Search Results

Search found 31839 results on 1274 pages for 'plugin development'.

Page 536/1274 | < Previous Page | 532 533 534 535 536 537 538 539 540 541 542 543  | Next Page >

  • GLSL billboard move center of rotation

    - by Jacob Kofoed
    I have successfully set up a billboard shader that works, it can take in a quad and rotate it so it always points toward the screen. I am using this vertex-shader: void main(){ vec4 tmpPos = (MVP * bufferMatrix * vec4(0.0, 0.0, 0.0, 1.0)) + (MV * vec4( vertexPosition.x * 1.0 * bufferMatrix[0][0], vertexPosition.y * 1.0 * bufferMatrix[1][1], vertexPosition.z * 1.0 * bufferMatrix[2][2], 0.0) ); UV = UVOffset + vertexUV * UVScale; gl_Position = tmpPos; BufferMatrix is the model-matrix, it is an attribute to support Instance-drawing. The problem is best explained through pictures: This is the start position of the camera: And this is the position, looking in from 45 degree to the right: Obviously, as each character is it's own quad, the shader rotates each one around their own center towards the camera. What I in fact want is for them to rotate around a shared center, how would I do this? What I have been trying to do this far is: mat4 translation = mat4(1.0); translation = glm::translate(translation, vec3(pos)*1.f * 2.f); translation = glm::scale(translation, vec3(scale, 1.f)); translation = glm::translate(translation, vec3(anchorPoint - pos) / vec3(scale, 1.f)); Where the translation is the bufferMatrix sent to the shader. What I am trying to do is offset the center, but this might not be possible with a single matrix..? I am interested in a solution that doesn't require CPU calculations each frame, but rather set it up once and then let the shader do the billboard rotation. I realize there's many different solutions, like merging all the quads together, but I would first like to know if the approach with offsetting the center is possible. If it all seems a bit confusing, it's because I'm a little confused myself.

    Read the article

  • How to show other characters in online 2D rpg

    - by Loligans
    I have Player 1 and Player 2 I am using Json to send and retrieve player data between the client and the server, but when another player logs in, and is in the same map, how would I send that data to both players to update the graphics engine to show there are 2 Players on the map? About my game it is a 2D RPG tile based game it is 24x15 Tiles it is Real time Action it should interact anywhere between 10-150 ping players interact with each other when in the same map and can see each other moving around the game world is persistent, and is saved when the server shuts down Right now the server just sends the player Only their information which is inside a Json Object Here is an example of what I am talking about If you notice there are 2 separate characters in 2 separate clients, but they are running on the same server. I am trying to get them to show up on both clients, but I don't know how I should accomplish this. Should I send it as an added value in the Json object? Also what is the name of this process so I can look it up and find more info on it?

    Read the article

  • slick2d missiles

    - by kirchhoff
    Hey I'm making a game in java with slick2d and I want to create planes which shoots: int maxBullets = 40; static int bullet = 0; Missile missile[] = new Missile[maxBullets]; I want to create/move my missiles in the most efficient way, I would appreciate your advises: public void shoot() throws SlickException{ if(bullet<maxBullets){ if(missile[bullet] != null){ missile[bullet].resetLocation(plane.getCentreX(), plane.getCentreY(), plane.image.getRotation()); }else{ missile[bullet] = new Missile("resources/missile.png", plane.getCentreX(), plane.getCentreY(), plane.image.getRotation()); } }else{ bullet = 0; missile[bullet].resetLocation(plane.getCentreX(), plane.getCentreY(), plane.image.getRotation()); } bullet++; } I created the method "resetLocation" in my Missile class in order to avoid loading again the resource. Is it correct? In the update method I've got this to move all the missiles: if(bullet > 0 && bullet < maxBullets){ float hyp = 0.4f * delta; if(bullet == 1){ missile[0].move(hyp); }else{ for(int x = 0; x<bullet; x++){ missile[x].move(hyp); } } }

    Read the article

  • circle - rectangle collision in 2D, most efficient way

    - by john smith
    Suppose I have a circle intersecting a rectangle, what is ideally the least cpu intensive way between the two? method A calculate rectangle boundaries loop through all points of the circle and, for each of those, check if inside the rect. method B calculate rectangle boundaries check where the center of the circle is, compared to the rectangle make 9 switch/case statements for the following positions: top, bottom, left, right top left, top right, bottom left, bottom right inside rectangle check only one distance using the circle's radius depending on where the circle happens t be. I know there are other ways that are definitely better than these two, and if could point me a link to them, would be great but, exactly between those two, which one would you consider to be better, regarding both performance and quality/precision? Thanks in advance.

    Read the article

  • Particle and Physics problem.

