next »

[ColossusEntertainment]

In my experience, the freely available "engines" are made by hobbyists, for hobbyists, with the aim of ticking a lot of next-gen features. It can be really hard work to try and get a game to a state where you can sell it, using one of those (but it seems like you've found that out already)

The commercially available engines (gamebryo, unreal, renderware when it was available) are done by professionals, for professionals. The good thing with them is that they are proven technology, and can certainly be used to make commercial games. But it also takes a lot of work, as their whole workflow is more adapted to the big studio way of doing things. If you have a team of people, it could work, but I wouldn't really recommend it for a single dev.

I'd say your best best is to either stick with your existing engine, but not worry about adding new, fancy features. Or, as was already mentioned, switch to 2D.

As I understand it, your motivation for this is commercial, you're aiming to sell the game, so I think it is important that you consider the areas where indie devs can compete with the AAA retail industry. Technologically or contentwise, we don't stand a chance. When it comes to unusual gameplay mechanics, niche markets, compatibility with low-end hardware, we have opportunities that are not just viable for the AAA guys.

Going 3D because it looks prettier, or to take advantage of peoples new expensive hardware... that, I feel, is the AAA guys job. To the average consumer (not hardcore gamer) I'd say 2D generally looks prettier than 3D.

[Garthy]

Thankyou Mattias for your thoughts on this one.

Re free versus commercial, I've heard similar sentiment expressed before. It does seem to be a very wide brush. Could I trouble you for some examples- as in the free engines you've had trouble with, and the commercial ones that worked well for you?

Of the commercial ones you have experience with, are there any within a sensible indie budget that you would recommend?

The secondary reason for considering a switch to a different engine is compatibility. I know my engine works fine on my cards, and a few others out there too. Unfortunately, it's never going to have the coverage of a common engine. So as tempting and self-gratifying as it would be to stick to my own engine, I have to weigh up the other options at least.

I'll keep your thoughts on the whole 3D versus 2D thing in mind. More detail on the dilemma there can be found in my previous posts.

And yes, my aim is commercial, so I will be taking careful aim at a safe niche. Thanks for the warning on that front, it is good advice.

[Ivan]

What's wrong with OGRE again?

[ColossusEntertainment]

Re free versus commercial, I've heard similar sentiment expressed before. It does seem to be a very wide brush. Could I trouble you for some examples- as in the free engines you've had trouble with, and the commercial ones that worked well for you?

Of the commercial ones you have experience with, are there any within a sensible indie budget that you would recommend?

I've had a look at ogre, irrlicht and C4, but quickly discarded them, for about the same reasons as you did... Compatibility and performance issues. Plus, I couldn't find any examples of commercial games done using them (this was a while back though). I haven't looked at Torque, but at least there's some games been done using that.

I've worked with Renderware, Gamebryo and Unreal engine, and I wouldn't say any of them worked well. They worked, but only because we threw a lot of programmers on the engine problems to make them go away. Generally, I experienced less headaches with in-house engines (though there certainly was some major issues there too). So even if those engines was in a reasonable indie price range, I wouldn't recommend them to indie-devs or small teams.

The secondary reason for considering a switch to a different engine is compatibility. I know my engine works fine on my cards, and a few others out there too. Unfortunately, it's never going to have the coverage of a common engine.

Maybe, maybe not. In my experience, compatibility is an issue even with the big commercial engines. It rarely "just works" everywhere... Speaking for myself, this is one of the major reasons I've chosen to go with 2D for everything I do. The compatibility issue becomes way more manageable.

As far as I'm aware, there's really no viable option for indies making 3D games. It is rather a lot of work to make your own engine. The ones that exist are either not suited for games (but more for tech demos) or requires a lot of work (= big teams) to get them to do what you want (plus, they're too expensive). Would be nice if there existed a solid, lightweight indie-engine for 3D games, design for the needs of an indie dev...

