Develop Reference

How to make 2d sprite move in a smooth wide arc in the direction of keypress?

I'm working on a small experimental project in Unity. I have a 2d sprite that moves forward with a velocity but I want it to turn left or right in a wide arc and keep moving in that direction on keypress.
I've tried to tweak its angular velocity to get the desired affect. Looks unnatural and it won't stop rotating.
Tried Lerping. Looks unnatural as well.
Code Snippet 1:
bool forward = true;
Vector3 movement;
void FixedUpdate()
{
if (forward)
{
//Moves forward
movement = new Vector3(0.0f, 0.1f, 0.0f);
rb.velocity = movement * speed;
}
if (Input.GetKeyDown(KeyCode.LeftArrow))
{
forward = false;
movement = new Vector3(-0.05f, 0.05f, 0.0f);
rb.velocity = movement * speed;
rb.angularVelocity = 30;
}
if (transform.rotation.z == 90)
{
movement = new Vector3(-0.1f, 0.0f, 0.0f);
rb.velocity = movement * speed;
rb.angularVelocity = 0;
}
}
Code Snippet 2:
void Update(){
if (Input.GetKeyDown(KeyCode.LeftArrow))
{
Vector3 target = transform.position + new Vector3(-0.5f, 0.5f, 0);
transform.position
=Vector3.Lerp(transform.position,target,Time.deltaTime);
transform.eulerAngles = Vector3.Lerp(transform.rotation.eulerAngles,
new Vector3(0, 0, 90), Time.deltaTime);
}
}
Can anyone point me in the right direction of what is the actual correct way to implement this?

Not entirely sure if this is what youre trying to accomplish but here's some pseudo code of what I came up with to get you started...
Essentially, when a direction is pressed, you want to increase the velocity in that direction until all of the velocity is pointed that way. At the same time you want to decrease the velocity in the direction you were previously going until it is zero.
This is a simplified formula however - if you truly wanted the velocity to be constant throughout the entirety of the arc you would have to use some geometry, knowing that V=(velX^2 + velY^2)^.5 but this would get you pretty close...
float yvel = 1f, xvel;
float t;
void Update()
{
GetComponent<Rigidbody2D>().velocity = new Vector2(xvel, yvel);
t += Time.deltaTime;
if (Input.GetKeyDown(KeyCode.D))
{
t = 0;
StartCoroutine(Move());
}
}
private IEnumerator Move()
{
while (t < 2) // at time t, yvel will be zero and xvel will be 1
{
yvel = 1 - .5f * t; // decrease velocity in old direction
xvel = .5f * t; // increase velocity in new direction
yield return null;
}
}

How to move 2D Object within camera view boundary

I have a scene that my camera doesn't follow my player. When player reaches the end of camera I want player to can't go further (out of camera view). How can I do this?
My codes for movement
public class PlayerBlueController : MonoBehaviour {
public float speed;
private float x;
// Use this for initialization
void Start () {
}
// Update is called once per frame
void FixedUpdate () {
x = Input.GetAxis ("Horizontal") / 100 * speed;
transform.Translate (x,0,0);
}
}
As you can see from this. It gets out of camera's view.

I noticed you used a Collider2D. You should be using Rigidbody2D.MovePosition instead of transform.Translate or you'll likely run into issues when transform.Translate is used.
1.Take the final move position and convert it to new position in ViewPortPoint with Camera.main.WorldToViewportPoint
2.Apply a limit with Mathf.Clamp to the result in #1.
3.Convert the ViewPortPoint back to world point with Camera.main.ViewportToWorldPoint.
4.Finally, move it with Rigidbody2D.MovePosition.
The code below is modified from this answer to include restriction to screen boundary.
Move without Rigidbody:
Use only if collision and physics are NOT required:
public float speed = 100;
public Transform obj;
public void Update()
{
float h = Input.GetAxis("Horizontal");
float v = Input.GetAxis("Vertical");
//Move only if we actually pressed something
if ((h > 0 || v > 0) || (h < 0 || v < 0))
{
Vector3 tempVect = new Vector3(h, v, 0);
tempVect = tempVect.normalized * speed * Time.deltaTime;
Vector3 newPos = obj.transform.position + tempVect;
checkBoundary(newPos);
}
}
void checkBoundary(Vector3 newPos)
{
//Convert to camera view point
Vector3 camViewPoint = Camera.main.WorldToViewportPoint(newPos);
//Apply limit
camViewPoint.x = Mathf.Clamp(camViewPoint.x, 0.04f, 0.96f);
camViewPoint.y = Mathf.Clamp(camViewPoint.y, 0.07f, 0.93f);
//Convert to world point then apply result to the target object
obj.position = Camera.main.ViewportToWorldPoint(camViewPoint);
}
Move Object with Rigidbody2D:
Use if collision and physics are required:
public float speed = 100;
public Rigidbody2D rb;
public void Update()
{
float h = Input.GetAxis("Horizontal");
float v = Input.GetAxis("Vertical");
//Move only if we actually pressed something
if ((h > 0 || v > 0) || (h < 0 || v < 0))
{
Vector3 tempVect = new Vector3(h, v, 0);
tempVect = tempVect.normalized * speed * Time.deltaTime;
//rb.MovePosition(rb.transform.position + tempVect);
Vector3 newPos = rb.transform.position + tempVect;
checkBoundary(newPos);
}
}
void checkBoundary(Vector3 newPos)
{
//Convert to camera view point
Vector3 camViewPoint = Camera.main.WorldToViewportPoint(newPos);
//Apply limit
camViewPoint.x = Mathf.Clamp(camViewPoint.x, 0.04f, 0.96f);
camViewPoint.y = Mathf.Clamp(camViewPoint.y, 0.07f, 0.93f);
//Convert to world point then apply result to the target object
Vector3 finalPos = Camera.main.ViewportToWorldPoint(camViewPoint);
rb.MovePosition(finalPos);
}