    - by Quincy
    This was originally a forum post so I hope you guys don't mind it being 2 questions in one. I am making a game and I got some basic physics implemented. I have 2 problems, 1 with particles being drawn in the wrong place and one with going through walls while jumping in corners. Skip over to about 15 sec video showing the 2 problems : http://youtube.com/watch?v=Tm9nfWsWfiM So the problem with the particles seems to be coming from the removal, as soon as I remove that piece of code it instantly works, but there shouldn't be a problem since they shouldn't even draw when their energy gets to 0 (and then they get removed) So my first question is, how are these particles getting warped all over the screen ? Relevant code : Particle class : class Particle { //Physics public Vector2 position = new Vector2(0,0); public float direction = 180; public float speed = 100; public float energy = 1; protected float startEnergy = 1; //Visual public Sprite sprite; public float rotation = 0; public float scale = 1; public byte alpha = 255; public BlendMode blendMode { get { return sprite.BlendMode; } set { sprite.BlendMode = value; } } public Particle() { } public virtual void Think(float frameTime) { if (energy - frameTime < 0) energy = 0; else energy -= frameTime; position += new Vector2((float)Math.Cos(MathHelper.DegToRad(direction)), (float)Math.Sin(MathHelper.DegToRad(direction))) * speed * frameTime; alpha = (byte)(255 * energy / startEnergy); sprite.Rotation = rotation; sprite.Position = position; sprite.Color = new Color(sprite.Color.R, sprite.Color.G, sprite.Color.B, alpha); } public virtual void Draw(float frameTime) { if (energy > 0) { World.camera.DrawSprite(sprite); } } // Basic particle implementation class BasicSprite : Particle { public BasicSprite(Sprite _sprite) { sprite = _sprite; } } Emitter : class Emitter { protected static Random rand = new Random(); protected List<Particle> particles = new List<Particle>(); public BaseEntity target = null; public Vector2 position = new Vector2(0, 0); public bool Active = true; public float timeAlive = 0; public int particleCount = 0; public int ParticlesPerSeccond { get { return (int)(1 / particleSpawnTime); } set { particleSpawnTime = 1 / (float)value; } } public float dieTime = float.MaxValue; float particleSpawnTime = 0.05f; float spawnTime = 0; public Emitter() { } public virtual void Think(float frametime) { spawnTime += frametime; if (dieTime != float.MaxValue) { timeAlive += frametime; if (timeAlive >= dieTime) Active = false; } if (Active) { if (target != null) position = target.Position; while (spawnTime > particleSpawnTime) { spawnTime -= particleSpawnTime; AddParticle(); particleCount++; } } for (int i = 0; i < particles.Count; i++) { particles[i].Think(frametime); if (particles[i].energy <= 0) { particles.Remove(particles[i]); // As soon as this is removed, it works particleCount--; } } } public virtual void AddParticle() { } public virtual void Draw(float frametime) { foreach (Particle particle in particles) { particle.Draw(frametime); } } } class BloodEmitter : Emitter { Image image; public BloodEmitter() { image = new Image(@"Content/Particles/TinyCircle.png"); image.CreateMaskFromColor(new Color(255, 0, 255, 255)); this.dieTime = 0.5f; this.ParticlesPerSeccond = 100; } public override void AddParticle() { Sprite sprite = new Sprite(image); sprite.Color = new Color((byte)(rand.NextDouble() * 255), (byte)(rand.NextDouble() * 255), (byte)(rand.NextDouble() * 255)); BasicSprite particle = new BasicSprite(sprite); particle.direction = (float)rand.NextDouble() * 360; particle.position = position; particle.blendMode = BlendMode.Alpha; particles.Add(particle); } } The seccond problem is the physics problem, for some reason I can get through the right bottom corner while jumping. I think this is coming from me switching animations but I thought I made it compensate for that. Relevant code : PhysicsEntity : class PhysicsEntity : BaseEntity { // Horizontal movement constants protected const float maxHorizontalSpeed = 1000; protected const float horizontalAcceleration = 15; protected const float horizontalDragAir = 0.95f; protected const float horizontalDragGround = 0.95f; // Vertical movement constants protected const float maxVerticalSpeed = 1000; protected const float verticalAcceleration = 20; // Everything needed for movement and correct animations protected float movement = 0; protected bool onGround = false; protected Vector2 Velocity = new Vector2(0, 0); protected float maxSpeed = 0; float lastThink = 0; float thinkTime = 1f/60f; public PhysicsEntity(Vector2 position, Sprite sprite) : base(position, sprite) { } public override void Draw(float frameTime) { base.Draw(frameTime); } public override void Think(float frameTime) { CalculateMovement(frameTime); base.Think(frameTime); } protected void CalculateMovement(float frameTime) { lastThink += frameTime; while (lastThink > thinkTime) { onGround = false; Velocity.X = MathHelper.Clamp(Velocity.X + horizontalAcceleration * movement, -maxHorizontalSpeed, maxHorizontalSpeed); if (onGround) Velocity.X *= horizontalDragGround; else Velocity.X *= horizontalDragAir; if (maxSpeed < Velocity.X) maxSpeed = Velocity.X; Velocity.Y = MathHelper.Clamp(Velocity.Y + verticalAcceleration, -maxVerticalSpeed, maxVerticalSpeed); lastThink -= thinkTime; DoCollisions(thinkTime); DoAnimations(thinkTime); } } public virtual void DoAnimations(float frameTime) { } public void DoCollisions(float frameTime) { Position.Y += Velocity.Y * frameTime; Vector2 tileCollision = GetTileCollision(); if (tileCollision.X != -1 || tileCollision.Y != -1) { Vector2 collisionDepth = CollisionRectangle.DepthIntersection( new Rectangle( tileCollision.X * World.tileEngine.TileWidth, tileCollision.Y * World.tileEngine.TileHeight, World.tileEngine.TileWidth, World.tileEngine.TileHeight ) ); Position.Y += collisionDepth.Y; if (collisionDepth.Y < 0) onGround = true; Velocity.Y = 0; } Position.X += Velocity.X * frameTime; tileCollision = GetTileCollision(); if (tileCollision.X != -1 || tileCollision.Y != -1) { Vector2 collisionDepth = CollisionRectangle.DepthIntersection( new Rectangle( tileCollision.X * World.tileEngine.TileWidth, tileCollision.Y * World.tileEngine.TileHeight, World.tileEngine.TileWidth, World.tileEngine.TileHeight ) ); Position.X += collisionDepth.X; Velocity.X = 0; } } public void DoCollisions(Vector2 difference) { CollisionRectangle.Y = Position.Y - difference.Y; CollisionRectangle.Height += difference.Y; Vector2 tileCollision = GetTileCollision(); if (tileCollision.X != -1 || tileCollision.Y != -1) { Vector2 collisionDepth = CollisionRectangle.DepthIntersection( new Rectangle( tileCollision.X * World.tileEngine.TileWidth, tileCollision.Y * World.tileEngine.TileHeight, World.tileEngine.TileWidth, World.tileEngine.TileHeight ) ); Position.Y += collisionDepth.Y; if (collisionDepth.Y < 0) onGround = true; Velocity.Y = 0; } CollisionRectangle.X = Position.X - difference.X; CollisionRectangle.Width += difference.X; tileCollision = GetTileCollision(); if (tileCollision.X != -1 || tileCollision.Y != -1) { Vector2 collisionDepth = CollisionRectangle.DepthIntersection( new Rectangle( tileCollision.X * World.tileEngine.TileWidth, tileCollision.Y * World.tileEngine.TileHeight, World.tileEngine.TileWidth, World.tileEngine.TileHeight ) ); Position.X += collisionDepth.X; Velocity.X = 0; } } Vector2 GetTileCollision() { int topLeftTileX = (int)(CollisionRectangle.TopLeft.X / World.tileEngine.TileWidth); int topLeftTileY = (int)(CollisionRectangle.TopLeft.Y / World.tileEngine.TileHeight); int BottomRightTileX = (int)(CollisionRectangle.DownRight.X / World.tileEngine.TileWidth); int BottomRightTileY = (int)(CollisionRectangle.DownRight.Y / World.tileEngine.TileHeight); if (CollisionRectangle.DownRight.Y % World.tileEngine.TileHeight == 0) // If your exactly against the tile don't count that as being inside the tile BottomRightTileY -= 1; if (CollisionRectangle.DownRight.X % World.tileEngine.TileWidth == 0) // If your exactly against the tile don't count that as being inside the tile BottomRightTileX -= 1; for (int i = topLeftTileX; i <= BottomRightTileX; i++) { for (int j = topLeftTileY; j <= BottomRightTileY; j++) { if (World.tileEngine.TileIsSolid(i, j)) { return new Vector2(i, j); } } } return new Vector2(-1, -1); } } Player : enum State { Standing, Running, Jumping, Falling, Sliding, WallSlide } class Player : PhysicsEntity { private State state { get { return currentState; } set { if (currentState != value) { currentState = value; animationChanged = true; } } } private State currentState = State.Standing; private BasicEmitter basicEmitter = new BasicEmitter(); public bool flipped; public bool animationChanged = false; protected const float jumpPower = 600; AnimationManager animationManager; Rectangle DrawRectangle; public override Rectangle CollisionRectangle { get { return new Rectangle( Position.X - DrawRectangle.Width / 2f, Position.Y - DrawRectangle.Height / 2f, DrawRectangle.Width, DrawRectangle.Height ); } } public Player(Vector2 position, Sprite sprite) : base(position, sprite) { // Only posted the relevant bit DrawRectangle = animationManager.currentAnimation.drawingRectangle; } public override void Draw(float frameTime) { World.camera.DrawSprite( Sprite, Position + new Vector2(DrawRectangle.X, DrawRectangle.Y), animationManager.currentAnimation.drawingRectangle ); } public override void Think(float frameTime) { //I only posted the relevant stuff if (animationChanged) { // if the animation has changed make sure we compensate for the change in with and height animationChanged = false; DoCollisions(animationManager.getSizeDifference()); } DoCustomMovement(); base.Think(frameTime); if (!onGround && Velocity.Y > 0) { state = State.Falling; } } void DoCustomMovement() { if (onGround) { if (World.renderWindow.Input.IsKeyDown(KeyCode.W)) { Velocity.Y = -jumpPower; state = State.Jumping; } } } public override void DoAnimations(float frameTime) { string stateName = Enum.GetName(typeof(State), state); if (!animationManager.currentAnimationIs(stateName)) { animationManager.PlayAnimation(stateName); } animationManager.Think(frameTime); DrawRectangle = animationManager.currentAnimation.drawingRectangle; Sprite.Center = new Vector2( DrawRectangle.X + DrawRectangle.Width / 2, DrawRectangle.Y + DrawRectangle.Height / 2 ); Sprite.FlipX(flipped); } So why am I warping through walls ? I have given this some thought but I just can't seem to find out why this is happening. Full source if needed : source : http://www.mediafire.com/?rc7ddo09gnr68zd (download link)

