Home » JAGGED ALLIANCE 3 » JA 3 Wish List » Engine discussion for next gen JA based mod/community game
|
|
| Re: Engine discussion for next gen JA based mod/community game[message #173433]
|
Mon, 28 January 2008 21:32  
|
|
 zed |
 |
Messages:42
Registered:November 2006 |
|
|
hi,
i did a little comparison between crytek 1.xxx OGRE and Irrlicht based on what crytek has to offer.
i deliberetly didn't add more engines since this ones are the most common (that i know of)
there are other engines such the one of the open source UFO game , fdd engine and more but these are not 3D and don't necesserly fits our needs.
you can see that the big difference is that all the features are under the same framwork in the case of Crytek. i think that if we will aim to a strong game editor we could make it work (even though this is a very challenging task to accomplish)
so here is the comparison:
1.CRYTEK - real-time game editor
OGRE - No
Irrlicht - irrEdit, 3D World editor
2. CRYTEK - Renderer:
integrates indoor and outdoor technology seamlessly.
Offers rendering support for OpenGL & DirectX 8/9.
OGRE - Yes
Irrlicht - Yes
3. CRYTEK Physics System: supports character inverse kinematics. vehicles, rigid bodies, liquid, rag doll, cloth and soft body effects.The system is integrated with the game and tools.
OGRE - has integration with several engines
Irrlich - has integration with ODE
3. CRYTEK Character Inverse Kinematics & Animation Blending: Allows a character model to have multiple animations while looking believable.
OGRE - skeletal morph and pos animation
Irrlich - skeletal and morph animation
4. AI System:
Enables team based AI and AI behaviors defined by scripts. Ability to create custom enemies and behaviors without touching the C++ code.
OGRE - No
Irrlich - No
5. Interactive Dynamic Music System: Tracks and responds to the player's actions and situations and offers CD Quality playback
in full 5.1 surround sound.
OGRE - has integration with other tools
Irrlich - irrKlang
6. Environmental Audio & SFS Engine: Ability to accurately reproduce sounds from nature with seamless blending between environments and interior/exterior locations in 5.1 audio. Includes EAX 2.0 audio support.
OGRE - didn't find matching tool
Irrlich - didn't find matching tool
7. Network Client and Server System: Manages all network connections for the multiplayer mode. It is a low-latency network system based on client/server architecture.
OGRE - has integration with several frameworks
Irrlich - has interation demonstation with RakNet
8. Shaders: A script system used to combine textures in different ways to produce visual effects.
Supports real time per-pixel lighting, bumpy reflections, refractions, volumetric glow effects, animated textures, transparent computer displays, windows, bullet holes, and shinny surfaces.
OGRE - has built in support
Irrlich - has built in support
9. Terrain:
Uses an advanced heightmap system and polygon reduction to create massive, realistic environments.
The view distance can be up to 2km when converted from game units.
OGRE - has built in support
Irrlich - has built in support
10. Lighting and Shadows: A combination of precalculated, real time shadows, stencil shadows and
lightmaps to produce a dynamic environment. Includes high-resolution, correct perspective,and volumetric smooth-shadow implementations for dramatic and realistic indoor shadowing. Supports advanced particles technology and any kind of volumetric lighting effects on particles.
OGRE - has build in support
irrlich - has built in support
11. Fog: Includes volumetric, layer and view distance fogging to enhance atmosphere and tension.
OGRE - didn't found specific support
irrlich - has support for FOG
12. Tools Integration: Objects and buildings created using 3ds max
[Updated on: Mon, 28 January 2008 21:52] by Moderator Report message to a moderator
|
Corporal
|
|
|
|
|
|
|
|
|
|
|
|
| Re: Engine discussion for next gen JA based mod/community game[message #173470]
|
Tue, 29 January 2008 00:33
|
|
| afp |
|
Messages:75
Registered:November 2007 |
|
|
Can anyone bring here some titles of the games made by fans? One of them would be the UFO variant brought here last week. Are there others too?
There is a big difference between a total conversion mod and a full game made from scratch, even if some mods took years of development. Maybe it worth analysing the faith and the success of that UFO game.
