Toasting with LibGDX Scene2D and Animation4j

For our latest Vampire Runner update we changed to use LibGDX scene2d instead Android GUI. The main reason for the change is that we wanted to use a common GUI API for Android and PC, and sadly we can't do that using Android API. With LibGDX scene2d we can code once and run in both platforms.

In particular, the toast feature of the Android API was really interesting to have and we want to share how we implemented it using LibGDX scene2d.

Toasting

A toast is defined as a scene2d Window that shows some text and disappear after a while, this is a pseudo code to give the idea of how to create that toast window:

Actor toast(String text, float time, Skin skin) {
	Window window = new Window(skin);
	window.add(new Label(text, skin));
	...
	window.action(new Action() {
		act(float delta) {
			// update the animation
			// if the animation is finished, we remove the window from the stage.
		}
	});
	...
	return window;
}

To animate the toast, we create a TimelineAnimation using animation4j defining that the window should move from outside the screen to inside the screen, wait some time and then go out of the screen again. The code looks like this:

TimelineAnimation toastAnimation = Builders.animation( //
	Builders.timeline() //
		.value(Builders.timelineValue(window, Scene2dConverters.actorPositionTypeConverter) //
			.keyFrame(0f, new float[] { window.x, outsideY }) //
			.keyFrame(1f, new float[] { window.x, insideY }) //
			.keyFrame(4f, new float[] { window.x, insideY }) //
			.keyFrame(5f, new float[] { window.x, outsideY }) //
		) //
	) //
	.started(true) //
	.delay(0f) //
	.speed(5f / time) //
	.build();

That code creates a new animation which modifies the position of the Window each time update() method is called.

Of course, you can animate the Window using LibGDX custom Actions or another animation framework like Universal Tween Engine, that is up to you.

If you want to see the code itself, you can see the Actor factory named Actors at our Github of commons-gdx.

In our subclass of Game, we added an empty Stage updated in each render() method, and a toast(string) method which creates a toast as explained before using default Skin and time.

MyGame extends Game {

	Stage stage;
	float defaultTime;
	Skin defaultSkin;

	render() {
		// all our game update and render logic
		...
		stage.act(delta);
		stage.draw();
	}

	toast(String text) {
		stage.add(Actors.toast(text, defaultTime, defaultSkin);
	}
}

So, if we want to toast about something, we only have to call game.toast("something") and voilá.

You can see a running example of this, you can run the Gui.Scene2dToastPrototype of our prototypes webstart (recommended), or watch the next video:

Conclusion

Despite being a bit incomplete and buggy yet, scene2d API is almost easy to use and it is great if you want to do simple stuff.

Using scene2d is great for our simple need of GUI interfaces because we can quickly test all the stuff in PC. In Vampire Runner we are using scene2d for the feedback dialog, the new version available dialog and for the change username screen.

An interesting thing to have in mind when using scene2d API is that you can make your own Skin to achieve a more integrated look and feel.

As always, hope you like the post and could be of help.

VN:F [1.9.22_1171]
Rating: 3.4/5 (5 votes cast)

How we use Box2D with Artemis

As you may know from our previous posts or from your personal knowledge (obviously), Box2D is a 2D physics engine and Artemis is an Entity System Framework. Box2D is used to add physics behavior to games however it could be used only to detect collisions if you want (that means no dynamic behavior). In this post, we want to share a bit how we are using both frameworks together.

Introduction

The main idea is to react to physics events, like two bodies colliding, to perform some game logic. For example, whenever the main character ship touches an asteroid, it explodes.

When you use Artemis, the game logic is done in an Artemis System or a Script (custom), if you use our customization. The ideal situation would be if you could check in your game logic which entities are in contact or not. In order to make that work, you have to find a way to link a Box2D contact with an Artemis Entity and vice versa.

Our solution

The first thing we do is, to each Artemis Entity we want to have a physics behavior, we add a PhysicsComponent the Box2D Body of the Entity and a Contacts instance where all the Box2D contacts for that Body are stored. Also, in order to get the Entity from the Body, we set the its userData pointing to the Entity.

