Removed…
The only real problem I’ve encountered with Java2D is performance. If you are going to scale or rotate those images, then you’ll definitely need LWJGL (or JOGL, but I prefer LWJGL).
Removed…
You are talking about the Java2D pipelines… still Java2D
unfortunately they aren’t… “good”
Your applet is using only 22% of your computer full capacity and you are worried about melting. Relax, remove all the odd sleeps and finish the game.
I started out using Java 2D for my game.
I found that I could paint 14000 tiny to small images (think 1x1 to 64x64) in about 200 ms or so. The performance was initially good enough but I noticed it started to chug with a larger canvas (anything larger than 600x600).
I switched to LWJGL. I can now paint 400000 quads of the same size with anti-aliased alpha-channel texture maps in a window of any size in 10 ms! This was accomplished using a single large texture map sub-divided into a grid of smaller textures, vertex arrays, and 32 bits for each color.
14000 images (=28000 triangles) of that size is nothing for OpenGL. Consider that you can also implement scaling and rotating, e.t.c. for free and it doesn’t make much sense to use Java2D.
Don’t be so hasty. Java2d can be really powerfull. It just do not work on all machines. If you remember the sprite shoot out test. Try java2d version against those. With my comp java2d beated slick2d by large margin. But with some comp slick2d produced 10 more times sprites before starting slowing down.
Link to test. http://code.google.com/p/protectcastle/source/browse/?r=81b0570d22e3f832422eb319a54227dc6b75f9de#hg%2Ftest%2Fpokeballjava
absolutely yes! however stretching java2d to its max performance is fun too.
Not when that “max performance” equals 4 FPS for a Final Fantasy style tile map. LWJGL gave me several hundred FPS, and that was 5 years ago.
Once I finish the current round of updates with my engine I will implement an OpenGL version of the graphics manager. I will run the two on the same machine and record the resulting FPS. Then I will upload it to my site so you can check it out for yourself.
Figured I better implement opengl now rather then later when I “need” to.
Guess I should define the game type context first before said that