Feedback on Game Engines

[gameboy]

Part II: The engines

First of all, let me go over what I will exclude from consideration. I won’t consider using for commercial development any GNU GPL (General Public License, or simply GPL) [http://www.gnu.org/copyleft/gpl.html] engines/tools, or even any that are licensed under GNU LGPL (Lesser General Public License, or LGPL) [http://www.gnu.org/copyleft/lesser.html]. Sadly, this will exclude engines like Tenebrae [http://tenebrae.sourceforge.net/index.php] or OGRE [http://ogre.sourceforge.net/] which is also used by PureBasic [http://www.purebasic.com/index.php3]. Tools licensed under these agreements are basically not practical for use in commercial development. In short, you have to open your commercial application up for dissecting or reverse engineering so users can freely modify either your code, or whatever you used that was licensed under these agreements, so they can recompile them with your code and run their own modified version of your game. Under the LGPL you could probably avoid this problem by furnishing an SDK to your application’s binaries so users could modify the LGPL portion of your game, and using the SDK to link the modified code back with your game application. I wrote to the open software foundation for clarification on the LGPL recently and they basically told me to either get a license giving me permission for commercial use from the creator of the tool licensed under the LGPL, or retain a lawyer who is expert in this field. Lawyers, ugh. No thanks, I’ll stay away, unless I can get the creator of the tool to give (or sell) me a full-blown commercial license IN WRITING.

I also excluded from consideration any technology that is obviously way too expensive for a typical individual to purchase – so we’ll have to exclude tools like the Jupiter or Lithtech Engine [http://www.touchdownentertainment.com/licensing.htm], or Aruan Jet [http://www.auran.com/jet/license.htm], etc. from consideration. Even Virtools [http://www.virtools.com/ ; http://www.id8media.com/framesets_3...o_dev_frame.htm] must be excluded, because by the time you add up the cost of the base product, add modules like physics, and then pay the royalties on the games you develop…well, you’ll need a wheelbarrow of cash to pay for it all. One clue that the tool you are considering is probably too expensive for an individual is if they don’t (or won’t) list their price. Sometimes, they won’t even answer your email inquiries (*cough* Vulpine [http://vulpine.de/index.html]). It’s one of those situations evidently where if you have to ask, you can’t afford it.

One discussion of game engines and development tools I came across that mentioned many different engines can be found on the 3D Game Studio’s forum: [http://www.conitecserver.com/ubbthr...part=1&vc=1]

I will try to keep the list relatively short, with only listing engines that seemed to have potential (IMHO) that I’ve run across. Perhaps you have others you’ve seen or have used them and can give us some more feedback about them as well.

----------------------------------------------------------
Quest3D [http://www.quest3d.com/]
----------------------------------------------------------

Price:
Professional Edition $749
Enterprise Edition $1999 (includes database and networking support).
No royalties or “per product” fees to pay.

Pros:
Performance of the engine is good and the graphics look gorgeous. The idea of linking channels (pre-compiled “chunks” of code) together graphically makes for rapid development works really well and you can see changes instantly as you tweak your project without having to recompile. If a channel doesn’t exist that has the capability you need, you can code one up in C++ using the Quest3D SDK. When you’re done, you’ve got a brand new channel that you can reuse in future projects if you like. The “channel” linking concept works really well and is both fast and flexible. The channel graphs can be nested as desired and displayed visually so that details can be hidden away to appropriate levels, allowing a design to be grasped and understood visually “at a glance” (well, once you get the hang of channel graphs, that is). Making games with this tool will still require work, but probably much less than if you used traditional methods.
The product has physics via the ODE library, and even support for rudimentary A.I., crowd animation, etc. Maps are portal based – no need for .bsp “compilation” step.
The demo games made with it look impressive.

Cons:
Getting animation data into Quest3D and working can be tricky. Although Quest3D provides the capability to bone and animate your own characters from within the product, its tools for doing so are rather primitive; your best bet is to find software that exports geometry and boned animation in the DirectX (.x) format. The good news is that many of the new 3d modeling and animation tools are starting to add support for games by doing this. Using channel graphs may seem a bit strange for programmers at first, and there aren’t any channel debugging tools available – about all you can do is visually see if a channel is being executed while running your project (there is a blue outline around the channel box). Also, I don’t know of any commercially released games that were developed using this product (yet). However, the same could be said about some of the other engines listed below as well.

----------------------------------------------------------
Crytek [http://www.crytek.de/]
----------------------------------------------------------

I initially wrote these guys off, because they didn’t list the price of their engine, and on other forums people were saying it was way too expensive for the indie developer. However, Renderman has said above that the price is about $900 that seems is reasonable. I wonder if that is a one-time charge, or a per game fee. I agree that this engine looks very good (based upon the screenshots I’ve seen).

----------------------------------------------------------
Amp II [http://www.4drulers.com/amp.html]
----------------------------------------------------------
$2000 for a commercial license. Not clear whether this is a one game or permanent license to develop any number of games.

----------------------------------------------------------
Aztica [http://www.aztica.com/modules/news/]
----------------------------------------------------------
Cost: Free until game is sold commercially, then there is a $195 licencse fee for each game developed.
Comments: A bunch of DLLs that can be used with Java, Delphi, C++ Builder, VB6, or .Net
Pros: Built in scripting language (called LUA). Supports MD2, MD3 (Quake2 & 3) models.
Cons: COM based. DX 7 only. (version 2 in the works).
Awkward website; no downloadable demos of games made with it.

----------------------------------------------------------
Cube [http://wouter.fov120.com/cube/index...f7e2c2d4483b273]
----------------------------------------------------------
Price:
Freeware, for any purpose. However, it uses Fmod and SDL (see below).

Pros:
Doesn't need any kind of map precompilation.
Map Editor is engine's map-making tool is a very quick and direct, WYSIWYG process.
Collision detection and physics included.
Good net code.

Cons:
Doesn't have sophisticated eye-candy of state-of-the-art 3d engines.
Engine is still in beta stages.
Uses Fmod [http://www.fmod.org], which requires a licensing fee ($100 per game developed, or $2000 per game developed if you sell a lot of copies).
Uses SDL [http://www.libsdl.org/license.php], which uses the LGPL license.
Not well documented.
Not the best choice for a commercial game, due to libraries being used, and the fact that the engine is only in beta form and not well documented.

----------------------------------------------------------
3D Game Studio [http://www.conitec.net/a4info.htm]
----------------------------------------------------------
Price:
$199 or $899 for commercial editions. No royalties or extra fees.

Pros:
Licensing, established track record, many users.
Built-in tools.
Can quickly produce a working game.
Visual (level and model editor) and scripting tools (C-based scripting language).
Latest version supports DirectX 8.1, boned animation, physics.
Many commercially released games have been made using this product.

Cons:
This engine simply does not perform nearly as well as other state-of-the art systems.
Graphics don't look great; engine is a bit behind state-of-the-art.
Might be difficult to get the engine to perform well.
Conitec's own documentation of what's new in version 6 says some things, like reading the mouse, take longer in version 6 than in version 5. In some ways version 6 is a downgrade from version 5. The manual also said many things had not yet been completed or tweaked yet.
However, the product is continually being updated and improved. For instance, boned-animation support has just been added.

----------------------------------------------------------
3D State [http://www.3dstate.com/]
----------------------------------------------------------
Price:
$399 (for developing one application).

Pros:
Engine looks good, the development package comes with integrated tools that all work together. The version 7 demos look great.

Cons:
Uses Fmod, which requires a fee for commercial use ($2000 for one application!)
Frame rate for the previous version demo was erratic; e.g. player could run much faster after a short time if not firing bullets at the same time.

----------------------------------------------------------
DarkBasic Pro [http://www.thegamecreators.com/]
----------------------------------------------------------
Price: $100

Pros:
Supports DX9, boned animation, many other modern features.
Easy to use programming language.

Cons:
Not too stable – the debugger for instance frequently crashes on my system.
Help file (i.e. command reference) is not completely searchable.
No physics (but could be added via DLL calls to the Tokamak [http://www.tokamakphysics.com/] physics engine).
Graphics don't look as sharp for some reason compared to other game engines. (Poor rendering?)

----------------------------------------------------------
Blitz3D [http://www.blitzbasic.com/]
----------------------------------------------------------
Price: $100

Pros:
Solid, fast programming language.

Cons:
Currently only supports DirectX 7.
No built in physics.

----------------------------------------------------------
The Nebula Device [http://sourceforge.net/projects/nebuladevice/]
----------------------------------------------------------
Description:
Nebula Device is an open source realtime 3D game/visualization engine, written in C++ and scriptable through Tcl/Tk, Python and Lua. Its rendering system supports both DirectX (8/8.1) and OpenGL. It runs under Linux and Window and is being ported to Mac OS X.

Price:
Free, for any purpose (even commercial). Must display copyright notices and include Nebula Licensing Terms with product. Nebula is released under the MIT license [http://www.opensource.org/licenses/mit-license.php]. Basically, you can use it for anything you want as long as the copyright notices stay in place.

Pros:
The demo game, Project Nomads (made with Nebula 1) looks great; visually stunning, very responsive.
Nebula 2 is in alpha stage at present.

Cons:
Lightly documented.
More suitable for people who already have a good understanding or experience in writing 3D engines.

----------------------------------------------------------
Origi-3d [http://www.origi3d.com/]
----------------------------------------------------------
Price:
Two plans. You have your choice of:
1. $2500 + 5% royalty on all games sold, or
2. $150,000 without royalty.

Pros:
Claim to have the world’s most powerful game engine.
The demo MOVIES look great.
Quote: “Magic Carpet's cloth simulation IS 360 TIMES FASTER THAN ALL OTHER REAL TIME CLOTH SIMULATIONS.”

Cons:
Was not able to get the engine demo to install successfully on any system.
Huge download (over 200MB). The documentation says running the demo takes about 4-5 minutes to load on a low-end Pentium 4!

----------------------------------------------------------
Power Render [http://www.powerrender.com/]
----------------------------------------------------------
Price:
(Power Render 5 engine).
$289 for a single programmer, Distribution of unlimited shareware and freeware products.
$1500 for a site license, Distribution of unlimited shareware and freeware products and Distribution of one commercial product.