JA3, Hired Guns, 7.62 are more or less based on the guidelines proposed. Each one of these took 2-3 years of development or even more and they were made by professional hired teams. It may be hard to keep a freelancer team for so long.
So I guess it's not about details, but about long term scheduling.
Report message to a moderator
|
Corporal
|
|
|
|
|
|
| Re: Engine discussion for next gen JA based mod/community game[message #173505]
|
Tue, 29 January 2008 07:02
|
|
| afp |
|
Messages:75
Registered:November 2007 |
|
|
I think it's gonna take much longer than 2-3 years, those team are working at least 8 hours / day which don't ussually happen if you work extra time.
Silent Storm engine also may be available out there, the destructive environment was great, plus it fits a turn based game.
Report message to a moderator
|
Corporal
|
|
|
|
|
|
| Re: Engine discussion for next gen JA based mod/community game[message #173523]
|
Tue, 29 January 2008 10:06
|
|
| Ar1z |
 |
Messages:33
Registered:January 2008
Location: Greece |
|
|
I've tried to mod ufo:ai in the past, to be more like ja2... I didn't modify any c++ code though. It was previous version (2.1). Although things changed since that version I have the following observations to make...
Besides that Ufo:ai is the same type of game like JA2 (in a different setting though - future), It has the following limitations:
No doors
No NPC interaction
No RPCs
NO IMP
NO websites
Different GUI
Cumbersome weapons system (at version 2.1 it was difficult to add weapons)
No vehicles (well it has airplanes)
No Arulco map (it has earth map though)
and the following advantages
3D engine
3D models (most human models are already done, no real weapons though)
support for wearing armour
Multilevel buildings!
It's OPEN SOURCE!
[Updated on: Tue, 29 January 2008 10:19] by Moderator Report message to a moderator
|
Private 1st Class
|
|
|
|
|
|
|
|
| Re: Engine discussion for next gen JA based mod/community game[message #173559]
|
Tue, 29 January 2008 14:32
|
|
| rasz |
|
Messages:38
Registered:August 2006 |
|
|
>you won't get support in new format such as .png.
>the graphic system in JA2 was written directly over the DirectShow function with everything made
>up from scrach so every new feature you want to add in the graphic level is very hard to do.
thats not true, just look at linux port or straciatella, running under SDL and supporting ton of new file formats (thru libSDL)
there is also a http://cubeengine.com/ with real time map editor and good documentation
[Updated on: Tue, 29 January 2008 14:34] by Moderator Report message to a moderator
|
Private 1st Class
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| Re: Engine discussion for next gen JA based mod/community game[message #173987]
|
Thu, 31 January 2008 21:07
|
|
| Madd_Mugsy |
 |
Messages:634
Registered:July 2005
Location: Canada |
|
|
Quote:6. can be adapted to some of today's code?
I don't think this is a reasonable goal, since the existing JA2 code is mostly crap. While it may worthwhile to keep some of the calculations intact, many of the algorithms and most of the code's structure should be jettisoned.
The biggest problems with the current code are that it's not extensible, scalable or moddable.
There are hard coded limits on everything. Our current solution is to try to externalize a lot of the values in the code. Prior to 1.13 there was a hard limit of 350 items and weapons only worked in the first 69 slots. I changed the limit to 5000 and externalized most of the item properties, but there are still many other limitations to overcome. Adding Mercs is still a dream, and creating larger maps (both tactical and strategic) is probably still just as a far away.
This is part of the reason why the code is so ugly. We're hacking at the existing system trying to make it do stuff it was never intended to do. Of course, it doesn't help that it was all written in C (not C++) and it was butt ugly to start 
We need to make sure we don't impose these kinds of limitations on the new system, and the first step towards that goal is to severely limit reuse of the JA2 code. I'd also argue that the xml file structure will have to change to accommodate the new system, so a lot of that work will need to be redone. Hopefully we can just run them through an XSLT, but at this stage it's all conjecture.
Anyway, I think that regardless of the base engine chosen, there's going to be a ton of work to make it into anything even resembling Jagged Alliance. Perhaps our best bet is to wait for JA3 and see what kind of mess they make of it before starting a new project?