[Garthy]

not all that far gamewise... but since 12-2007 i'm completely hooked on games . so i made some prototypes with my own framework, which really can't be considered as an engine. it's just a set of independent classes to handle windows on win/linux, do basic interaction, draw stuff with opengl, do calcs with vectors and make some noise. it's usable for prototyping and learning, but for nothing more. simple games are possible that way tough, but if it gets real 3d, things definitely get more complicated. drawing is only one aspect, like we all know. - the point, when i was completely fed up with all that engine tech stuff was just a couple of days ago, so i can't really tell you if it's going to work out with Panda3D (which is really a professional but easy to learn/use engine) or any other ready-made package.

Yeah, I've developed a bunch of useful classes as well. Fully intend to keep a hold of them. I haven't found 3D in general to be that hard, mostly because I keep it simple. 3D graphics on the other hand, is hard work.

the last prototype i've made was just a simple indoor scene. i mean: simple as it can get: only one model with some seperated rooms drawn with hard-baked lightmap + diffuse texture plus a skybox. i just wanted to see how good my collision method will work in such an environment using basic opengl rendering and realtime sound for collision response. - but it didn't work as expected. like expected . although collision was quite easy (i did it directly on the polygons of the model, which were split up first in a regular grid), i saw that there were some problems with the camera movement. sometimes it lagged hard, and screen flickered. so i turned on vsync with a spcial ogl extension. but all was slowed down by that, and the sound synth began to crackle. so i had to compute the sound i a seperated thread (linux incompatible). but the hangers were still there, sometimes the framrate dropped instantly under ~10 fps. i had to add simple frustum culling with an additional octree (besides the regular collision grid, hmm). but there were still some unsmooth parts in the cam movement, i had to seperate the simple "physics" in its own thread and to update it only in regular timesteps, cause delta_t physics simply don't work out (what i had to realize after testing both, euler and verlet method with dt). but the jittering was still there sometimes, just because the fillrate of my graphics cards isn't high enough. and the scene is really very simple, just a couple of overlapping polygons. i would have to learn how to do occlusion culling/portals/bsp now. then all the scene graph thing with matrix/quaternion transformations would have to be done, realtime lighting for non-lightmaped models, writing editing tools and so on. simply not the thing i would like to do anymore. NOOO! 

It gets very complicated very fast, doesn't it. :}

I haven't got any heavy scene optimisations in my engine, so it'll never run as fast as others. It's mostly just a big node tree, with a whole bunch of funky effects that I like.

but it's really a point where all indie devs have to decied by themselves imho. - what do you really want? - the techy path is as equally deep as the more creative one, inf == inf... i just think, it depends a lot on what type of games you are interested in too. 3d is only that complicated if the whole gameplay is 3d. (but i don't really know yet, i'm still learning). all the rendering isn't so crazy hard in my opinion. it makes sense to restrict it to a basic level, like standard opengl, for compatibility reason too. that's compilable and runnable on almost every computer without too much hassle.

If you were still on your own engine, I'd ask if you've had much of a play around with your engine of different PCs. My experiences were that it starts getting tricky, as every card and driver has its quirks.

I fear I only have a passing interest in the 3D graphics side- or at least, I'm not happy about only being able to dedicate a small amount of time to it.

@Irrlicht/Ogre: after some tests a while ago, i think they are simply too crappy and overloaded to use in commercial projects. i couldn't rely on them.

That's interesting that you say that. I keep telling myself that I'm being far too harsh, and I'm far too biased toward my own code, when I think that maybe they aren't up to scratch.

@MinGW: that compiler always worked for me, even for these gamey prototypes... in combination with Mingw Developer Studio it's quite a useable alternative to visual studio (one good thing with that: you can export gcc makefiles there, easy modifiable to use under linux). code::blocks is ok too, and available under linux. - but at the end, MSVC is simply the best ide i know. and its compiler produces mostly smaller, faster binaries, and the build process is faster too, the debugger better. but sometimes it can even make sense to use both - msvc and mingw - in parallel. i did that when i made my synth lib, which should run and compile under linux and windows equally well. so mingw was a good way to produce fully gcc compatible code. sometimes there are still differences between them (esp. with templates, although it improved a lot since msvc version 6). and that way, it was nice to have an easy double check mechanism for cleaner more compatible code...

