Printable View
I am very new to game programming (new to all programming, and have done nothing with games), but I am the best my small organization has, so I am going to need to do all the programming and am going to need some help. We are going to use Blender, and anyone who has used it to make a game before and would help me would be much appreciated. Thanks!
My first question is: where do I even start? We have a few blend files and we have all the planning done, but now where do I begin?
Do you want to learn programming, or do you just want to make a game? Many (I'd even dare to say most) people that jump into programming just to make a game don't tend to stick with it (it's a lot harder than people think).
Many people nowadays use pre-coded libraries to take care of the 'hard stuff.' This means you don't have to worry about the implementation details required behind hard stuff like rendering and physics.
By the time I'd finally gained the skills necessary to actually write 'games' (e.g. the hard rendering and physics stuff) I'd gotten pretty much sick of programming in general, changed my major and have a job in a field completely unrelated to programming (just to give you my personal experience).
A good way to see how other people are doing this is to go to gamedev.net and check out the image of the day. Luckily in the past few years it has become increasingly easier and easier to produce stunning games using many free and quite sophisticated game engines and libraries.
I want to learn the programming. I normally wouldn't be making games, but I'm the only one my organization new who can pretend to start to program. I will check out some libraries and see where that gets me (maybe I could get some ideas) but I want to do a bit of the hard stuff. DO you know what blender is?
Making games is hard. Doing it with a game engine is easier then starting from scratch, but it is still really hard. When I first started I wanted to make a fluid solver plugin to a 3d software. Took me one year to make anything else then console programs.
What I am trying to say is that making something as advance as a game before you know the basic is almost impossible.
Blender is a 3d program. One that I dont like very much. Programming and 3d animation has close to nothing in common.
If you want to start making games, start simple.
Try making a program where you try to guess a random number.
Lol, yes, yes I do know what blender is. Have you ever thought about how you would actually take the models you've produced in Blender and actually use them in your own game? That may be a good thing for you to read online about. Simply put, most of your basic questions can be answered from reading a lot online.Quote:
DO you know what blender is?
I second that. BTW, have you checked out gamedev.net's image of the day yet? You ought to find a game that looks like something you'd like to make and ask them what libraries they'd used to make it, how long it took them, etc.Quote:
If you want to start making games, start simple.
Try making a program where you try to guess a random number.
Writng games is much more difficult than writing normal applications simply because it invovles so many different aspects of programming. But it is definately the thing to flex yoru muscles on.
I can make better games on my graphing calculator! It's not that I'm unappreciative, I'm very thankful for your advice, but I might have come off wrong about this, I think I found something out, and it seems that I can manipulate objects with blender without to much trouble. I will get back to you once I do something, then I'll ask for opinions and advice. Thanks a lot.Quote:
Originally Posted by h3ro
Making the step from manipulating models in Blender to actually getting those models into your engine or game is a HUGE step. I had to essentially write a mesh class that encapsulated a mesh interface in Direct3D. In OpenGL you will likely have to write your own mesh class and your own loader as well.
If you know very little about core 3D principles and mathematics you will be lost when it comes to doing more than just loading the mesh. Don't get us wrong you can make games in your own time but most of us here have come to realize that by the time you actually have the skills to get the job done you usually don't have the time. Making any game is requires an enormous time investment and making a 3D game multiplies that by about a hundred. You can use other engines out there but without understanding 3D it is unlikely you will ever create a game even with the best 3D engine money can buy.
Also if you want to make a 3D game you should have a very very good grasp of C++, data structures, design patterns and anti-patterns, algorithms, optimizations, etc, etc. In short it requires a lot of knowledge to make a 3D game. Most of us here on this board have pretty much been reduced to small tech demos and/or small renderers that do this or that.
