3d Engines of choice

[unreason]

Hi, all. I'm working on a 2D game now using HGE, but I eventually plan to do some 3d games. I've used both OpenGL and DirectX, but I'd really like to use an engine to minimize the work load. I've looked into most of the engines out there, like Ogre, Torque, and Irrlicht, but I'm curious what people here actually prefer, for those of you who have finished a 3D game and used a 3D engine to do it, what did you use? Would you recommend it?

[Sybixsus]

When I finished a 3d game, I used Blitz3d, but that's getting long in the tooth now. It's still DX7 and doesn't even offer texture compression, so I wouldn't use it again now, and indeed I'm not.

I'm currently working on a game ( though far from finished ) in TrueVison3D which hits my three main requirements, being screamingly fast, ridiculously easy to use, and overflowing with features. It's also pretty cheap. Responsive developers, good communication and regular updates are also a great bonus.

Irrlicht seems good and easy to use, but it's just a bit too rough around the edges for my taste. Same with Ogre. No ( serious ) experience with Torque, so I can't comment.

[unreason]

When I finished a 3d game, I used Blitz3d, but that's getting long in the tooth now. It's still DX7 and doesn't even offer texture compression, so I wouldn't use it again now, and indeed I'm not.

I'm currently working on a game ( though far from finished ) in TrueVison3D which hits my three main requirements, being screamingly fast, ridiculously easy to use, and overflowing with features. It's also pretty cheap. Responsive developers, good communication and regular updates are also a great bonus.

Irrlicht seems good and easy to use, but it's just a bit too rough around the edges for my taste. Same with Ogre. No ( serious ) experience with Torque, so I can't comment.

Heh. Yeah, I put an emphasis on finished because I know a lot of people who started using an engine for a game, liked it at first, but weren't as satisfied with it after they had been working on their game for awhile. I kinda had that trouble with Irrlicht. Seemed easy at first, but as I experimented more with it I kept having to modify the code to get it to do the things I wanted it to do. Regarding truevision, I notice on their page that they have an odd licensing scheme; $150 for a license for only one game, $500 for a multigame license. Do you know if there's an easy conversion path between the two licenses?

[Sybixsus]

No, I'm afraid I don't now about an upgrade path. I'm working on a big project which is going to take a very long time, so I don't really expect to be making a second project any time soon. Best thing would be to email them and ask. They're a pretty talkative bunch when you need information.

[TamLin]

Do you know if there's an easy conversion path between the two licenses?
Yes, you can get a "coupon" for the $150 license when you upgrade to the $500 one. The developers have stated this on the forums more than once.

[Evak]

were using Aurora that one of our team mates created. Its a Blitzmax framework that wraps several C++ modules and relies on the Ogre3D engine for the rendering.

So far were very pleased with the results. There are several options

http://dreamspace.banishedstudios.com/

[TimS]

If you're not shy about 3D engines, and you're NOT developing for the casual market (they stopped supporting DX7 entirely), Ogre3D is a pretty good choice. If you're serious about your own project and you're not willing to change it to fit an engine, you'll always have to modify the engine here and there to get exactly the results you want (or to get them faster / easier, etc.).

Don't think I'd use Ogre for a low-poly game that didn't require all (or most) of the good stuff that the latest Ogre builds offer... there's too much overhead.

I'd say the decision of which engine to use would be VERY particular to the project you were about to create -- if you're in the middle of making some 2D games and thinking about doing 3D games later, then I'd say hold off the decision and ask again for suggestions later -- (and post details about the project's 3D needs... )

Good luck with the quest...

-Tim

[Evak]

yeah were targeting less casual indie gamers about a year from now, and currently everythings going well on my old 1200mhz GF3 system I bought about 5 years ago which I plan to have as a minimum spec with the current base material techniques. One nice thing in the later ogre version is that you can easily create up to 4 levels of material techniques or LOD levels for your range of system specs.

The other nice thing is the art path if your using 3dsmax lends itself well to fast development of art assets with a lot of things engine related set up by the artist within max without requiring a programmer.

Were currently using the ofusion max exporter, which comes with its own sceneloader library including source, and WYSIWYG preview within 3dsmax own viewports, by rendering the viewports with the ogre engine itself.

http://ofusion.inocentric.com/index.html

Initialy it was the art pipeline that sparked our interest, as it was one of the most important features for us if we were going to migrate from the excellemt B3D pipeline that is avaliable for Blitz3D, which we have been using previously.

After digging in past the pipeline the engine has been exceptional, and with new non LGLP licensing and the ability to port to consoles should we wish it it opens a lot of possibilities for our games on other platforms, like 360 and the C# version of our engine, should our game get the right kind of interest.

[Sean Doherty]

I have used True Vision 6.5 Beta for about six months. The engine is good but at the time it was still along way away from being complete. Is there a release date? Also, it generally targets higher end machines (well not the casual market I guess).

[Backov]

I just took a look at TV3D 6.2, and the samples don't compile at all on VS.Net 2005. There's some fixes on their forums to get them to compile on .net 2005, but they seem pretty ugly.

Does the 6.5 beta work fine with VS.Net 2005?

Also - that's a pretty short list of 3d engines. What happenned to the giant list of 3d engines from a few years back - they all fall by the wayside?

And if you want JUST a 3d engine they're not hard to find - but harder to find is a good 3d engine with good GUI support. Ogre's the only one I've seen so far that can promise that with CEGUI, but as mentioned - it's a bit of a pain to setup. Getting a working sample isn't too hard, but making a decent skeleton app that doesn't suck (ie - isn't structured like their example framework) is a bitch.

[Sysiphus]

There's a good engines database here :

http://www.devmaster.net/engines/

[Savant]

That list would be a lot more useful if you could check off boxes and filter it like that. "I'd like OpenGL support with physics and a software rendering fallback - GO". As it is, it's pretty hard to navigate...