Pros:
Comes with a fairly complete set of tools for game creation, with support of most major 3d file formats. They claim they have an extremely easy API developed over 6 years of actual usage with commercial applications, and full DirectX 9 support.
The movies for their upcoming prototype engine look impressive. They don’t apply to this product, however.

Cons:
No physics (but has its own collision detection). Could use Tokamak physics engine.
Price of $1500 applied to each game developed. Price for next version (Prototype) will be prohibitively expensive, at least for the source code versions (15-50K). Non-source code versions of the engine (to be released later) are supposed to be cheaper.

----------------------------------------------------------
Qube [http://www.qubesoft.com/q/overview.php]
----------------------------------------------------------
Price:
Q for windows is currently free (for now).
However, pricing for support can be very expensive (5,000 pounds for only 3 months!).

Pros:
This engine has been used for development on other platforms (game consoles).

Cons:
Frame rate seemed slow compared to some of the other engines, and the rendering did not look as good.

----------------------------------------------------------
Sylphis 3D [http://rainbow.cs.unipi.gr/~harkal/sylphis/index.php]
----------------------------------------------------------

Not too much information is available yet about this engine. But the playable demo and demo movies look impressive. The website doesn’t get updated too often. But it is something to keep an eye on.

Price:
Unknown.

Pros:
Stunning performance. Includes physics.

Cons:
Not yet released.

----------------------------------------------------------
Torque Game Engine [http://www.garagegames.com/pg/product/view.php?id=1]
----------------------------------------------------------

Price:
$100. But if you have a successful game and your revenue is over $500,000, it costs you $10,000 per game (up front!) to develop using their engine and tools.

Pros:
Decent performance. Comes bundled with a set of tools to assist in making games.

Cons:
The tools that come bundled with the engine can be purchased or downloaded on your own. The demo games were OK, but the engine is not as state-of-the-art graphically as other engines.

----------------------------------------------------------
Truevision 3D [http://www.truevision3d.com/html/]
----------------------------------------------------------

Supported by .NET.

Price:
Single commercial product: $150
Multiple commercial products: $500

Pros:
Full DirectX 8.1 and 9 support.
This engine may be an option if C++ programming is not desired, or .NET development is desired.

Cons:
Tech demo was OK, but inferior to other engines. However, improvements are being made: http://www.truevision3d.com/html/modules.php?op=modload&name=phpBB2&file=viewtopic&t=3240

----------------------------------------------------------
Cipher [http://www.cipherengine.com/index.php]
----------------------------------------------------------
Is probably worth a look. It comes with a complete set of tools, and a really good price (just $100, period). The downloadable demos look OK, but may not be quite as visually stunning as some of the better tools listed above.

Pros:
Looks to be a solid engine with great licensing for commercial development. You won't find a better price.

Cons:
Some of its game development tools are designed to work with 3ds max's .ASE format, which no other tools apparently export to. Little documentation on this scene format is available, making writing a reliable conversion program to its format difficult.

So, what would you recommend for game development?

[jud]

Im not sure if this is what you are looking for as your thread is way too long for me to read all of it,but this application has great format support for all game engines and will probably make working with different platforms a breeze,and it has full rendering capabilities and loads of other tools for game developement and it use's Python scripting.
Hers the link
http://www.caligari.com/gamespace/

Oh! and it is very reasonably priced at $199

jud:surprised

[Renderman_XSI]

gameboy, thanks for the long reply.Its great now i have more options to look at.

i really wish Softimage would intergrate a game engine, in XSI.

[gameboy]

Actually Renderman, there is an all-in-one modeling/animation/rendering tool that includes a game engine; it's called Blender. http://www.blender.org/ It is open source however, and therefore its licensing is not really suitable for commercial development.

[DonMeck]

Quest3D and Virtools
----------------------------
The schematic scripting/programming metaphor in Quest3D looks very low-level for me therefore i don't believe that rapid development of complex games can be done in it. Because everything is a channel and you need to go through a couple of folders (a grouping mechanism for channels) just to see whats going on.

At least this is my impression when i did some short glances at it.

Virtools costs a lot more and it asks for royalisties, but their schematic programming metaphor is much more high-level and therefore better (in my opinion) as you can see what is going on more quickly than using the quest3D metaphor. Additionally they addressed some draw-backs that occur when using a visual/schematic programming metaphor by introducing a scripting language. Code-programming is simply cleaner and easier to handle when it's about creating low-level logic (many tests and loops).

Open source
-----------------
open source doesn't mean it doesn't fit for commercial use. It depends on the license. Is it GPL or LPGL etc. LPGL allows you to not release your source code if it's not part of the engine itself. Ie. game logic, plugIns etc. For example you use OGRE and you modify/improve OGRE itself, then you are asked to provide the sourcecode for those modifications to the community. But your game logic - which is a part from the OGRE engine - doesn't need to be open source. Same for plugIns.

Some more free 3d engines
-----------------------------------
NEON: http://neoengine.sourceforge.net/
Irrlicht: http://irrlicht.sourceforge.net/

More criterias
-----------------
Very important for choosing an engine is also if it provides you with everything you need - and if not how long or how easy is it to add the needed functionality. Same for the pipeline for importing media. So both comes back to time. Then there is the question of support (and community?). etc.

cya
dom

[Renderman_XSI]

keep in mine im no programmer, when asking this questiion. Im looking for a commercial game engine, that have at least DX8 support, and would be up to speed to create PS2,XBOX titles. Price i can afford 1000 USD per license. Something that lets us implement our own tools with ease. Cipher is nice, but it seems to lack the texturing ability(base on their base demo).

the genre is adventure game, and graphics are important. it would include a good number of combo's, attacks,etc with sword.

BTW: virtools seems nice.i'll look into it.

[gameboy]

Thanks for your comments, DonMeck. They were just the kind of feedback I was looking for.

I agree that Virtools is more suitable and a better tool for producing games than Quest3D, but it's price is beyond my means at present.

I also found Quest3D channel programming to be somewhat awkward at first. Channels that have been hidden in "folders" (which can be expanded/contracted by pressing the space bar) help overall program comprehension in that a folder appropriately named describes the functionality inside of it, at a glance. A subtree of Channel graphs can be saved as a "template", further abstracting the development process. In short, you can compose "higher level" code (or channels, in this case) with templates, and string them together. The Quest3D designers hesitate to call it the "ideal" tool for making games, saying some programming logic may be too complex to code using channel-logic and channels are limited with respect to debugging methods. But this can be overcome through the use of the SDK, which works with MSVC++, and can be used to create your own custom channels when the need arises. The Quest3D developers say that games can be made with it, and some of the demos like the Monster Mayhem truck game demo look impressive. I am comfortable and used to programming in the "traditional" way, but channel graphs can speed up development (for example, you can get a 3D animated scene up and running with a few mouse clicks by dragging a template and importing an animated model).

Regarding the LGPL license, the problem I see with it is that you must allow users to be able to change the public code covered under the LGPL license and then recompile it with the rest of your application, and then run the modified application. This has at least two drawbacks; one is that unless you provide some sort of SDK for your application, you must "open" it up somewhat so they can recompile it. The second drawback is that users would be able to create their "customized" version of your game, and if you want players to be able to compete against one another online, these customized versions of the game would be equivalent to "cheating" (unless the point of the game was to allow users to compete against one another using their own customized modifications).

Other open source licenses, like the MIT, BSD, and zlib-libpng, for example, are just fine for commercial development.

[Blud_Stane]

you guys seem like a knowledgable group so here goes:

Which engine would you suggest using for...

-Equal quality to "Jak & Daxter" games by Naughty Dog
-bump mapping
-cell shading
-DX9
-ragdoll physics


Thank you as I'm still too confused with all the "hype" all those engines put on their sites... i wanna know what is realistic.

[gameboy]

Blud,

I would like to mention one engine that has impressed me the more I play around with it, and that is the TrueVision 3D engine. I initially wrote it off somewhat because the techdemo video presentation didn't seem all that great to me, but as I work my way through their programming tutorials, I am impressed by its power and ease of use. I'm getting frame rates of several hundred frames per second, even when programming in Visual Basic!:eek: (At least on an Athlon XP 2800+ running at 2GHz, and an ATI RADEON 9700 Pro video card, at 1280x1024 resolution, 32bit color). Feature-wise, it has support for DX8.1 and DX9, and supports DOT3 Bumpmapping, etc. You can find a list of some of its features here [http://www.truevision3d.com/html/modules.php?op=modload&name=News&file=article&sid=2]. The engine is continually being updated, and the next major release (version 6.5) will be optimized for .NET managed code. Right now, version 6 of the TrueVision 3D engine is COM based, so VB6/VC++6 slightly outperform their .NET counterparts. This is expected to change with version 6.5. Its licensing cost for commercial game development is very good ($150 for one game, $500 for an unlimited number of game releases. The engine is free for non-commercial uses). The engine also comes with a set of tools to aid you in quickly making games. This engine and its toolset may be the fastest way to create games that contain a high degree of visual quality.

As far as ragdoll physics, you can use the Tokamak physics library [http://www.tokamakphysics.com/] which is free.

[Blud_Stane]

how did you even get the thing working.. we've almost given up cause our programmer is trying to just get it started but has like 90+ errors. Maybe theres a simple setup out there, and we'd like to use Truevision as its powerful.... but im leanin toward Torque unless we get it going :(

[gameboy]

Blud,

I've been using VB6 and TrueVision3D, going through the tutorials and modifying them for experimentation. It has been about the easiest thing I've done yet.

The C++ file has a line you must change in the tv3dcpp.h file (line 154, I believe) that must point the path to the truevision3d.dll file to match the location on your computer. However, regarding the VC++ file compilation, I was getting "fatal error C1084: cannot read type library file: ...\truevision3d.dll". I'm not sure if it is because this library somehow isn't registered correctly for VC++. I made sure my Directories were pointed to the correct locations, etc. :banghead: If you go to their website, you can open up a private chat with one of the developers and ask them (I was going to eventually do this).

