how to make 3D stuff - some may say it's really difficult to get started with, but I say it's sometimes better to just do things with weight on your feet =) - and it has a really simple 3D game engine that's basically "join stuff with your mouse". You can script it pretty easily in Python. Blender also exports stuff pretty widely, so you can use it to model stuff for "real" 3D engines. (I've heard Blender + GtkRadiant + CrystalSpace rocks.)
And yes, 3D modeling for games is difficult when you start. Don't give up. I'm not a gigantic big expert either, but Blender is simple enough and I've seen people do amazing things with it.
For 100$ you can get torque. It was the best investment I ever made.
.DTS format which is what torque uses, and you can do all of your mapping with Valve's Hammer Editor, which is a very nice editor.
I'm a gaming and simulation major at college, and I needed something for my senior project. I still have a year left before I have to start on it, but I decided to get a jump start.
99% of things you want to do can be done with the scripting langauge in torque, which in my opinion is very C like. Milkshape will let you export your models to
I'm working on a project with 2 friends, ones a character modeler, ones a mapper, and I'm the progamming guy to put it all together. most of what I want to do has already been done before, so all I have to do is look on the torque forums to find step by step directions on how to do everything from adding flashlights, adding vehicles, night visions, or anything else I could possibly want.
And, if you buy the upgraded lighting pack for 50$, the torque engine looks very pretty.
You know you can't just pick up some tool and press 10 buttons to get a game; that's what beginners want to do. 3D Game Construction Kit doesn't exist because the dynamics of a 3D game are so HUGE.
:)
Writing scripts in a game, as you say you can do, is possible only because someone wrote huge swathes of code behind it, including tying a scripting engine into hundreds or thousands of classes and objects.
If you got a "simple" tool, it would be too simple to give you the environment you are currently comfortable and competant in coding in. You need to get more complicated
Quake 1-3's engines have been GPL'd. You can find them on ID's ftp server. If you're looking for the best available free engine I think Q3's would be near the top of the list.
You should try Irrlicht: [ http://irrlicht.sourceforge.net/ ]
It's relativily easy and yet versatile and powerfull
http://www.garagegames.com/ It's $99, but a great engine and has some very good support from the company and community. If you're really serious about it, $99 is incredibly cheap for the value it offers.
I agree wholeheartedly with the parent. This is a hobby project, right? The point is to have fun, and to learn something? In that case, I think you'll have more fun if you don't start with a premade "game engine."
Start from scratch. It sounds intimidating, but that's just because you haven't tried it: You will be amazed at what you can do.
You don't need to spend any money. Compilers are freely-available. Between that, the web as a reference, and your own intelligence and creativity, that's all you need!
Do you know a programming language? If not: You say you can write scripts. If you can learn to write scripts, you can learn to write code. If you need to choose a language, I'd recommend C++ or Java; C++ is probably the most commonly-used language in game development. This site was an important reference for me when I was getting started. Go through all the tutorials. Don't rush to get out of the console: It's a great place to focus on your logic and your algorithms without worrying about interfacing with graphics APIs. It's where you learn how to think about programming.
Then, graphics! I'd suggest you start with some OpenGL programming. That's for a number of reasons. It's not just that OpenGL code is widely portable: Compared to Direct3D code, it's a breeze to write. Carmack himself wrote a piece about how OpenGL programming is a straightforward, enjoyable experience.
Where to start? Try NeHe's tutorials, and its parent site, gamedev.net. That's how I started writing OpenGL code, and I had a lot of fun doing it.
Then, think about what sorts of worlds you want to represent. Landscapes? Indoors? Look up the relevant algorithms and data structures. You care about spatial partitioning schemes, occlusion culling, and LOD. This is where it starts to get advanced. Here, you'll learn more than just to code; you'll learn some actual Computer Science.
