Using my own camera

I'm using my own camera code so I'm setting up the modelview- and projection-stacks myself.

To render I need a SIO2 Camera so I set one up like this:

Code:

sio2->_SIO2camera = sio2CameraInit("GameCamera");

And then to render the resources:

Code:

sio2CameraUpdateFrustum( sio2->_SIO2camera );
sio2ResourceCull( sio2->_SIO2resource, sio2->_SIO2camera );
sio2ResourceRender( sio2->_SIO2resource, sio2->_SIO2window,   sio2->_SIO2camera, SIO2_RENDER_SOLID_OBJECT | SIO2_RENDER_CLIPPED_OBJECT );


Now, I created a scene in Blender with just the monkey head, no materials, didn't change any settings. It looks like every other face is backface-culled (kind of like when the vertex winding is messed up when stripping for instance). I'm thinking it must be the data but I just thought I'd get some advice here before I start debugging...

sio2CameraSetPerspective ???

and do NOT set your camera look at {0 , 0 ,0 }

I'm setting up the modelview- AND projection-stacks with my own code. All my debug drawing works just fine.

My projection matrix is different than the SIO2 projection matrix but that shouldn't give me the issue I'm seeing.

I am just begin to explore SIO2, seeing if it makes sense for my project. I am wondering if it is possible to design portions of the game in blender and others outside? As the first poster seems to have alluded to, I think that I would like to use an opengl camera not defined by a blender camera. Is this possible?

Thank you.

Absolutely the culling system is based on the reconstruction of frustum planes, simply render your camera, then use a dummy camera to store the current frustum.

Code:



// Render your camera

// Use a dummy camera to hold the frustum data
sio2CameraUpdateFrustum( sio2->_SIO2camera );
      
// Cull the objects in the resource base on that frustum
sio2ResourceCull( sio2->_SIO2resource, sio2->_SIO2camera );

// Render the objects   
sio2ResourceRender( sio2->_SIO2resource, sio2->_SIO2window, sio2->_SIO2camera, SIO2_RENDER_SOLID_OBJECT | SIO2_RENDER_ALPHA_OBJECT );



If you need to use other SIO2 features that are based on the SIO2camera data (in example the billboards), just fill in the data manually from your camera data and that's it

"do NOT set your camera look at {0 , 0 ,0 }"


Why?


-j

I am trying to alter Tutorial2 to use a projection matrix defined by glfrustumf. I've added these lines per Mr. SIO2's instruction. What i've done causes the program to break when it reaches SIO2's getprojectionmatrix method. This is the function where it breaks. This makes me think that there is no projection matrix to get.

First: Does anyone know what I'm doing wrong?
Second: How do I position and orient an opengl camera relative to an SIO2 scene?

Code:


         //Set the OpenGL projection matrix
         glMatrixMode(GL_PROJECTION);
         glFrustumf(-.5,.5,-.5,.5,.01,1000);
         //glViewport(0, 0, 320, 240);
         
         // Use a dummy camera to hold the frustum data
         sio2CameraUpdateFrustum( sio2->_SIO2camera );
         
         // Cull the objects in the resource base on that frustum
         //sio2ResourceCull( sio2->_SIO2resource, sio2->_SIO2camera );
         // Render our camera.
         sio2CameraRender( _SIO2camera );


I'm hoping to get a hold of working with projection matrices and the SIO2 structure.

Thanks for the help,
Jesse

Put back the matrix mode to GL_MODELVIEW before sio2CameraUpdateFrustum

Thanks for the response. Setting it back to model_view doesn't fix the problem. The code breaks at the same point.

Have you initialize SIO2?

I am working from tutorial2 so where in tutorial2 is the sio2 object initialized? I don't see it in template.mm

I used this in template.m in tutorial 15 to use my own camera.
I just ignored sio2->_SIO2camera. I guess you could initialize it to zero.

Code:

void templateRender( void )
{
   glMatrixMode( GL_MODELVIEW );
   glLoadIdentity();

   glClear( GL_DEPTH_BUFFER_BIT | GL_COLOR_BUFFER_BIT );

/*
   if( !sio2->_SIO2camera )
   {
   SIO2camera *_SIO2camera = ( SIO2camera * )sio2ResourceGet( sio2->_SIO2resource, SIO2_CAMERA, "camera/Camera" );
   if( !_SIO2camera )
   { return; }

   sio2->_SIO2camera = _SIO2camera;
      
   sio2CameraSetPerspective( _SIO2camera, sio2->_SIO2window );
   }
*/   
   
   // Setup Camera Manually
   glMatrixMode( GL_PROJECTION );
   glLoadIdentity();
   
   glFrustumf( -0.0005705f, 0.0005705f, -0.000761f,  0.000761f, 0.1f, 10000.0f );
   cam_lookAt(8.904, -23.491, 10.53, 0, 0, 0, 0.0f, 0.0f, -1.0f);
   
   glMatrixMode( GL_MODELVIEW );