I've been working most recently in VB6 with their product, putting it through its paces.

[DonMeck]

Blud_Stane,

sounds like an ambitious project you are planing.

You should consider that - i think - only state of the art physique engines have out-of-the-box rag-doll physqies. Contraining the characater parts is one thing, but limiting etc might be something not very trivial - if it should look convincing.

"Jak & Daxter" did have large levels and background loading (if i remember correctly). Nothing that every engine has. Indeed it was a special feature of the game (i think) - like for Dungeon Siege.

How much can you invest in an engine?
I wanted to propose the copperhead engine, but i can't find info about it anymore. Maybe it has beeen sold? Was the successor of the jupiter engine, i believe.
http://www.touchdownentertainment.com/

[Blud_Stane]

Gameboy: also do you know about implementing netcode into Truevision? And what projects are you working on, i'd like to just talk to you about Truevision and what you know?

thanks

steve (bludstane)
aol: royers24
icq: 83646867
yahoo: sroyer_hcc

[gameboy]

Blud,

I'm just learning TrueVision by working throught the tutorials; I'm not a guru on it or anything. I've been looking at a lot of software tools lately to determine what I should use for game development. I haven't tried to implement a client-server application using TrueVision3D.

The networking capabilities of TrueVision3D, according to their website, should be as follows:

Net Engine :
- DirectPlay8.1 support.
- Peer 2 Peer, Server, Client
- Send custom messages or predefined 3d messages.
- Works on LAN or on Internet.
- Unlimited number of players with server component.
- Easy player management with events (join, quit...)

Hope this helps. :D

[Blud_Stane]

Originally posted by DonMeck
Blud_Stane,

sounds like an ambitious project you are planing.

You should consider that - i think - only state of the art physique engines have out-of-the-box rag-doll physqies. Contraining the characater parts is one thing, but limiting etc might be something not very trivial - if it should look convincing.

"Jak & Daxter" did have large levels and background loading (if i remember correctly). Nothing that every engine has. Indeed it was a special feature of the game (i think) - like for Dungeon Siege.

How much can you invest in an engine?
I wanted to propose the copperhead engine, but i can't find info about it anymore. Maybe it has beeen sold? Was the successor of the jupiter engine, i believe.
http://www.touchdownentertainment.com/

Thanks for the help Don,

The motto for my small team is "more with less" I've been on and off contract projects with people that couldn't run a team, and been through a couple mods. I am planning a small title, with the most advanced but cheapest engine because that's more money i dont need to ask for from a publisher. I don't want to spend more than $100 (maybe $200 max) on an engine, but ultimately i believe there are resources out there for a starting group like mine to make a demo for free. I'd like to talk more but gotta run my contact info is above in the last post if you wouldnt mind talkin over things like this, any help is "help"

steve

[gameboy]

TrueVision update:

Their site has changed (it may have been hacked :shame: - the forum is down and the chat area displays Chinese (or Japanese?) characters instead of english now).

There no longer appears to be any mention of networking capability for their product in the features section, and the demo does not contain any classes or objects for networking.

It is possible to add networking to a project with third party tools, but this could be expensive (e.g. http://www.replicanet.com/pricing.html).

I may have to give up on TrueVision if the networking features have been dropped...just when I thought I had found the right tool. :sad:

[Blud_Stane]

well gameboy my small team and I have gone with:

Torque engine

...and we're gonna try and incorporate ODE physics

seems like the cheapest way to get state of the art games

[DonMeck]

If you use the torque engine, u nd to publish through their site (garage games), not? You wanna do that?

[gameboy]

I was just visiting the TrueVision3D website, and they're evidently experiencing problems due to a change in servers. They DO have network support evidently, in an object called "TV3DNetEngine". However, I did not receive this object in the download of their demo.

Blud, if you want Garagegames to handle your distribution, and take 35% of the revenue, then maybe that is the way to go. (Oh, and by-the-way, if you have a really successful game and your revenues are over $500,000, you must pay them $10,000 to develop with their software [per title, up front from then on]). I am also not convinced that their engine is as state-of-the-art graphically as some of the other engines out there, but I guess the net code is supposed to be pretty good.

Good luck. Hope it all works out. :thumbsup:

[Blud_Stane]

DonMeck: Well, I do not plan on using Garage Games if I can help it. I will be attempting to hit up some mainstream publishers that I think our Genre and budget are aimed for. Garage Games will be a fall back if several attempts are shopping around do not get anywhere.

gameboy: Like I mentioned above, Garage Games is a cushion, and more of a garauntee to get published if a more stable publisher does not pick up on it. I have unlimited webspace, ftps use, private mIRC channels...and luckily just recieved an email of yet another programmer that has a Torque engine license and has played around with the engine for a year.

Thanks for the kind words, and luck... I am becoming more and more optimistic :) Truevision did not work out because our only programmer at this time, uses C++ and the engine is in pieces he cant even get it running.

[DonMeck]

make sure htat you understand this paragraph from the eula:

"
(b) Licensee may not use the Software in whole or in part to create products for competing game publishing companies, commercial websites, or any other commercial or non-commercial entity, whether public or private if the sum of the annual revenue of the Licensee and publishing entity exceeds $500,000, without obtaining a COMMERCIAL LICENSE from the Licensor.
"

otherwise, good luck! :)

[Blud_Stane]

well if we make over $500k with sales I'll be happy to stick with GG or anyone because that is enough success right there :)

but thanks!

[Fox Mc Cloud]

Here some really nice cutie a state of the art engine that
gone Open Source the tech demo looks really nice:
http://catmother.sourceforge.net/
3D Max 5 Exporters and some other Tools are included
but you gona need Visual C++ experiance to get startet
with this Engine.


Greetz Fox Mc Cloud

[gameboy]

Thanks Fox.

Yeah, I just saw a mention about that on www.gamedev.net. It uses the BSD license, which makes it good even for commercial development.

[DonMeck]

i just donwloaded the demo, but its sooo dark - cant see anything!

Animation is too stiff and the story is really cliché pur. But great that they realease the whole stuff for free!!!

[gameboy]

TrueVision (http://www.truevision3d.com/html/) network code update:

Evidently the networking code portion of TrueVision isn't ready yet, but it will eventually be there in version 6.5 (right now they are at version 6.0). Here is a quote from darqSHADOW (TrueVision staff member) on their website:

"For 6.5 we are going to be taking my networking technology from Nada Games and creating a very solid, and flexible networking system. (A subset of the same system that NG sells to clients for $10k+ will be available inside of TV3D.)"

Hope this clears up any confusion I might have caused. :blush:

[gameboy]

More info on TrueVision:

Evidently the engine was originally written in VB, but is being recoded in C++. This has caused some problems with those trying to code the engine in C++ (VB seems to work fine).

The port will not be completed until version 6.5, or so it has been said on their forums; so you might want to hold off C++ development using this engine until then.

[jaydmax]

Here is another website for researching low cost game engines
http://www.salleurl.edu/~manuellv/Public/english.cg_gameengines.html

[gameboy]

Thanks for the info, jaydmax. It looks like they've got a ton of info on that website. :thumbsup:

[jaydmax]

Thank you Gameboy for the research you've passed on. I myself am currently playing with Quest3D (im still having issues getting animated files from maya to import properly. Models I got but not animation yet.) and and a few others but I really want virtools. I got a evalution copy a few years ago at Siggraph and its a sweet program for non-programmers like myself. I think im gonna save my pennies for virtools.
Does anyone know the royalties virtools charge?
One thing though, the Crytek engine that you priced at $900 bucks in not for the engine but for the Polybump plugin. The engine itself is priced up there with Unreal. I emailed them and they didnt bother to reply. oh well... but the Jupiter engine which looks really sweet can be evaluated upon signing a NDA. But its still 10K.




www.jaydmax.com

[DonMeck]

and dont forget the coppermine from them. Yes it's 10k but it has physiques and networking ... this costs extra with virtools and its isn't cheap. In the end , its maybe about the same price ... but of course ... virtools can also be used to a certain degree by non-coders

[jaydmax]

what happened to coppermine on the website? it is no longer advertised. did they sell it or something?

[gameboy]

It appears that 3D GameStudio (http://www.conitec.net/a4info.htm), Conitec's latest engine A6, has come a long way. Two of the latest demos based on the A6 engine, FX Ball (http://www.conitecserver.com/down/fxball.zip) and Crane (http://www.conitecserver.com/down/crane.zip) , look pretty good, and demonstrate realistic physics. :eek: The graphics and lighting seem to look better now, with some real-time shadowing effects in evidence.

I may have to give this product a try. With its 30-day trial download, you can't go wrong. You can evidently save your work, but not in a form that doesn't require the GameStudio itself. Visit their download (http://www.conitec.net/a4update.htm) page for more demos.

[gameboy]

Another RAD game-making tool that should give 3D GameStudio a run for its money is Radish Work's Cosmos Creator (http://www.radishworks.com/CosmosCreatorInfo.htm). It touts the ability to rapidly create and edit 2D or 3D scenes while play testing a game.

It contains its own internal tools for creating models, textures, animation (yes, even skinned/boned animation) but also supports import of 3ds and .X (with boned animation for the .X files!). It also has a game engine that you can script via LISP or C++, and has built-in physics and other various special effects like real-time shadowing and bump-mapping! You can create standalone .exe's or interactive web presentations/games that can be played via the Radish Works viewer (http://www.radishworks.com/viewer/GameViewer.htm) (988K). Your creations can be distributed royalty free, so the licensing is great for commercial development.

It stores all its files in a special internal format from which it creates all geometry procedurally on-the-fly, making for some extremely small file sizes. :eek:

A free trial download is available, with tutorials to get you started. The product can be purchased in one of three levels of features: Silver ($49.95), Gold ($99.95), and Platinum ($499.95). To get the C++ API you've got to purchase the Platinum edition. :D

This tool might very well be the best all-in-one package for the money. :thumbsup:

[DonMeck]

Ah interesting link! I didn't know that one although i've seen a screenshot from it on the ODE site.