The Contacts concept gives us useful methods to get information about contacts and the API looks like this:

    getContactsCount() : int - returns the contacts quantity
    getContact(index: int) : Contact - returns the contact information

And our Contact concept API, returned by the Contacts getContact() method, looks like this:

    getMyFixture() : Fixture - returns the fixture in contact of the Contacts owner Entity.
    getOtherFixture() : Fixture - returns the fixture of the other Entity.
    getNormal() : Vector2 - returns the normal of the contact.

(note: we decided to make a deep copy of the contacts information since it is recommended in the Box2D manual if you use a ContactsListener)

Then, we have a ContactsListener (named PhysicsListener) which, whenever a contact is reported (begin or end), it gets the bodies from the contact and gets the entities from each body userData and then adds or removes the contact data to/from each Entity's PhysicsComponent using its Contacts instance.

(note: we decided to use a custom ContactListener since it is recommended in the Box2D manual)

Finally, in each Artemis System or Script, we use the Entity's PhysicsComponent to get the contacts data and we proceed to do the logic we want, for example, destroy the character or enable some special ability, etc.

Here is an example of how we use it inside a Script from our Leave me Alone game:

public void update(World world, Entity e) {
	PhysicsComponent physicsComponent = Components.getPhysicsComponent(e);
	
	Contacts contacts = physicsComponent.getContact();
	
	if (!contacts.isInContact())
		return;
	
	boolean shouldExplode = false;
	
	for (int i = 0; i < contacts.getContactCount(); i++) {
		
		Contact contact = contacts.getContact(i);
		Entity otherEntity = (Entity) contact.getOtherFixture().getBody().getUserData();
		
		GroupComponent groupComponent = Components.getGroupComponent(otherEntity);
		
		if (groupComponent == null)
			continue;
		
		if (groupComponent.group.equals(Groups.EnemyCharacter)) {
			shouldExplode= true;
			break;
		}
		
	}
	
	if (shouldExplode)
		eventManager.dispatch(Events.MainExploded, e);
}

If you use Box2D and you are starting to use Artemis or vice versa, hope this post could help you. Otherwise, I hope you like it.

Also, if you use Artemis with Box2D in another way, would be great to have your point of view.

Thanks.

VN:F [1.9.22_1171]
Rating: 3.5/5 (11 votes cast)

Modifying textures using libGDX Pixmap in runtime - Explained

We have previously shown a bit how we were using LibGDX Pixmap to modify textures in runtime here and here for a game prototype we were doing. In this post I want to share more detail of how we do that. The objective was to make destructible terrain like in Worms 2.

Introduction

When you work with OpenGL textures, you can't directly modify their pixels whenever you want since they are on OpenGL context. To modify them you have to upload an array of bytes using glTexImage2D or glTexSubImage2D. The problem is you have to maintain on the application side an array of bytes representing the modifications you want to do.

To simplify working with byte arrays representing images, LibGDX provides a useful class named Pixmap which is a map of pixels kept in local memory with some methods to interact with a native library to perform all modifications with better performance.

Moving data from Pixmap to OpenGL Texture

In our prototypes, we wanted to remove part of the terrain whenever a missile touches it, like a Worms 2 explosion. That means we need some way to detect the collisions between the missile and the terrain and then a way to remove pixels from a texture.

We simplified the first problem by getting the color of the pixel only for the missile's position and checking if it was transparent or not. A more correct solution could be using a bitmap mask to check collisions between pixels but we wanted to simplify the work for now.

For the second problem, given a radius of explosion of the missile, we used the pixmap fillCircle method by previously setting the color to (0,0,0,0) (fully transparent) and disabled Pixmap blending to override those pixels.

But that only modified the pixmap data, now we needed to modify the OpenGL texture. To do that, we called OpenGL glTexImage2D using the bytes of the pixmap as the new texture data and that worked correctly.

Transforming from world coordinates to Pixmap coordinates

One problem when working with pixmaps is we have to map world coordinates (the position of the missile for example) to coordinates inside the Pixmap.


This image shows the coordinate system of the Pixmap, it goes from 0 to width in x and 0 to height in y.


