Jump to content

  • Log In with Google      Sign In   
  • Create Account

View Matrix Decomposition (Direct3D)


Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.

  • You cannot reply to this topic
5 replies to this topic

#1 lucky6969b   Members   -  Reputation: 634

Like
0Likes
Like

Posted 18 December 2012 - 11:19 PM

I want to retrieve back the eye position of the camera when a view matrix is given.
How can I achieve that? Any good articles that I can take a look at?
Thanks
Jack

Sponsor:

#2 Solid_Spy   Members   -  Reputation: 452

Like
0Likes
Like

Posted 19 December 2012 - 01:10 AM

do you have variables that are being plugged into the x, y, and z values of the camera's position? If you do, then it should be easy to just display it with Directx text or GDI text. If you do have variables plugged in, are they global? If not(which they shouldn't, globals are evil ;) ) then do you have a way of Setting and Getting them?

#3 BornToCode   Members   -  Reputation: 948

Like
0Likes
Like

Posted 19 December 2012 - 01:11 AM

Depending on how your matrix is set up. If you are using row major it will be the last row of your matrix. if you are using Column major it will be the last column of your matrix. Or to put it in better perspective. If you your matrix is represented as a single dimensional array it will be index 12,13,14.

Edited by BornToCode, 19 December 2012 - 01:54 AM.


#4 L. Spiro   Crossbones+   -  Reputation: 14248

Like
0Likes
Like

Posted 19 December 2012 - 03:31 AM

Extending off the above post, the view matrix is actually the inverse of the camera matrix. You should first invert the view matrix to get the camera matrix and then extract the position from that.
Since this is DirectX, once you have the camera matrix, the position will be in the _41, _42, and _43 members of the matrix.

But world, view, and projection matrices are concepts generally reserved for graphics processing and nothing else. Meaning you are not meant to extract anything from them except things directly related to graphics, such as view frustums etc. In other words, whatever the reason you have for getting the camera’s position this way is invalid and a hack. It indicates you have a flaw in your design that should be addressed.


L. Spiro
It is amazing how often people try to be unique, and yet they are always trying to make others be like them. - L. Spiro 2011
I spent most of my life learning the courage it takes to go out and get what I want. Now that I have it, I am not sure exactly what it is that I want. - L. Spiro 2013
I went to my local Subway once to find some guy yelling at the staff. When someone finally came to take my order and asked, “May I help you?”, I replied, “Yeah, I’ll have one asshole to go.”
L. Spiro Engine: http://lspiroengine.com
L. Spiro Engine Forums: http://lspiroengine.com/forums

#5 lucky6969b   Members   -  Reputation: 634

Like
0Likes
Like

Posted 21 December 2012 - 08:34 PM

Comment on this code snippet?
The eye pos is very weird.
Thanks
Jack
	//// Start exporting View and projection matrices
   Interface *ip2 = GetCOREInterface();
   ViewExp * pView = ip2->GetActiveViewport();  //  Get the viewport in question
   GraphicsWindow *gw = pView->getGW();  //  Get the GraphicsWindow context
   gw->getCameraMatrix( mat, &invTM, &persp, &hither, &yon); // getting these values to work with ... see above for their types
   float oneOverDepth = 1.0f / (yon - hither);
   // Set the Direct3D Camera View Position and Camera Projection Transforms.
   //
   // The first matrix is the full projection transformation matrix that
   // converts World Coordinates into NPC.  This means that the matrix is the
   // product of the Camera View Position transformation matrix and the Camera
   // Projection matrix.  The second matrix is the inverse of the Camera View
   // Position transformation matrix so if we multiply this second matrix by
   // the first, we get the Camera Projection matrix.  If we take the inverse
   // of the second matrix, we get the Camera View Position matrix.
   //
   // The Camera View Position transformation converts World coordinates into
   // Camera View Position coordinates where the camera is located at the
   // origin.  We have been given the inverse of the Camera View Position
   // matrix so the first step is to take the inverse of this transform to
   // obtain the Camera View Position matrix.
   // General conversion from 3ds max coords to Direct3D coords:
   //
   // 3ds max:  (Up, Front, Right) == (+Z, +Y, +X)
   //
   // Direct3D: (Up, Front, Right) == (+Y, +Z, +X)
   //
   // Conversion from 3ds max to Direct3D coords:
   //
   // 3ds max * conversion matrix = Direct3D
   //
   // [ x y z w ] * | +1  0  0  0 | = [ X Y Z W ]
   //               |  0  0 +1  0 |
   //               |  0 +1  0  0 |
   //               |  0  0  0 +1 |
   //
   // The View transform below accomplishes this.  The standard View transform
   // received makes the rotation about the X axis because the assumption was
   // to transform to RH coords with the XY plane being the vertical plane
   // instead of the XZ plane.  The negation of the the Z column does the RH
   // to LH flip.  Thus, the View transform makes the transition from RH 3ds
   // max coords to LH Direct3D coords.
   // View Matrix in 3ds max, inverse's inverse to become original [Jacky Luk]
	Matrix3 camTM = Inverse(invTM);