    Read the article

  • Workaround the flip queue (AKA pre-rendered frames) in OpenGL?

    - by user41500
    It appears that some drivers implement a "flip queue" such that, even with vsync enabled, the first few calls to swap buffers return immediately (queuing those frames for later use). It is only after this queue is filled that buffer swaps will block to synchronize with vblank. This behavior is detrimental to my application. It creates latency. Does anyone know of a way to disable it or a workaround for dealing with it? The OpenGL Wiki on Swap Interval suggests a call to glFinish after the swap but I've had no such luck with that trick.

    Read the article

  • Pixel perfect collision with paths (Android)

    - by keysersoze
    Hi I'm writing a game and I'm trying to do some pixel perfect collisions with paths. The player's character has a bitmask which looks for example like this: Currenly my code that handles player's collision with path looks like this: private boolean isTerrainCollisionDetected() { if(collisionRegion.op(player.getBounds(), terrain.getBottomPathBounds(), Region.Op.INTERSECT) || collisionRegion.op(player.getBounds(), terrain.getTopPathBounds(), Region.Op.INTERSECT)) { collisionRegion.getBounds(collisionRect); for(int i = collisionRect.left; i < collisionRect.right; i++) { for(int j = collisionRect.top; j < collisionRect.bottom; j++) { if(player.getBitmask().getPixel(i - player.getX(), j - player.getY()) != Color.TRANSPARENT) { return true; } } } } return false; } The problem is that collisions aren't pixel perfect. It detects collisions in situations like this: The question is: what can I do to improve my collision detection?

    Read the article

  • How can I generate a texture that looks like left-over tea leaves?