I've had mixed results with MinGW. It can be great to work with, but an incredible pain sometimes.

From what I've heard: Smaller, hell yes. Faster, maybe not. The template situation in MSVC 6 was apparently really bad. I'm a bit rusty. I believe that was the one I used to use?

Thanks for sharing your experiences.

[Garthy]

Heya.

What's wrong with OGRE again?

Paragraph nine in my original post.

[Crackerblocks]

My brain dump...

You sound too clever for your own good.

My advice: simplify your life. And enjoy what you're coding.

1. Simplify your game design to something that can be completed by one person before that person gets bored with the design.

2. Pick an engine that you're familiar with and would enjoy using. ie. your own suite. Can it draw a mesh in three-dee? ok good, you're set.

I'm not going to come close to the technical capabilities of any of the better engines out there

Uh huh.

And where's the team of artists that will make content for this cutting edge 3D engine?

3. Lose the multi-platform idealism for a while. Just pick whatever platform you like using. Completed your kick ass game? Ok, fine, then you can port it.

4. Make a fun game.

I enjoy developing the gameplay side of things much more.

Then develop the gameplay side of things much more.

Worry about all the bullpoop later.

[Garthy]

I've had a look at ogre, irrlicht and C4, but quickly discarded them, for about the same reasons as you did... Compatibility and performance issues. Plus, I couldn't find any examples of commercial games done using them (this was a while back though). I haven't looked at Torque, but at least there's some games been done using that.

Last time I looked at Torque I had a bit of trouble finding reliable information short of paying, and this time around I keep hearing "awful doco". I've kind-of ruled it out already, short of a few people piping up and say "yeah, I've used it, and it's great".

Haven't looked at C4 at all though.

I've worked with Renderware, Gamebryo and Unreal engine, and I wouldn't say any of them worked well. They worked, but only because we threw a lot of programmers on the engine problems to make them go away. Generally, I experienced less headaches with in-house engines (though there certainly was some major issues there too). So even if those engines was in a reasonable indie price range, I wouldn't recommend them to indie-devs or small teams.

That's quite worrying.

Maybe, maybe not. In my experience, compatibility is an issue even with the big commercial engines. It rarely "just works" everywhere... Speaking for myself, this is one of the major reasons I've chosen to go with 2D for everything I do. The compatibility issue becomes way more manageable.

I had suspected as such- it's hard for me to know though, as I have zero experience with any of the larger commercial engines.

As far as I'm aware, there's really no viable option for indies making 3D games. It is rather a lot of work to make your own engine. The ones that exist are either not suited for games (but more for tech demos) or requires a lot of work (= big teams) to get them to do what you want (plus, they're too expensive). Would be nice if there existed a solid, lightweight indie-engine for 3D games, design for the needs of an indie dev...

That's... disturbing. I've seen a bundle of 3D games floating around, surely not all using their own hand-made engines though? But I'm honesty not sure. I guess I've also been hoping that someone will post "I've used X to create a game, and it's great".

Many things to think of, and not a lot of good news. Regardless, thanks for the extra input.

[ColossusEntertainment]

I've seen a bundle of 3D games floating around,

Really? I've found commercial indie 3d games to be quite rare... Most indies making money off of what they do seems to be making 2D stuff. All the 3D stuff I come across seems to be "in development" (which means it doesn't count). Though I do know of a few Torque games which have been completed and sold, which was why I mentioned Torque in the first place.

If I was going to make a 3D game to sell (not likely, I think 2D rocks :D) I'd get a Torque license and investigate its capabilities myself. Could well be that it's good enough to get the job done, at least there's proof that others have succeeded.