Its defenitly an intersting thing, but still not developed enough - in my eyes. BumpMapping doesn't look bumpy. The physique in the physiques doesn't feel very well - but from other applications i know that ODE can do better - maybe a settings issue. The LensFlare and the LightGlow demo isn't convincing at all. Unfortunally the webPlayer crashed quite often.

But the file sizes are impressive. Realtime boolean is great and lightmap calculation is also very nice. It's something to keep an eye on. :)



About the engine from touchdownentertainment. It's copperhead and not coppermine - i remembered it wrongly. It's not accessible through their homepage but still can be found on their site:

http://www.touchdownentertainment.com/CInfo.htm

It really looks very sophisticated ...

[gameboy]

Another new game engine has emerged, called Artificial Heart (http://www.3dlevel.com/artificialheart.php). It is a work in progress, written in VB.NET. It's physics demo (http://www.3dlevel.com/ardemo.php) is interesting, and demonstrates real-time shadowing as well.

Information as to how to use the engine is not given through an SDK, but a series of tutorials (http://www.3dlevel.com/forum/viewforum.php?f=21) .

I believe the engine is free for non-commercial use, and for commercial use, you must contact them via email.

http://www.3dlevel.com/

If you've been looking for another option for developing games using VB.Net, this is worth looking into. :D

[tburbage3]

gameboy,

Many thanks for getting this discussion started and the obvious amount of research you've already put into it. I had thought to start a similar thread when I came here and found yours.

As you've reported there are lots of possibilities out there -- some more complete solutions, some basically just rendering systems, some only suitable for arcade-style games, others trying to push the envelope on current state of the art. It is hard to sort through.

The ideal solution to me would sound like this:
- Provides a nominally complete set of design-time tools to bridge the gap between one of the good 3D content apps and what the engine needs;
- Most ideally, would allow use of said 3D app as the primary level design tool;
- High quality (for me, OpenGL so Linux or OSX could be targeted) renderer oriented toward realistic output where the developers are shooting for state of the art including full vertex/pixel shader support;
- Some C++ programming may be required, but would not require a full staff of programmers to maintain and enhance -- allowing development focus to remain for the most part on content and game logic;
- Provide decent animation and effects support -- particle system, physics/collision detection, procedural animation (some crowd, flocking, etc. support);
- Total cost within reach for a *small* group of artist/developers (say 1-5) -- as you mentioned -- within the same scope as the other necessary components such as 3D content app. Non-restrictive license -- *perhaps* with royalty provided the amount is *very* reasonable. These criteria let out the big names in commercial game engines;
- Scriptable animation/game logic using a standard language such as Python or perl.

Again, great thread. Keep it going!

[gameboy]

Thanks tburbage3.

Another place I came across during my "research" for this thread that you might want to check out is Neuro's 3D Engine List (http://www.3dhangout.com/cgi-bin/Blah/Blah.pl?b=Gen_3DEngines,v=display,m=1046376474,s=0). He lists a few engine that have not yet been mentioned on this thread as well, and has even put them into price categories. One thing I'm beginning to realize is that if you really want a well documented and supported (i.e. frequently updated) engine, you generally have to pay for it, and the larger the installed base of customers, the better off you'll be in this regard. I guess this is probably obvious. The fewer the installed base of customers, the more you'll probably be paying for a quality engine and support (i.e. add-ons, etc.) although some of the licensing agreements can be rather ridiculous at times when comparisons are made between the various options out there.

Good luck in your search. Let us know if you find that "ideal" solution. :thumbsup:

[atenyotkin]

There is one engine that I didn't see on your list, that may be of interest. It's called Tenebrae. The first version requires you to have Quake 1, which wouldn't work out for you since you want to sell that game. Tenebrae 2, however, is a stand alone game-engine that should come out in the near future. It uses the per-pixel shaders that AMP II, Power Render X, Doom III engine use. So it looks good. You need a pretty quick machine to run it though. The engine is free, and you can do whatever you want with it.

www.tenebrae2.com


My group is using Cipher Engine right now. It is the most powerful out of all the ones you listed. We tryed Torque, and we tryed the 3D Game Studio, and it just wasn't what we were looking for graphically. Cipher Engine really cought my eye though when I saw the demos. The only problem with using it though, is that you have to have experienced people making the game. All geometry is made in MAX, even the level. Everything else including entities (light, button, etc.,) and event scripts are done in c/c++. It's really really though in the beginnging, but even those who barelly knew how to program managed to understand it. Now, because everything is coded, you have unlimited possibilities for every entitiy and the scripts. And also becuase all models and levels are made in MAX, you can animate anything. For example you can have a terrain that will deform, or face morph out of a wall. Pretty cool. You just gotta know the software.

We were actually considering purchasing licenses for the Source engine used on the Half Life 2 game. But once we saw Cipher, we realized that we can add all the features we were interested in.

One other thing about cipher. it has two tools that come with it that really help you. Cipher Shader Tool and Particle Tool. They are very powerful.

So, even if you're not going to be using Cipher Engine now, I would give it a try for your next game. The engine is evolving way quicker than AMPII or 3D Game Studio.

The forums are always full of people to answer your quesions, and the developer of the engine is usually the first one to anwser.

[DonMeck]

gameboy,

very good link, thx.

atenyotkin,
yes Cipher looks really good. Especially foe that price. But i only wonder how much OOP is it when its only in c? How far developed is the charater system? secondary animations?

[atenyotkin]

I am not sure what you mean by how developed? Everything is animated in MAX. You can check out what cipher supports as far as animations on their site.

http://cipherengine.com/features.php



Download the Dancing Alien Girl movie to see the quality of the in-game animation.

http://cipherengine.com/download.php


There are no tools that come with the engine that help you animate something in a certain way.

The only other thing you can do to the model, is load it into a shader tool, and see how the shaders are applyed in real-time. For example with the dancing alien girl, you can change the shaders on the skin texture as she is dancing.

I hoped I helped.

[DonMeck]

Talking about mixing animations at runtime. I like primary animation (walk cycle) + secondary animation (something with the torso and arms or so)

[Mussi]

haven't read most of the thread :P, but i would recommend you dark basic profesional, I've worked with quite a few gamemaking packages but this is the one I like most. it is an easy to understand language, It has amazing powers and the thing is... it's still growing :). unlike the click and drag game making packages you really have the freedom to make any type of game with any kind of features. but ofcourse it takes some time to learn it and to be able to make some cool games, but I promise you, if read a lot, you will be able to make some real good looking games with it within months

[Mussi]

haven't read most of the thread :P, but i would recommend you dark basic profesional, I've worked with quite a few gamemaking packages but this is the one I like most. it is an easy to understand language, It has amazing powers and the thing is... it's still growing :). unlike the click and drag game making packages you really have the freedom to make any type of game with any kind of features. but ofcourse it takes some time to learn it and to be able to make some cool games, but I promise you, if read a lot, you will be able to make some real good looking games with it within months

[gameboy]

My current favorite game development engine is TrueVision3D (http://www.truevision3d.com/html/). It is fast, seems very easy to work with, and has a good licensing scheme for commercial development (i.e. inexpensive and fair). :D It is also free for non-commercial development.

I found DarkBasic Pro to be far slower and "buggier" than TrueVision3D. TrueVision3D's rendering quality is also superior.

With TrueVision3D, you can pick which language/IDE you want to develop in. They have tutorials to bring you up to speed using different programming languages (e.g. Visual Basic 6.0, VB.NET, C++, C#, Delphi). Currently, I'm using Visual Studio .NET and C# (along with Whole Tomato Software's Visual Assist .NET (http://www.wholetomato.com/) - a great little tool by the way). It is nice to have professional quality tools to work with; in comparision, the DarkBasic Pro IDE seemed rather toy-ish.

However, of all those Basic-language all-in-one programming-engine tools, (e.g. Blitz3D (http://www.blitzbasic.com/), PureBasic (http://www.purebasic.com/index.php3), IBasic (http://www.pyxia.com/), etc.), DarkBasic seems to have the most features and is updated the most at present.

But none of those Basic-language based tools really compare favorably to TrueVision3D's power. TrueVision3D recently released version 6.1 of their SDK and tools. (http://www.truevision3d.com/html/modules.php?op=modload&name=Downloads&file=index) It's worth checking out. :thumbsup:

[Mussi]

checked it out, there is no feature that can't be done in darkbasic pro, is there lightmapping support? bump mapping? detail mapping? blend mapping? realtime shadow shading? alpha mapping? Nvidea effect supprt(.cg or .fx)? it doesn't say anything about those things... does that mean you have to program them your self?

btw, you can also program DLLs in an other language for DBpro :)

[gameboy]

Mussi, you probably didn't see this (http://www.truevision3d.com/html/modules.php?op=modload&name=News&file=article&sid=2) page. You can look at it and judge for yourself. Or better yet, download the SDK and try out some of the examples.

I'm hoping this thread won't turn into a flamefest as to which engine is the best, but just a place to provide feedback on engines to help others decide which one is right for them.

DarkBasic Pro is getting better, but IMHO TrueVision3D offers better performance than DarkBasic Pro, in terms of FPS and rendering quality. I have yet to see a demo of DarkBasic Pro that blows away what TrueVision3D can do. Yes, programming skill and knowledge play a major role in this, but my point is that I like the programming IDE toolset available to me in TrueVision3D over that of DarkBasic Pro, at least for the moment. With improvement, DB Pro may someday change my mind. :D

[Mussi]

I'm hoping this thread won't turn into a flamefest as to which engine is the best
did not intend to do that :)

have yet to see a demo of DarkBasic Pro that blows away what TrueVision3D can do
I'm working on it ;)

[anim8]

Me and the rest of the development team just went thru the ordeal of chosing an engine.....
We went with Cipher, and so far everything seems to be going well.
To clarify, you do the modeling and animation in 3D Max (or Maya) and export to .ase files. ASE is not uncommon. The export is the trick. You get vertex colors, etc with .ase... Cipher comes with a tool that is drag and drop to convert the .ase files to .mdl files. Couldnt be easier.
Normally, after you build your models, textures, etc, you export them to ase. Afterwards you build your map in Max/Maya, group your segments (rooms, environments, ect) and export that file (using a script that comes with Cipher) to a .seg file. This file is opened by the engine as a map/level. The groups help with BSP, lighting, etc...
You'll notice there is no game editor. We chose to build one.
The reason we went with cipher is the toolset, the source code and the license. Cipher has shader and particle designers that make the job pretty easy and expandable.
You will need a decent programmer! I lucked up and got two great coders which makes my job easier.
The Cipher site also has good documentation and help from the developer.