    - by Jedidja
    We are working on a project for iPhone and Windows Phone 7 where we'd like to be able to generate tea leaves at the bottom of a cup. It doesn't have to look photo-realistic, and actually cartoon-y is ok. What sort of techniques should we research to accomplish this? Are there any libraries (preferably in C, but we can translate) that would be helpful? Here are some samples pulled from a Google Image search

    Read the article

  • Newton Game Dynamics: Making an object not affect another object

    - by Boreal
    I'm going to be using Newton in my networked action game with Mogre. There will be two "types" of physics object: global and local. Global objects will be kept in sync for everybody; these include the players, projectiles, and other gameplay-related objects. Local objects are purely for effect, like ragdolls, debris, and particles. Is there a way to make the global objects affect the local objects without actually getting affected themselves? I'd like debris to bounce off of a tank, but I don't want the tank to respond in any way.

    Read the article

  • how to label a cuboid using open gl?

    - by usha
    hi this is how my 3dcuboid looks ,i have attached complete code , i want to label this cuboid using different name across sides how is it possible using opengl in android...plz help me out public class MyGLRenderer implements Renderer { Context context; Cuboid rect; private float mCubeRotation; // private static float angleCube = 0; // Rotational angle in degree for cube (NEW) // private static float speedCube = -1.5f; // Rotational speed for cube (NEW) public MyGLRenderer(Context context) { rect = new Cuboid(); this.context = context; } public void onDrawFrame(GL10 gl) { // TODO Auto-generated method stub gl.glClear(GL10.GL_COLOR_BUFFER_BIT | GL10.GL_DEPTH_BUFFER_BIT); gl.glLoadIdentity(); // Reset the model-view matrix gl.glTranslatef(0.2f, 0.0f, -8.0f); // Translate right and into the screen gl.glScalef(0.8f, 0.8f, 0.8f); // Scale down (NEW) gl.glRotatef(mCubeRotation, 1.0f, 1.0f, 1.0f); // gl.glRotatef(angleCube, 1.0f, 1.0f, 1.0f); // rotate about the axis (1,1,1) (NEW) rect.draw(gl); mCubeRotation -= 0.15f; //angleCube += speedCube; } public void onSurfaceChanged(GL10 gl, int width, int height) { // TODO Auto-generated method stub if (height == 0) height = 1; // To prevent divide by zero float aspect = (float)width / height; // Set the viewport (display area) to cover the entire window gl.glViewport(0, 0, width, height); // Setup perspective projection, with aspect ratio matches viewport gl.glMatrixMode(GL10.GL_PROJECTION); // Select projection matrix gl.glLoadIdentity(); // Reset projection matrix // Use perspective projection GLU.gluPerspective(gl, 45, aspect, 0.1f, 100.f); gl.glMatrixMode(GL10.GL_MODELVIEW); // Select model-view matrix gl.glLoadIdentity(); // Reset } public void onSurfaceCreated(GL10 gl, EGLConfig config) { // TODO Auto-generated method stub gl.glClearColor(0.0f, 0.0f, 0.0f, 1.0f); // Set color's clear-value to black gl.glClearDepthf(1.0f); // Set depth's clear-value to farthest gl.glEnable(GL10.GL_DEPTH_TEST); // Enables depth-buffer for hidden surface removal gl.glDepthFunc(GL10.GL_LEQUAL); // The type of depth testing to do gl.glHint(GL10.GL_PERSPECTIVE_CORRECTION_HINT, GL10.GL_NICEST); // nice perspective view gl.glShadeModel(GL10.GL_SMOOTH); // Enable smooth shading of color gl.glDisable(GL10.GL_DITHER); // Disable dithering for better performance }} public class Cuboid{ private FloatBuffer mVertexBuffer; private FloatBuffer mColorBuffer; private ByteBuffer mIndexBuffer; private float vertices[] = { //width,height,depth -2.5f, -1.0f, -1.0f, 1.0f, -1.0f, -1.0f, 1.0f, 1.0f, -1.0f, -2.5f, 1.0f, -1.0f, -2.5f, -1.0f, 1.0f, 1.0f, -1.0f, 1.0f, 1.0f, 1.0f, 1.0f, -2.5f, 1.0f, 1.0f }; private float colors[] = { // R,G,B,A COLOR 0.0f, 1.0f, 0.0f, 1.0f, 0.0f, 1.0f, 0.0f, 1.0f, 1.0f, 0.5f, 0.0f, 1.0f, 1.0f, 0.5f, 0.0f, 1.0f, 1.0f, 0.0f, 0.0f, 1.0f, 1.0f, 0.0f, 0.0f, 1.0f, 0.0f, 0.0f, 1.0f, 1.0f, 1.0f, 0.0f, 1.0f, 1.0f }; private byte indices[] = { // VERTEX 0,1,2,3,4,5,6,7 REPRESENTATION FOR FACES 0, 4, 5, 0, 5, 1, 1, 5, 6, 1, 6, 2, 2, 6, 7, 2, 7, 3, 3, 7, 4, 3, 4, 0, 4, 7, 6, 4, 6, 5, 3, 0, 1, 3, 1, 2 }; public Cuboid() { ByteBuffer byteBuf = ByteBuffer.allocateDirect(vertices.length * 4); byteBuf.order(ByteOrder.nativeOrder()); mVertexBuffer = byteBuf.asFloatBuffer(); mVertexBuffer.put(vertices); mVertexBuffer.position(0); byteBuf = ByteBuffer.allocateDirect(colors.length * 4); byteBuf.order(ByteOrder.nativeOrder()); mColorBuffer = byteBuf.asFloatBuffer(); mColorBuffer.put(colors); mColorBuffer.position(0); mIndexBuffer = ByteBuffer.allocateDirect(indices.length); mIndexBuffer.put(indices); mIndexBuffer.position(0); } public void draw(GL10 gl) { gl.glFrontFace(GL10.GL_CW); gl.glVertexPointer(3, GL10.GL_FLOAT, 0, mVertexBuffer); gl.glColorPointer(4, GL10.GL_FLOAT, 0, mColorBuffer); gl.glEnableClientState(GL10.GL_VERTEX_ARRAY); gl.glEnableClientState(GL10.GL_COLOR_ARRAY); gl.glDrawElements(GL10.GL_TRIANGLES, 36, GL10.GL_UNSIGNED_BYTE, mIndexBuffer); gl.glDisableClientState(GL10.GL_VERTEX_ARRAY); gl.glDisableClientState(GL10.GL_COLOR_ARRAY); } } public class Draw3drect extends Activity { private GLSurfaceView glView; // Use GLSurfaceView // Call back when the activity is started, to initialize the view @Override protected void onCreate(Bundle savedInstanceState) { super.onCreate(savedInstanceState); glView = new GLSurfaceView(this); // Allocate a GLSurfaceView glView.setRenderer(new MyGLRenderer(this)); // Use a custom renderer this.setContentView(glView); // This activity sets to GLSurfaceView } // Call back when the activity is going into the background @Override protected void onPause() { super.onPause(); glView.onPause(); } // Call back after onPause() @Override protected void onResume() { super.onResume(); glView.onResume(); } }

