I am trying to do collision detection between a rectangle and a circle. I came up with this method:
-(BOOL) isCollidingRect:(CCSprite *) spriteOne WithSphere:(CCSprite *) spriteTwo {
float diff = ccpDistance(spriteOne.position, spriteTwo.position);
float obj1Radii = [spriteOne boundingBox].size.width/2;
float obj2Radii = [spriteTwo boundingBox].size.width/2;
if (diff < obj1Radii + obj2Radii) {
return YES;
} else {
return NO;
}
}
and this is how I check it:
if ([self isCollidingRect:player WithSphere:blocker] == true) {
[self playerdeathstart];
}
This seems to work properly on the side of the rectangle but it doesn't above or below it. On the top and bottom, the collision occurs too early.
Is there a way I can get this collision to detected properly? Thank you for your help.
You can use CGRectIntersectsRect to achieve this.
-(BOOL) isCollidingRect:(CCSprite *) spriteOne WithSphere:(CCSprite *) spriteTwo {
return CGRectIntersectsRect([spriteOne boundingBox],[spriteTwo boundingBox]);
}
It is not pixel perfect but as i understand that is not necessary in this case.
This is not a solution for those who use Cocos2d-ObjC, but will help for Cocos2d-x devs (for instance, personally I found this topic because was searching for the same for my c++ game).
Cocos2d-x has method "intersectsCircle" for Rect class.
Here is how I solved in my c++ project almost the same problem as one described by you:
bool ObstacleEntity::hasCollisionAgainst(cocos2d::Sprite *spr)
{
cocos2d::Rect rect = cocos2d::Rect( spr->getPositionX(), spr->getPositionY(), spr->getBoundingBox().size.width, spr->getBoundingBox().size.height);
float rw = this->getBoundingBox().size.width / 2;
float rh = this->getBoundingBox().size.height / 2;
float radius = ( rw > rh ) ? rw : rh;
cocos2d::Vec2 center( this->getPositionX() + rw, this->getPositionY() + rh );
return rect.intersectsCircle( center, radius );
}
Passed Sprite here is rectangle, while ObstacleEntity always is almost ideally round.
Note that anchor points for all entities are set to lower left corner in my case.
Related
The only way I can think to do it is to check velocities for all physics bodies during every collisions.
- (BOOL)ccPhysicsCollisionBegin:(CCPhysicsCollisionPair *)pair piece:(CCNode *)pieceA piece:(CCNode *)pieceB{
float x = 0;
float y = 0;
for (int i = 0; i < [[_physicsWorld children] count]; i++) {
x = x + [[[_physicsWorld children][i] physicsBody] velocity].x;
y = y + [[[_physicsWorld children][i] physicsBody] velocity].y;
}
if ( x == 0 && y == 0 ) {
NSLog(#"stopped");
}
return YES;
}
This logs “stopped” multiple times when the scene first loads, then doesn’t log “stopped” again, even after physics bodies have clearly started moving and colliding and then come to a stop.
Ideally I'd like a delegate method that would notify me when all physics bodies have stopped moving, but I can't seem to find one.
FYI: I'm using the standard Chipmunk physics engine that's baked into Cocos2d V3.0
Chipmunk has a internal mechanism, which can, if activated, automatically deactivate physics bodies. My approach (I am using cocos2dx 3.11.1 and not -obj version with chipmunk 7.0.1) is:
activate the chipmunk idle mechanism (0.5 second - meaning, if an object is not moving for longer than 0.5 second it will be deactivated):
cpSpaceSetSleepTimeThreshold(space, 0.5f);
You do not need to use
cpSpaceSetIdleSpeedThreshold(space, <speed>);
because chipmunk calculates the threshold speed for you (according the gravitation used).
use this code for determination if all objects are not moving (static and kinetic bodies never sleep):
bool isAnyPhysicsBodyMoving(){
int i = 0; bool isMoving = false;
const Vector<PhysicsBody*>& bodies = getPhysicsWorld()->getAllBodies();
while( i < bodies.size() && !isMoving){
PhysicsBody *body = bodies.at(i);
isMoving = cpBodyGetType(body->getCPBody()) == CP_BODY_TYPE_DYNAMIC
&& !body->isResting();
i++;
}
return isMoving;
}
use static (and not kinetic) body for walls, in order to let objects sleep:
// wall
Size visibleSize = Director::getInstance()->getWinSize();
Vec2 origin = Director::getInstance()->getVisibleOrigin();
float border = 10.0f;
Size wallBodySize = Size(visibleSize.width+2*border, visibleSize.height+2*border);
PhysicsBody *wallBody = PhysicsBody::createEdgeBox(wallBodySize, PhysicsMaterial(1.0f, 1.0f, 0.5f), border);
Node* wall = Node::create();
wall->addComponent(wallBody);
wall->setAnchorPoint(Vec2(0.5f, 0.5f));
wall->setPosition(Point(visibleSize.width/2+origin.x, visibleSize.height/2+origin.y));
cpVect tt;
tt.x = wall->getPosition().x; tt.y = wall->getPosition().y;
//set position manually and BEFORE adding the object into the space
cpBodySetPosition(wallBody->getCPBody(), tt);
cpBodySetType(wallBody->getCPBody(), CP_BODY_TYPE_STATIC);
addChild(wall);
Any dynamic body connected to a kinetic body (for example laying on) will never sleep.
test it with DEBUG activated
getPhysicsWorld()->setDebugDrawMask(PhysicsWorld::DEBUGDRAW_ALL);
the boxes (their content) must become grey (=sleeping) and not red (=active):
In order to let it work, I have:
added an access method (to get cpSpace) in CCPhysicsWorld.h:
inline cpSpace* getSpace() const { return _cpSpace; }
Fix call of
cpBodySetTorque(body, 0.0f);`
in CCPhysicsBody.cpp to
if (body->t != 0.0f){
cpBodySetTorque(body, 0.0f);
}
Fix call of
cpBodySetPosition(_cpBody, tt);`
in CCPhysicsBody.cpp to
if (!cpveql(tt, cpBodyGetPosition(_cpBody))){
cpBodySetPosition(_cpBody, tt);
}
Steps 2. and 3. are necessary to avoid setting of the same physics body properties, which wake up a sleeping body.
The advantage of this approach is, that the chipmunk does not make any calculations for such physical bodies - saving CPU and battery.
I found something that works.
tl;dr
The basic idea is to keep track of the positions of the sprites myself, and then periodically check them to see if any of them have moved since they were last checked.
Longer version
I created a subclass of CCNode with the class name Piece.
These are my objects that are added to the physics world.
#implementation Piece {
float _previousX;
float _previousY;
}
-(void)updatePreviousScreenXandY{
_previousX = self.position.x;
_previousY = self.position.y;
}
-(BOOL)hasntMoved{
float currentX = self.position.x;
float currentY = self.position.y;
if ( currentX == _previousX && currentY == _previousY ) {
return TRUE;
}else{
return FALSE;
}
}
This is in my CCNode that acts as the game scene
-(void)doStuffAfterPiecesStopMoving:(NSTimer*)timer{
BOOL noPiecesHaveMoved = TRUE;
for (int i = 0; i < [[_physicsWorld children] count]; i++) {
if ( [[_physicsWorld children][i] hasntMoved] == FALSE ) {
noPiecesHaveMoved = FALSE;
break;
}
}
if ( noPiecesHaveMoved ) {
[timer invalidate];
NSLog(“Pieces have stopped moving”);
}else{
NSLog(“Pieces are still moving”);
[self updateAllPreviousPiecePositions];
}
}
-(void)updateAllPreviousPiecePositions{
for (int i=0; i < [[_physicsWorld children] count]; i++) {
Piece *piece = (Piece*)[_physicsWorld children][i];
[piece updatePreviousScreenXandY];
}
}
All I have to do is
[NSTimer scheduledTimerWithTimeInterval:TIME_BETWEEN_CHECKS
target:_gamePlay
selector:#selector(doStuffAfterPiecesStopMoving:)
userInfo:nil
repeats:YES];
and it’ll run whatever code I want after all Piece nodes have stopped moving.