You don't need to worry about this now, but when you get to the point where you're ready for it, there's stuff you can google. Spatial partitioning: BSP (older algorithm, computationally beautiful, generally used for indoors), Octtrees (more modern approach, conceptually simpler, lets you efficiently throw stuff at the graphics card), Quadtrees (variant for outdoor environments where the map extends mainly in two dimensions), K-D trees (special case of BSP which behaves more like Oct or Quad-trees). LOD: For outdoor scenes in particular: ROAM, geomipmapping. Occlusion culling and visibility: Portals, precomputed PVS, image-space techniques with occluders.
Just explore, experiment, and have fun. You'll learn a lot.
Try the Blender Game Engine. You have all your 3d tools right there in one package. The engine is quite versatile, and now supports GLSL shaders, multi view ports, and physics. http://www.blender3d.org/
Soya3d is a 3d engine written as a Python module (really). Being a python beginner, and someone who hasn't written a lick of C in ten years, I can tell you it's EXCELLENT. It even comes with a sample 3rd person game, Balazar, that has been fun to play (although it's not quite done) //and// to tweak under the hood.
I highly recommend it.
This is over and above the other mentions of Ogre3d, Nevrax, Cube/Sauerbraten, the Quakes, Nexuiz, the Worldforge projects, etc.
There are a couple of free compilers to start with as well. The gcc/g++ ones to start with (http://gcc.gnu.org/). I find those to be a tad harder to setup but are completly free and you can look at the code under it (if thats what you want). There is also the visual studio compiler. There is a free learner edition (express) that comes with the gui and the SDKs are fairly straight forward to get and setup (http://msdn.microsoft.com/vstudio/express/default .aspx). I recomend the vs one if you are going to do any sort of directx stuff. Also if you are going to do that I would get a few COM tutuorials to get you into the groove of making and destroying objects.
I also recomend getting a book or two. Even if you just go to the library to get one. There are also a few companies that have released the source to their older stuff (Quake I/II/II, Freespace, etc...).
Above all have fun with it.
I don't think 3d is that hard, sure it uses university-level maths but what you're lacking there you can read up on. Most of the problems you are going to encounter have been solved in traditional math already, you just need to transscribe the algorithms into your programming language (and know where and how to use them, of course). The basic transformations and lighting are done by the API, the hardest part is optimizing stuff and even that has been explored thoroughly. Of course you're not going to churn out a state-of-the-art engine on your first try but you'll get something more than workable easily.
Justice is the sheep getting arrested while an impartial judge declares the vote void.
Cheers!
--clint
I don't think it is really worth doing 3d programming in terms of a learning experience until you know the fundimentals. If there were a tool that one could use to write graphics stuff without understanding matrix maths one would probably come out of the experience still not understanding matrix maths since one would not need it with that tool. Go to a university book shop, you should be able to find a decent mathamatics text book that covers vector spaces and linear maps, read that. Then check out the NeHe OpenGL tutorials and learn OpenGL. Now you will know the fundimentals of 3d graphics and you can just practice until you are good.
When Argumentum ad Hominem falls short, try Argumentum ad Matrem
As other posters have noted, if you don't have the skill set then you shouldn't be worrying about engines quite yet. How familiar are you with vector and matrix mathematics? If you ever plan on doing ANYTHING in 3D, learn that. Trying to do anything without knowing the basics is just plain ludicrous. Once you've gotten the hang of those, you can start coding. The mathematics are usually the biggest block to get over when programming in 3D. After learning that stuff, OpenGL or Direct3D will seem alot less daunting. Alternately, since there's alot more to 3D game programming than just learning a graphics API, you could then try making a mod or playing with one of the many engines listed in previous posts. My own personal opinion though: Engines are wrong. They're usually general purpose solutions when I want a specific solution. Engines are evaulated by companies/individuals based on features that they want. They'll usually go with the engine that has a balance between features and ease of use. One could argue that learning an engine and adding the features you want would be just as time consuming as writing your own engine from scratch (it's an argument that I ascribe to). As a learning environment, I really can't recommend anything better than reading tutorials; getting comfortable with the concepts; and then doing everything yourself. That's the approach I've taken, and I've found it to be a rather useful approach (it's the same with programming languages - did you learn the science behind it or did you only learn the syntax?).