I would recommend this engine over A6 (no source code) and TGE(no shader support, and everything looks like a T2 mod..)

BTW- The vision engine people emailed me back...:
"The regular investment for a 18 month, core devlopment cycle with full support is 250,000 $ for a commercial game production heading for a 5-10 million ROI"

:hmm:

Cipher for me!

Hope this helps!

www.fathomthegame.com (http://www.fathomthegame.com)

[gameboy]

Actually, I've found the .ase file format to be rather uncommon. There seems to be little information out there about its file structure.

Unfortunately I'm not using 3ds or Maya (I'm using Lightwave), so I gave up on Cipher because of this. I'm finding that often the workflow between digital content creation and importation of that content into game engines chokes because of incompatible, unsupported, or only partially supported file formats.

I have to admit though, that for the power you get for the price, Cipher (http://www.cipherengine.com/licensing.php) seems to be one of the best options out there.

DarkBasic Pro (http://darkbasicpro.thegamecreators.com/) seems to have among the most features supported at a good price, but I'm not sure that it performs all that fast. Blitz3D (http://www.blitzbasic.com/) seems to run at higher frame rates, but it is only DirectX 7 enabled at the moment as far as I know.

[Mussi]

there is currently a compo running for db users. it ends at 31 december, there will be some stunning entries so why don't you give a look at the games when they are done, I'm sure you'll find it a good tool then. hope I can finish myn in time :)

[anim8]

Even if .ase is uncommon, the final output is .mdl. So regardless of how you get there, ase to mdl, or straight to mdl, you have options. Especially with the new Blizzard tools...:thumbsup:

[gameboy]

ani8, Do these Blizzard tools convert to .ase, or from .ase? Can you give a link for the Blizzard tools?

Thanks. :)

[anim8]

Neither, they convert to .mdl....

http://www.battle.net/war3/files/index.shtml#arttools

[Noxerus]

Of course, you can also always use Deep Exploration to convert between file formats.

[Scottie]

ShaderWorks XT is not a game engine, but may be a GREAT fit for most engines that use effect files. It's a shader development program using a graph-based approach at writing shaders...... it's currently in alpha and free for download

www.ShaderWorks.com

have fun

[Scottie]

Here's an intersting tool...

ShaderWorks XT is not a game engine, but may be a GREAT fit for most engines that use effect files. It's a shader development program using a graph-based approach at writing shaders...... it's currently in alpha and free for download

www.ShaderWorks.com

have fun

[Mussi]

VERY interesting tool, downloading it now, can't wait to experiment with it

[MadMax334]

i have used darkbasic pro...to tell u the truth i find that it gets to slow when u try to do tomany things at once......also its a hassel to get any kind of OOP out of it.......the include file is virtualy useless....u must either compile ur own dll's or us tpc.....i would either stick with truevision or any of the sdk by eclipse (wildtangent for web, genesis3d).....even if they dont have some features......or maybe some that gameboy illisutrated like cipher or cube engine.....but i dont have oney for licenses so i stick to gpl's

[anim8]

One of the code guys (Section8) on my team has teamed up with the code guys from Out of The Box and built an editor for the Cipher engine.:bowdown:
I've been using it for a few weeks now and it will be released to the masses in the next few days if all goes as planned.
This, in addition to the shader and particle editors, makes Cipher a good choice for indie developers. The price is right and its higher quality than alot of the engines out now.

If you have any questions feel free to post them in our forums, the coders will be happy to help you guys out.


FathomTheGame.com (http://www.fathomthegame.com)

[gameboy]

Well, after much searching and some tinkering, I've decided at this time to go with 3D GameStudio (http://www.conitec.net/a4info.htm).

The engine has come a long way since I first came across it a couple of years ago. I think it has now reached a point where some great looking games giving good performance can now be made with it. It has a large support staff backing it, and probably the largest installed base of developers for any indie development system. It is frequently updated to keep up with both bug fixes and changes in state-of-the-art technology. The company behind the technology, Conitec, regularly publishes a forecast (http://www.conitec.net/forecast.htm) for upcoming upgrades to their engine and game development tools, even setting priorities for them and giving their current status, something which is nearly nonexistent from most other organizations.

The engine is specifically designed for making games, so that many features that make game development quicker and easier are built in - this cuts down on development time. Bundled tools ensure you have a way to create content for the game without having to worry about incompatibilities between content creation tools and the game engine (yet it also works with output from other stand-alone tools that may be better suited for their particular task, like 3D modeling and animation). It has a powerful scripting language and support for developing games using other programming languages as well (see their forecast (http://www.conitec.net/forecast.htm) page - this will become even easier in the future). The engine is not only very complete, supporting online play and the latest technology (e.g. DX9), but it can fall back automatically to run in a state which is compatible with older systems that do not support such technology.

I'm not trying to start any flame wars, neither am I saying 3D GameStudio outperforms all other game creation tools in every area -- but for me at least, when you take ALL the criteria I set forth at the beginning of this thread (e.g. overall engine performance, technological feature set, cost, documentation, speed and ease of development, tech support, game and content creation workflow, suitability for a 1-man development team, licensing issues), 3D GameStudio would appear to be at the very least not only an excellent choice, but also one that perhaps best satisfies the requirements initially specified in this thread at this time.

:thumbsup:

[Renderman_XSI]

I'd to add one more game engine to the list. for future reference.

X-engine:

http://www.thexengine.com/index.html

[Tinny_McFinny]

EDITED

[gameboy]

Another one to add to the list:

3Impact 3D Engine (http://www.3impact.com/)

[mastermesh]

http://www.crack.com/golgotha/ Golgatha.... old, but from what I understand is totally free to use and you don't have to reveal your source in commercial use... if you do get it get
http://www.dynabits.com/ivcon/index.htm to get converter... EDIT - URGH, THAT GOLGATHA DOWNLOAD LINK IS NOW DEAD! IT WAS NOT DEAD ABOUT 2 WEEKS AGO - GLAD I DOWNLOADED BEFORE NOW! :)


http://www.3drad.com - this is the brother app to 3Impact...

and http://cg.cs.tu-berlin.de/~ki/engines.html is the oldest and probably best list of Engines that I've come across... OTHER lists do exist like http://fraktali.849pm.com/graphengines.html if you google enough.

[mastermesh]

http://sv3.3dbuzz.com/vbforum/showthread.php?s=&threadid=53435

[mastermesh]

http://66.102.7.104/search?q=cache:O7iealq9x30J:www.crack.com/golgotha/+&hl=en&start=1&ie=UTF-8

[Mastakojo]

Most of the people here probably are programmers. Im an 3d artist and planning to make a game with some people too, does any of these game engines supports normal mapping, good network system and data management? I know AMP2 and tenebrea2 supports normal mapping, but their others features other then graphics seems to be behind....

[Mussi]

DarkBasicPro can handle all of that, check it out atTheGameCreators (http://www.thegamecreators.com/)

[Sashelas]

After reviewing each tool on your list, the best bet seems to the combination of the Torque indie license at $100 US, then the Caligari gamespace pack: http://www.caligari.com/gamespace/ for $300.

The community for support is large enough to flush out most of the bugs you would find in using any product. Having purchased a variety of tools in the past and getting burned, I think the two tools above would give you what you need in order to meet your break even point when you can sell your resulting products.

Custom writing everything yourself is easier in some cases. Use VB, use directX, postpone .NET until everyone has the run time. Find a good free or cheap installer (Clickteam or Nullsoft) and get something out. After you have done one or two round trips with this approach, compare it with the time you would have to spend understanding the eccentricities of someone elses engine.

However there are enough games out there written with Torque, that you should be able to arrive at a quick win with your game with the same effort. If your skills are more design related, you will want a tool with a very strong developer following. At the moment that appears to be Torque. If you have doubts, check out the Radiosity mods from the forums on their site. You'll be pleasantly surprised.

[SuperDre]

Originally posted by gameboy
----------------------------------------------------------
Torque Game Engine [http://www.garagegames.com/pg/product/view.php?id=1]
----------------------------------------------------------

Price:
$100. But if you have a successful game and your revenue is over $500,000, it costs you $10,000 per game (up front!) to develop using their engine and tools.

Pros:
Decent performance. Comes bundled with a set of tools to assist in making games.

Cons:
The tools that come bundled with the engine can be purchased or downloaded on your own. The demo games were OK, but the engine is not as state-of-the-art graphically as other engines.

I don't know where you got the $10,000 but that's not correct, it's only $495 per programmer, just the same as the $100 which is also per programmer and requires you to show at least the Torque logo, and can only be used for games (Yeah like you want to do something else with it.. LOL).. Also they will soon release the Torque Shader Engine.. Which will also cost about the same..
The tools are also free, and there is also a great community..

[operativem]

If you're looking for a real solution, I would make scene graph flexibility an extremely *serious* priority, especially if you want to make a Jak and Daxter-themed title. This is going to save you a lot of time and trouble in the long run. Engines with poorly designed scene graphs (inflexibility is one harbinger of poor design) tend to make development a total nightmare. Look at OpenFlight (http://www.multigen.com) if you want to see a well-implemented scene graph.

OpenFlight, or a similar level of quality and flexbility, is the minimum you should expect. A lot of people and large companies don't really do their homework with respect to scene graph quality when they research solutions. I know of numerous examples at large, very-well known companies, where scene graph problems caused significant problems on high-profile projects. (For reference, I think OpenFlight was the basis for the scenes in both Mario64 and San Francisco Rush 64.)

The only possible downside -- if it's really even a downside -- to very flexible scene graphs is that they can be more work to optimize, and typically, you have to use elision. However, a flexible scene graph gives you lots of control over mesh granularity, which makes it easier to elide the scene based on whatever conditions (distance) you want.