This image shows how we normally need to move, rotate and resize the Pixmap in a game.

To solve this, we are using a LibGDX Sprite to maintain the Pixmap transformation, so we can easily move, rotate and scale it. Then, we can use that information to project a world coordinate to Pixmap coordinate by applying the inverse transform, here is the code:

	public void project(Vector2 position, float x, float y) {
		position.set(x, y);

		float centerX = sprite.getX() + sprite.getOriginX();
		float centerY = sprite.getY() + sprite.getOriginY();

		position.add(-centerX, -centerY);

		position.rotate(-sprite.getRotation());

		float scaleX = pixmap.getWidth() / sprite.getWidth();
		float scaleY = pixmap.getHeight() / sprite.getHeight();

		position.x *= scaleX;
		position.y *= scaleY;

		position.add( //
				pixmap.getWidth() * 0.5f, //
				-pixmap.getHeight() * 0.5f //
		);

		position.y *= -1f;
	}

(note: it is the first version at least, it could have bugs and could be improved also)

To simplify our work with all this stuff, we created a class named PixmapHelper which manage a Pixmap, a Texture and a Sprite, so we could move the Sprite wherever we wanted to and if we modify the pixmap through the PixmapHelper then the Texture was automatically updated and hence the Sprite (since it uses internally the Texture).

The next video shows how we tested the previous work in a prototype were we simulated cluster bombs (similar to Worms 2):

Some adjustments to improve performance

Instead of always working with a full size Pixmap by modifying it and then moved to the OpenGL texture, we created smaller Pixmaps of fixed sizes: 32x32, 64x64, etc. Then, each time we needed to make an explosion, we used the best Pixmap for that explosion and then we called glTexSubImage2D instead glTexImage2D to avoid updating untouched pixels. One limitation of this modification is we have to create and maintain several fixed size pixmaps depending on the modification size. Our current greater quad is 256x256 (almost never used).

Then, we changed to store each modification instead performing them in the moment the PixmapHelper erase method was called, and we added an update method which performs all modifications together. This improvement allow us to call Pixmap update method when we wanted, maybe one in three game updates or things like that.

Conclusion

Despite using LibGDX Pixmap for better performance, moving data to and from OpenGL context is not a cheap operation, on Android devices this could means some pauses when refreshing the modified textures with the new data. However, there is a lot of space for performance improvement, some ideas are to work only with pixmap of less bits instead RGBA8888 and use that one as the collisions context and as the mask of the real image (even using shaders), between other ideas.

Finally, the technique looks really nice and we believe that it could be used without problems for a simple game but it is not ready yet to manage a greater game like Worms 2.

Hope you like the technique and if you use it, maybe even also share your findings.

P.S.: In case you were wondering: yes, I love Worms 2.

VN:F [1.9.22_1171]
Rating: 3.9/5 (7 votes cast)

Reusing Artemis entities by enabling, disabling and storing them

As we mentioned on a previous post, we were having some performance issues in Vampire Runner and we were trying different approaches to improve its performance.

Introduction

One limitation of Android when making games is you have to avoid generating garbage whenever you can since the garbage collection would generate pauses on your games and that leads to a bad user experience. Then, we should try to reuse already created object instead of creating new ones.

In Vampire Runner, one problem we were having was that we were creating a lot of entities at a specific moment of the game, when we detected a new obstacle should be created, and that was making some pauses on the Android version.

As we use Artemis, we should try to reuse some entities when we can. For example, if we make a shooting game (like the Jetpac prototype I made) it seems a good idea to reuse bullets since their life cycle is really short. Ziggy made two blog posts about this topic some weeks ago here and here, however we followed a slightly different approach and we will explain it in this post.

Storing entities to reuse them

We created a concept named Store (similar to LibGDX Pool class) which let us easily store objects, in this case entities of one kind (for example bullets).

	free(T t) // returns an entity to the Store to be reused later

	get() : t // returns an entity from the Store, it reuses an object from the free 
			collection if there is one or creates a new object otherwise.

The idea is to, for example, instead of creating a new bullet when a weapon is fired, calling store.get() and set the component values as they should be, and when the bullet collides with something call the store.free(e) instead of deleting the entity, so we can reuse it later.

