C++ Rotation function works in 2D perspective but fails in 3D

Click For Summary

Discussion Overview

The discussion revolves around a C++ rotation function intended for rotating points in 3D space. Participants explore issues related to the function's performance in 3D compared to its success in 2D, particularly focusing on the behavior of points during rotation and the handling of coordinate transformations.

Discussion Character

  • Technical explanation
  • Debate/contested
  • Mathematical reasoning

Main Points Raised

  • One participant describes a problem where points rotate towards the center of the object in 3D, despite the function working correctly in 2D.
  • Another participant suggests that the issue may stem from passing modified values of coordinates to the rotation function instead of the original values.
  • A participant acknowledges the oversight regarding the order of operations in the rotation calls and expresses gratitude for the clarification.
  • A later post indicates that after rewriting the rotation function, the participant encounters a new issue where the line shrinks during rotation when using values from an array, despite fixed values working correctly.
  • The updated rotation function is presented, which uses different logic based on the angle of rotation, but it is unclear if this resolves the shrinking issue.

Areas of Agreement / Disagreement

Participants generally agree on the need to address the order of operations in the rotation calls, but multiple competing views and unresolved issues remain regarding the effectiveness of the updated rotation function and the behavior of points during rotation.

Contextual Notes

Participants discuss various assumptions about coordinate transformations and the implications of using integer versus floating-point arithmetic in the rotation calculations. There are also unresolved mathematical steps in the updated rotation function that may affect its performance.

dot_binary
Messages
6
Reaction score
0
I'm making a 3D rotation of a poligon.The problem is when I rotate some points in 2D, it works, but if I use the same function to rotate the same points in a 3D perspective , while rotating the vertices move to the center of the object and remain there.
I also wrote the 3D rendering part.At first I thought that my rotation function is not precise, I made it more precise and still it it doesn't work.The 3D rendering function only takes the values stored in the objects array and puts vertices on screen.

The values stored in o[0][0] and o[0][1] represent the X and Y coordonates of the origin
and the ones stored in o[1][0] and o[1][1] are the coordonates of a point.
For now I'm only rotating on X and Y axis.

This is the rotation function:

int rotate(int t,double Ox, double Oy , double Ox2, double Oy2 , double Ub ){
double pi = 3.14159265 ;

if( Ub > 359 ){ Ub = Ub-360; }
if( Ub < 0 ){ Ub = 360-abs(int(Ub)); }

double U = (Ub*pi)/180 ;
double Tx = Ox2 - Ox ;
double Ty = Oy2 - Oy ;
double x = ( Tx *cos( U ) ) - ( Ty *sin( U ) );
double y = ( Ty *cos( U ) ) + ( Tx *sin( U ) );

Ox2 = floor( (Ox + x) + 0.5 ) ;
Oy2 = floor( (Oy + y) + 0.5 ) ;

if ( t == 1 ){
return int(Ox2) ;
}else{
return int(Oy2) ;
}
}


This is how I'm doing the rotation of one point around another point:

o[1][0] = rotate(1,o[0][0],o[0][1],o[1][0],o[1][1],Ub2);
o[1][1] = rotate(2,o[0][0],o[0][1],o[1][0],o[1][1],Ub2);

where Ub2 is the angle of rotation.
 
Technology news on Phys.org
You didn't explain why you are doing all the conversions between int to double so I don't really understand what the routine is doing, but I'll take a guess at the problem.

The first call to rotate() changes the values of o[1][0].
In the second call, should you pass in the original value of o[1][0], not the changed value?

Ignoring the int-to-double conversions, it would probably be better to change the routine to
Code:
void rotate(int t,double Ox, double Oy , double& Ox2, double& Oy2 , double Ub )
and update both of Ox2 and Oy2 with one call
 
The first call changed the value of o[1][0] , that was the problem.I overlooked that.
Thank you for pointing that out.
 
I still have a problem with my rotation function.

I rewrote the rotation function because the previous one didn't work as I wanted.
Now if I rotate a line in a 2D space, if I give it fixed values linke this:

line(400,400,rotate(1,400,400,400,600,Ub2),rotate(2,400,400,400,600,Ub2));

...it works but if the values come from an array, the line shrinks while rotating.And I'm shure I didn't made the mistake I previously made.
This is how I do the rotation now:

rx = rotate(1,aq[1][1],aq[1][2],aq[2][1],aq[2][2],Ub2) ;
ry = rotate(2,aq[1][1],aq[1][2],aq[2][1],aq[2][2],Ub2);
aq[2][1] = rx ;
aq[2][2] = ry ;
line(aq[1][1],aq[1][2],aq[2][1],aq[2][2]);

The rotation function:

int rotate(int u ,int Ox, int Oy , int Ox2, int Oy2 , float Ub ){
float PI = 3.1415926535;

if( Ub > 359 ){ Ub = Ub-360; }
if( Ub < 0 ){ Ub = 360-abs(int(Ub)); }

int Tx = abs(Ox2 - Ox) ;
int Ty = abs(Oy2 - Oy) ;

// get size of line
int hkj = int( round( sqrt( (Tx*Tx) + (Ty*Ty) ) ) );

if( ( Ub >=0 ) and ( Ub < 90 ) ){
Ox2 = int( sin((Ub*PI)/180) * hkj );
Oy2 = int( cos((Ub*PI)/180) * hkj );
}
if( ( Ub >=90 ) and ( Ub < 180 ) ){
Ox2 = int( cos(((Ub-90)*PI)/180) * hkj );
Oy2 = int( (sin(((Ub-90)*PI)/180) * hkj)* -1);
}
if( ( Ub >=180 ) and ( Ub < 270 ) ){
Ox2 = int( (sin(((Ub-180)*PI)/180) * hkj )* -1);
Oy2 = int( (cos(((Ub-180)*PI)/180) * hkj )* -1);
}
if( ( Ub >=270 ) and ( Ub < 360 ) ){
Ox2 = int( (cos(((Ub-270)*PI)/180) * hkj )* -1 );
Oy2 = int( sin(((Ub-270)*PI)/180) * hkj ) ;
}

if ( u == 1 ){
return Ox+Ox2 ;
}else{
return Oy+Oy2 ;
}
}
 

Similar threads

Replies
1
Views
2K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 2 ·
Replies
2
Views
4K
  • · Replies 2 ·
Replies
2
Views
3K
Replies
5
Views
4K
  • · Replies 5 ·
Replies
5
Views
14K
  • · Replies 2 ·
Replies
2
Views
1K
  • · Replies 10 ·
Replies
10
Views
13K