Draw a single sprite to Pixmap - LibGDX - java

I am currently working on a game in which i need pixel perfect collision with objects. The only obstacle i had gotten to however, is with actually getting a mask of the object. I thought about rendering a sprite of an object (which is properly scaled and transformed) to FrameBuffer, which then i'd convert to a pixmap, but i had no luck doing it (tried many methods, including this)
So, my question is: Is there a way of rendering a single sprite to a pixmap in LibGDX?

Sounds like putting the horse behind the cart. A Sprite is a region of a Texture, which in turn is an image loaded from file aka Pixmap. If you just want to load the image from file then you can do: new Pixmap(Gdx.files.internal("yourfile.png"));. You can also scale and transform your coordinates without rendering to a FBO first.
That said; getting the Pixmap of a FrameBuffer is as "simple" as taking a screenshot while the FBO is bound, because it is exactly that.
fbo.begin();
Gdx.gl.glClear(...);
... //render to the FBO
Pixmap pixmap = ScreenUtils.getFrameBufferPixmap(0, 0, fbo.getWidth(), fbo.getHeight());
fbo.end();
Note that this will look like it is up-side-down.
But you are probably only interested in the pixel data, so you might as well skip the Pixmap step in that case and use the ScreenUtils.getFrameBufferPixels method.
However, that said; using the pixel data for collision detection is most likely not the best solution for whatever it is you are trying to achieve. Instead I'd advise to look into other options, for example have a look at this tool which can be used to convert your images into a collision shape.

Related

Thickness of OpenGL 2D Textures

What is the easiest way to give 2D texture in OpenGL (lwjgl) some kind of "Thickness". Of course i could get the border of the texture somehow and add Quads, orriented by the normal of the quad that the texture is drawn on, in the color of the adjacent texture pixel. But there has to be an easier way to do it.
Minecraft is using lwigl as well and there are the (new) 3D Items, that spin down on the ground and don't cause as much of a performance issue, as is if they were drawn of dozends of polygons. As well, when you hold an item in your hand, there is that kind of "stretched" Texture in depth, which also works with high resolution textures.
Does anyone know how that is done?
A 2D texture is always infinitely thin. If you want actual thickness (when you look edge onto it) you need geometry. In Minecraft things look blocky, because they've been modeled that way.
If you look at some angle and ignore the edges you can use parallax mapping to "fake" some depth in the texture. Or you can use a depth map and use a combination of tesselation shaders and vertex shaders to implement a displacement map, that generates geometry from the texture.

LibGDX Scaling and Rendering Sprites from TextureAtlas

I am currently experiencing issues when drawing and scaling sprites.
I am loading my assets from a texture-atlas, which I packed no problem with the LibGDX texture packer gui tool. My texture atlas image currently looks like this.
These images are supposed to be buttons, but as you can see, the image is very small, so when the sprites are loaded, they load a sprite of say, 34x16 pixels. When I render these buttons, on a canvas of 1920x1080, they are much too small. I use sprite.scale(int scale) to scale the sprites, but when I scale them, they appear blurry. What I would like to happen, is when they are scaled, each pixel is scaled proportionally, keeping the pixelated effect on the button, rather than a blurry resized image from a really small texture. I currently render the sprites using sprite.render(SpriteBatch batch). Is this the proper way of rendering a sprite, after they are loaded using atlas.createSprite(String name)? I am new to using sprites, and loading textures from a texture-atlas, so i am wondering if this is the correct way of doing things.
Also, when I initialize my game, I load numerous different Sprite objects from a TextureAtlas. Each sprite holds a texture that will represent a game object, however it is my understanding that you render a sprite using sprite.render(SpriteBatch batch), so therefore I could only use a sprite, loaded from the TextureAtlas for one game object, because I would also have to set the scale, and position of the sprite, as it represents the game object. I am used to loading a Texture, then rendering this texture using batch.render(), at a given position, but I don't see how this is possible if I am using a sprite. Even if I use batch.render(Sprite, x, y), I am unable to scale the sprite properly, because as I mentioned before, I would like to scale the sprite while maintaining a pixelated effect, and even so, using the Sprite.scale() method, this would scale the Sprite object as a whole, making it impossible to use the Sprite's texture multiple times for numerous game objects.
Any suggestions would be greatly appreciated.
The code I am currently using to render/load the sprites is as follows:
Loading from TextureAtlas:
public static TextureAtlas atlas = new TextureAtlas(Gdx.files.internal("data/texture/pack/output/pack.pack"));
public static Sprite sprite = atlas.createSprite("buttonUp");
sprite.setScale(10);
Rendering Sprite: GdxGame.WIDTH/HEIGHT are 1920x1080. Though the Desktop window is scaled down from that size. Thus, everything is rendered as if the screen were 1920x1080.
batch = new SpriteBatch();
camera = new OrthographicCamera(GdxGame.WIDTH, GdxGame.HEIGHT);
camera.position.set(GdxGame.WIDTH/2, GdxGame.HEIGHT/2, 0);
camera.setToOrtho(false, GdxGame.WIDTH, GdxGame.HEIGHT);
public void render(float delta){
Gdx.gl.glClear(GL10.GL_COLOR_BUFFER_BIT);
batch.setProjectionMatrix(camera.combined);
batch.begin();
camera.update();
sprite.draw(batch);
batch.end();
}
After further investigation, I have discovered that I can use Sprite.set(Sprite sprite) to make Sprite x a copy of Sprite y, and therefore render each sprite multiple times. However, this does not solve the issue of scaling the sprite. I must emphasize further that when rendering a scaled sprite, the scaling is not done by pixel, meaning that it is blurry. But, when rendering a TextureRegion, like: batch.draw(TextureRegion, x, y, width, height) if the width and height are greater than that of the original texture, it will scale each pixel, rather than blur the whole thing to try and make it look better. The following is an image of the blurriness I am talking about:
Notice how this sprite is scaled to be blurry, even though the original image is small, and pixelated.
What TextureFilter settings are you using in your code or in the texturepacker? Try the "Nearest" filter. If you have set it to "Linear" or alike, it will always take 4 texture pixels (texels) and interpolate them to get the color of the pixel to be drawn.
That might help against the blur, but I am not sure if it will produce exactly that 8-bit look you are aiming for...