This is a generic approach and we can use different stores to reuse different kind of entities but it has a big problem, those entities keep being in Artemis world, that means they keep being processed (collisions, render, etc). A basic solution to this problem was adding a new state to the entity, and we explain that in the following section.

Enabling and disabling Artemis entities

Artemis supports reuse of entities by internally caching created entities inside the World class, however their state (which components their have) is not easily reused, and that was one of the big problems when creating a new entity, we wanted to reuse their state.

Our current solution to the problem was adding a new state to the entities, if they are enabled or not. Being enabled means the entity is processed by all interested EntitySystems, being disabled means the entity is still in the Artemis world but it is not processed by any system.

So, in our customization of Artemis we added three new methods to Entity to be called whenever you want to enable or disable an entity:

	disable() : disables an entity to avoid it to be processed on EntitySystems

	enable() : enables again an entity to let it be processed on EntitySystems

	isEnabled() :  returns true if the entity is enabled, false otherwise.

Then, we added new methods to EntitySystem API to let each EntitySystem to be aware an entity of interest was enabled or disabled:

	disabled(Entity e) : called whenever an entity of this EntitySystem was disabled

	enabled(Entity e) : called whenever an entity of this EntitySystem was disabled

In our case, we are using them to enable and disable Box2D bodies in our PhysicsSystem, and also to remove them from our render layers in our RenderSystem.

As an example, we have a nice video of Vampire Runner we made by changing the zoom of the camera to see the behind the scenes:

As you can see, when entities like wall, fire and Christmas stuff are behind the main character, they disappear. That is because they are disabled and moved again to their stores so they stop being processed by Artemis, in particular, stop being rendered.

Conclusion

By combining both solutions, we have an easy way to reuse created entities of one kind, like our obstacles tiles in Vampire Runner, while at the same time we can disable them when they are on a store to avoid them being processed.

In case of Vampire Runner, this solution improved Vampire Runner performance since we now pre create a lot of entities we need during the game and then disable them and enable them only when needed, in this way, we could avoid creating a lot of entities in one update after the game was started.

This is a first approach solution to the problem and seems good for our current games but it may not fit other type of games or bigger games, we don't know that yet.

If you use Artemis and you had this problem too, hope this blog post is helpful to you.

VN:F [1.9.22_1171]
Rating: 3.4/5 (5 votes cast)

Basic frustum culling to avoid rendering entities outside screen

As we were having some performance issues with Vampire Runner and we didn't have a clear idea of what was happening, we started trying some improvement techniques. The first one we implemented was a basic frustum culling technique to avoid trying to render objects outside of the screen.

Basic implementation

First, we created an Artemis component named FrustumCullingComponent with a Rectangle representing the bounds of that entity to easily detect if the entity is inside the screen or not. For now, as it is a basic implementation, the rectangle was only modified when the entity was created. So, for example, if we know an entity was able to rotate during the game, then we create a bigger bounding box using box diagonal.

Then, we added a method to our custom 2d Camera implementation to get the camera frustum (by making the corresponding transformations).

Finally, we modified our Artemis render system to check before rendering if an entity has or not a FrustumCullingComponent, if it hasn't one, then we perform the render logic as we always did. If it has one, then we check if the bounds of that entity overlaps with the camera frustum, if it does, then we render as we always did, if it doesn't, then we avoid rendering that entity.

Here is an example of the bounds and the frustum of the camera:

In the image, the element (a) and (b) are rendered because their bounds overlaps with the camera frustum. The element (c) is not rendered because its bounds are totally outside the camera frustum.

Conclusion

For Vampire Runner, we didn't notice the difference of having this technique enabled or not since the game always render fast (on our devices) and we had no metrics of the render process time. However, as it was really easy to implement this basic version of the technique, we believe it should help to maintain render performance, and we can reuse the logic for all of our games.

As always, hope you like it.

VN:F [1.9.22_1171]
Rating: 4.5/5 (2 votes cast)

Simulating parallax scrolling for 2d game

In this post we want to share how we simulated parallax scrolling in Vampire Runner, a simple technique that could be used in any 2d game.