I was going to use python so I could utilize blenders game logic things. I will have all the blends made for me by our master blender. (I have no idea how good he really is because I haven't seen much blender, I would have include a blend of his just to show you a bit, but it says blend files are invalid.) I can use the blender stuff to have keyboard sensors, and I can python script the fancy things, so I am confused as to what will be so hard. You keep saying I need to have knowledge about "C++, data structures, design patterns and anti-patterns, algorithms, optimizations, etc, etc.", but I don't know why I would need any of those. Someone please tell me why I'm the only one who thinks this will be easy!
Okay, if you were going to code your own engine you'd need to have the knowledge mentioned by Bubba. Realize most of the people that frequent this site are aiming to code, in C/C++ (not Python or GameBlender's logic bricks) the actual engine themselves. That's how it came across in your original post, and that's what we meant by 'the hard stuff' (coding the engine yourself at a very low level). I was under the impression you'd use Blender as a platform to export models, which you'd then turn around and load into your engine.
Otherwise you are right, using GameBlender's Python scripting and logic bricks is an easy way for people with little to no programming experience to make games.
Also note that Blender in and of itself is for making models, not games. GameBlender is an engine, coded in C++, which encapsulates the logic bricks and Python scripting aimed for making games. This falls inline with what was mentioned before, in that you either need to a)spend a lot of time coding your own engine or b)use a pre-built engine [GameBlender in this case].
Enthusiasm and lack of experience? Nothing about making games in someone else's engine or in your own is easy.Quote:
Someone please tell me why I'm the only one who thinks this will be easy!
Do you remember when you first started with 3D animation? You probably wanted to create matrix 4 or lord of the rings 11 as your first project. If not, you have probably seen people who wanted to.Quote:
Someone please tell me why I'm the only one who thinks this will be easy!
Trying to make a game without knowing how to program is the same thing. Learn to walk before you want to fly
Oh, so you are saying that I should make my own engine now? I see that as a more complicated route... I think I will do this one in blender gaming to get the hang of it.
If it was easy, everyone would do it and noone would pay $60 for a game. Not trying to discourage you, just trying to get you to accept some realistic expectations.
If none of your team have experience with game programming then why not make a mod for an existing game instead? Fair enough you arent going to have quite the freedom as if you were to create a game from scratch, but it would be a whole lot more productive. Your media producers should be able to add all the stuff they make, and it would not involve anywhere near as much coding. It would also be a good experience for you to see how much is involved in making a game, as well as (to some extent) how it is done.
If you havent made even a simple game yet then I predict one of two outcomes: If you are lucky you will screw up early on and give up. If you are unlucky you will keep at it for several months until you realise that you are only part-way through it and your code is a hideous mess, then you will give up. Sorry to sound like such a pessimist; thats what happened to me. I still have an unfinished RPG and an unfinished strategy game. If I ever get back to either I will be starting again from scratch. For something big you really need experience and a well thought out design to start with. Thats one of the reasons I generally don't attempt big projects anymore.
To be fair to mcotter he is referring to GameBlender, which makes it possible to create interactive games using Blender models and animations using Python scripting and logic bricks.
"Game Blender is used by inserting "logic bricks," "controllers" and "actuators" to control the movement and display of objects in the engine. Game Blender is also able to be extended via the Python programming language. Game Blender is bundled with Blender 3D which can be downloaded freely from Blender's official website, http://www.blender.org/."
http://en.wikipedia.org/wiki/Gameblender
He's not interested in coding the game engine, he's simply interested in learning how to use Python and logic bricks inside GameBlender. This is akin to modding or using a pre-built engine, so I really don't think it's useful to keep telling him how hard it is to code an engine from scratch.
With that said, I don't really know how much use we will be. The experienced people on these forums tend towards writing the actual engine themselves, relying upon more 'universal' mathematical conceps. Using another program requires somebody with experience specific to that application (GameBlender in this instance), and I personally haven't created anything with GameBlender.
The only reason I even have Blender is to have a wider range of file formats that I can fiddle with (e.g. to export and import, such that I can convert models to my own .ntmdl format).