		 
   // We now have an affine matrix (4x3) with no perspective column (it is
   // understood to be (0, 0, 0, 1)).  We add the fourth column and flip the
   // Z-axis because Direct3D uses a left-handed coordinate system and MAX
   // uses a right-handed coordinate system.
   // Copy the affine view matrix data


   int ki, kj;
   MRow *pcvm = camTM.GetAddr();
   for (ki = 0; ki < 4; ki++) {
 	for (kj = 0; kj < 3; kj++) {
		d3dViewXform.m[ki][kj] = pcvm[ki][kj];
	}
   }
   // Assign the fourth column (perspective terms)
   d3dViewXform.m[0][3] = d3dViewXform.m[1][3] = d3dViewXform.m[2][3] = 0.0f;
   d3dViewXform.m[3][3] = 1.0f;

    	
   // Scale the Z-axis (third column) by -1 to flip to left-handed Direct3D
   // coordinate system

   for (ki = 0; ki < 4; ki++) {
	d3dViewXform.m[ki][2] *= -1.0f;
   }


   // Calculate the Direct3D Camera Projection transformation matrix.
   //
   // First, multiply the MAX full projection matrix by the inverse of the MAX
   // Camera View Position matrix to obtain the MAX Camera Projection matrix.
   //
   // This gives us a correct Direct3D Camera Projection matrix except for the
   // lower right quadrant.
   //

   MRow *pa = invTM.GetAddr();
   for (ki = 0; ki < 4; ki++) {
     float val = (float)(ki==3);
     for (kj = 0; kj < 4; kj++) {
	d3dProjXform.m[ki][kj] = pa[ki][0] * mat[0][kj] +
	pa[ki][1] * mat[1][kj] +
	pa[ki][2] * mat[2][kj] +
	val		 * mat[3][kj];
    }
  }

  // Now calculate the lower right quadrant of the Camera Projection matrix
  // using the facts that MAX uses an NPC Z-axis range of +1 to -1 whereas
  // Direct3D uses an NPC Z-axis range of zero to +1.
  //
  // For ease of reference, the general forms of the Direct3D Projection
  // matrix for perspective and orthographic projections are given below.
  //
  // Please note that the matrices are specified in row-major order.  This
  // means that the translate terms are located in the fourth row and the
  // projection terms in the fourth column.  This is consistent with the way
  // MAX, Direct3D, and OpenGL all handle matrices.  Even though the OpenGL
  // documentation is in column-major form, the OpenGL code is designed to
  // handle matrix operations in row-major form.
  if (persp) {
      // Perspective projection.  The general form of the Direct3D Camera
      // Projection matrix is:
      //
      // |    2n/(r-l)       0            0            0       |
      // |      0          2n/(t-b)       0            0       |
      // | (r+l)/(r-l)  (t+b)/(t-b)      f/(f-n)       1       |
      // |      0            0         -fn/(f-n)       0       |
      // 
      // Construct the lower right four terms correctly for Direct3D.
      //
      d3dProjXform.m[2][2] = yon*oneOverDepth;
      d3dProjXform.m[2][3] = 1.0f;
      d3dProjXform.m[3][2] = -(yon*hither*oneOverDepth);
      d3dProjXform.m[3][3] = 0.0f;
		
   } else {

    	
      // Orthographic projection.  The general form of the Direct3D Camera
      // Projection matrix is:
      //
      // |     2/(r-l)       0            0            0       |
      // |      0           2/(t-b)       0            0       |
      // |      0            0           1/(f-n)       0       |
      // | (r+l)/(r-l)  (t+b)/(t-b)     -n/(f-n)       1       |
      // 
      // Construct the lower right four terms correctly for Direct3D.
      //

      d3dProjXform.m[2][2] = oneOverDepth;
      d3dProjXform.m[2][3] = 0.0f;
      d3dProjXform.m[3][2] = -(hither*oneOverDepth);
      d3dProjXform.m[3][3] = 1.0f;
  }
  D3DXVECTOR3 v;
  D3DXQUATERNION q;
  D3DXVECTOR3 s;

  D3DXMatrixDecompose(&s, &q, &v, &d3dViewXform);
  DOMElement *elem1 = doc->createElement(L"Eye-Pos");

Edited by lucky6969b, 21 December 2012 - 08:45 PM.


#6 Medo3337   Members   -  Reputation: 680

Like
0Likes
Like

Posted 21 December 2012 - 09:15 PM

Check out this link:<br />http://stackoverflow.com/questions/6030933/get-eye-target-and-up-vectors-from-view-matrix




Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.



PARTNERS