Just my 25 cents.

P.S. I don't work for MultiGen.

[DonMeck]

can you lay out what the criteria are for a good scene-graph design?

Do you mean a render engine with different scene-graph-possibilities (like plug-ins)?

[operativem]

In my last post, I referenced OpenFlight, which is a commercial scene graph made by MultiGen. (I’d like to reiterate: I’m not associated with them in any way.)

When I use the word flexible, I am talking about a combination of extensibility and user control. This subject is probably too complicated for a quick post, but I'll try and cover exactly what I mean.

Extensibility is important for obvious reasons – most applications have custom data types that need to be stored somewhere, many times in the scene graph. If you look at the design of OpenFlight, you’ll probably get a good idea about what I mean. (http://www.openscenegraph.org is another good example; I’m not associated with them either.)

A user-defined scene graph (like OpenFlight or the scene graph offered in Maya) provides many features critical to optimized content creation. For example, object grouping and the ability to control mesh granularity are two of the most important. Here’s an example why:

Let’s say we’re using portals to optimize a scene with a large castle and a large outdoor area. In this case, we’re not using exact portal clipping, we’re just using the portals to cull objects that don’t overlap the portal. For reference, the camera is inside the castle, looking out through a window. The landscape is visible, as are a few stands of trees.

If we have the landscape as a single mesh (say a height field, which is a technique common to many game engines), then we have to render the entire height field, except what’s clipped to the primary and secondary viewports, where primary viewport is the screen and secondary viewport is the portal.

This means, for all intents and purpose, the portal culling has no effect on the landscape, which probably has thousands of polygons, especially if it’s an inefficient height-field. (Instead of a TIN.) Anyway, depending on your implementation of portal technology, this tends to require two passes of database traversal, so the performance cost of the optimization is significant enough that it’s a shame to waste it by getting no optimization on the landscape when it’s viewed through the window.

Without “groups” of trees, each tree has to be tested against the primary and secondary viewports. There could be hundreds or thousands of trees, each requiring per-frame tests for overlap against the primary and secondary viewports. If we have user-defined groups and a high-degree of control over what constitutes a mesh, you can create groups of trees and efficiently remove irrelevant groups of trees from the rendering list. Instead of testing thousands of trees, you might test 10 groups and find that only one group of trees overlaps the primary and secondary viewports.

Here's a scene graph with 1000 trees, I've omitted most of them for obvious reasons.

- Scene
-- Landscape
-- Tree 1
-- Tree 500
-- Tree 1000

Would you rather test 1000 trees or test 10 groups of trees when half the groups aren't even onscreen? Does the scene graph of your engine allow this?

Going a step further, user-controlled mesh granularity allows us to separate the terrain into groups and meshes, which can be efficiently culled, creating an even more optimal rendering list. Thusly, only the portions of the landscape that are actually visible through the window are rendered. There will be some overdraw, but you can greatly limit your costs in that respect.

An example of an efficient scene graph:

- Scene
- Group 1
-- Landscape Mesh Quadrant 1
-- Tree Group 1
-- Tree Group 2
-- Tree Group 3
-- Tree Group 4
- Group 2
-- Landscape Mesh Quadrant 2
-- Tree Group 1
-- Tree Group 2
-- Tree Group 3
-- Tree Group 4
- Group 3
-- Landscape Mesh Quadrant 3
-- Tree Group 1
-- Tree Group 2
-- Tree Group 3
-- Tree Group 4
- Group 4
-- Landscape Mesh Quadrant 4
-- Tree Group 1
-- Tree Group 2
-- Tree Group 3
-- Tree Group 4

Again, does the scene graph of your engine support the structure shown above? Or is it really just a naive list of objects that have to be culled each time the renderer draws a frame?

I don’t think these types of considerations are frequently made when people evaluate engine technologies. (I have yet to see one question about scene graph quality in this thread, except for the one I raised.)

Not enough emphasis is placed on the scene graph because people tend to be concerned with rendering quality and performance. However, in many ways rendering performance is directly and indirectly tied to the scene graph, so I find it unusual that more thought would not be given to this area.

It is simply a fact that even if an engine has a super-efficient renderer, a comparably inefficient rendering list will quickly overwhelm and make useless any and all rendering optimizations you can implement. If you have little or no control over how the scene is constructed, then you have little or not control over how the scene is culled. The result of that is that you have little or no control over the rendering list, and are essentially stuck with how your scene performs inside whatever scene graph the engine designer has created. Of course, you can always re-engineer major portions of the product, but then what's the point of licensing to save time and money?

This is the root of complaints about engines being suitable for some types of content but not for others. (i.e. using a shooter engine for racing game or platformer.) It's also the reason that many companies spend thousands of extra development dollars trying to work around the performance problems they encounter when they try and fit their content into a scene graph that was poorly engineered.

The task of optimizing a graphics engine is rarely, IF EVER, solved by a single magic bullet. It's a process that requires attention to the structure of the scene/database and at all points along the geometry pipeline. No use feeding bloated mesh soup to your geometry pipeline, that won't do much good. (This is something that many engine programmers know, but they often don't have time to focus properly on this because they're too busy with other things.)

Developers should be concerned with how much control their artists have over the efficiency of the rendering list, because it saves time and money. I really doesn't cost much more to create a totally efficient scene from the ground up.

This explanation would be better with diagrams but I just don’t have the time.

[DonMeck]

Thank you for clearig this up. I read that in ORGE the scenegraph is acutally a plug-in or so - so it can be replaced easily.

The engine I am using allows groupings, hierarchies, places and portals. This allows the user to organize the scene to unload the culling-process.

[Antimode Siker]

Originally posted by gameboy
Regarding the LGPL license, the problem I see with it is that you must allow users to be able to change the public code covered under the LGPL license and then recompile it with the rest of your application, and then run the modified application. This has at least two drawbacks; one is that unless you provide some sort of SDK for your application, you must "open" it up somewhat so they can recompile it. The second drawback is that users would be able to create their "customized" version of your game, and if you want players to be able to compete against one another online, these customized versions of the game would be equivalent to "cheating" (unless the point of the game was to allow users to compete against one another using their own customized modifications).

I am not sure how you mean with having to open the application up. You would have to make available the unlinked objects and a make file such that the project could be relinked with another version of the LGPL licensed objects. This is similiar to how the dotXSI plugin for Maya comes in unlinked object files.

As far as the second point goes though, I think it is important to realize that security by obscurity is not. If the concern is to prevent 'cheating' by disallowing people from changing anything in the game executable, you would have an enormous challenge ahead because you would be unable to even use DirectX or OpenGL, both which rely on DLLs and vendor provided drivers. To retain compatibility with many different computers, it is probably near impossible to lock these from being changed.

In fact, if you use any kind of dynamic standard libraries, they can and will be changed depending on the user's system - compare what DLLs are available on Windows 98 vs Windows XP for example.

So unless you wish to implement your own operating system for the game to run on, you are stuck with changeable parts and a "customizable" game wether you use LGPL software or not, I believe.

More importantly, to prevent cheating, entirely different measures are necessary than statically linking the executable. These might include internal software encryption and CRC checksums compared to an internet database for online games. These measures will not stop a professional but can deter most casual cheaters.

Using or not using OGRE on the other hand probably has very little effect on determent if these meassures are in place.

[arli]

Hi..

Iam one of the developers over at TV3D (Truevision3D, LLC.).
I just saw this article and thought it might be of intrest to get some updated information on the TV3D SDK.

Some of you might have already noticed we have a new and redesigned webpage now, http://www.truevision3d.com , so some of the links that point to us in the posts from here might come out with a 404 error.

By visiting our new webpage you are able to see that we are reaiming our goals by moving our position in the 3D Engine market upwards.

We are atm working on the TV3D SDK 6.5, which will include many many new and exciting features. All of which can be expected from the big names such as HL2, Doom3 etc.

If you read the info on 6.5 on our webpage you see we have now developed plugins for the diffrent modelling packages to have support for our own formats (which are introduced in 6.5). This means smooth integration between maya, max, gamespace etc. right into the tv3d-project you are working on.

With the new 6.5 SDK we are totally redesigning the SDK aswell, some of which can be expected is new tools. ShaderBoy (a pixel- vertex- shader editor), ModelView (a Model Viewer/Editor/Processor .. assign effects to your exported models) and totally new samples which will introduce you to all of these new features.

All of the engines are being rewritten from scratch with a very optimized layout and multilanguage support (ATL/COM, .NET, Direct Library Support (C++)).

Very exciting stuff to look forward to, we are currently in the 6.5 Beta stadium. We have a closed beta atm, and we are progressing fast. We update the page pretty often with news on our progress and a development video.

I hope this brings us a bit up2date with this post, and keep looking back into our page, Im currently working on a video that will demonstrate how easy it is to get your model from your modelling package into your game with full perpixel light and bumpmapping using normalmaps (generated from a highpoly model, this is done in ModelView) and with Vertex- Pixel- Shader effects such as Glow effects etc.

PS: Heres a lil screenshot of a model in ModelView that is around 9000 poly, the normalmap is generated of a highpoly model that is at around 600 000 poly, this is all applied with a smooth sense of specular. And this is rendering very fast, due to its real polycount is only 9000!

EDIT: Our artist had an updated and nicer model so i updated the image on the site!

http://www.truevision3d.com/images/normalmapSS.jpg

Ah and also, anyone who has the 6.x license has of course automatically the 6.5 license as well, so the prices arent changing with the new SDK!

Have a nice day everybody!

Cheers.

- Arli Mujkic
Senior Developer
http://www.truevision3d.com

[SuperDre]

Oh damn, that's looking rather nice.. Looking forward on trying the new TV3D 6.5 engine out..

[ICO88]

Out of all the engines that i would have used I found it was a toss up between virtools and Quest 3d. Virtools was way too expensive for me. Quest Graphics wise is the equal to virtools. The problem is that it is pretty low level and there are not enough game tutorials. That being said Im able to do things in quest easily that took me tons of time in gamestudio. If you keep doing the quest tutorials then Quest just clicks. The core engine is good, its just those fools at quest need to make more templates and tutorials for game makers. then quest will surpass virtools.

