btMatrix3x3 getEulerZYX singularity wrong values

[GoogleCodeExporter]

What steps will reproduce the problem?
1. create a btMatrix3x3 with m_el[2].x = +1 or -1

What version of the product are you using? On what operating system?
2.75

Please provide any additional information below.

With theese formulas
http://www.gregslabaugh.name/publications/euler.pdf

it works

the corrected method is like this: in mentioned paper yaw = ф, pitch = ψ,
roll = ψ. seems that yaw & roll are switched. 

void getEulerZYX(btScalar& yaw, btScalar& pitch, btScalar& roll, unsigned
int solution_number = 1) const
  {
    struct Euler{btScalar yaw, pitch, roll;};
    Euler euler_out;
    Euler euler_out2; //second solution
    //get the pointer to the raw data

    // Check that pitch is not at a singularity
    if (btFabs(m_el[2].x()) >= 1)
    {
      euler_out.yaw = 0;
      euler_out2.yaw = 0;

      // From difference of angles formula
      btScalar delta = btAtan2(m_el[0].y(),m_el[0].z());
      if (m_el[2].x() < 0)  //gimbal locked up
      {
        euler_out.pitch = SIMD_PI / btScalar(2.0);
        euler_out2.pitch = SIMD_PI / btScalar(2.0);
        euler_out.roll = euler_out.yaw + delta;
        euler_out2.roll = euler_out.yaw + delta;
      }
      else // gimbal locked down
      {
        euler_out.pitch = -SIMD_PI / btScalar(2.0);
        euler_out2.pitch = -SIMD_PI / btScalar(2.0);
        euler_out.roll = -euler_out.yaw - delta;
        euler_out2.roll = -euler_out.yaw - delta;
      }
    }
    else
    {
      euler_out.pitch = - btAsin(m_el[2].x());
      euler_out2.pitch = SIMD_PI - euler_out.pitch;

      euler_out.roll = btAtan2(m_el[2].y()/btCos(euler_out.pitch), 
                   m_el[2].z()/btCos(euler_out.pitch));
      euler_out2.roll = btAtan2(m_el[2].y()/btCos(euler_out2.pitch), 
                m_el[2].z()/btCos(euler_out2.pitch));

      euler_out.yaw = btAtan2(m_el[1].x()/btCos(euler_out.pitch), 
                  m_el[0].x()/btCos(euler_out.pitch));
      euler_out2.yaw = btAtan2(m_el[1].x()/btCos(euler_out2.pitch), 
                               m_el[0].x()/btCos(euler_out2.pitch));
    }

    if (solution_number == 1)
    { 
        yaw = euler_out.yaw; 
        pitch = euler_out.pitch;
        roll = euler_out.roll;
    }
    else
    { 
        yaw = euler_out2.yaw; 
        pitch = euler_out2.pitch;
        roll = euler_out2.roll;
    }
  }

Original issue reported on code.google.com by plbar...@gmail.com on 13 Jan 2010 at 11:54

[GoogleCodeExporter]

Thank for the report.

We probably remove all euler support in Bullet 3.x and provide some separate 
utility 
functions for developers.

Original comment by erwin.coumans on 20 Jan 2010 at 4:14

• Added labels: Milestone-3.x

[GoogleCodeExporter]

well I have already done the functions i need, you mean something like that, 
isn't it?

Original comment by plbar...@gmail.com on 22 Jan 2010 at 2:15

Attachments:

• btMatrix3x3Euler.h
• redondeo.cpp

[GoogleCodeExporter]

a new problemm appeared at gimbal block at pitch = - PI/2.

cos ( yaw + roll ) = -M13
-sin ( yaw + roll ) = M12

then

cos ( yaw + roll ) = -M13
sin ( yaw + roll ) = -M12

tan ( yaw + roll ) = M11 / M13. 

is not valid whwen using atan2,

yaw  +  roll = atan2 ( M12,  M13) incorrect values
yaw  +  roll = atan2 (-M12, -M13) correct values

Sorry last time i misleaded the correct files.

Original comment by plbar...@gmail.com on 24 Feb 2010 at 4:41

Attachments:

• btMatrix3x3Euler.h
• btMatrix3x3Euler.cpp

[GoogleCodeExporter]

Closing issues that we are not fixing anymore, our focus is on Bullet 3.x at 
http://github.com/erwincoumans/bullet3

Thanks for the contribution anyway,
Erwin

Original comment by erwin.coumans on 11 Sep 2013 at 12:12

• Changed state: WontFix