   // sio2WindowEnterLandscape3D();
   
   
   {
      // Use a static time step for the physic simulation.
      sio2PhysicRender( sio2->_SIO2physic, 1.0f / 60.0f, 0 );         
   
      // sio2CameraRender( sio2->_SIO2camera );

      //sio2CameraUpdateListener( (SIO2camera*) sio2ResourceGet( sio2->_SIO2resource,
       //                                         SIO2_CAMERA,
       //                                            "camera/Camera" ) );

      // We have an ambient sound that is streaming from
      // memory, this command will automatically update
      // the stream and swap the buffer. Toggle the "UseAlpha"
      // button in the Map Image panel withing blender to
      // activate it. However when using double buffer the
      // soundbuffer cannot be indepently shared among
      // sound source. Since the sound buffer will be
      // dynamically updated on a global scale, it is impossible
      // to have an independent playback on a sound source
      // basis. Use it with care... my suggestion is that the
      // ambient music should be streamed and all other
      // sound FX should be directly pushed into memory.
                           
      sio2ResourceRender( sio2->_SIO2resource,
                     sio2->_SIO2window,
                     sio2->_SIO2camera,
                     SIO2_RENDER_SOLID_OBJECT | SIO2_UPDATE_SOUND_STREAM );

   }
   
   //sio2WindowLeaveLandscape3D();
}

Thanks for the help. Would you mind showing me your camlookat function?

Its probably the same as sio2LookAt

meteors wrote:"do NOT set your camera look at {0 , 0 ,0 }"


Why?


-j

I tried this and searched around - you can't look directly up or down the Z axis. I would think it is something to do with the translation/rotation matrices, but without delving _too_ deep, I can't answer for sure. If you're going to lock the axis to look straight down (i.e. 2D) then look for the details on this forum about setting 2D mode.
I did what someone else suggested - locked it to the X or Y axis, as I'm not using physics but still want 3D objects in a 2D plane.

Yeah, it is pretty much the same as sio2LookAt, but it is an all in one function (other than opengl and math calls).
Based on the mesa 3d lib version.

/

Code:

/ ====================================================================================== //
// // Adapted from the gluLookAt Mesa Graphics lib ( http://www.mesa3d.org) and the Oolong3d engine.
// ====================================================================================== //
void cam_lookAt(GLfloat eyex, GLfloat eyey, GLfloat eyez, GLfloat centerx, GLfloat centery, GLfloat centerz, GLfloat upx, GLfloat upy, GLfloat upz) {
      GLfloat m[16];
      float fmult, magf;
      GLfloat z1, z2, z3, y1, y2, y3, x1, x2, x3;
      
      z1 = eyex - centerx;
      z2 = eyey - centery;
      z3 = eyez - centerz;
      magf = sqrtf( z1 * z1 + z2 * z2 + z3 * z3);
      if ( magf != 0 ) {
         fmult = ( 1.0f / magf);
      } else {
         fmult = 0;
      }
      z1 *= fmult;
      z2 *= fmult;
      z3 *= fmult;
       
      x1 = upy * z3 - upz * z2;
      x2 = -upx * z3 + upz * z1;
      x3 = upx * z2 - upy * z1;
      y1 = z2 * x3 - z3 * x2;
      y2 = -z1 * x3 + z3 * x1;
      y3 = z1 * x2 - z2 * x1;
   
      magf = sqrtf(  x1 * x1 + x2 * x2 + x3 * x3);
      if ( magf != 0 ) {
         fmult = (1.0f / magf);
      } else {
         fmult = 0;
      }

      x1 *= fmult;
      x2 *= fmult;
      x3 *= fmult;

      magf = sqrtf( y1 * y1 + y2 * y2 + y3 * y3 );
      if ( magf != 0 ) {
         fmult = ( 1.0f / magf);
      } else {
         fmult = 0;
      }
      y1 *= fmult;
      y2 *= fmult;
      y3 *= fmult;
   
      m[0] = x1;
      m[4] = x2;
      m[8] = x3;
      
      m[1] = y1;
      m[5] = y2;
      m[9] = y3;
      
      m[2] = z1;
      m[6] = z2;
      m[10] = z3;
      
      m[12] = m[13] = m[14] = m[3] = m[7] = m[11] = 0.0f;
      
      m[15] = 1.0f;

      glMultMatrixf( m );
      glTranslatef(-eyex, -eyey, -eyez);
   
} // cam_lookAt




Thanks so much, I got it working. I am curious what I was doing wrong with the SIO2 code, maybe that will become more obvious when I am more familiar.