Btw, about poor documentation... Some of the big commercial engines I have worked with have had really poor documentation. But I think that when you're a developer using third-party stuff, you always feel like you want more documentation, and then some more again. Though what is really needed is that you spend lots and lots of time to understand the inner workings of the library/engine, so you know what's going on and how to best use it.

A lot of people expect an engine to just take care of stuff for you. In my experience, no engine does, and you can choose to use your time either to write your own tech, or to become intimately familiar with someone elses. I reckon it adds up to about the same amount of time in the end (though using a ready-made engine will get you stuff on screen quicker, while with your own tech that bit takes longer. When it comes to finishing of the product, you'll work quicker with your own tech, and much slower with the third-party). At least in my experience.

[Garthy]

Hi Crackerblocks,

There's a lot to read through in the thread and it would be unreasonable to expect you to have read all of it (especially true due to my verbosity), but have a skim through the fingers and toes comment I made above. Completion is not an issue for someone who has completed multiple games and projects. The advice, whilst good in general, is mistargeted.

I've already got a cross-platform framework that I'm using that I developed as part of my game suite. That's not a problem. I can't take your advice on sticking to one platform and porting at the end as it is not consistent with my experience in this area. Unless by "end" you mean the end of each month. Perhaps this is a debate for another day. For the time being, I'm evaluating different 3D engines to see what I can gain. I am comfortable spending time exploring the options. What are your experiences in regard to 3D engines?

Re a team of artists- I learnt from that mistake last time. The design is such that the artistic requirements are minimised, and for what is left, I intend to throw money at a pro. I have no artistic ability but saleable skills. I work, I get money, I give it to a pro, and they make it purty. There's a lot to developing 3D engines, as I have discussed through the thread, and if it is something I can leave to the more dedicated, I wish to do so.

I realise there is only so much advice that can be offered with only the tiniest slice of my actual circumstances. Thanks for the input

[Garthy]

I've seen a bundle of 3D games floating around,

Really? I've found commercial indie 3d games to be quite rare...

I see them float around on various indie-type sites, usually in the feedback sections. Unfortunately, if I'm going to make an assertion like I did, it's up to me to back it up, and I'll have trouble doing so. Hence the following sentence talking about the engine in use. And also...

All the 3D stuff I come across seems to be "in development"

I suspect you're quite right there. If I could rattle off a list of strong, completed 3D games, I might then be able to dig around and find the engines they were using. More research needed by me, I think.

If I was going to make a 3D game to sell (not likely, I think 2D rocks :D) I'd get a Torque license and investigate its capabilities myself. Could well be that it's good enough to get the job done, at least there's proof that others have succeeded.

I've considered just dropping the money and then trying it out. I just don't hear enough people saying "damn, this *is* good" to overcome my concerns. I'll have to keep thinking on this one.

Btw, about poor documentation... Some of the big commercial engines I have worked with have had really poor documentation. But I think that when you're a developer using third-party stuff, you always feel like you want more documentation, and then some more again. Though what is really needed is that you spend lots and lots of time to understand the inner workings of the library/engine, so you know what's going on and how to best use it.

I feel quality documentation achieves three things- makes it easier to evaluation, makes it easier to get started, and makes it easier to maintain through its life. I've had to struggle through frameworks with shocking documentation before, and if the standard is low enough, sometimes it's easier just to write your own from scratch. In my experience, poorly documented systems tend to be the first to fall into disrepair and into obsolescence- unless something external is keeping them alive (eg. being the only solution). Thus my sensitivity to documentation.

And yes, I know some people will just complain about anything, and sometimes a framework is so good that you just bite the bullet and don't fret about the doco.

A lot of people expect an engine to just take care of stuff for you. In my experience, no engine does, and you can choose to use your time either to write your own tech, or to become intimately familiar with someone elses. I reckon it adds up to about the same amount of time in the end (though using a ready-made engine will get you stuff on screen quicker, while with your own tech that bit takes longer. When it comes to finishing of the product, you'll work quicker with your own tech, and much slower with the third-party). At least in my experience.