The key to getting it to work well is to get the values for the Chipmunk space’s sleepTimeThreshold and the timer above’s time as low as possible.
My experimenting suggests the following settings work okay, but anything lower will cause problems (i.e. collisions not taking place properly):
sleepTimeThreshold = 0.15
my timer = 0.05
If anyone has a different/better solution or improvements to the above code, please post.
I am trying to experiment with UIViews on screen and using pan gestures. So I got some open source code from another project that I am looking at - and trying to learn a few things from it.
-(BOOL)isPointContainedWithinBezelRect:(CGPoint)point {
CGRect leftBezelRect;
CGRect tempRect;
CGFloat bezelWidth = 20;
CGRectDivide(self.view.bounds, &leftBezelRect, &tempRect, bezelWidth, CGRectMinXEdge);
return CGRectContainsPoint(leftBezelRect, point);
}
I understand that CGRectDivide function "Slices up a rect", but thats as far as I can make out.
I hope to get more clarification regarding the function. Also, how does the function return value vide a false / true value?
void CGRectDivide(
CGRect rect,
CGRect *slice,
CGRect *remainder,
CGFloat amount,
CGRectEdge edge
)
The CGRectDivide method splits a CGRect into two CGRects based on the CGRectEdge and distance from the rectangle side amount provided to the method.
Source
You should check
https://developer.apple.com/library/ios/documentation/graphicsimaging/reference/CGGeometry/Reference/reference.html#//apple_ref/c/func/CGRectDivide
and
http://nshipster.com/cggeometry/
But it seems that this method could be simplified to
-(BOOL)isPointContainedWithinBezelRect:(CGPoint)point {
CGRect leftBezelRect = self.view.bounds;
leftBezelRect.size.width = 20;
return CGRectContainsPoint(leftBezelRect, point);
}
or even to
-(BOOL)isPointContainedWithinBezelRect:(CGPoint)point {
return CGRectContainsPoint(self.view.bounds, point) && (point.x <= 20);
}
I'm trying to make my first game using Spritekit, so i have a sprite that i need to move around using my accelerometer. Well, no problem doing that; movement are really smooth and responsive, the problem is that when i try to rotate my sprite in order to get it facing its own movement often i got it "shaking" like he has parkinson. (:D)
i did realize that this happens when accelerometer data are too close to 0 on one of x, y axes.
So the question: Is there a fix for my pet parkinson?? :D
Here is some code:
-(void) update:(NSTimeInterval)currentTime{
static CGPoint oldVelocity;
//static CGFloat oldAngle;
if(_lastUpdatedTime) {
_dt = currentTime - _lastUpdatedTime;
} else {
_dt = 0;
}
_lastUpdatedTime = currentTime;
CGFloat updatedAccelX = self.motionManager.accelerometerData.acceleration.y;
CGFloat updatedAccelY = -self.motionManager.accelerometerData.acceleration.x+sinf(M_PI/4.0);
CGFloat angle = vectorAngle(CGPointMake(updatedAccelX, updatedAccelY));
_velocity = cartesianFromPolarCoordinate(MAX_MOVE_PER_SEC, angle);
if(oldVelocity.x != _velocity.x || oldVelocity.y != _velocity.y){
_sprite.physicsBody.velocity = CGVectorMake(0, 0);
[_sprite.physicsBody applyImpulse:CGVectorMake(_velocity.x*_sprite.physicsBody.mass, _velocity.y*_sprite.physicsBody.mass)];
_sprite.zRotation = vectorAngle(_velocity);
oldVelocity = _velocity;
}
}
static inline CGFloat vectorAngle(CGPoint v){
return atan2f(v.y, v.x);
}
i did try to launch the update of the _velocity vector only when updatedAccelX or updatedAccelY are, in absolute value >= of some values, but the result was that i got the movement not smooth, when changing direction if the value is between 0.1 and 0.2, and the problem wasn't disappearing when the value was under 0.1.
i would like to maintain direction responsive, but i also would like to fix this "shake" of the sprite rotation.
