i'm using SDL2.0/OpenGL + Box2D with C++ to learn some stuff about games, so i created two rectangles bodies: a small box(dynamic) wich will fall into the second body: the ground(static), i draw them on the screen with OpenGL primitives, as recomended i'm not using "pixel thinking", i'm using meters to represent everything, well, the problem is that i can't get a good physic simulation, just almost, when rendering i can see the small box falling into the ground but it don't touch the ground, it gets real close but don't touch, i really can't understand why, so i came here to see if can get some help:

Here we create the bodies(and do other things):

void MyLoop::onCreate(){
world = new b2World(b2Vec2(0, -10));

player = new Player(world);
//creates the player body

//Creates the ground
b2BodyDef bdef;
b2PolygonShape* pshape = new b2PolygonShape();
b2FixtureDef fdef;

bdef.type = b2_staticBody;
ground = world->CreateBody(&bdef);

pshape->SetAsBox(40.0f, 1.0f);
fdef.shape = pshape;

ground->SetTransform(b2Vec2(20.0f, 0.5f), 0);

//configurating the apropriate scale
glOrtho(0, 40, 0, 20, 1, -1);

delete pshape;

The most important here is the bodies creation and the glOrtho call wich will set the camera to our coordinate system.

The creation of the player body:

b2BodyDef bdef;
bdef.type = b2_dynamicBody;

pbody = world->CreateBody(&bdef);

b2PolygonShape* pshape = new b2PolygonShape();
pshape->SetAsBox(0.5f, 1.0f);

b2FixtureDef fdef;
fdef.shape = pshape;
pbody->SetTransform(b2Vec2(19.0f, 10.0f), 0);

So now we go to the render loop: i get the bodies position with getPosition() method and use them to draw the bodies:

glClearColor(1.0f, 0.0f, 0.0f, 1.0f);


float x = ground->GetPosition().x;
float y = ground->GetPosition().y;

renderer.setColor(0.0f, 0.0f, 1.0f);
renderer.drawRectangle(x - 20.0f, y + 0.5f, 40.0f, 1.0f);


world->Step(1 / 20.0f, 8, 3);

The player render:

float x = pbody->GetPosition().x;
float y = pbody->GetPosition().y;

renderer.setColor(0.0f, 1.0f, 0.0f);
renderer.drawRectangle(x - 0.25f, y + 0.5f, 0.5f, 1.0f);

1 Answer 1


I think that there is no problem with your code. It's Box2D's thing, that it creates a little bit of skin around the polygons.

According to the Box2D manual, a little gap between polygons is being intentionally created to make continuous collision resolution more stable by preventing tunneling, eg. in stacking scenarios. Manual also says one can make their visual shapes slightly bigger to make up for this artificial gap. I think you could go with that.

  • \$\begingroup\$ i dont know why but i have faced three situations(including the code in this post) where dividing the bodies sizes(width and height) by two and using this values for rendering gets a good simulation \$\endgroup\$ Commented Sep 23, 2015 at 23:14

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .