plib-users

[Plib-users] OpenGL primitives

[Stafford G. <su...@ma...>]

Hi,

Im trying to debug some physics code i have, and i would like to draw a 
'force vector' (GL_LINE) indicating where the force is applied. However, when 
i do, it does not show. (Yes i did it between glBegin/glEnd pairs) I think it 
is because im also using the ssg library to draw the main scene.

Is it possible to draw using both OpenGL and ssg scenes at the same time?

Also, the reason i need to debug this code so intensely is because it seems 
that the axes are different depending on if you are rotating or translating... 
for translating x appears to be -, y is |, and z is . (into the screen), but 
for rotations, x is around |, y is around -, and z is the same. 
For example:

an object at (0, 0, 0):  
                                         -.
an object at (0, -2, 5):
                                              -.
an object at (0, -4, 0), rotated by (180, 0, 0):
                                         .-
                         .
the last one should be '- ' if the coordinate systems were the same.

Is this a bug in plib, or some strange 'feature' of OpenGL? I have tried both 
plib 1.0.20, and 1.1.11, and they are the same in this respect. This anomaly 
is giving me much trouble in creating rotations relative to an object (as one 
needs in physics). For example i can only get a helicopter to rotate around the 
global y axis (|) not its own y axis, which may be totally different.
See http://marys.dyndns.org/~surge/geome/ for some pictures of this.

Thanks for any help.

--
Stafford Goodsell  <su...@po...>                                     _
                                                                       C")
  Programmer, administrator, avid gamer and all-round computer geek    (_)
                 http://www.marys.dyndns.org/                          -"-

Re: [Plib-users] OpenGL primitives

[Land T. S. <lan...@am...>]

Hello Stafford,

I don't know about 2d drawing, but I have a starfield in the back of one of my
menus -- here is the code I use:

void StarField::Draw()
{
    int i;
    float *px, *py, *pz;
    static float c = 0.0f;

    glMatrixMode( GL_PROJECTION );
    glLoadIdentity();
    gluPerspective(90.0f, 640.0f/480.0f, 1.0, 500.0);


    glMatrixMode( GL_MODELVIEW );
    glLoadIdentity();
    gluLookAt( 0.0, 0.0, -5.0f,
           0.0, 0.0, 0.0,
           0.0, 1.0, 0.0 );

    px = x;
    py = y;
    pz = z;

    glDisable(GL_LIGHTING);
    glPointSize(1.0f);
    glBegin( GL_POINTS );

    for( i=0; i<numStars; i++ )
    {
 glColor3f( 1.0f - *pz / STAR_MAXZ,
     1.0f - *pz / STAR_MAXZ,
     1.0f - *pz / STAR_MAXZ);

  glVertex3f( *px, *py, *pz );
 px++; py++; pz++;
    }
    glEnd();
    glEnable(GL_LIGHTING);
    glDisable(GL_BLEND);
}

Stafford Goodsell wrote:

> Hi,
>
> Im trying to debug some physics code i have, and i would like to draw a
> 'force vector' (GL_LINE) indicating where the force is applied. However, when
> i do, it does not show. (Yes i did it between glBegin/glEnd pairs) I think it
> is because im also using the ssg library to draw the main scene.
>
> Is it possible to draw using both OpenGL and ssg scenes at the same time?
>
> Also, the reason i need to debug this code so intensely is because it seems
> that the axes are different depending on if you are rotating or translating...
> for translating x appears to be -, y is |, and z is . (into the screen), but
> for rotations, x is around |, y is around -, and z is the same.
> For example:
>
> an object at (0, 0, 0):
>                                          -.
> an object at (0, -2, 5):
>                                               -.
> an object at (0, -4, 0), rotated by (180, 0, 0):
>                                          .-
>                          .
> the last one should be '- ' if the coordinate systems were the same.
>
> Is this a bug in plib, or some strange 'feature' of OpenGL? I have tried both
> plib 1.0.20, and 1.1.11, and they are the same in this respect. This anomaly
> is giving me much trouble in creating rotations relative to an object (as one
> needs in physics). For example i can only get a helicopter to rotate around the
> global y axis (|) not its own y axis, which may be totally different.
> See http://marys.dyndns.org/~surge/geome/ for some pictures of this.
>
> Thanks for any help.
>
> --
> Stafford Goodsell  <su...@po...>                                     _
>                                                                        C")
>   Programmer, administrator, avid gamer and all-round computer geek    (_)
>                 http://www.marys.dyndns.org/                          -"-
>
> _______________________________________________
> plib-users mailing list
> pli...@li...
> http://lists.sourceforge.net/mailman/listinfo/plib-users

