"The First King" - My next project, Concept phase

[JP]

Hi.
JP: I've checked the IRRwizard. The last release seem to be dated june. And seem inactive right now. But there seem to be some basic AI in there. They are using "MicroPather" for path tracing. I could try to study the code and perhaps do something.

I was able to compile the full framework and try it. The ennemy seem to be invicible, by seeing us(the player) from everywhere (they see thru everything) and their bullets goes thru wall. So not sure they have implemented the AI for "seeing" and a basic collision for their bullet. I've also tried it in DirectX and the AI code seem disabled.

I was able to make one flee and that worked well.(So path tracing seem to work in there) So learning the framework code is a good idea. But it's sure I cannot use the framework as it is now (That was the author intend in the beginning). I need IRRScenes in there (They are using meshes directly) and would also need to use IRRKlang.

Perhaps with deep modifications. It's well structured. Could have some benefit of using it in the long term.

The AI, I would like for ennemy. Is that the ennemy character is seeing (about 45-50 degree wide) and when the chararter will be really visible (It will not see it if the player is behind a wall). The attack state will start. This is already there but the "see" code is not working. (see thru walls).

Also if the player is firing; the ennemy could heard the sound and will go and search the source of it (Then attack when it see the player). So it's basicaly a path trace from the ennemy to the source of the sound.

In my dissertation project at Uni i used (a java based version of) irrlicht to do a simple FPS framework to implement AI agents in and i made them have a limited field of view which was impeded by the walls of the environment and they would weigh their odds of beating an opponent when they spotted them and flee if they didn't fancy their chances or make a fight, aim nicely, collect resources by searching the environment, have different skill levels, turn to find enemies if they were hit by one they couldn't see etc.

As i say, that's all in java so not all that useful for you but if you like i'll certainly give you a hand with any AI questions you have and if you wanted to see the source code then i could offer that up to you. Maybe i could even help out with coding some of it if you wanted the help, i understand if you want to do the whole project alone though of course!

[christianclavet]

Hi! JP!

... to do a simple FPS framework to implement AI agents

Wow! Just what the project need!

Maybe i could even help out with coding some of it if you wanted the help, i understand if you want to do the whole project alone though of course!

That's really great! I would like to complete this project in a team, I would not be able to complete this only by myself. I prefer to focus on the level design and modeling, and if some people could concentrate on developping the engine it could be great. The source code will be made available on the first attemps.

One of the things that will be needed on the developpement is that all the codes will be put clearly (comments and includes filenames for the functions). So your code for AI could be used on other projects. (Perhaps spark again the IRRwizard project) The thing I consider important, is that once we start working on this, the whole community will benefit from it.

I'm currently thinking of starting the project in "PHASES", so that we don't put on ourself too much pressure:

- Phase one: Buildind the first level (Cairo Egypt), and navigate thru it. (Basicaly load an IRRScene with the elements)
Occlusion culling is needed to have fast framerates because there will be lots of nodes. Frustum culling will not be good enough. Also, I'll have to have models with normal maps details so that they really look good. (Still figuring out the baking of this) Will have to do some tests first to see how the normal maps and lighting work in Irrlicht. Shadow map implementation could be good also. Some interactivity there for accessing the tomb. (Hidden door mecanism we'd activate and use of a map). This phase is one of the most important ones, it will build the framework for the game.

- Phase two: Buildind the "Stargate" and have the tunnel animation, that will load the second level (Nibiru) Interactivity to activate the Stargate.

- Phase three: Building the Sumerian City and the Ziggurat (Sumerian Pyramid). Be able to navigate thru it and find the Artefact. Some interactivity for doors and things to access the Artefact. One thing that could be interesting is that we determine multiple places where could be the artefact and this could defined randomly at runtime.

- Phase four: Creating the Jackal Warriors, animate them. Put the AI in the system so that they attack the player. Have some weapon dissemined in the Sumerian city. Once we have phase one completed, the other phase could be worked on by some team members. Like that important phase for AI and FPS gaming.

- Phase five: Put an intro and an outtro for the game, clean up. Fix some issues. Put the last addons (Perhaps some tourist walking near the Pyramids). Working on a full gui. Setup, Key Mapping, etc.

- Phase six: Clean up the codes, put some comments. Fix the last seen bugs. Release the final source to the community.

If you see all theses phase, on a guy alone would take forever. We'll need to have a team. That's why I'm working on planning it carefully. If we're gonna work on this, we need to know what will we have to do.

I think doing this with a team would take us about 6-9 month. Depending on the time we give onto the project. Alone would take me 1-2 year surely.

[JP]

That sounds cool. As I work full-time i don't have a huge amount of time to dedicate to a project like this but i think that my time on AI for the project could be quite minimal as i've got the whole system designed already in some respect so it's just a case of converting it to C++ and making it a sort of add-on module in a way which shouldn't be too much effort.

Luckily at work i have a huge amount of free time to work on personal projects (currently got two of my own PS3 games in development) so technically i should be ok to work on the project whilst at work (I have no time at home to work on programming unfortunately). But i may have quite a big Sony project coming my way soon which could take up a lot of time but from the look of your timeline you don't really need me for a while as you're starting off basically doing your old project again with an egyptian theme which you should be able to cope with yourself and then when that's done i'll probably be freed up to do the AI and such.

[christianclavet]

Hi. Here are some updates on the project:

- Requested official project site on Sourceforge today.