Report message to a moderator
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| Re: Engine discussion for next gen JA based mod/community game[message #174805]
|
Tue, 05 February 2008 21:41
|
|
| zed |
|
Messages:42
Registered:November 2006 |
|
|
ok,
i have looked (only viewed the code not even played the game) the render of the ufo:ai.
they are using a propritery engine based on quake II engine they have created over SDL.
it is supporiting sprites (don't now to what level), basic animation (didn't figure what type of animations), the are working with md2 md3 models (2d model and 3d model maybe?), lights, shadows, practicals.
very little documentation (doxy gen is not realy documentation...) , very little comments inside the code. using their own scripting format. active forum.
don't know how is the game experience itself. the code is all in C code.
i think for future engine it has to support as much common tools and formats as possible to maximize the amount of contributers.
BTW,
searched allot for free game editors (to has the same game design experience as the FarCry editor) and beside Irrlicht editor
http://www.ambiera.com/irredit/
i didn't found any more - this editor is not open source but is can be used freely and can be extendent using API so we can add to it additional content (AI for example)
[Updated on: Tue, 05 February 2008 21:43] by Moderator Report message to a moderator
|
Corporal
|
|
|
|
|
|
|
|
|
|
| Re: Engine discussion for next gen JA based mod/community game[message #179514]
|
Sun, 30 March 2008 06:08
|
|
| Kaerar |
 |
Messages:2021
Registered:January 2003
Location: Australia :D |
|
|
BirdFlusome thoughts and questions about the 3D engine
1. considering graphics ONLY, what would or should be the difference to, let's say, Brigade E5?
2. what should be the minimum requirements for a possible 3D engine?
3. should the actual game (mapwise) be in 3D or just a subset of it or should it be a whole new mapset?
-> if it is the whole game, then it can't be fully released until ALL maps are converted
4. if there is a 3D engine, then fixing the camera for isometric view is quite simple. It is actually
much simpler than creating an interface for arbitrary viewing directions.
5. don't underestimate the memory requirements for large 3D environments, especially as memory on
graphics board is quite limited (alse see point 2). In 2D you just have a bunch of textures, but
in 3D you have at least the same amount of textures and you also have a lot of geometry data.
6. If you go the 3D way, then you have to convert almost every graphical element (maps, weapons, animations)
to its 3D counterparts. If you stay in 2D you could reuse old stuff and upgrade or replace it gradually.
I would say, that the gradual approach is more motivating for the creators (i.e. US) as you see
that development is moving forward (slowly, but usable all the time) in contrast to the "leap"-approach,
where you have only 20% of the work done and still have to do 80% until you can call it remotely usable.
7. A 3D engine will probaly require a modified or even a new physics engine. Don't know how much work
this would require.
8. I would say a combination of "2D-Background" with 3D foreground elements (weapons, mercs, etc.) is an idea worth
thinking about. I mean we have limited resources and using a simpler approach would probably be more fruitful.
---------------------------
Quote:
Quote:Have you read what Birdflu said earlier about his engine? He is thinking of implementing a Supreme Commander style of zoom interface. That
I believe would solve all the range issues.
That is impossible with 2D. I imagine he was talking about something else.
Why would that be impossible? You have a bunch of texture that you render on a bunch of quads. Zooming in and out
would require some scaling of the quads and use of a mipmapping-mechanism for the textures (so it doesn't look like shit).
Are we talking about the same thing?
Quote:
but then, there would still be the issue with the very limited and outdated JA2 2D engine and especially the very limited color palettes.
it
Report message to a moderator
|
|
|
|
|
| Re: Engine discussion for next gen JA based mod/community game[message #179653]
|
Mon, 31 March 2008 21:52
|
|
| Stromb |
|
Messages:40
Registered:March 2007
Location: France |
|
|
For a short term project (a few months), using Combat Mission Shock Force to build a small campaign based on JA Universe should work. Of course, you'll hardly have individual characters for instance, as the game is squad based.