image not respond .
but you can check player location
x = Input.GetAxis ("Horizontal") / 100 * speed;
if(gameobject.transform.x > someValue)
x=0
gameobject will be OBJECT in scene that u attach class to it.
another way is place 2 empty gameobject with collider as invisibleWall and get collider to player

Possible solution is:
1.Get the coordinates of your screen cornes (top left, top right, bottom left, bottom right). You can get this coordinates using Screen.height and Screen.width.
2.Convert this coordinates using the camera you need with Camera.ScreenToWorldPoint.
3.Get your player coordinates and check that they are inside the rect which is formed by 4 coordinates of the screen corners.

You need to limit your transform's position based on the edges of the camera. Here is an answer describing the different coordinate systems in unity
You're probably looking to do something like this:
float xMin = Camera.main.ViewportToWorldPoint(Vector3.zero).x;
float xMax = Camera.main.ViewportToWorldPoint(Vector3.one).x;
Vector3 currentPos = transform.position;
float dx = Input.GetAxis ("Horizontal") / 100 * speed;
Vector3 desiredPos = new Vector3(currentPos.x + dx, currentPos.y, currentPos.z);
Vector3 realPos = desiredPos;
if(desiredPos.x > xMax)
realPos.x = xMax;
else if(desiredPos.x < xMin)
realPos.x = xMin;
transform.position = realPos;
Read up here for more info on ViewportToWorldPoint(), it's extremely useful to become comfortable with the different coordinate spaces and how you can convert between them.

Mathf.clamp() function in Unity

I have a Cube(Player) in my game scene. I have written a C# script to constrain the movement(using Mathf.Clamp()) of the Cube so it does not leave the screen.
Below is my FixedUpdate() method from the Script
private void FixedUpdate()
{
float moveHorizontal = Input.GetAxis("Horizontal");
float moveVertical = Input.GetAxis("Vertical");
Vector3 movement = new Vector3(moveHorizontal, 0.0f, moveVertical);
rb.velocity = movement * speed;
rb.position = new Vector3(
Mathf.Clamp (rb.position.x, x_min, x_max),
0.5f,
Mathf.Clamp(rb.position.z, z_min, z_max)
);
}
Values of x_min, x_max, z_min, z_max are -3, 3, -2, 8 respectively inputted via the unity inspector.
PROBLEM
The script is working fine but my player(Cube) can move up to -3.1 units in negative X-AXIS (if I keep pressing the left arrow button) which is 0.1 units more in negative X-AXIS(This behavior is true for other axes too). It obviously clamps -3.1 to -3 when I stop pressing the button.
Why is this happening? Why doesn't it(Mathf.Clamp()) restrict the Cube to -3 units in first place?
Why am I getting that extra 0.1 units?
And Why is this value 0.1 units only? Why not more or less?

Your issue could be caused because you are setting velocity and then position, but before the next frame unity adds the velocity to your objects position. Which is why it ends up 0.1 units off.
To fix this, try resetting the velocity of the object to zero if it's about to leave your boundaries.

This is happening because the internal physics update moves the cube after you have put it in its clamped position. Mathf.Clamp() does restrict the cube to the expected position in the first place, as you can verify by putting a Debug.Log directly after you assign the position
You are getting the extra .1 units because you give the object a speed that gets applied after you clamp
It is 0.1 units because of the speed and the setting for Fixed Timestep (in Project Settings > Time). If you would set a higher speed it would be more. If you would lower the Fixed Timestep it would also be a bit more.