--
+-------------------------+--------------------------------------------+
| Landshark (dave)        | Visio clone for *nix? It's in the works:   |
| lan...@am... | http://queesio.sourceforge.net             |
+-------------------------+--------------------------------------------+
| X10 GUI Device Controller for *nix? http://q10.phlan.net             |
+----------------------------------------------------------------------+

RE: [Plib-users] OpenGL primitives

[Norman V. <nh...@ca...>]

Stafford Goodsell writes
>Sent: Thursday, June 08, 2000 9:28 PM
>To: pli...@li...
>Subject: [Plib-users] OpenGL primitives
>
>
>
>Hi,
>
>Also, the reason i need to debug this code so intensely is 
>because it seems that the axes are different depending on 
>if you are rotating or translating... 
>for translating x appears to be -, y is |, and z is . (into 
>the screen), but for rotations, x is around |, y is around -, 
and z is the same. 

My guess is that you are confused by the difference between the
coordinate systems used by SSG and OpenGL.

In OpenGL Z is going in and out of the screen
whereas in SSG Z is up down or Y in OpenGL

To convert from  SSG to OpenGL you can use the following
// hack
sgMat4 copy_of_ssgOpenGLAxisSwapMatrix =
{
  {  1.0f,  0.0f,  0.0f,  0.0f },
  {  0.0f,  0.0f, -1.0f,  0.0f },
  {  0.0f,  1.0f,  0.0f,  0.0f },
  {  0.0f,  0.0f,  0.0f,  1.0f }
} ;

sgMat4 vm_tmp, view_mat;
sgTransposeNegateMat4 ( vm_tmp, current_view.VIEW ) ;
sgCopyMat4( view_mat, copy_of_ssgOpenGLAxisSwapMatrix ) ;
sgPreMultMat4( view_mat, vm_tmp ) ;
glLoadMatrixf( (float *)view_mat );

Nice Looking Helicopter :-))

Have you checked out FlightGear.
http://www.flightgear.otg

It is using SSG for the majority of its rendering
and comes with source. 

Norman Vine

Re: [Plib-users] OpenGL primitives

[Steve B. <sjb...@ai...>]

Stafford Goodsell wrote:

> Im trying to debug some physics code i have, and i would like to draw a
> 'force vector' (GL_LINE) indicating where the force is applied. However, when
> i do, it does not show. (Yes i did it between glBegin/glEnd pairs) I think it
> is because im also using the ssg library to draw the main scene.

Well for starters, you need GL_LINES  the 'S' is important!  It's *really*
annoying that OpenGL defines a valid token 'GL_LINE' - but it's not the token
you pass to glBegin() - always use 'GL_LINES' - even if you are only drawing
one line!

However, that may not be your only problem because SSG generally leaves OpenGL
in an unspecified state - texture might be enabled - or lighting might be
turned on with the glMaterial specifying an alpha of zero (hence invisible
lines)...you really don't know...but we've thought about this, so...
 
> Is it possible to draw using both OpenGL and ssg scenes at the same time?

Yes - I do that all the time.

You'll probably need to do things like disabling texture, make sure that
the modelview and projection matrices are what you want (SSG has functions
to help with that).

The easiest way to get the OpenGL state to be what you want is to define
an ssgSimpleState object for your GL_LINES and to call:

     my_state->apply () ;

...just before you start drawing.  That way, SSG will change the current
state to do what you want without you having to worry about what states
SSG might have messed with.
 
> Also, the reason i need to debug this code so intensely is because it seems
> that the axes are different depending on if you are rotating or translating...
> for translating x appears to be -, y is |, and z is . (into the screen), but
> for rotations, x is around |, y is around -, and z is the same.

SSG uses a Z-is-up, X-is-right and Y-is-into-the-screen coordinate
system, with heading, pitch and roll defined appropriately for that
coordinate system.

> For example i can only get a helicopter to rotate around the
> global y axis (|) not its own y axis, which may be totally different.
> See http://marys.dyndns.org/~surge/geome/ for some pictures of this.

I think you need to read my paper:

  http://web2.airmail.net/sjbaker1/matrices_can_be_your_friends.html

-- 
Steve Baker                  http://web2.airmail.net/sjbaker1
sjb...@ai... (home)  http://www.woodsoup.org/~sbaker
sj...@ht...      (work)