I have top-down game for iPhone (no gravity) and I'm using Cocos2d and Box2d.
I try to move a bullet by this code:
// 'targetPosition' is a point of touch
b2Vec2 touchInWorld = b2Vec2(targetPosition.x/PTM_RATIO, targetPosition.y/PTM_RATIO);
b2Vec2 direction = b2Vec2(touchInWorld.x - ballBodyDef.position.x, touchInWorld.y - ballBodyDef.position.y);
b2Vec2 force = b2Vec2(direction.x, direction.y);
force.Normalize();
ballBody->ApplyLinearImpulse(force, ballBodyDef.position);
The problem is that the ball moves waaaay to fast if the sprite is small (10x10 px).
If the sprite is 50x50, then the speed is smaller and looks okay.
It's driving me crazy because I can't control the speed at all.
And not only that, if I don't put force.Normalize() the the speed is different depending on the direction of touch ...
Everything was working great when I was using just Cocos2d and animations. I tried using Box2d to implement collisions but it seems such a HUGE amount of effort that I am considering doing the physics myself :(
First thing , i am not sure but why are you dividing by PTM_Ratios here, i don think you need that, may be you are handling it some other way.
Second, 10x10 travels faster than 50x50 because you are setting SetMassFromShapes, so assuming you are creating a 50/PTM_RATIOSx50/PTM_RATIO body for 50x50 sprite, the body has greater mass than 10x10 body. so it travels slow.
Third, you are using Distance for Impulse so definatly it you are not going to Normalize you are going to get a Proportional Impulse to your distance
and for the Hackish Solution , if nothing else works for you just Decrease the PTM_RATIO!
Related
I have a physics body that gets hit and rotates a lot. I want to limit the amount it can rotate (only 45 degrees in each direction or something) and I wanted to try to do that by applying torque if the direction goes past a certain point. I'm ok with the sprite wobbling back and forth (it's a bird).
I'm open to other suggestions on how to limit the rotation as well.
Thanks in advance!
Edit: I've already tried using zRotation (I printed it out and it was always 0.) I also tried using the vector but that gave me the direction the sprite was going.
I'm trying to create a chain in unity3d. A player should be capable of grabing one side of it and pull it to different location. So i created some grids and connect them together. It all works fine, the problem is only when user pulls a bit faster, then I got some spaces bettwen seperate grids. Is there anyway to set max distance on that?
Btw. I'm doing these in 2d so i have 2d Rigidbody and 2d Hinge joint.
Thanks!
The solverIterationCount of rigidbodies affects the smoothness of physics when they are moving fast. You should try increasing it or dynamically adjusting it according to the speed of the rigidbodies to increase stabilty
http://docs.unity3d.com/Documentation/ScriptReference/Physics-solverIterationCount.html
There a number of ways to tackle this issue, all with the up sides and down sides:
increasing accuracy:: this is usually the first place new game designers go, and crank up accuracy to the max. But then end up playing the price later when performance plummets. So be gentle, try to find a good balance, and if it does not cut it make up the rest with other tricks.
The main ways to increase accuracy are, increasing fixed times interval, and increasing ridged body Iteration count
.
Increasing restrictions::This often requires the most time, but cutting corners allows for smoother more predictable physics and can increase performance
small example:: the top link of the chain only needs to simulate rotation on the ridged body
.
Cheating:: find any way to make things easier, fake it
for example:: does the image really need to match the physics? Why not make sure the sprites stick together, but alow the physics to have small gaps
playing with the same concept more or less this week, experimented some with 2D Unity joints, I think the issue you are having is identical to one I had yesterday, the force you are applying 'breaks' joints for some frames, hence that almost 'spring' joint effect, make sure the mass and/or force applied are not too heavy, also maybe increase the chain parts mass, makes the joints more solid too.
In Editor, Hinge Joint -> Use Spring = true; Spring/Damper/Target Position = 0.
I have been working on a Unity ping pong game using the Leap Motion. I use rigidbody.MovePosition() to move the paddle. However, when I hit the ball (which uses gravity), the paddle launches it too far. Even when I change the masses of both, it doesn't do anything.
What variable should I change to reduce this energy going into the ball?
From the following link.
"MovePosition will put your object at the target location, no matter what. It may push aside other objects in a realistic way, or may blast them out of the way, or may just pass right through them. But it will gladly move you through a solid wall or a mountain.