But as far as i've seen, there's some trick you could use to simulate one or several mercenaries fighting alongside militia against various type of adversaries.
As for gfx, there's already a few skins available (SWAT, Specnacz, ...) to change from the modern US troopers look. Also, as you play more often with a global aerial view, the lack of details is not that much a problem.
For environment, it's ok to simulate more or less some Arulco areas (like Tixa prison or small villages and their surroundings).
For weapons, you'll just have to use basic one (assault rifle, rocket launchers, grenades, mortars, etc...), but in this case, the goal would be to take advantage of the tactical battlefield simulation (spotting, ambushing, hunting, assaulting, sniping, ...) , not of the weaponry simulation.
For vehicles (jeep, truck, tanks), well, you just have to prevent them from moving, or not putting them in a mission. A lone tank against a group of AT soldiers is in great danger if not covered by infantry (i tried, cf. Africa mini mod).
For infantry, there's several combinations possible between basic soldier (militia) and elite unit (mercs) for instance.
Of course the engine is not completely open, but still, the Battlefront team has provided basic tools to change what's necessary (though they don't seem to favor modding currently, probably because the CMSF engine is not finished ; the next add-on will certainly had a lot of veriety and possibilities). After that, you can also easily change music, probably some basic sounds (so using some dialogs lines of some JA characters does not seem impossible), some part of the interface (splash screen) it seems, and the description / name of units.
To have an idea of what missions could be done in the JA spirit, look at this AAR => http://www.battlefront.com/cgi-bin/bbs/ultimatebb.cgi?ubb=get_topic&f=71&t=000190#000000
I played it, it's great. Just imagine that instead of Taliban you raid a Deidranna outpost or whatever 
To have an idea of what can be done in other area of the game, look at this mini african mod =>
http://www.battlefront.com/discuss/ultimatebb.php?ubb=get_topic;f=52;t=003892#000000
I tried it, it's interesting. In a way it may simulates for instance a Deidranna attack on a mine, or again, whatever we can imagine.
Finally, it seems that the map editor is rather easy to use, and the IA scripting ask for a bit of practice, but it does not seem too complicated (and part of it is Wysiwyg if i'm not wrong)
I don't have the necessary time to work on this idea currently, but i'll try to adapt a CMSF mission to a JA one later. If anyone is interested, or has tried CMSF, i'd appreciate any comments / idea.
Here's an online version of the scenario design manual for those curious : http://homepage.mac.com/gibsonm1/CMSF/FileSharing24.html
[Updated on: Mon, 31 March 2008 22:15] by Moderator Report message to a moderator
|
Corporal
|
|
|
|
|
|
| Re: Engine discussion for next gen JA based mod/community game[message #185498]
|
Sat, 17 May 2008 14:05
|
|
| Privni |
|
Messages:4
Registered:May 2007 |
|
|
I would be quite glad if the game logic wasn't coded in c++ but I have 0 experience in game development.
-graphics, physics: c++
(torque engine? quite cheap for an engine, but i'm not sure it's easy to use)
then a webdev oriented architecture could be:
-having a network/http layer between c++ and the game logic engine.
I suppose it could make easier to do a multiplayer game.
-an embedded database accessible from both renderer and logic engine. I think it's easier for configuration.
-a logic engine interfaced with the renderer engine via a strong api. It will be very modular as its only need to answer to http requests. A good choice of language here. Difference with web is that's bi directionnal communication.
or interfacing with high levels languages like OCaml, Haskell, Erlang or such...
It could be easier to have a web development architecture as I suppose it's easier to find web developers familiar with an architecture with this kind of separation of concerns.
The problems is for the real-time part of the game. Will it be too slow?
The first idea of having real web pages to hire mercenaries, or managing the ja laptop is quite good because it's very easy to write webapps nowadays.
[Updated on: Sat, 17 May 2008 16:09] by Moderator Report message to a moderator
|
Civilian
|
|
|
|
Goto Forum:
Current Time: Thu Sep 12 05:33:58 GMT+3 2024
Total time taken to generate the page: 0.02403 seconds
|
Blog
Search
Help
Members
Register
Login
Home
