Slick Forums

How do I go about combining roughly 3000-9000 small (32*32) small images together, into one image?
The source is an Image[w][h];
It's much appreciated!

You shouldn't combine them into a single texture/image for it would be too large for the driver. 2048x2048 should be a maximum single texture size to aim for, or lower for older systems (512x512 is safest). You can query the driver's max texture size with BigImage.getMaxSingleImageSize().

It's not hard to combine images into a single question, but there is most likely a better way of doing whatever you're attempting.

Why do you need to combine that many small images together? Why do they need to be in a single image (i.e. instead of rendered as tiles)? Do you actually have 3000-9000 different images, or are you simply repeating them (thus wasting memory and texture space)? Do you plan to save/export this image or are you simply using it for rendering in-game?

Well it's too costly to render all of these tiles at once, and I'm honestly unsure how to render just the tiles around the player, that would probably a better alternative actually. Well there's usually 3000-9000 tiles (since I have a reasonably large randomly generated map). There's only going to be roughly 10 different tile images, but these will be repeated lots of time etc.

I'm just looking for a faster alternative to render the tiles
Thanks, Harry.

What exactly is a texture atlas? Is it just an image/spritesheet of lots of iamges? Why do you begin/end with it?

I need to figure out how to do that, because I have an offsetX and an offsetY and no real way of knowing which tile I'm on, unless I do a loop through every tile checking for a hitbox, since mathematically it doesn't seem very easy/possible.

That is exactly what it is.

The begin/end stuff is needed due to the way openGL works, normally you need to tell it when to start a task and when to end it, these operations is a bit expensive and a waste if you are drawing the same image many times, such as when drawing the sub-Images of a texture atlas, where you draw the parts of an image.

When you draw your image with the normal draw method, it tells openGL to start and end the task of drawing the image every time its drawn.

drawEmbedded doesn't call the begin and end methods, so when you beginUse/drawEmbedded/endUse you control yourself when to tell openGL to start and end the task.

Here's an example of what happens.
This code:

Is the same as this:

But if you control it yourself, it could look like this:

This way is faster, especially if you are drawing the image alot of times.

_________________
For every new problem, a new source of solutions has come to exist.