[vtkusers] What co-ordinate system does the return value ofcamera->GetViewTransformMatrix() represent?

[Veerapuram Varadhan]

What for the 4th row of the view matrix is used?

If my math looks ok, here is another question.

Suppose after transformation, my original point (100,100,0) gets
transformed into (-210,320,112).  What does the negative value
signify?

> I'd post back to the mailing list, but I just switched email and can't
> get my return address correct for a little bit, so the mailing list
> won't accept my posts.
> 
I am keeping all our conversation and cc'ing it to the mailing list as well.

V. Varadhan.


On Fri, 16 Jul 2004 15:08:08 -0400, Andrew J. Dolgert
<ajd27 at tc.cornell.edu> wrote:
> Your math looks correct.  You can look in vtkMatrix4x4.cxx for
> vtkMatrixMultiplyPoint in order to see it typed out.  The only
> difference is that the code uses 4x4 matrices instead of the 3x3 you
> show.   Each tM->Element[i] is a row of the view matrix.
> 
> I'd post back to the mailing list, but I just switched email and can't
> get my return address correct for a little bit, so the mailing list
> won't accept my posts.
> 
> Drew
> 
> > -----Original Message-----
> > From: Veerapuram Varadhan [mailto:v.varadhan at gmail.com]
> > Sent: Friday, July 16, 2004 2:31 PM
> > To: Andrew J. Dolgert
> > Cc: VtkUsers
> > Subject: Re: [vtkusers] What co-ordinate system does the
> > return value ofcamera->GetViewTransformMatrix() represent?
> >
> > Hi Drew,
> >
> > Thanks for your time and explanation.
> >
> > Though, I am confused with the values in the transformation matrix.
> >
> > For ex:  Let us consider a point in world co-ordinate system,
> > say, (100,100,0).  Now, I perform a rotation.  Now, I want to
> > see the transformed position of the point (100,100,0).  Now,
> > I get the transformation matrix from the active camera.
> >
> > Suppose, if I get values like 
> > tM->Element[1] = {-0.900863, 0.2854167, -0.327081} Element[2] =
> > tM->{-0.196284, 0.9398594, -0.279532} Element[3] = { 0.381930,
> > tM->-0.1876199, -0.902707}
> 
> 
> >
> > When I do,
> > newPoint[0] = tM->Element[0][0] * 100+tM->Element[0][1] *
> > 100+tM->Element[0][2]*0;
> > newPoint[1] = tM->Element[1][0] * 100+tM->Element[1][1] *
> > 100+tM->Element[1][2]*0;
> > newPoint[2] = tM->Element[2][0] * 100+tM->Element[2][1] *
> > 100+tM->Element[2][2]*0;
> >
> > will newPoint have the correct transformed co-ordinate of the
> > point (100,100,0)?
> >
> > Am I missing anything?
> >
> > TIA,
> >
> > V. Varadhan.
> >
> > On Fri, 16 Jul 2004 14:02:58 -0400, Andrew J. Dolgert
> > <ajd27 at tc.cornell.edu> wrote:
> > >  Maybe I can answer this while my contour filter renders...
> > >
> > > There are three coordinate systems you usually work with:  camera,
> > > world, and actor.  The view transform expresses world
> > coordinates in
> > > the view space.  The actor transform expresses coordinates
> > in the view
> > > space.  You specify the location of actors and cameras using world
> > > coordinates, though.
> > >
> > > In math lingo, you might picture the transformation
> > matrices, M, like
> > > this:
> > >
> > > Camera -  c_M_w - World - w_M_a - actor
> > >
> > > Call our coordinate vector by v_w for a world coordinate, v_c for
> > > camera, and v_a for actor.
> > >
> > > v_c = c_M_w * v_w     and v_a = a_M_w * v_w
> > >
> > > The view transform is c_M_w.  It's inverse we call w_M_c, but it's
> > > really (c_M_w)^-1.
> > > The actor transform is a_M_w.  When VTK displays a point on the
> > > screen, it multiplies the actor coordinate by c_M_w *
> > w_M_a, which is
> > > (view) * (actor)^-1, and then it multiplies by other stuff
> > you don't
> > > seem to be delving into now.
> > >
> > > I don't think of the transformation matrix as being in any
> > particular
> > > coordinate system unless you want to think of that matrix
> > as composed
> > > of vectors for dot products.
> > >
> > > In answer to your TIA, I HTH.
> > > Drew
> > >
> > > > -----Original Message-----
> > > > From: vtkusers-bounces at vtk.org
> > > > [mailto:vtkusers-bounces at vtk.org] On Behalf Of Veerapuram Varadhan
> > > > Sent: Friday, July 16, 2004 1:21 PM
> > > > To: VtkUsers
> > > > Subject: [vtkusers] What co-ordinate system does the return value
> > > > ofcamera->GetViewTransformMatrix() represent?
> > > >
> > > > Hi All,
> > > >
> > > > I am using camera->GetViewTransformMatrix() to get the current
> > > > tranform matrix of the view.  According to vtkCoordinate
> > > > documentation, camera values are represented in "VIEW"
> > co-ordinate
> > > > system.  Is it correct? or Did I mis-interpret it?
> > > >
> > > > To find out the transformation of a point in World-Co-ordinate
> > > > system, should i need to convert the
> > > > camera->GetViewTransformMatrix() values to
> > World-Co-ordinate system?
> > > >
> > > > Any help is deeply appreciated.
> > > >
> > > > TIA,
> > > >
> > > > V. Varadhan.
> > > > _______________________________________________
> > > > This is the private VTK discussion list.
> > > > Please keep messages on-topic. Check the FAQ at:
> > > > <http://public.kitware.com/cgi-bin/vtkfaq>
> > > > Follow this link to subscribe/unsubscribe:
> > > > http://www.vtk.org/mailman/listinfo/vtkusers
> > > >
> > >
> >
>