Introduction

The basic idea to simulate parallax is to work with different render layers and to move them at different speeds.

To easily understand the post, we will work with a basic example based on Vampire Runner graphics where we have three layers, the first layer is the background which is static, the second layer are the mountains and the third layer is the world layer where the main character runs.

In Vampire Runner, the main character is the center of the world, despite being running, he is in a fixed position on the screen while all the other objects move accordingly to give the correct idea.

To simulate the parallax, each layer of objects moves at different speeds to give the correct idea of depth. In case of Vampire Runner, the background layer is static, the mountains layer moves at 1/12 the speed of the main character and the world layer moves at the speed of the main character.

Simulating movement with cameras

We said before that the character is in a fixed position on the screen, however the real entity of the character is moving around the world, even more, he is running!.

To simulate him being in a fixed position, we have a camera following the character, so the player always see the character in the same position.

However, as all the other objects (obstacles, trees, etc) have a fixed position in the world when the camera follows the main character they move in the opposite direction. That is how we move the world layer and the mountains layer. The only difference is we move the mountains layer camera slower.

In conclusion, we have one camera for each layer and two of them move following the main character.

Using OpenGL model view matrices to configure the cameras

As we are using OpenGL through LibGDX (and LWJGL on desktop), we need to translate all the previous camera behavior to OpenGL.

In OpenGL, to simulate a world camera what you typically modify the model view matrix and then render the objects.

So, for Vampire Runner, we had three different matrices to render each layer, and the render method was something like this:

    1.1 configure background layer camera
    1.2 render background layer

    2.1 configure mountains layer camera
    2.2 render mountains layer

    3.1 configure world layer camera
    3.2 render world layer

In the update method we only had to update the cameras depending on the character position.

To simplify the work, if you are using LibGDX, you can use the Camera class which comes with a proper class for 2d games named OrthographicCamera. In our case we used a custom implementation which provided a nicer API to work with.

Conclusion

Working with different cameras with OpenGL is not so hard and it let you achieve a correct feeling when moving across the world.

As always, hope you like the post.

VN:F [1.9.22_1171]
Rating: 2.0/5 (4 votes cast)

Resizable Applets and Applications with LWJGL

Some posts before, in the comments section, Gornova asked me to write about resizable Applets if I had some time. As I never used resizable LWJGL Applets, I had to do some research. This post is to show what I found.

(note: LWJGL version 2.8.2 was used for all the tests I made).

Introduction

LWJGL Applets are similar to LWJGL Applications, they work using a class named Display which uses an AWT Canvas component. On desktop, you create a JFrame or similar and add inside a Canvas, on Applet you just create the Canvas and add it to the Applet. Unless you want some custom behavior, I believe that is the common usage.

(note: using fullscreen is a different case but has no meaning for this post)

How to resize an Applet

To resize an Applet on the browser, two possible ways are, use Javascript to change the applet tag sizes or just configure the HTML applet tag to adapt to its container size, so if you resize the browser window then all corresponding HTML components are resized.

The other way around, suppose you have a way inside your game to choose the game resolution, you will have to communicate that to the Browser, probably calling a Javascript function from your Applet to let the Browser react to the change.

How to handle Applet resize

If you are working with LWJGL directly, then it depends a lot of how which kind of layout are you using for the container of the Canvas object. For example, if you chose a layout which grows to the parent container size, then each time you resize the container, the Canvas will be resized. If you chose another layout, then maybe you will call Canvas.setSize() by hand.

However, you need to react to that change in your game, to do that the best way (imho) is use a similar approach LibGDX does. What LibGDX does is, its stores the current Canvas size on each iteration of the main application loop, on the next iteration, if the Canvas size is different to the stored, then it calls a resize() method to tell your game the screen size changed.

Another way is to create a ComponentListener and add that to the Canvas component so each time the componentResized() method is called you resize your game. However as all the AWT stuff runs in another thread you need to store the new size and process the resize in your LWJGL thread to avoid problems.