    Read the article

  • Is there a simple way to stop enemies standing in the same spot?

    - by Iain
    So: top-down game, my enemies chase the player, when they get within a certain distance they stand still and fire. If they're all coming from the same direction they all end up standing in the same spot (i.e. standing "within" each other), as I'm not currently doing collision detection between enemies - they are free to pass over each other. What's a simple way around this? Either some form of collision detection or some ai?

    Read the article

  • Isometric Camera trouble - can't rotate or move correctly

    - by Deukalion
    I'm trying to create a 3D editor, but I've been having some trouble with the Camera and understanding each component. I've created 2 camera that works OK, but now I'm trying to implement an Isometric Camera in XNA without success on the rotation and movement of the camera. All I get working is Zoom. (Cube with x=3f, y=3f, z=1f in center) And this is the constructor for my IsometricCamera (inherits from ICamera, with methods for Rotation, Movement and Zoom, and Properties for World/View/Projection matrices) public IsometricCamera3D(GraphicsDevice device, float startClip = -1000f, float endClip = 1000f) { matrix_projection = Matrix.CreateOrthographic(device.Viewport.Width, device.Viewport.Height, startClip, endClip); rotation = Vector3.Zero; matrix_view = Matrix.CreateScale(zoom) * Matrix.CreateRotationY(MathHelper.ToRadians(45 + 180)) * Matrix.CreateRotationX(MathHelper.ToRadians(30)) * Matrix.CreateRotationZ(MathHelper.ToRadians(120)) * Matrix.CreateTranslation(rotation.X, rotation.Y, rotation.Z); } Problem is when I rotate it, all that happens is that the Cube gets more or less shiny and nothing happens. What is wrong and how should I create my View matrix to move it / rotate it correctly? Rotate, Move and Zoom looks like: MethodName(Vector3 rotation/movement), Zoom(float value); and just increases the value, then calls an update to recreate the View Matrix according to the code in the constructor. Currently, in my editor I use MiddleButton + Mouse Movement to rotate the camera, but it's not working as the other camera. But in my default camera I use World Matrix to move, but I guess that's not the best way to go which is why I'm trying this.

    Read the article

  • Impact of variable-length loops on GPU shaders

    - by Will
    Its popular to render procedural content inside the GPU e.g. in the demoscene (drawing a single quad to fill the screen and letting the GPU compute the pixels). Ray marching is popular: This means the GPU is executing some unknown number of loop iterations per pixel (although you can have an upper bound like maxIterations). How does having a variable-length loop affect shader performance? Imagine the simple ray-marching psuedocode: t = 0.f; while(t < maxDist) { p = rayStart + rayDir * t; d = DistanceFunc(p); t += d; if(d < epsilon) { ... emit p return; } } How are the various mainstream GPU families (Nvidia, ATI, PowerVR, Mali, Intel, etc) affected? Vertex shaders, but particularly fragment shaders? How can it be optimised?

    Read the article

  • Understanding math used to determine if vector is clockwise / counterclockwise from your vector

    - by MTLPhil
    I'm reading Programming Game AI by Example by Mat Buckland. In the Math & Physics primer chapter there's a listing of the declaration of a class used to represent 2D vectors. This class contains a method called Sign. It's implementation is as follows //------------------------ Sign ------------------------------------------ // // returns positive if v2 is clockwise of this vector, // minus if anticlockwise (Y axis pointing down, X axis to right) //------------------------------------------------------------------------ enum {clockwise = 1, anticlockwise = -1}; inline int Vector2D::Sign(const Vector2D& v2)const { if (y*v2.x > x*v2.y) { return anticlockwise; } else { return clockwise; } } Can someone explain the vector rules that make this hold true? What do the values of y*v2.x and x*v2.y that are being compared actually represent? I'd like to have a solid understanding of why this works rather than just accepting that it does without figuring it out. I feel like it's something really obvious that I'm just not catching on to. Thanks for your help.