I started programming with Torque Game Engine, and I have nothing but good things to say about it (except that it will make you want to do a fps, which is a bad idea). I would point anyone looking to start out in that direction (maybe start with Torque2D now that it's around). I was able to switch from Torque Script directly to c++ with the help of CPlusPlus.com (a book would have been better in recollection ). I've never heard of Blender Gaming, and that alone makes me skeptical about it. (Thanks to BobMcGee for pointing out that he is using that btw). If you're going to use an engine, I would start out with a game like missile command, or another arcade game. Ballistics aren't too hard (especially if the engine is doing it for you), and that will get you started on something.
Build your games iteratively!
Oh, finally, something constructive. Thanks, bob mcgee! Okay, so now that we have gotten of the pessimistic page; if I need help, this would be a bad place to ask then? Also, I'm thinking that here might not be te best place for python questions then? Okay. I will get expirience from this game and then get advice on how the real hard core stuff works when we are done. We should eventually have a website where you can play it. I will post it up here for criticism and to deter future pessimism. "The greatest workers are those who see into their future and gaze upon success." I also have to add though, that I pointed out that I was using blender game, and I got more pessimism, and that even if I wasn't, I wouldn't give up. So about my questions, if this isn't the place to be asking for blender/python help, I guess I'll see you guys when I have a bit more experience.
There is nothing wrong with any of the posts in this thread. All of them are very truthful and if we told you this was easy or what you want to do inside of blender or whatever you use was going to be easy we would be lying to you.
If you really want something easy then I suggest you forget about game programming. If you don't mind researching, reading, coding, and ultimately failing quite often then you just might have a chance. We see your type a lot around here and enthusiasm is great but without the skills and knowledge to back it up it will get you just past the opening title screen and thats about it.
Nothing about programming and/or creating a game is easy. However it is very rewarding, can be fun, and during your journey you will learn a great deal. Programming games is really still a black art that has few manuals that show how to do it right. It is a different brand of programming that has much different challenges than what you may face in a non-game oriented code line.
Even using someone else's engine goes far beyond just plopping down objects and throwing scripts at them. Eventually you will have to get into some 3D mathematics to do what you want. After all the research and time invested I have a fairly decent terrain renderer, a decent start to a 4x space game, a 2D tile engine, and several little 2D games like Asteroids and side scrollers. Keep in mind all that took a great deal of time and I'm still dissatisfied with most of the results. Others on here have had similar experiences so we are not just being pessimistic we are being realistic. Pessimism is being negative without due reason or just for the sake of being negative. Realism is not blind ignorance of the facts but rather facing them head on, addressing them, realizing their impact or relevance, and using creative solutions to resolve the issues they present.
I personally do not have experience with GameBlender, or even the Python programming language.Quote:
if I need help, this would be a bad place to ask then?
The misconception was mostly your fault, as you don't currently know enough to properly ask your questions (e.g. your opening statement should have clearly said 'I am not interested in coding the engine from scratch'). Also, when you refer to 'Blender,' not 'GameBlender,' you are technically making a reference to a modeling/animation package, not something for making interactive software. We're not out of the 'pessimistic page' (which I found to be a rude comment), we're out of the page where you now know the difference between coding your own engine and using somebody else's. Bubba does in fact use Blender (not GameBlender) for producing models. If he wanted to, jumping into Python and logic blocks would be a trivial task compared to the heavy duty programming he's used to. If you got on his good side I'm sure he'd still be able to help you.Quote:
I also have to add though, that I pointed out that I was using blender game, and I got more pessimism
...
