3D Game Engine From Scratch

WIP games, tools & toy projects
#1

Here are some pictures of the 3D Game Engine I’m developing. I made it all the way from scratch using only plain old Java. (Don’t ask me why, I did it because I can).

I’m not the most experienced programmer. I have about a year experience with java and a year and a half of object oriented programming.

Any ideas, suggestions or usefull articles are welcome. Feel free to give some feedback.

Currently I can rotate from left to right, up and down and I can walk around. I have the ability to make some objects and place them in the world. That’s about it :wink:

UPDATE:
I’ve been working a few hours last night and today to clean up the code. I’m now starting my list of features I want to implement. As soon as I finish my list I’ll see if I can upload it so you can download it.

I have a laptop with an AMD quad core 2.3GHz Processor, AMD Radeon 1GB dedicated VRAM graphics card and 6GB RAM.
I can reach up to 130fps sometimes, but generally it runs between 60fps and it’s cap 111fps. It runs on 60 ups.

#2

Cool! It’s a real challenge to tackle writing an engine. Just to warn you: if you want to make games, don’t start with making an engine. Creating an engine will eat up a year or two of development and get you not as far as the already established engines. That being said, it is a great way to learn about the inner workings of games and the engines that power them. The knowledge you gain from it will help you down the road.

#3

Thanks for the advice! I know it’s indeed ambitious to start with an engine right away, but I already learned a ton of it. I’ve had some experience making games in java and C#, especially 2D games. I also worked with some engines, but I’m very eager to learn make games from scratch including their engines etc. that’s why I started this project. I thought it would be usefull to post it for some help and motivation to finish it.

#4

Is it open source?
VERY cool by the way, this is really impressive.

#5

Not yet, it’s still in it’s roots. I have a little checklist in my head and when most of that is done I’ll put it online. There are still a few issues I need to resolve before releasing it.

#6

Pretty cool engine youve got there. Using java2d correct?

Also, the [img] tags should work if you put the url inside of them. If that doesn’t work, try uploading to imgur.

#7

It’s very cool that you’re using Java2D for this, at first I assumed it was done in OpenGL, and I wasn’t too impressed, but then I realized my mistake! Good job!

#8

Just plain Java ;). It took me about 3 weeks to figure out how to work with matrices and stuff. Then it took me another 3 weeks to build it as far as I am now. Long live wikipedia and google :stuck_out_tongue:

#9

Oh, matrices suck, I seriously despise them! I’ll have to eventually learn them someday, but I’m very impressed you learned them that quickly :slight_smile: Very good job, I can’t wait to see what else you can do!

#10

http://www.sjbaker.org/steve/omniv/matrices_can_be_your_friends.html

#11

Very nice. Can you post your system specs and the fps you are achieving? Just can’t believe that Java2D can be that fast to tackle a 3D engine. Good luck with your project.

#12

Very impressive, look forward to seeing how far you take this. Huge challenge to set yourself!

#13

Why shouldn’t it? Softwares renderers for Java exist for ages. On a current i5/i7 machine, you can render a fully textured Quake3 level in 1280*1024 @ 30-60fps in software.

#14

I have a laptop with an AMD quad core 2.3GHz Processor, AMD Radeon 1GB dedicated VRAM graphics card and 6GB RAM.
I can reach up to 130fps sometimes, but generally it is between 60 and it’s cap 111fps. Meanwhile it runs on 60 ups.

#15

Sounds like you have a list of features that you’re planning to investigate, design and implement all at the same time - big challenge! but an interesting one.

Having attempted to do something very similar myself (3D using OpenGL) one approach I would suggest is to follow an iterative development style:

Decide what feature(s) you’re going to deliver next and stick to it. Once you’ve got the next feature up-and-running have a pause and take the time to:

  • tidy-up your code and documentation or notes.
  • extract any generic code into reusable classes/packages that can be used elsewhere (and checking you haven’t broken anything).
  • re-review your list of features, you might have to change the order or priorities depending on what you’ve learnt.

I like to call this the perspiration-then-consolidation approach :wink:

  • stride
#16

Hey that’s a good one! Thx for the advice.

#17

What kind of performance tests have you done . Good ones are where you set it to render a certain number of verticies and then check the fps.

Also what do you use to render is the render function for your objects called in a tick function?

#18

Well , that just changed my perception on matrices, thank you very much!!

#19

I haven’t done alot of tests, however I have tested it on performance when using a load of vertices. When the amount of vertices increases significantly (and then I’m speaking about hundreds of thousands) the framerate starts to drop. However when using 5 grids (so that is 20 * 20 + 20 * 10 * 4 = 1200 points in the case of the pictures in my post) it doesn’t effect the performance alot.

#20

How are you going to do the rasterization its currently what im stuck on.