    Read the article

  • Storing game objects with generic object information

    - by Mick
    In a simple game object class, you might have something like this: public abstract class GameObject { protected String name; // other properties protected double x, y; public GameObject(String name, double x, double y) { // etc } // setters, getters } I was thinking, since a lot of game objects (ex. generic monsters) will share the same name, movement speed, attack power, etc, it would be better to have all that information shared between all monsters of the same type. So I decided to have an abstract class "ObjectData" to hold all this shared information. So whenever I create a generic monster, I would use the same pre-created "ObjectData" for it. Now the above class becomes more like this: public abstract class GameObject { protected ObjectData data; protected double x, y; public GameObject(ObjectData data, double x, double y) { // etc } // setters, getters public String getName() { return data.getName(); } } So to tailor this specifically for a Monster (could be done in a very similar way for Npcs, etc), I would add 2 classes. Monster which extends GameObject, and MonsterData which extends ObjectData. Now I'll have something like this: public class Monster extends GameObject { public Monster(MonsterData data, double x, double y) { super(data, x, y); } } This is where my design question comes in. Since MonsterData would hold data specific to a generic monster (and would vary with what say NpcData holds), what would be the best way to access this extra information in a system like this? At the moment, since the data variable is of type ObjectData, I'll have to cast data to MonsterData whenever I use it inside the Monster class. One solution I thought of is this, but this might be bad practice: public class Monster extends GameObject { private MonsterData data; // <- this part here public Monster(MonsterData data, double x, double y) { super(data, x, y); this.data = data; // <- this part here } } I've read that for one I should generically avoid overwriting the underlying classes variables. What do you guys think of this solution? Is it bad practice? Do you have any better solutions? Is the design in general bad? How should I redesign this if it is? Thanks in advanced for any replies, and sorry about the long question. Hopefully it all makes sense!

    Read the article

  • OpenGL ES rotate texture

    - by 0xSina
    I just got started with OpenGL ES... I have a fragment: const char * sFragment = _STRINGIFY( varying highp vec2 coordinate; precision mediump float; uniform vec4 maskC; uniform float threshold; uniform sampler2D videoframe; uniform sampler2D videosprite; uniform vec4 mask; uniform vec4 maskB; uniform int recording; vec3 normalize(vec3 color, float meanr) { return color*vec3(0.75 + meanr, 1., 1. - meanr); } void main() { float d; float dB; float dC; float meanr; float meanrB; float meanrC; float minD; vec4 pixelColor; vec4 spriteColor; pixelColor = texture2D(videoframe, coordinate); spriteColor = texture2D(videosprite, coordinate); meanr = (pixelColor.r + mask.r)/8.; meanrB = (pixelColor.r + maskB.r)/8.; meanrC = (pixelColor.r + maskC.r)/8.; d = distance(normalize(pixelColor.rgb, meanr), normalize(mask.rgb, meanr)); dB = distance(normalize(pixelColor.rgb, meanrB), normalize(maskB.rgb, meanrB)); dC = distance(normalize(pixelColor.rgb, meanrC), normalize(maskC.rgb, meanrC)); minD = min(d, dB); minD = min(minD, dC); gl_FragColor = spriteColor; if (minD > threshold) { gl_FragColor = pixelColor; } } Now, depending on wether recording is 0 or 1, I want to rotate uniform sampler2D videosprite 180 degrees (reflection in x-axis, flip vertically). How can I do that? I found the function glRotatef(), but how do i specify that I want to rotate video sprite and not videoframe? Thanks

    Read the article

  • Explicit resource loading in Ogre (Mogre)

    - by sebf
    I am just starting to learn Mogre and what I would like to do is to be able to load resources 'explicitly' (i.e. I just provide an absolute path instead of using a resource group tied to a directory). This is very different to manually loading resources, which I believe in Ogre has a very specific meaning, to build up the object using Ogres methods. I want to use Ogres resource management system/resource loading code, but to have finer control over which files are loaded and in what groups they are. I remember reading how to do this but cannot find the page again; I think its possible to do something like: Declare a resource group Declare the resource(s) (this is when the actual resource file name is provided) Initialise the resource group to actually load the resource(s) Is this the correct procedure? If so, is there any example code showing how to do this?

    Read the article

  • Basic AI FSM - Handling state transition

    - by Galvanize
    I'm starting to study on how to implement game AI, and it seems to me that a very simple FSM for my Pong demo would be a nice way to start. My vision on implementing this would be to have a basic state interface and a class for each state, then the NPC would have an instance of the current state. The class should have an update method and directions on wich state to go next, depending on the event received. The question is: How do I handle this event? Should I have a regular addEventListener and a costum event system? Or should I check on update for the things that could change the current state? I'm feeling a bit lost, I feel I have a good grasp on the FSM concept but a good implementation seems tricky, thanks in advance.

    Read the article

  • Accelerating 2d object collision with other objects [on hold]