This is good advice and is consistent with my experience in other (non-3D-engine) areas, with one exception- that the own-tech versus learning pre-existing is roughly comparable in time. I've found it varies a lot, and not always in the same direction. But then, we're talking 3D engines, and not in general, and my experience in working with and evaluating different engines is extremely limited.

[Ivan]

Just write your own damn engine if you're that picky. Though I skimmed through your comments and I couldn't find a single decent argument, which as to why OGRE is an unacceptable engine (except for your extremely unreasonable desire to have software rendering, which no decent engine has). I think OGRE is a fine engine, written by professionals over the course of many years and many freeware and commercial games were done with it. I don't mean to sound harsh, but I think you're just being a smartass about it.
Write something with it! And come up with some concrete technical reasons as to why you don't like it, and if you really don't, I recommend you write your own engine, even if only to gain an appreciation of engine design.

[ColossusEntertainment]

and commercial games were done with it.

cool, could you name a few, I'd like to check them out 

[Ivan]

http://www.fun-motion.com/physics-games/de-blob/ (this was the old freeware one, i dont know if the wii version uses OGRE)
http://www.jackkeane.com/
http://www.ankh-game.com/
http://www.aridocean.com/
http://www.devbump.com/nimblebit.com/

im sure tehre's more

[ColossusEntertainment]

So, that Ankh game is made with OGRE? cool, hadn't seen it before, shame there's no demo available (or did I just miss it?)

As for your other links, they're either under development or not games, so not really of interest... but thanks anyway 

[moi]

I think ankh was a retail game.

[Garthy]

toastie,

A read through my initial comment will indicate to you that I have written my "own damn engine", as well as why I am looking at other ones.

If you have pent-up outrage, may I suggest that you read first, then rage later.

I won't address your other points. You are extremely rude.

[Ivan]

I apologize if I came across as rude. It was not my intention. I was simply pointing out that for all of your deliberation, you didn't seem to give OGRE (which should be one of your top considerations) a decent trial and dismissed it offhand because you couldn't get it set up correctly because you were running it in VMWare. It's a good engine and you should judge it based on its actual merits, rather than the fact that it has no software fallback and you couldn't use it through VMWare.
Your decision of whether to use another engine or write your own should be based solely on how comfortable you are with using the engine in question. Please actually try writing something with it, and a lot of your questions will be answered.

That said, I would like to reiterate the point already stated by others here that no engine is perfect. The only way to have a perfect engine (for you) is to write one yourself and you should do that if you feel technically capable. In either case, I would urge you to take a good look at OGRE as it is at least a very good engine architecture reference and for me it has been invaluable in my own engine design.

[Garthy]

you didn't seem to give OGRE... a decent trial and dismissed it offhand

... I've got a PC with an ATI 3D card set aside for a Windows testing, and occasional access to a couple of others .... Perhaps I should set it up today and see how well the samples behave on real hardware ...

toastie, I propose that we not begin an argument and instead mutually acknowledge that we may not be able to help each other much in this particular discussion at this particular time.

[ColossusEntertainment]

I think ankh was a retail game.

Yeah? Well, that makes sense I guess...

I'm still struggling to find released, commercial indie-games done using OGRE. I wish they had something like this page: http://www.garagegames.com/products/torque/tge/games/ (or maybe they do and I just missed it?)

My major concern with OGRE though, is its LGPL license. It's rare that you can use an engine to make a game into a product without changing it, sometimes extensively, and with OGRE, that means those changes are not yours, but you need to give them away to anyone who wants them. That's not a restriction I would want to work within if I was running a business... I just wish people would either charge for their stuff, or give it away completely free (no restrictions), and not mess about with all that in-between stuff... would make it easier for everyone, I think