I'm sorry for my bad english, and thanks in advance for any advice.
You can try a low pass filter (cf. to isolate effect of gravity) or high pass filter (to isolate effects of user acceleration).
#define filteringFactor 0.1
- (void)accelerometer:(UIAccelerometer *)accelerometer didAccelerate:(UIAcceleration *)acceleration {
//low pass
accelerX = (acceleration.x * filteringFactor) + (accelerX * (1.0 - filteringFactor));
//idem … accelerY
//idem … accelerZ
//or high pass
accelerX = acceleration.x - ( (acceleration.x * filteringFactor) + (accelerX * (1.0 - filteringFactor)) );
//idem … accelerY
//idem … accelerZ
}
I'm writing a 2D ball game with sprite kit on iOS 7 and currently struggling on one physic simulation.
To explain the expected behavior: if a ball is dropped into a tea cup, it will circle around, loosing speed and finally stand still in the center of the cup.
I've tried to archive this with gravity, but gravity in sprite kit only applies to vertical X and Y axis, not Z-axis. I also tried to use level gravity by switching gravity values with small physic bodies on beginContact depending on the current ball position in the tea cup. But some contacts are dropped and the result is far away to look realistic.
I think I need to solve this in the update: method, but I have no idea which way to go.
Any advice greatly welcome and I need to mention that I'm not an expert on math, please explain your path to go. :-)
Since there's no built-in support for this kind of behavior in SpriteKit, rather than trying to hack existing functions to get what you want, you're probably better off integrating some published 2D physics formulas in your x,y 2D world. I would think that something like simulating magnetic or a homing behavior might be right for this.
A simple example would be something like (in the scene's -update: method):
CGFloat strength = 0.5; //(some scaling value)
CGPoint ballLocation = ball.position;
CGPoint cupLocation = cup.position;
[ball.physicsBody applyForce:CGVectorMake((cupLocation.x - ballLocation.x) * strength,
(cupLocation.y - ballLocation.y) * strength)];
following Joshd great idea, I have created an NSArray with like explained in my comment above. Hope this snippets does help some others...
The result could be found on youtube: http://youtu.be/Uephg94UH30
Sorry for the bad Airplay frame rate, it runs perfectly smooth on my iPad
The -update: functions does the work but only triggered if _meditationIsActive. This bool is set in -didBeginContact: when any ball gets in contact with a hole.
if (_lastCheck > 0.005)
{
if (_meditationIsActive)
{
CGFloat strength = 0.1; //(some scaling value)
CGPoint ballLocation;
CGPoint holeLocation;
for (MeditationHole * holeObj in _meditationHoles)
{
if (holeObj.connectedMeditationBall != nil)
{
ballLocation = holeObj.connectedMeditationBall.position;
holeLocation = holeObj.position;
[holeObj.connectedMeditationBall.physicsBody applyForce:CGVectorMake(
(holeLocation.x - ballLocation.x) * strength, (holeLocation.y - ballLocation.y) * strength)];
}
}
_meditationIsActive = [self doesMeditationApplies];
}
_lastCheck = 0;
}
At the end I'm checking if there is a valid ball out of the array in contact with a hole to avoid checking during every update. This is done with the following function where position check +/- 48 detects a ball close to a hole and +/-1 ball stands still
- (bool)doesMeditationApplies
{
bool isInArea = NO;
int perfectMatchCount = 0;
for (MeditationHole * holeObj in _meditationHoles)
{
if (holeObj)
{
if (holeObj.connectedMeditationBall != nil)
{
MeditationBall * ballObj = holeObj.connectedMeditationBall;