If you're using MovePosition on a rigidbody to add from where you currently are, it looks like AddForce. With AddForce, the physics step does all the work: applies your velocity, sees the collision and handles the bounce. With MovePosition, the physics step sees you're mysteriously overlapping a solid object. If it isn't too much, it will bounce you apart."
You won't need to use MovePosition. Instead, you can figure out the direction of the shot (based on the position of the ball relative to the paddle). Then you can add an impulse force in that direction to the ball.
Pseudo-code (from the following link):
Vector3 shootDir = ballPosition - paddlePosition; // Calculate direction of the shot
shootDir.Normalize(); // Normalize the shot vector
ball.AddForce(shootDir * speed, ForceMode.Impulse); //Add impulse force in correct direction.
Credit due to Owen Reynolds and Tim Michels.
I use Unity3D and I have a player with a rigidbody. I add force tto the body for moving the player. My player walks over a terrain but is able to walk up mountains that are to steep to climb. I want to limit the player so it cannot walk up a slope that is to steep.
I know there is a CharacterController component that has this functionality, but I have to use the rigidbody, so I want the same but on my rigidbody.
I can get the normal of the triangle I am standing on, and calculate its angle, but I cannot seem to make the player stop moving up the slope. Only make the player stop moving (which makes the player unmovable once it hits a angled slope)
Any ideas how to solve this problem?
It's difficult to answer without more details on how you're using the physics engine. How/Are you using friction? What angle are you applying the force? Is it always horizontal or at the angle of the floor? Does the player have a mass?
Anyway I can think of a few ways to solve this
Go the pure physics route. Using player mass, friction, force angle, gravity, etc. Get the physics to handle these situation for you. This may take a fair amount of time and programming.
Keep the rigid body but fake the forces. Scale the force you are applying to the body of the player with the angle of the triangle the player is on. You can either use trigonometry to work out what you should apply or your own mapping. By your own mapping I mean set an angle where 0 force is applied (say 45 degree) and do a linear(or non linear) scale on the force applied so on flat ground force is 1 and at 45 force is 0.
Don't use rigid bodies. There is a reason most games don't use rigid bodies to control characters. It's hard and complicated and most of the time not worth the time it would take. Of course I don't know the details of your project so if this isn't an option, fine.
Hope that gives you some things to think about.
I was wondering if anyone could help me get started or pointed in the right direction. I'm looking to make a simple driving game using core animation. I'm a pretty good obj-c programmer, but when it comes to movement and math, I fail. I want a steering wheel to control the direction of the car, and a forward/backward to control acceleration & deceleration. If anyone can help me with the steering wheel code, I would be greatly appreciated! So basically I need help making a circle that can rotate with 1 finger drag, and i'll pass it's transform values to my car view ( i think i can handle the accel/deccel code) Any takers? :) Also, I have 2 invites for dribbble.com to give away, anyone who helps me, id be glad to give one out to.
grady, totowtwo, thanks for both your answers, I appreciate it. I;ve gotten to the point where I can rotate a "steering wheel" and it will rotate my "car", I also got it to move forwards and backwards. Just need to moving to be more realistic... Here is a link to the xCode project, http://www.cl.ly/7VBU, kept it very simple so if anyone looks at it - it will be easy to change / add code. So, if anyones wants to look at it,and help me make the movement more realistic, i'd be forever in your debt! :) thanks!
A starting point would be to recognize that a typical car steering wheel has 900 degrees of movement, whether you want that for the finger wheel or not is a design decision. The wheels in a car typically point about 30 degrees off +z at maximum rotation of the wheel. You could use a simple linear transformation from steering wheel angle to wheel angle.
From there, Car delta angle is a function of wheel angle times forward distance. As the car moves forward, you can imagine tracing the tangent of a circle with an infinite radius when you are pointing straight ahead, and a circle of smaller radius (~4x car length) when steering wheel is full tilt.
Of course, feel free to discard the design rules of typical vehicles. Give your users fun drive trains like holonomic robots have!
I am no expert but you could probably do something like:
CGPoint centerOfWheel;
CGPoint currentTouchPoint;
float radians = atan2(currentTouchPoint.x-centerOfWheel.x, currentTouchPoint.y-centerOfWheel.y);
then apply the transformation to the wheel ... of course you will want to make the transformations cumulative on the view over a distance, as continued turning of the wheel will result in continued turning.
also to make it feel more real, and less responsive you may want to use a lesser degree of turn, so that turning the wheel in once circle doesn't result in a whole circle turn of the vehicle. you will have to play around.