    - by Silent Cave
    Making my very first attempt at game programming with SDL/OpenGL. So I made an object Actor witch can move in all four sides with acceleration. And there are bunch of other rectangles to collide to. the image Movement and collision detection alghorythms work just fine by itself, but when combined to prevent the green rectangle from crossing black rectangles, it gives me a kind of funny resault. Let me show you the code first: from Actor.h class Actor{ public: SDL_Rect * dim; alphaColor * col; float speed; float xlGrav, xrGrav, yuGrav, ydGrav; float acceleration; bool left,right,up,down; Actor(SDL_Rect * dim,alphaColor * col, float speed, float acceleration); bool colides(const SDL_Rect & rect); bool check_for_collisions(const std::vector<SDL_Rect*> & gameObjects ); }; from actor.cpp bool Actor::colides(const SDL_Rect & rect){ if (dim->x + dim->w < rect.x) return false; if (dim->x > rect.x + rect.w) return false; if (dim->y + dim->h < rect.y) return false; if (dim->y > rect.y + rect.h) return false; return true; } movement logic from main.cpp if (actor->left){ if(actor->xlGrav < actor->speed){ actor->xlGrav += actor->speed*actor->acceleration; }else actor->xlGrav = actor->speed; actor->dim->x -= actor->xlGrav; if(actor->check_for_collisions(gameObjects)){ actor->dim->x += actor->xlGrav; actor->xlGrav = 0; } } if (!actor->left){ if(actor->xlGrav - actor->speed*actor->acceleration > 0){ actor->xlGrav -= actor->speed*actor->acceleration; }else actor->xlGrav = 0; actor->dim->x -= actor->xlGrav; if(actor->check_for_collisions(gameObjects)){ actor->dim->x += actor->xlGrav; actor->xlGrav = 0; } } if (actor->right){ if(actor->xrGrav < actor->speed){ actor->xrGrav += actor->speed*actor->acceleration; }else actor->xrGrav = actor->speed; actor->dim->x += actor->xrGrav; if(actor->check_for_collisions(gameObjects)){ actor->dim->x -= actor->xrGrav; actor->xrGrav = 0; } } if (!actor->right){ if(actor->xrGrav - actor->speed*actor->acceleration > 0){ actor->xrGrav -= actor->speed*actor->acceleration; }else actor->xrGrav = 0; actor->dim->x += actor->xrGrav; if(actor->check_for_collisions(gameObjects)){ actor->dim->x -= actor->xrGrav; actor->xrGrav = 0; } } if (actor->up){ if(actor->yuGrav < actor->speed){ actor->yuGrav += actor->speed*actor->acceleration; }else actor->yuGrav = actor->speed; actor->dim->y -= actor->yuGrav; if(actor->check_for_collisions(gameObjects)){ actor->dim->y += actor->yuGrav; actor->yuGrav = 0; } } if (!actor->up){ if(actor->yuGrav - actor->speed*actor->acceleration > 0){ actor->yuGrav -= actor->speed*actor->acceleration; }else actor->yuGrav = 0; actor->dim->y -= actor->yuGrav; if(actor->check_for_collisions(gameObjects)){ actor->dim->y += actor->yuGrav; actor->yuGrav = 0; } } if (actor->down){ if(actor->ydGrav < actor->speed){ actor->ydGrav += actor->speed*actor->acceleration; }else actor->ydGrav = actor->speed; actor->dim->y += actor->ydGrav; if(actor->check_for_collisions(gameObjects)){ actor->dim->y -= actor->ydGrav; actor->ydGrav = 0; } } if (!actor->down){ if(actor->ydGrav - actor->speed*actor->acceleration > 0){ actor->ydGrav -= actor->speed*actor->acceleration; }else actor->ydGrav = 0; actor->dim->y += actor->ydGrav; if(actor->check_for_collisions(gameObjects)){ actor->dim->y -= actor->ydGrav; actor->ydGrav = 0; } } So, if the green box approaches an obstacle from up or left, everything goes as planned - object stops, and it's acceleration drops to zero. But if it comes from bottom or right, it enters into obstacles inner space and starts strangely dance, I'd rather say move in inverted controls. What do I fail to see?

    Read the article

  • How to make Box2D bodies automatically return to a initial rotation

    - by sm4
    I have two long Box2D bodies, that can collide while moving one of them around with MouseJoint. I want them to try to hold their position and rotation. Blue body is moved using MouseJoint (yellow) towards the Red body. Red body has another MouseJoint - Blue can push Red, but Red will try to return to the start point thanks to the MouseJoint - this works just fine. Both bodies correctly rotate along the middle. This is still as I want. I change the MouseJoint to move the Blue away. What I need is both bodies return to their initial rotation (green arrows) Desired positions and rotations Is there anything in Box2D that could do this automatically? The MouseJoint does that nicely for position. I need it in AndEngine (Java, Android) port, but any Box2D solution is fine. EDIT: By automatically I mean having something I can add to the object "Paddle" without the need to change game loop. I want to encapsulate this functionality to the object itself. I already have an object Paddle that has its own UpdateHandler which is being called from the game loop. What would be much nicer is to attach some kind of "spring" joint to both left and right sides of the paddle that would automatically level the paddle. I will be exploring this option soon.

    Read the article

  • Is it safe StringToHash() to use in Unity?

    - by Sebastian Krysmanski
    I'm currently browsing through the Unity tutorials and saw that they're recommending to use Animator.StringToHash("some string") to created unique ids for animation properties (see here). Since I'm a programmer, to me the word "hash" doesn't represents something unique. Like the Java documentation for hashValue() states: It is not required that if two objects are unequal [...], then calling the hashCode method on each of the two objects must produce distinct integer results. So, according to this (and my definition of "hash"), two strings may have the same hash value. (You can also argue that there are an infinite number of possible strings but only 2^32 possible int values.) So, is there a possibility that StringToHash() will give me an id that actually belongs to another property (than the one I requested the hash for)?

    Read the article

  • Progress bar in Super Hexagon using OpenGL ES 2 (Android)

    - by user16547
    I'm wondering how the progress bar in Super Hexagon was made. (see image, top left) Actually I am not very sure how to implement a progress bar at all using OpenGL ES 2 on Android, but I am asking specifically about the one used in Super Hexagon because it seems to me less straightforward / obvious than others: the bar changes its colour during game play. I think one possibility is to use the built-in Android progress bar. I can see from some Stackoverflow questions that you can change the default blue colour to whatever you want, but I'm not sure whether you can update it during the game play. The other possibility I can think of for implementing a progress bar is to have a small texture that starts with a scale of 0 and that you keep scaling until it reaches the maximum size, representing 100%. But this suffers from the same problem as before: you'll not be able to update the colour of the texture during run-time. It's fixed. So what's the best way to approach this problem? *I'm assuming he didn't use a particular library, although if he did, it would be interesting to know. I'm interested in a pure OpenGL ES 2 + Android solution.

    Read the article

  • Why do we use Pythagoras in game physics?

    - by Starkers
    I've recently learned that we use Pythagoras a lot in our physics calculations and I'm afraid I don't really get the point. Here's an example from a book to make sure an object doesn't travel faster than a MAXIMUM_VELOCITY constant in the horizontal plane: MAXIMUM_VELOCITY = <any number>; SQUARED_MAXIMUM_VELOCITY = MAXIMUM_VELOCITY * MAXIMUM_VELOCITY; function animate(){ var squared_horizontal_velocity = (x_velocity * x_velocity) + (z_velocity * z_velocity); if( squared_horizontal_velocity <= SQUARED_MAXIMUM_VELOCITY ){ scalar = squared_horizontal_velocity / SQUARED_MAXIMUM_VELOCITY; x_velocity = x_velocity / scalar; z_velocity = x_velocity / scalar; } } Let's try this with some numbers: An object is attempting to move 5 units in x and 5 units in z. It should only be able to move 5 units horizontally in total! MAXIMUM_VELOCITY = 5; SQUARED_MAXIMUM_VELOCITY = 5 * 5; SQUARED_MAXIMUM_VELOCITY = 25; function animate(){ var x_velocity = 5; var z_velocity = 5; var squared_horizontal_velocity = (x_velocity * x_velocity) + (z_velocity * z_velocity); var squared_horizontal_velocity = 5 * 5 + 5 * 5; var squared_horizontal_velocity = 25 + 25; var squared_horizontal_velocity = 50; // if( squared_horizontal_velocity <= SQUARED_MAXIMUM_VELOCITY ){ if( 50 <= 25 ){ scalar = squared_horizontal_velocity / SQUARED_MAXIMUM_VELOCITY; scalar = 50 / 25; scalar = 2.0; x_velocity = x_velocity / scalar; x_velocity = 5 / 2.0; x_velocity = 2.5; z_velocity = z_velocity / scalar; z_velocity = 5 / 2.0; z_velocity = 2.5; // new_horizontal_velocity = x_velocity + z_velocity // new_horizontal_velocity = 2.5 + 2.5 // new_horizontal_velocity = 5 } } Now this works well, but we can do the same thing without Pythagoras: MAXIMUM_VELOCITY = 5; function animate(){ var x_velocity = 5; var z_velocity = 5; var horizontal_velocity = x_velocity + z_velocity; var horizontal_velocity = 5 + 5; var horizontal_velocity = 10; // if( horizontal_velocity >= MAXIMUM_VELOCITY ){ if( 10 >= 5 ){ scalar = horizontal_velocity / MAXIMUM_VELOCITY; scalar = 10 / 5; scalar = 2.0; x_velocity = x_velocity / scalar; x_velocity = 5 / 2.0; x_velocity = 2.5; z_velocity = z_velocity / scalar; z_velocity = 5 / 2.0; z_velocity = 2.5; // new_horizontal_velocity = x_velocity + z_velocity // new_horizontal_velocity = 2.5 + 2.5 // new_horizontal_velocity = 5 } } Benefits of doing it without Pythagoras: Less lines Within those lines, it's easier to read what's going on ...and it takes less time to compute, as there are less multiplications Seems to me like computers and humans get a better deal without Pythagoras! However, I'm sure I'm wrong as I've seen Pythagoras' theorem in a number of reputable places, so I'd like someone to explain me the benefit of using Pythagoras to a maths newbie. Does this have anything to do with unit vectors? To me a unit vector is when we normalize a vector and turn it into a fraction. We do this by dividing the vector by a larger constant. I'm not sure what constant it is. The total size of the graph? Anyway, because it's a fraction, I take it, a unit vector is basically a graph that can fit inside a 3D grid with the x-axis running from -1 to 1, z-axis running from -1 to 1, and the y-axis running from -1 to 1. That's literally everything I know about unit vectors... not much :P And I fail to see their usefulness. Also, we're not really creating a unit vector in the above examples. Should I be determining the scalar like this: // a mathematical work-around of my own invention. There may be a cleverer way to do this! I've also made up my own terms such as 'divisive_scalar' so don't bother googling var divisive_scalar = (squared_horizontal_velocity / SQUARED_MAXIMUM_VELOCITY); var divisive_scalar = ( 50 / 25 ); var divisive_scalar = 2; var multiplicative_scalar = (divisive_scalar / (2*divisive_scalar)); var multiplicative_scalar = (2 / (2*2)); var multiplicative_scalar = (2 / 4); var multiplicative_scalar = 0.5; x_velocity = x_velocity * multiplicative_scalar x_velocity = 5 * 0.5 x_velocity = 2.5 Again, I can't see why this is better, but it's more "unit-vector-y" because the multiplicative_scalar is a unit_vector? As you can see, I use words such as "unit-vector-y" so I'm really not a maths whiz! Also aware that unit vectors might have nothing to do with Pythagoras so ignore all of this if I'm barking up the wrong tree. I'm a very visual person (3D modeller and concept artist by trade!) and I find diagrams and graphs really, really helpful so as many as humanely possible please!

    Read the article

  • Multiplayer game communication framework for mac/ios

    - by ishaq
    (Cross post from stackoverflow) I am creating a multiplayer 2D game for Mac and iOS devices. I'll be using cocso2d for graphics/game engine, however I am largely blank on what to use for multiplayer communication. Please note that I cannot use central severs e.g. SmartFox, RedDwarf, etc since I want the players to "host" games for others and be able to play it on their LAN, VPN or my own servers. Any pointers? I checked lidgren but it's for .NET only and hence not an option for me. EDIT: just in case it wasn't clear, the messaging has to be real time hence it's probably going to be over UDP

    Read the article

  • Kinect Hand tracking in xna

    - by N0xus
    I'm trying to create an application using the Kinect to simulate the following project: Kinect Hand Tracking I want my project to have similar usability with the Kinect tracking hand and finger positions for use in a menu system, or to navigate another system. What I would like to know is; is it possible for the exact same to be accomplished in XNA using Kinect? I know that it can be done in Winform / C#, but I know XNA / C# a lot better and would (ideally) prefer to use that.

    Read the article

< Previous Page | 532 533 534 535 536 537 538 539 540 541 542 543  | Next Page >