Ray Casting a transparent PNG in Java with LWJGL

Im making a game and id like to implement raycasting for the hero's laser (and other stuff in the future), i have my sprites in a sprite sheet which i bind in the beggining and access when i draw since each element knows how to draw itself, but the spritesheet is a PNG, and thus some elements posess transparency, which works ok in openGL. i know each element's position, size etc but if some sprites have transparency, the position and size arent enough for the ray cast to be perfect since it would only hit the "bounding box". So is there a way to throw a ray using Bresenham algorithm (i believe it is the lightest way, correct me if im wrong) and make it pixel perfect in openGL, so that i can acquire the collision point of the ray with the actual non-transparent zone of the first sprite it appears in the way?
There is no easy way to do this. You would have to create a custom collision checker for your raycast to see if it would pass through or if it would collide with part of the sprite.
However it might be a better idea to use a smaller bounding box, or a circle to represent it, or both. These are much easier and faster to calculate then checking every pixel within the texture.

Simple image rotation

I am making a game in Java. I made a planet seen from outer space and I want to make it appear like the planet is slowly rotating. But I don't know how to rotate a image. I need a simple command that rotates my image 1 degree around its own center once. Any help?
This is what I want to do:
Image
Take a look at these tutorials:
Java2D: Have Fun With Affine Transform
Coordinate Translations and Rotations: Example Code
Transforming Shapes, Text, and Images
What you are describing is not rotating an image, but changing an image to represent a 3D rotation of the object in the image.
Ideally you wouldn't be working with this as an image but rather as a 3D object with a different camera angle. Then you would simply rotate the camera around the object and display the resulting image to the user.
However if you're set on doing this as an image, then you need to create a different images representing various states of rotation of your planet and have a separate thread that would replace the displayed image with the next one in sequence, at repeated intervals. Search the web for "java image animation" - there are plenty of tutorials on how to do this.
If you want to rotate an image in 2d space, you can use something like this:
Image image = ...
Graphics2D g2d = ...; //
g2d.translate(170, 0); // If needed
g2d.rotate(1); // Rotate the image by 1 radian
//or g2d.rotate(180.0/3.14); to rotate by 1 degree
g2d.drawImage(image, 0, 0, 200, 200, observer);

Why does my smudge image algorithm make the image darker?

I'm trying to write a graphical effect where a circle moves around an image smudging the image as it goes (like the way the smudge tool in Gimp or Photoshop would work). The basic algorithm I'm using is:
the circle moves from position A to position B on the bitmap
copy a circle of pixels from position A into a temporary bitmap
draw this circle of pixels from the temporary bitmap to position B using alpha of about 50%.
This works fine and looks like what I would expect, where the image will look like it's getting smudged if the circle moves 1 pixel at a time over the image.
I now want to add some texture to the smudge effect. I have a bitmap that contains a picture of a paint blob. The algorithm from the above is modified to the following so the smudge takes the shape of this paint blob:
as before
replace the temporary bitmap pixels with the paint blob texture then copy the circle of pixels from position A into the temporary bitmap but only keep the pixels that match up against paint blob pixels (i.e. use Porter-Duff "source in destination" mode when drawing the circle into the temporary bitmap).
as before
This almost works and it looks like it's fine initially but gradually the smudging makes the colors in my image darker! If the circle passes over the same area several times, the colors eventually change to black. Any ideas what I could be doing wrong?
I've implemented the above in Android. I happened upon this post about bitmaps in Android (like my paint blob texture) being loaded with "premultiplied alpha", where the author says it caused his images to become darker because of it:
http://www.kittehface.com/2010/06/androidbitmap-and-premultiplied-alpha.html
I suspect I'm suffering from a similar problem but I don't understand what's going on well enough and don't know how to fix it. Does anyone have hints at what might be going on?
Well from first glance the reason the image is getting darker is because #3 in the first three steps. You overlaying a pixel over an existing pixel at 50%. You might want to consider using the mean of the original pixel value and the new pixel value. You might want to research some blurring algorithms.

Categories