Okay, so now that we have gotten of the pessimistic page
In the meantime I've found some tutorials for you:
http://ltc.smm.org/gamestudio/tutorials/blender
http://www.ncsu.edu/project/stemgaming/blender.html
http://www.ingiebee.com/tutorials/Bl...e%20Engine.htm
the key in programming is to learn new peices of code each day , its along process but there is alot to learn in this field , just practice with small code and try and think of creating something , and if you can , take 1 step further and add more things to it , thats how i am learning to program games.
i started with first
createdevice
render vertices
transformations
camera
lights
materials
textures
mesh
i still need to learn pixel shader , but the rest above i learnt very slowly
i dont use blender , i use 3dx max and export my .x with panda exporter
Anddos, he's not looking to write the engine.
After what I understand you are trying to create a large game, not just a simple 2d puzzle or something like that?
Creating something like that without any programming experience is going to be hard. Using python and a game engine makes it easier, but you still have to program.
I know there are several "click and drag" game maker programs out there. I think Game Maker or something is one of them. That would be the fastest and easiest way to get started.
If not, best of luck :)
I'd say too many people dive in to the rendering without actually understanding the maths. That's like building a car without knowing how to build.
I made a Ray Tracer without knowing any math at all. When I started I could not tell the difference between dot product and E=MC2 :PQuote:
I'd say too many people dive in to the rendering without actually understanding the maths. That's like building a car without knowing how to build.
Debugging was a bit crap though :P
It can be done. IMO, it really only becomes a problem when you want to move beyond what the API can do without you knowing.Quote:
Originally Posted by zacs7
Oh, I had never heard of gameblender before. It sounds easier, but then the games i was making in the past were in a version of Basic designed for games. So its still possible to make a mess even with very high level languages. Or maybe thats just me, lol.Quote:
To be fair to mcotter he is referring to GameBlender, which makes it possible to create interactive games using Blender models and animations using Python scripting and logic bricks.
Okay, so I henceforth conclude arguing on the basis of mutual benefit through forgiveness. Though arguing can create strong debates and lead to the passion of challenge, growth is best found when sought by a whole.
So, anything constructive will be appreciated and anything otherwise will be promptly ignored. Here is what I'm gonna need to do.
The guy needs to move (blender can easily manipulate coordinates), and that shouldn't take too long.
The guy is gonna need to shoot. I know how to get the barrel fire, but I have no idea how to have the bullet land where he is aiming.
The camera angling will be parented to his gun, and that needs to be controlled by the mouse. Again I have little idea, but I think my friend might be able to help with that.
I will need enemies who shoot back. There I am utterly clueless.
And I need to do health, witch will be easy with Python.
Thanks again to everyone, and one more apology. (Once again I owe Bob for hitting the mark dead on.)
Gameblender has its own community, forums and tutorials, have you posted any of your Qs there? I'd bet those familiar with this engine would be of far better help to you.
Ok?Quote:
Okay, so I henceforth conclude arguing on the basis of mutual benefit through forgiveness. Though arguing can create strong debates and lead to the passion of challenge, growth is best found when sought by a whole.
Anyways.
Then it shouldn't take long to figure it out in Python if you know the language. Normally you have an up, right and look vector. Moving along the look vector equates to walking forward and backward. Moving along the right vector equates to strafing left and right. Moving along the up vector equates to flying up or flying down.Quote:
1. The guy needs to move (blender can easily manipulate coordinates), and that shouldn't take too long.
This is a simple vector operation. You can do this several ways. One way is to raycast the bullet instead of having it become an actual object in the world. The raycast is the simple parametric form of a line or P(final) = P(origin) + shoot_vector * distance. Normally you could use the shoot vector to perform a ray intersection test with the bounding volume of your object. Once that passed you would then do a triangle intersection test on the mesh itself to see which ray the triangle hit. Once you computed that you would then compute the barycentric coordinates of the ray intersection within the triangle. From there you would then fire off some type of 'hit' message which would then deduct health, etc.Quote:
2. The guy is gonna need to shoot. I know how to get the barrel fire, but I have no idea how to have the bullet land where he is aiming.
