[vtkusers] vtkHardwareSelection produces bad a bad result

[David E DeMarle]

This is an inherent limitation in the algorithm.

The hardware selector, and it predecessor vtkVisibleCellSelector, render an
image and determine what cell was drawn behind each pixel. I said cell not
cells because there is no way to recover more than one cell from the color
of any given pixel. The only workaround I know of is to increase the screen
size as you are already doing.

There might be some sort of topological walk algorithm that would produce an
exact result but I am not aware of one and suspect that it will be orders of
magnitude slower than the rendering based algorithm we have now. A frustum
selection will produce an exact result and do it fairly quickly, but it does
not and can not determine which cells are visible and which are not.

David E DeMarle
Kitware, Inc.
R&D Engineer
28 Corporate Drive
Clifton Park, NY 12065-8662
Phone: 518-371-3971 x109


On Mon, Oct 18, 2010 at 9:10 AM, <s61500 at htw-dresden.de> wrote:

> I think pictures will be the best illustration of the problem.
>
> http://www2.htw-dresden.de/~s61500/example_sphere.png<http://www2.htw-dresden.de/%7Es61500/example_sphere.png>
>
> 1. window --> viewport
> 2. window --> result with RenderWindow size of [300, 300]
> 3. window --> result with RenderWindow size of [1000, 1000]
> 4. window --> result with RenderWindow size of [1000, 1000]
>
> http://www2.htw-dresden.de/~s61500/example_testobject.png<http://www2.htw-dresden.de/%7Es61500/example_testobject.png>
>
> 1. window --> viewport
> 1. row --> result with RenderWindow size of [300, 300]
> 2. row --> result with RenderWindow size of [1000, 1000]
>
> http://www2.htw-dresden.de/~s61500/example_testobject2.png<http://www2.htw-dresden.de/%7Es61500/example_testobject2.png>
>
> - detailed view of the problem
>
> I hope that the screenshots clarify the problem in the best way.
>
> Stefan
>
> > Can you clarify what you mean by the result isn't right?
> >
> > David E DeMarle
> > Kitware, Inc.
> > R&D Engineer
> > 28 Corporate Drive
> > Clifton Park, NY 12065-8662
> > Phone: 518-371-3971 x109
> >
> >
> > On Sun, Oct 17, 2010 at 1:12 PM, <s61500 at htw-dresden.de> wrote:
> >
> >> Hi,
> >>
> >> upon finding the error in the example code I noted that the result
> isn't right.
> >>
> >> The example code can you find here:
> >>
> >>
> http://www.vtk.org/Wiki/VTK/Examples/Cxx/Broken/Filtering/ExtractVisibleCells
> >>
> >> I find out that you get a better result when you set the size of the
> vtkRenderWindow higher.
> >> But with a size of [1000,1000] I don't get a perfect result by my test
> sphere object with 27.000 triangles. To set the window size high enough
> couldn't be the right way, because the program will work with nets of 1
> million triangles.
> >>
> >> So my question, is there another way to get a better result without set
> the window size higher?
> >>
> >> Stefan
> >>
> >> _______________________________________________
> >> Powered by www.kitware.com
> >>
> >> Visit other Kitware open-source projects at
> >> http://www.kitware.com/opensource/opensource.html
> >>
> >> Please keep messages on-topic and check the VTK FAQ at:
> >> http://www.vtk.org/Wiki/VTK_FAQ
> >>
> >> Follow this link to subscribe/unsubscribe:
> >> http://www.vtk.org/mailman/listinfo/vtkusers
> >>
> >
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.vtk.org/pipermail/vtkusers/attachments/20101018/ead68ebc/attachment.htm>