[DonMeck]

Well, not sure if it is all that it takes.
You got scripting language (VSL) and now you can also assign those scripts to some keys, so you can speed up your workflow.

Additional the SDK has now an inerface for integrating the source/media control solution you want to use. Alien nxn interface - or so - is availble by default.

They also allow now, to use their UI within the SDK - allow to stay in look&n feel when adding new plug-ins.

Moreover, the channel thing is a really basic concezpt in terms of information-organisation.
Virtool's differentiation into BuildingBlock and Parameter, ParameterOperation and so on, gives a higer order of information. Kommunikating the concept faster to the developer.

But well - this is my personl impression from only looking very shortly into quest3D. No doubt that good things can be done and I am hearing good things. Price is really good. Only for Virtools users it is hard to go to quest3d, because the now know the other concept.

[The Weed Killer]

I tried the truevision3d engine out about a year ago and wanted to see how they have progressed since then but I can't get to their site, anyone else having this problem?

[ICO88]

You can deal with things in a higher order in quest by using the templates. But youre right virtools users would not find any real reason to switch. Either way for the price Quest wins overall maybe Virtools edges out quest. They are both on the top.

[struve]

Wow, this thread is very handy. Thanks guys!

[raven3000]

What engine would you recommend for a team trying to make a survivor horror game that take place in feudal japan.The engine that we're looking for needs to be able to produce good visual,since we want the environment of the game to look real.
Also,the engine needs to be able to handle multiple NPC without slowing down.Plus,good collision detection and works with software like 3D studio max

[DonMeck]

ICO88,
I should take a look onto the template thing.

raven3000,
depends on many other facts like: what is your budget for the engine, do you have programmers to extend/cutomize it to your needs ... ? Then it goes into technically stuff like: it should have p/v shader support, AI, character engine etc. Even this can be break down. So, what do you have, what do you need. Then responses can be more accurate.

[Tocpe]

Do any of these programs allow you to develop games for both the PC and Mac? Or how do you even do that?

[raven3000]

what is your budget for the engine:600-700 dollars
do you have programmers to extend/cutomize it to your needs:Yes,we have 2

We basically need an engine that has accurate AI control to make it easier for to make the enemies and NPC act more humanlike.Also,it being compatible with Visual C++ 6 is a plus.

[gameboy]

An engine which I've mentioned I've liked before in this thread isTrueVision3D (http://www.truevision3d.com/home.php). One of their programmers, Arli, has posted an update about it above. I've found TrueVision3D easy to use, and it supports the use of various programming languages in addition to C++ like Borland's Delphi, or any of the .NET compatible languages. In some preliminary testing using this engine I was able to get frame rates in the thousands of fps for simple graphics, and some of the more sophisticated demos they've come out with have had frame rates around 100 fps on a system with an Athlon XP 2800+ and Radeon 9700 Pro. The version 6.5 update of this engine should be quite significant, and is something to keep an eye on.

[Sashelas]

Operativem has an very important point. Poor scene graph management breaks most engines that attempt to provide a degree of programmatic control and allow loading of large numbers of objects.

**Has anyone attempted to implement a scene graph inside an RDMS?

A pre-calculated scene graph cache on clients that is refreshed from a b-tree indexed set of related tables might work. The cache would be pushed down to the client on login. The list of quadrants and groups to render could be sent downstream to the client based on the result of a heartbeat driven procedure call.

I thought maybe PlayNC was going down this path after I read some of their developer's writings.

--

I've been doing engine reviews again for a new project I'm on that has a 'game-like' need for a high performance graphics engine. The engine would need to load scene graph elements from a large dataset.

Here are my current requirements for the engine:
1) .NET support a must since most of the code on a current project is C#
-I plan to interface with my C# code which periodically retrieves datasets from six SQL Server Wintel boxes over the internet at anywhere from 28.8 - T1 speed.
My data size is fairly large - 330 million records, 900 GB.
3D scene graph data and business record data is retrieved to cache on remote client machines. I would write a C++ plugin or wrapper to do this if necessary.

2) Some way to hit remote RDMS over TCP/IP for ACID transactions. UDP, Directplay or custom multiplayer code could also be used but not for required save points and transaction management.

3) Cost under $500. If it is much more would use DirectX w/o an add'l engine

4) Working samples on install. If bugs occur off the bat, forget it.

5) Sizable forum community or diverse product offering
I've been burned by the folding of engines in the past (Intel 3dr for example)

6) Easy import/export of digital assets. I use Photoshop, Caligari, Bryce, 3DSMax, and some in-house tools that store and interchange data using XML and BLOB fields in SQL Server.

7) Windows 2003 compatible with an array of video cards. XP Home & Pro compatible with same. I have ten machines that represent different combinations. Some engines have performance differences or strange crashes on 2003 and I'm not sure why. Some engines also run faster on 2003 surprisingly if given ample RAM.

--


Here are my recent test results looking for an engine that fits the requirements:

Quest3d Lite
-$89.00
http://www.quest3d.com/index.php?id=8
-House demo failed under Win 2003
-Rendering quality is fantastic
-Fur and crowd demos are amazing
-Useful but overpriced add-ons. The multi screen flight demo is wild.
-$2000 for Enterprise which includes all add-ons. Still not very programmable
-Poor RDMS support - better off hitting RDBMS (OLEDB, ADO.NET) than using ODBC interface.
-This could be the best if I can create an interface layer or wrapper. Check out Monster Mayhem demo.


Truevision3D
-$150.00 single product, $500 multi product
http://www.truevision3d.com
-Flexible. I write alot of code so this appeals to me.
-Visual Basic.Net, C# support
-Unresolved bugs in Visual Studio .NET 2003 such as crash on close
-Graphics are a lacking due to lagging DirectX support.
-New version coming soon which supports DirectX9.
-Spooky and dated humankind demo
-Mesh soup. Performance problems with lots of objects at 1920x1440.
-Might be easier to use C#/DirectX tutorial code from DirectX SDK

Torque Game Engine SDK "Indie" License
-$100.00
http://www.garagegames.com/pg/product/view.php?id=1
-No .NET support. Would need to write a wrapper object in C++.
-Nice editing environment for menus. Created a usable menu in 15 minutes.
-Great demos. Very playable.
-Problems connecting to master server from Windows 2003. Worked from XP.


gameSpace1.5 (best tool add-on)
-$299
-MAP export
-WAD import/export
-Have purchased Truespace several times in the past and used it with DirectX

Gamestudio / A6
-$199.
http://conitec.net/a4info.htm
-mp3 and mpg playback are nice to haves
-4 players is limited in $199 version
-CScript, Would need to write a wrapper object in C++.
-DLL interface, COM in future (.NET integration looks somewhat difficult)
-Pyrotechnics demo was nice.
-Great plug in support.

**I notice one Gamestudio plugin from Mythic Entertainment. Does anyone know if Dark Age of Camelot uses this engine? Do Mythic's other games use it?

3dlevel
-Free, VB.NET based
http://www.3dlevel.com/index.php
-Might be easier to use C#/DirectX tutorial code from DirectX SDK
-Still evaluating

Cipher
-$100
http://www.cipherengine.com/licensing.php
Very nice demo graphics quality and speed
-Still evaluating to decide if useful

DarkBasic
-$90
http://darkbasicpro.thegamecreators.com/
-Cheap. Lots of Samples.
-Poor language support and integration capabilities.
-Difficult to integrate RDMS and multiplayer networking code.
-Still evaluating to decide if learning Darkbasic would be a waste of effort


3Impact
-$100
http://www.3impact.com/
-Simple C++ (very elegant)
-Decent performance. Nice graphics.
-Small following.
-Write own C++ multiplayer code.
-Still evaluating

-Sash

[DanSilverman]

For the record, I use 3D GameStudio and have been doing so for years ... since version 3.9 (and we are now on version 6.31). I have also used (though to a lesser degree) several other engines and have licenses to Torque, Cipher and DarkBASIC Pro. I have a healthy respect for all of these engines (including the ones that I do not currently have a license to or do not use).

When evaluating an engine there are so many things to consider. I have seen people "slam" one engine or another and the fault may not have been the engine at all. Instead, the fault may have been that the end user simply chose the wrong engine for the project. For example, if I choose an engine that is primarily an in-door engine and attempt to make an expansive terrain-based game with it ... well ... then I am in for trouble. This seems to happen quite a bit and is, in my oppinion, a lot of the reason for negative input from people on many of these game engine's boards. So matching a game engine to your needs and the needs of your project is paramount.

I am not a programmer, but a CG artist. As such, using an engine that requires programming is not something I can do if I am to work alone. I have licenses to Torque, Cipher and DarkBASIC, all of which require programming to use, because of projects I have had to work on in the past and because of possible future projects. But, for the most part, my business is made from 3D GameStudio.

With 3D GameStudio I can create a complete level and have it up and running with absolutely no programming at all. I don't even have to tweak a script or any of that. I simply build the level, place my entities, build the BSP and go. I can also use the new scripts (and tweak them without ever once touching code) to do basic things like player movement, doors opening and closing, create lifts that will carry the player or an object, create basic enemy AI, etc. And all this without touching code (and be able to custimize it as well ... either without touching code or by creating custom code).

The scripting language, called C-Script, is certainly advancing. It is going to be changing a bit more and become more like C. But if someone does not like the scripting langauge then they can program with C, C++ and Delphi using the SDK. So there is flexability there for both the non-programmer and the programmer.

The engine is currently BSP based, which some people do not like nor do they tend to understand. BSP is restrictive when it comes to the types of geometry you can use (no concave surfaces) but it makes for very fast frame rates. Even so, the company that makes GameStudio is planning to implement Octrees in the near future. The GameStudio user will have the choice of using either BSP or Octrees for their level. Once Octrees are in place, then arbitrary geometry can be used for level geometry.