You already got the reason why this happens.
In order to fix the problem, you can use this code to prevent the cube to go outside your boundaries:
private void FixedUpdate() {
Vector3 movement = new Vector3(Input.GetAxis("Horizontal"), 0f, Input.GetAxis("Vertical"));
// Get new position in advance
float xPos = rb.position.x + movement.x * speed * Time.fixedDeltaTime;
float zPos = rb.position.z + movement.z * speed * Time.fixedDeltaTime;
// Check if new position goes over boundaries and if true clamp it
if (xPos > x_max || xPos < x_min) {
if (xPos > x_max)
rb.position = new Vector3(x_max, 0, rb.position.z);
else
rb.position = new Vector3(x_min, 0, rb.position.z);
movement.x = 0;
}
if (zPos > z_max || zPos < z_min) {
if (zPos > z_max)
rb.position = new Vector3(rb.position.x, 0, z_max);
else
rb.position = new Vector3(rb.position.x, 0, z_min);
movement.z = 0;
}
rb.velocity = movement * speed;
}

Unity - jerky motion when going down the slope

Im trying to make simple 3rd person character controller using unitys character controller component instead of rigidbody. I have problem when making my character sliding down the slope, the motion is jerky, just as if the character was going down stairs.
I move my character using normal to the ground by reversing its y axis, then i apply some additional gravity and put this vector to charactercontroller.move() function.
Here is some of the code where i apply slide and gravity:
void ProcessMotion(){
MoveVector = transform.TransformDirection (MoveVector);
if (MoveVector.magnitude > 1)
MoveVector = Vector3.Normalize (MoveVector);
ApplySlide ();
MoveVector *= MoveSpeed;
MoveVector = new Vector3 (MoveVector.x, VerticalVel, MoveVector.z);
ApplyGravity ();
TP_Controller.CharacterController.Move (MoveVector*Time.deltaTime);
}
public void Jump(){
if (TP_Controller.CharacterController.isGrounded) {
VerticalVel=jumpSpeed;
}
}
void SnapAlignCharacterWithCamera(){
if (MoveVector.x != 0 || MoveVector.z != 0) {
transform.rotation = Quaternion.Euler(transform.eulerAngles.x, Camera.main.transform.eulerAngles.y, transform.eulerAngles.z);
}
}
void ApplyGravity(){
if (MoveVector.y > -TermVel) {
MoveVector = new Vector3 (MoveVector.x, MoveVector.y - Gravity * Time.deltaTime, MoveVector.z);
}
if (TP_Controller.CharacterController.isGrounded && MoveVector.y < - 1) {
MoveVector = new Vector3 (MoveVector.x, -1, MoveVector.z);
}
}
void ApplySlide(){
if (!TP_Controller.CharacterController.isGrounded) {
return;
}
SlideDirection = Vector3.zero;
RaycastHit hitInfo;
if (Physics.Raycast (transform.position , Vector3.down, out hitInfo)) {
if(hitInfo.normal.y < SlideTreshold){
SlideDirection = new Vector3(hitInfo.normal.x, -hitInfo.normal.y, hitInfo.normal.z)*10;
}
}
if (SlideDirection.magnitude < MaxMagnitude) {
MoveVector += SlideDirection;
//Debug.DrawLine (transform.position,transform.position + new Vector3(hitInfo.normal.x*0.5f,-hitInfo.normal.y,hitInfo.normal.z*0.5f), Color.red,1.0f);
}else {
MoveVector = SlideDirection;
}
}
And here are screens with gizmos that show path of the object:
Sliding slowly
Sliding 10xfaster
In advance thanks for your help!

I'm not sure the scale of your objects, but it looks like your raycast is coming from inside of your capsule, and possibly colliding with the bottom. The resulting RaycastHit would report that information, which is a normal that points downward, so when you reverse it, you are heading upwards, creating the sawtooth effect.

Keeping my object from rolling out of the screen

I am using the tilt function on a phone to control an object to roll left and right. I do not want it roll anything beyond the screen width.
As in the simple illustration below, the dotted lines represent the width of the screen. The 'O' is the object and the Max signs indicate the maximum point the object is allowed to roll to.
Max--------O--------Max
But currently using my code, the object still rolls out of the screen. Also tried testing both height n width and ended up the same result where the object rolls out of the screen. Please advice what I am doing wrong. Thank you.
public float speed = 10.0F;
void Update()
{
Vector3 dir = Vector3.zero;
dir.x = Input.acceleration.x;
if (dir.sqrMagnitude > 1)
dir.Normalize();
dir *= Time.deltaTime;
if (!(Mathf.Round(dir.x) > Mathf.Round(Screen.width/2)) || !(Mathf.Round(dir.x) < -Mathf.Round(Screen.width/2)))
{
transform.Translate(dir * speed);
}
}
**Updated
public float speed = 10.0F;
void Update()
{
Vector3 dir = Vector3.zero;
dir.x = Input.acceleration.x;
if (dir.sqrMagnitude > 1)
dir.Normalize();
dir *= Time.deltaTime;
//transform.Translate(dir * speed);
if (transform.position.x < 0)
{
transform.position = new Vector3(0, this.transform.position.y, this.transform.position.z);
}
else if (transform.position.x > Screen.width)
{
transform.position = new Vector3(Screen.width, this.transform.position.y, this.transform.position.z);
}
else
{
transform.Translate(dir * speed);
}
}

I think thereare several ways you can go about checking the transfomrs position.
first off if you were using a 2d camera you could use a method like
leftBounds = Camera.x - (Camera.pixelWidth/2);
however because ortho camera angles are not set at any particulare size at x distace from camera they are hard to calculate.
i have seen some instances were coliders on the camera just outside the render rand were placed as camera children
adding a colision mask to only affect the appropriate game object would be best.