if ((ballObj.position.x >= holeObj.position.x - 48) &&
(ballObj.position.x <= holeObj.position.x + 48) &&
(ballObj.position.y >= holeObj.position.y - 48) &&
(ballObj.position.y <= holeObj.position.y + 48))
{
isInArea = YES;
}
else
{
holeObj.connectedMeditationBall = nil;
}
if ((ballObj.position.x >= holeObj.position.x - 1) &&
(ballObj.position.x <= holeObj.position.x + 1) &&
(ballObj.position.y >= holeObj.position.y - 1) &&
(ballObj.position.y <= holeObj.position.y + 1))
{
perfectMatchCount++;
isInArea = YES;
}
}
}
}
if (perfectMatchCount == _oxydStonesMax)
{
if (_sound)
{
self.pauseMusicPlaybackBlock(YES);
NSLog(#"PlaySound Meditation");
[OxydScene PlaySystemSound:#"Win2"];
}
isInArea = NO;
[self showPauseScreenWithWin:YES andPauseOnly:NO];
}
return isInArea;
}
There is a class associated with the program Physics Editor called GB2ShapeCache that loads shapes that I make in the program. I noticed that it is not currently possible to change the scale of the shapes on the fly so I would like to be able to scale the fixtures for the shapes that I made in Physics Editor. Now the scale of my CCSprite in my app can be random so currently in the addShapesWithFile method, I do this for polygons:
vertices[vindex].x = (offset.x * sprite.scaleX) / ptmRatio_;
vertices[vindex].y = (offset.y * sprite.scaleY) / ptmRatio_;
and this for circles:
circleShape->m_radius = ([[circleData objectForKey:#"radius"] floatValue] / ptmRatio_) *sprite.scale;
I also changed the method so that I can pass in my sprite so I can get the scale to:
-(void) addShapesWithFile:(NSString*)plist forSprite:(CCSprite*)sprite
so that I can pass in my sprite so I can get the scale.
HOWEVER, I find this to be inefficient because I should not have to reload ALL my shapes in my plist since they are already added.
So is there any way to do what I am doing now but in the addFixturesToBody method? This way I do not re-create the already added plist shapes and I only scale the fixtures when it is ready to be added to my body.
If anyone needs to see more code or needs more info, feel free to ask. I know this issue must be simple!!!
Thanks!
I would recommend implementing it in the addFixturesToBody method.
(see https://github.com/AndreasLoew/GBox2D/blob/master/GBox2D/GB2ShapeCache.mm)
Try this method below, this should scale the shapes accordingly to the sprite's they are for. Just pass in your CCSprite and this method will handle the rest.
- (void)addFixturesToBody:(b2Body*)body forShapeName:(NSString*)shape forSprite:(CCSprite*)sprite {
BodyDef *so = [shapeObjects_ objectForKey:shape];
assert(so);
FixtureDef *fix = so->fixtures;
if ((sprite.scaleX == 1.0f) && (sprite.scaleY == 1.0f)) {
// simple case - so do not waste any energy on this
while(fix) {
body->CreateFixture(&fix->fixture);
fix = fix->next;
}
} else {
b2Vec2 vertices[b2_maxPolygonVertices];
while(fix) {
// make local copy of the fixture def
b2FixtureDef fix2 = fix->fixture;
// get the shape
const b2Shape *s = fix2.shape;
// clone & scale polygon
const b2PolygonShape *p = dynamic_cast<const b2PolygonShape*>(s);
if(p)
{
b2PolygonShape p2;
for(int i=0; i<p->m_vertexCount; i++)
{
vertices[i].x = p->m_vertices[i].x * sprite.scaleX;
vertices[i].y = p->m_vertices[i].y * sprite.scaleY;
}
p2.Set(vertices, p->m_vertexCount);
fix2.shape = &p2;
}
// clone & scale circle
const b2CircleShape *c = dynamic_cast<const b2CircleShape *>(s);
if(c) {
b2CircleShape c2;
c2.m_radius = c->m_radius * sprite.scale;
c2.m_p.x = c->m_p.x * sprite.scaleX;
c2.m_p.y = c->m_p.y * sprite.scaleY;
fix2.shape = &c2;
}
// add to body
body->CreateFixture(&fix2);
fix = fix->next;
}
}
}