[Backov]

There's a good engines database here :

http://www.devmaster.net/engines/

I didn't ask for a list - I know they're out there.. That wasn't my point. The point was in a thread about 3d engines on a game board, only what, 3 were mentioned? That tells me that the rest aren't so popular. :)

[Sysiphus]

Or that ..I'm unsure that really a lot of developers that aim to casual gamers do use 3d realtime. I think the majority is 2d, be it hires 2d, pixel art, or prerendered stuff.

To me that blitz3d would be a safe way to go, though.

[oldschool]

To:
Backov :
I've been looking for 4 years and the best thing I've found (for my needs) is the Unity Enginehttp://otee.dk/unity/. Now I just have to save up for a good Mac and I can buy it

To Savant :
I overlook things sometimes:) Here is the advanced search http://www.devmaster.net/engines/search.php
you can check off graphics api, physics, language, scripting, built in editors, textureing, lighting, AI, etc. :) Now all I need a Unity Engine clone(dirty word?) for the PC that ports to a mac without me having to own a Mac:D for free no less!


not going to happen, I still dream.


Evak:

What I liked is the pc demo. The mac game gooball they are selling meant to me they had a engine that indeed worked. And from what I read(spiel), it looked like it suited my needs perfectly. I'm a scientist not a programmer. I work within the *Constraints* of chemistry and biology. I can't make stuff up. So when I see a constraint I don't say "NOOOooo" Its just a way of life. I must try to understand the system given to me(reverse engineering anyone). I think you programmers want to build everything. Can you imagine if I had to make benzene, acetonitrile, methanol, B.S.A., or Monoclonal antibodies all on my own? I wouldn't get much done would I? Not to mention I don't have the skill to do all of them in a timely manner.

That said, control feels good, but whats the trade off. Couple days or weeks to understand someone elses engine or couple months building one.
Middle ground for me Pygame:)

good luck

[Savant]

Whoah, neat!

[Evak]

have you tried unity though, I assume since you don't have a mac, you just like the spiel on the web page. I wondered how good it was, but most of what I hear is its a bit simplistic and restrictive for serious developers. Probably fine for simple casual game projects though

[Backov]

Nice looking tool. I do have a Mac here, and it seems like a good tool.. But the initial first impression is that it's really a "game creator" thingy and might not be suitable for what I am doing. I'm downloading the indie trial now to take a look at it.

EDIT: Yep, if you've got a game that has "levels" and a very simple UI, this will work for you. Not for me unfortunately.

[Coyote]

I'll give a *reserved* recommendation for Torque. If you are doing a multiplayer 3D game, it is really your best choice. It takes a lot of the pain out of multiplayer.

But it puts a lot of new pain into single-player. If you aren't taking advantage of the multiplayer features, I think the choice becomes a lot more open.

All my released titles thus far have been with in-house engines. After Void War I decided to bag it and work with an existing, mature, WORKING game engine. If I were to start on Void War *TODAY*, I'd go with Torque in a heartbeat. However, it takes several months of homework on it to start to understand the engine.

[Backov]

The only thing I don't like about Torque is the fact that they really want you to do everything with Torquescript, which blows goats.

I bought one of their books, but it's basically a big tutorial on how to use Torquescript with TGE.. The second "advanced" volume isn't any better.

So far I really, really like Ogre3d.. It's just getting my head around how it works is taking a little longer than I'd like.

[CabalDoug]

If you want to focus on content, I highly recommend Torque.

[Evak]

I used Torque for a few months, just the regular version and it's not a bad engine. It does try to be a jack of all trades through scripting it's true. And outside of killer networking its not a master of any of those trades.

It relies on the MP even for SP games, unless you really butcher the engine, and this restricts you a lot, The art path is pretty poor, and lacking in really basic things like multitexturing unless it's on the terrain, or BSP compiled lightmaps. only 1UV set limits what you can do with the DTS format, even in TSE, unless your only doing single skinned meshes.

In the end we decided to make our own engine using ogre, which gives us more flexibility than torque offers, since using torque typiicaly feels like modding, and creating something new and exciting that stands out with the torque source seems pretty difficult.

Vehicles don't work all that well, the FPS aspect is very generic and would take a lot of effort to make something new and innovative in either of those areas.

What I do like about torque is how easy it is to throw characters in, and the animation blending id excelent.

[Coyote]

I do a lot in Torque with the engine itself in C++, but I do try to dump any game-related tasks into Torquescript where possible. I know that Ken Finney's books deal with TorqueScript exclusively - it's far better documented than the underlying C++ code. But it's hard to do much with Torque (except a generic FPS) without getting your hands dirty with the underlying C++ code.

The problem is once you get in there, there is a lot to understand in order to be productive. You'll rapidly end up reinventing the wheel, or you'll spend hours trying to figure out why messing around in some no-longer-used code is not having an effect.

The advantages that Torque offers are:
* A VERY NICE terrain engine, with a nice water system (TSE is supposed to be even better)
* A nice BSP-style engine fully integrated with the terrain engine, capable of using .MAP-style levels
* Good animation system
* Totally kick-butt networking code.
* Scripting system (not a huge plus, as I'd just as soon hook up Python or LUA into another engine...)
* Support for several content tools for getting shape and map data into the engine.
* Really NICE UI system and built-in UI editor
* Nice built-in terrain editor.
* Mac OS-X and Linux support

It's got a lot of cons, too. Overly complex in some points, the engine is "mature" and can look a little bit dated (but is that a big deal to an indie?). But those were the deciding factors for me. The biggest hurdle is the complexity, IMO... it is only simple for initial prototyping. After that, it gets complex very fast.