What to do inside your game

The very basic approach is to just modify the OpenGL viewport of your game, then the game is scaled to the new size and, in some cases, that should be enough.

	glViewport(0, 0, width, height);

However, You may want to take care about aspect ratio since just scaling the game could look ugly. One option is to modify the viewport in values to maintain the aspect ratio you want, for example:

	glViewport(0, 0, width, width / aspectRatio);

You could also provide different common aspect ratios (4:3, 16:10, 16:9, etc). However, I believe the best option is to provide fixed commonly used resolutions like 640x480 or 800x600 instead.

In some cases you could want to adapt your game elements, for example GUI stuff, where you may want to maintain the element sizes and move them accordingly over the screen instead just scaling them.

Then, you have to take care of the user input. As mouse location is dependent on the viewport size, you will have to take that in mind when interacting with part of your application, like the GUI stuff, after a size change. For example, LibGDX Camera uses the current viewport size when using project and unproject methods (which allow you to transform world coordinates to viewport coordinates and vice versa).

Not so sure as I am not an OpenGL expert but if you work different cameras (model + projection matrices) you probably need to update them in order to render everything in the correct way. I have to go deeper in this issue.

Optionally, you may want make other changes when reacting to the resize like loading HD assets if you go to a higher resolution and vice versa.

Some issues

  • You have some glitches if you do resize in real time (tested myself, but cant find references right now). However, in general nobody does this so it is not a big problem.
  • I believe Applet resizing is not working on mac yet.

Conclusion

Handling all the mentioned things when resizing LWJGL Applets and LWJGL Applications depends a lot on the game you are making, in some cases the basic approach work fine and should be easy to implement.

Despite the article was focused mainly on LWJGL Applets, great part of the findings can be applied to LWJGL Applications as well.

This post is only to show some basic experimentation I made on the subject, as I said before, I am not an OpenGL nor LWJGL expert, so some assertions could be wrong. If you have more experience and want to contribute your opinion, feel free to do it.

Thanks and hope you like the post.

VN:F [1.9.22_1171]
Rating: 5.0/5 (1 vote cast)

Making buttons using Inkscape - part 3

This is the third part of the buttons tutorial series, in this case I will show how to make more complex shapes to be used inside the buttons.

As I did on previous posts, I will show images with the steps to follow to make each icon.


This image explains the steps to follow to make a question mark icon.

This image explains the steps to follow to make an icon for high scores.

Finally, a time lapse video showing how I made all the stuff:

Here is the link to download the question mark SVG and here is the link to download the high scores icon SVG.

As always, hope you like it.

VN:F [1.9.22_1171]
Rating: 5.0/5 (2 votes cast)

Making buttons using Inkscape - part 2

Following the buttons tutorials, in this case I want to show how to make square buttons and how to make simple inner shapes to be used over the button base.

First, the next image tutorial explains a bit how to make buttons to increase/decrease values by using plus and minus symbols.

Here is an example of using it to modify the volume of a game:

The next image shows how to make square buttons with a similar style used for the circular buttons of the previous tutorial.

And here is a time lapse to show how I made the square button:


(note: watch it in 720p, youtube killed the video when converting it to 360p and 480p)

If you want to use this templates, you can access here for the SVG file used for the tutorial, and here is the SVG with several shapes to be used with these buttons.

Finally, I made a new icon to show our android market games at the icons widget in the right side bar, if you like it, feel free to use it. Here is how it looks like:

Hope you like the post. Next time make complex shapes like the high scores goblet/cup.

VN:F [1.9.22_1171]
Rating: 5.0/5 (5 votes cast)

Making buttons using Inkscape - part 1

Inspired by the Gamasutra article named 2D Game Art For Programmers part 1, I was making some stuff using Inkscape and I want to share my learnings.

In this case, I want to share how I made the background for buttons to be used in games. As the article's author did, I created an image explaining step by step:

Creating a button step by step
(note: click the image for full resolution)

After adding content to the background, we could create stuff like this:

Button examples

I will create another post explaining how to make those button content shapes.

You can download the SVG template from this link if you want to use it, and or modify it.

Hope you like this post, because I am planning to make more :D.

VN:F [1.9.22_1171]
Rating: 5.0/5 (4 votes cast)