In an engine this is handled by the camera class. The player's gun is pointing down the look vector of the camera. To translate the gun to the correct world position you would have a vector in local space that represented the location of the gun relative to the local origin. Then you would multiply the inverse world transpose matrix of the object with the local translation matrix of the gun to move it to the correct location.Quote:
3. The camera angling will be parented to his gun, and that needs to be controlled by the mouse. Again I have little idea, but I think my friend might be able to help with that.
Enemies will have to 'rotate to face' your player and fire. The rotation can be done via quaternion slerp or it can be done via euler angles and axis-angle rotations. To do this you first find the up vector. This can be done by creating a ray from the player to the target and crossing it with the player's look vector. This will give you the up vector relative to the current player orientation. Once you have this you compute the dot product between the look vector of the target and the ray to the player. This is the amount of rotation or the angle of rotation you will need. All that is left now is to find the axis of rotation - which just happens to be the up vector. So if you rotate (target.look dot rayToTarget) about the up vector you just computed you will snap to or rotate to the player and vice versa. To smoothly rotate you linear interpolate between the current vector and the target vector you just computed. Once this is done you can set a flag that the target is now ready to fire. You fire the gun of the target at the player and use the bullet calculations I provided above in the previous answers. To add some randomness you would add a vector offset to the actual vector to the player. The offset would be a direct representation of the accuracy or skill level of the attacking object.Quote:
4. I will need enemies who shoot back. There I am utterly clueless.
Health can be done via simple scalars. Reduce the health on hits by a constant amount or by an amount based on which bounding volume of the player was hit. The head bounding volume would obviously reduce or completely eliminate any health thus killing the player. Other bounding volumes would have a lesser damage amount.Quote:
5. And I need to do health, witch will be easy with Python.
K, the first one I got, but lets take these others a bit slower.
The terms raycast and parametric mean nothing to me. Vectors I get, but how would I calculate the Shoot_Vector? Other than that, this also makes sense.Quote:
Originally Posted by Bubba
Vectors agian? Okay, this is starting to come together! So how does my mouse affect the vector?Quote:
Originally Posted by Bubba
The rest I can tackle after theese two.
A 2D raycast is:Quote:
The terms raycast and parametric mean nothing to me. Vectors I get, but how would I calculate the Shoot_Vector? Other than that, this also makes sense.
A 3D raycast in vector form:Code:
float final_x = start_x + cosf(angle) * distance;
float final_y = start_y + sinf(angle) * distance;
A unit vector is a vector that has range 0.0f to 1.0f in all three components. This is also known as a normalized vector. You could test final for collision however this is not all that accurate. Actually all you need to do is determine if the unit_vector intersects the bounding volume of the target. Remember that a vector stretches from infinity to infinity so the intersection math works just fine regardless of the magnitude of the vector. If you want to make the bullet 'fly' along the shoot vector you simply normalize the shoot vector and and then multiply it by a distance scalar. The distance traveled is good old distance = rate * time where rate is the speed of the bullet and time is the frame delta or the time it took to render the frame. Keep in mind if you want to 'fly' the bullet to the target you will have to do a completely different collision test to determine if the bullet could hit target X in time interval T. If the interval is infinite then you know the bullet did not hit the target.Code:Vector3 final = start + unit_vector * distance;
One way to do it:Quote:
Vectors agian? Okay, this is starting to come together! So how does my mouse affect the vector?
The rest I can tackle after theese two.
DirectInput will return mouse coordinates in relative by default so this works fine. If your particular mouse driver or API returns them in absolute you will have to derive the relative via a simple subtraction from the last known mouse x and y.Code:
m_pCamera->Yaw(m_pMouse->GetX() * reduction_factor);
m_pCamera->Pitch(m_pMouse->GetY() * reduction_factor);
There are other more sophisticated methods such as converting the mouse x,y to spherical coordinates or mapping them to a unit sphere and then adjusting the up, look, and right vectors accordingly.
Okay, I'll see where this gets me and get back to you.