EDIT: The project was almost immediately approved!
Here is the link to the sourceforge page:
http://sourceforge.net/projects/first-king/

- Phase 1 of the project will be started in a few days:

Phase 1 is to start developpement on the engine, it will refer to the IRRwizard framework. Environnement modeling will not start fully until the engine have a minimum functionnality. (So it will mostly basic shapes or low version without maps)

Objectives of phase 1:
- Building the core of the engine with Irrlicht 1.4 (Will perhaps use the BETA in the first phase developpement)
- Have support for loading an IRRscene in a extensible framework (Using IRRWizard as a reference point)
- Have a FSM and entities support for interaction (ladders, doors)
- Have shader support for normal maps/specular maps for characters and environnements. (90% chances using Nadro wrapper for CG)
- Builing the first environnement (Great Pyramids & Sphinx) but with basic details, will be refined in the end of the phase.
- Working or have someone working in ways to optimize the framerate (basic occlusion culling for the environnement). First step will be to build in "blocks" so that they can be hidden when they are not in view.
- Adding IRRKlang for sound and music. Having some events that will release sound. (Wall hit, footsteps, etc)

JP volunteered to work on the architecture of the AI engine for phase IV and is willing to help me in is spare time. Anybody willing to join is welcome.

The needs for this project are the following:

Design:
Game Design: Me (suggestions accepted, will need to follow story)
Environnement design: Me
Characters design: Me, ?
Weapon design (phase IV):

IRRLicht Engine programming:
AI Architecture programming: JP, I'll surely assist a little
Engine framework (FSM, Entities): Me, JP and with the help of the IRRWizard
Occlusion culling system for environnements: ?
Shaders programming: will use Nadro wrapper for CG
Plugins/includes architecture for extensible framework: ?
First Person / Third view cameras system: Starting with std FPS Camera.
Sound system: (Will use IRRKlang.) ?
Particles: Me, basic use of it (sand, effects)

Modeling / Animation:
Environnement modeling: Me,
Character modeling: Me,
Weapons modeling & animation:?
Character animation, cycles: ?
Player bodyparts modeling, animation: ?

Sound:
Music: (Will take open source music for starting)
Sound Effects design: (Sound based on events)

[Woozley]

I'm very impressed with this & your previous demo ! What modeller software do you intend to use ? I'm fiddling about with blender but the export scripts seem to be a bit fickle.

[cassini]

After learning a lot on my first project. Here is the start of my concept for the next "DEMO". This one is more ambicious than the first project. I think I could do it.

EVERYTHING IS SUBJECT TO CHANGE. This is only a concept.
...

How about it? Is it too crazy?

It is not crazy, I love the idea.
Looks like you invested lots of work, time and money in concept art already.
Good luck

[christianclavet]

Hi.

I should start PHASE I very soon. I think 1 or 2 week from now.
The Sourceforge project is up now. Got to figure some other things (start of the framework, and also put the code on sourceforge (with SVN version for developpers))

cassini: Thanks! Yes. I have to put some time there first so the objectives are "reasonnable" and the project to be solid. Having a bad preparation would have us drop the project because it's too big or have not suspected major developping problem when we started. I've also defined it in "phases". So, If we complete a phase, we completed the objectives of that phase.

This project source code and the models will be available so this could be expanded further (more levels, physics engines, network wrappers etc.) But the "first king" will only be single player with AI NPC (JP will help me a lot on this AI implementation). We plan on working on the framework so that it could be easily expanded to allow expansion.

Woozley: I will model, in Lightwave, 3DS Max, and Modo (Perhaps others! ). The tool we use is not that important. The features we need from the model are the most important:

- A version in high details that is used to produce a low version with a normal maps. So the mesh need good UV maps and someone to produce the Low level mesh from the High detail one (unless you can do it all).

- The final model will be in .B3D or .X format depending on the need and availability of tools. A exchange format will have to be used if there is more than one modeler/animator on the project.

Until Phase 4, we'll need most objects to be for the environnements.

Here are the modeling/animation objectives for the project:

Phase 1: Pyramids (Inside/outside), Sphinx, some desert landscape with rocks. With some objects

Phase 2: A fully detailled Portal (2) with some Egyptian/Sumerian design like the STARGATE (Will need to it be a little different). A wormhole to use for the "travel" animation. Water animation will also need to be done (water object that will be modeled and animated for the opening of the "STARGATE")

Phase 3: Modeling inside and outside the Ziggurat, building a temple and city buildings, some great walls, and evidently a place where will be hidden the "ancient knowledge artefact". Will need some doors and ladders there. Need the city to be a little big. If phase III is completed. I will consider that the project is a success. Phase IV will be more of the "crazy" parts that will make it in a full FPS game. References for these Sumerian builings is very sparse, will have to imagine the details.

In phase 4, we'll need to have the Anubis Jackal Warriors type character with animations poses. Also in that phase, we'll model the player body parts. If possible, I'd like to have animations of the player healing himselft. (Removing bullets or objects from the body). Like in FAR CRY 2 (UBISOFT). For Phase 4 to be successful. We'll need to be in team. Without it the project will make a nice game but adventure only, not a full FPS.

Phase 5: More of a polishing part, remove most the bugs from the game engine. Clean up/optimize what we can from the geometry. etc. Particles could be done there to add to the environnement, like dust and sand in the wind. Also some dust to fall down when we open doors...

Heres are the problems we will try to overcome:

Modeling: Creating the model with high detail with the normal/specular maps (Mostly when we will try to do the baking part and import it correctly into IRRLicht from IRRedit)
Animation: Cycles that work correctly in IRRlicht and blending working
Rendering: Framerate optimisation. At the moment there is no OCCLUSION CULLING in IRRlicht. We'll have to find a way to get some good framerates. Also this is the first time, I will have to rely almost completely on shaders (Nadro have made a good wrapper for Cg and JP have developped already a good shader, but this will have to be tested, there could be surprises. Especially on the Lighting part.