One of the best things about GameStudio is it comes with a suite of tools. While MED, the Model Editor, is out-dated, at least there is a modeler there to use. Its main purpose is really for converting models from other programs over to GameStudios' own MDL format. The other tools are great. WED, the World Editor, is a lot like Hammer/Worldcraft, and is specifically designed for BSP level creation. There are also plugins that allow you to use 3D Studio MAX to create both models and level geometry as well as for Lightwave. And WED will also take MAP files. There is also a script editor called SED so you have an IDE for coding as well. Other tools are in the words such as a shader designer to help the non-programmer (as well as the programmer) create custom shader effects for their games.

One of the nice things about the tools is that you don't have to look for tools to use and you don't have to worry about a restrictive EULA. For example, the EULA for Hammer/Worldcraft does not allow you to use Hammer for commercial games. WED can be used with no restrictions what-so-ever (as with the other tools).

GameStudio is also advancing. Some would say that it is advancing a bit more slowly than other engines in its class and, to a degree, I would have to agree. But there is a difference. Many of these other engines seem to release updates with fairly fresh code and, while this allows the user base to get updates and features more quickly, the users tend to become a beta testers group and encounter bugs and instabilities. This often can lead to user frustration. The programmers for 3D GameStudio have their own beta test team and all features go through various iterations and tests before going to a public beta (and all users are warned that the release is a beta). Once the official release is out, it is fairly bug-free and extremely stable. So, development with current releases is a dream. I rarely have problems using a current release of GameStudio in my projects and I use GameStudio professionally.

Some have stated the graphical output of GameStudio is a bit low. I would disagree and agree :) . The graphical display of GameStudio is bright, crisp and accurate, not muddy as I have witnessed with some other engines. On the other hand, only recently did we get 24-bit textures for level geometry and shaders. So, the graphical output has increased. The weakest area graphically, in my oppinion, is the lightmapping tool. While I like WED because it will automatically generate your lightmaps, you currently have no control over the quality or size of the lightmap. The result can be banding from time to time if things are not well thought out. I believe that Conitec is planning to work on this in the future, but it is not even on their forecast list at the moment.

There is a lot more that I could say, but this post is long enough as it is. I will say this, though: I am using GameStudio as the core of my business and have been doing so for a few years. Therefore, GameStudio can certainly be used professionally and holds up in a professional environment.

[spakman]

I haven't seen this solution mentioned yet. http://www.renderware.com.

I've worked with this engine personally. It's being used on a game I worked on that's due out this Christmas. I've seen some good results and some surprising results, and was wondering what others thought.

Thanks!

[Tocpe]

I haven't seen this solution mentioned yet. http://www.renderware.com.

Hmmm....maybe it's just the lack of sleep, but I couldn't find a price on their site. :shrug:

...course, then again, like they say: "If you gotta ask how much it costs, you probably can't afford it".

[Oblivionbringa]

This might be of interest..

www.artificialstudios.com (http://www.artificialstudios.com)

http://www.devmaster.net/engines/engine_details.php?id=26

http://artificialstudios.com/licensing.php

[gordon7up]

just spent the last hour or so reading this thread, very informative, I am a Maya user, I seen a few references to maya in this thread, is there any engine which works particularly well with maya models, textures, maps etc or are they much the same??

thanks,
Gordon

[spakman]

Anybody have any personal inisight on Microsoft's new XNA standard? (I'm just reading throught this link (http://www.microsoft.com/xna/) right now).

peace d=^)

[gameboy]

One concern about XNA that may portend it being another attempted standard that bites the dust is that Microsoft says the DirectX API will continue to be a baseline development technology for the standard. My understanding is that DirectX 9 will be the last iteration of DirectX, to be replaced by a newer API (Avalon) once Windows (Longhorn) is shipped, although it sounds like from this (http://www.pcmag.com/article2/0,1759,1586741,00.asp) article that the existing DirectX may continue to be a subset of Avalon's graphics technology:


Avalon will be the replacement for the existing Windows desktop graphics technologies, and is built on XML and DirectX. Avalon itself is an API framework, much like the GDI interface on which the current Windows desktops are built.

[spakman]

*bump*

(mods, u guys should prolly make this thread a sticky)

Thanks gameboy!

peace d=^)

[spakman]

just spent the last hour or so reading this thread, very informative, I am a Maya user, I seen a few references to maya in this thread, is there any engine which works particularly well with maya models, textures, maps etc or are they much the same??

thanks,
GordonI'd wager they're all much the same. I've only worked at one game company that didn't use Maya. (At that place we used SoftImage 3.8). Maya has all sorts of things to make game assets engine friendly, like individual f-curve output control for example. (I only bring that up, because I've heard horror stories about Max and f-curve data wrangling - don't know much about Max personally...)

peace d=^)

[EVIL]

This might be of interest..

www.artificialstudios.com (http://www.artificialstudios.com)

http://www.devmaster.net/engines/engine_details.php?id=26

http://artificialstudios.com/licensing.php
I am constantly visiting their website if there are new stuff or anything about a licencing price

[DevilHacker]

I have always looked at DarkBasic, but it seems that everyone is bashing it...
Why? I really want to know why I should not get it; it looks to be very good?
Any one work with it first hand?
Also, I am not trying to start a flame war, just find out more info, everyone seems to just be saying its not good... Please explain.
:shrug:

[gameboy]

Well, in the past, they've had a poor reputation for producing buggy software (and somewhat slow - although I've seen some efforts that used it produce high frame rates - at least with DarkBasic Pro). I got rather fed up when the debugger kept crashing on me! :rolleyes: In times past, fixing errors wasn't a high priority on their list. It seems they would rather rush to put out the next great new feature of DirectX 9, etc. but would tend to leave many previous bugs unfixed.

However, a few months ago they added some people and finally took some time to fix outstanding bugs. Judging from some of the comments their developers posted in their "diaries", they seemed surprised the code was so buggy. Perhaps now they've learned from their mistakes and are putting out better tools. :D Their upcoming FPS Creator (http://www.thegamecreators.com/data/newsletter/newsletter_issue_22.html) product does look rather promising. Check out the videos on it there (http://www.thegamecreators.com/data/newsletter/newsletter_issue_22.html) also.

[DevilHacker]

Any word if there will be another update to the DarkBasic engine...
Or is it stopping at Pro right now?
-----------------------------------------------------------------
Also, has anyone used the Nebula Device (http://nebuladevice.cubik.org/) engine?
It looks really good… :shrug:

[gameboy]

The 3Impact engine has recently released a new version, 3.7, which now supports bump-mapping and per-pixel lighting, in addition to its already existing features such as car and ragdoll physics, volumetric shadows, reflective surfaces, alpha transparency, network gaming, etc. This engine is coming along quite nicely, and I've yet to come across another engine which is as elegant or quite as easy to use. It also doesn't cost a lot. It currently supports development using C++, Delphi, and IBasic. (A port to BlitzMax (http://www.blitzbasic.com/logs/userlog.php?user=1&log=1) would be great when it is released for the PC).

The 3Impact engine is definitely worth checking out! You can view information on its features and download a demo of it from here (http://www.3impact.com/). :thumbsup:

@DevilHacker: The creators of DarkBasic Pro have released a new SDK that allows you to program the game engine behind DarkBasic Pro with another language like C++, and since they say it is also compatible with Visual Studio .NET 2003, it should be possible to make it work with any .NET compatible language as well. Check it out in their latest newsletter here (http://www.thegamecreators.com/data/newsletter/newsletter_issue_23.html).

[pixeloddity]

how did you even get the thing working.. we've almost given up cause our programmer is trying to just get it started but has like 90+ errors. Maybe theres a simple setup out there, and we'd like to use Truevision as its powerful.... but im leanin toward Torque unless we get it going :(

[gameboy]

3Impact comes with several tutorial code files, and searchable reference files for programming 3Impact in several languages (C++, Delphi, IBasic). I suggest your programmer start out by playing around with modifying the demo code that comes with it. One thing to keep in mind though; the prefix name of the .dll file you produce should match the name of the 3Impact executable. For example, if you are running the 3Impact debug version (3Impactd.exe) the name of the .dll file should be 3Impactd.dll. When you're ready to run the release version of 3Impact (3Impact.exe), the .dll name should be 3Impact.dll. :)

[PeterTable]

This is a very informative thread!!
About 3 weeks I bought Indie Torque engine and returned it becouse if you want to get the juice you have to program with their scripts and buy the other plugins they sell.

Now I am deciding between Cipher and 3impact.
Both have almost same feature levels, are in the same budget and allow me to do all in C++

I hadn´t bought either of them, which do you prefer?

I plan to work on a persistent world 3D space RTS game and a persistent world multiuser 3rd person RPG.

[Strang]

I skimmed this thread to make sure. I didn't see NeL engine up here so here it goes

http://www.nevrax.org/tiki-index.php?page=NeL

NeL developed by nevrax is a mmorpg based engine, although its open to any interpretation.

NeL is designed around a typical client / server setup

it was used to develop The Saga Of Ryzom (http://www.ryzom.com)

It is developed for a GNU General Public License...

I haven't seen many MMO engine solutions so I think this is something good to post about.

BTW.. if you make it to the site.. go down to the bottom right and vote for more 3dsmax examples for water/tree/sky/skin

getting content into the engine is still a task because of lack of examples.

thanks

-Steven

[PeterTable]

Nel gots a lot of features but its GPL licensed.
In short words our game´s code must be released to the public, thus allowing skilled people to develop network "pluggins" or 3rd party tools that would help others cheat.

I know that some skilled persons could reverse engineer any network code but opensource GPL licenses serve the network protocols on a silver platter to malicious hackers.

I´d better use RakNet (rakkarsoft.com) for networking, it has lots of features only rivaled by Torque Networking Library and other higher priced libs.

[Strang]

I am glad you mentioned this...

I hang out on the forums and irc channel and have heard of going commerical with NeL.. so it would still be open source, but if you purchase you would then be protecting your intelecutal property. thus making your Network code that is based on NeL's but obviously for hacking/cheating reasons you would refine and rework to fit your need.

I hope this will also bring more people to NeL, now that it could be of interest to larger developers

stay tuned on

www.nevrax.org

[raimo]

Awesome thread!

I have done quite a lot of research conserning different game engines in the past few days. I don't have any formal game developing experince (At the moment