[Paraview] issue with Contour on ImageData & UnstructuredGrid...

[Karl König]

Hi,

While trying to check out Ken's instructions I accidentally choose a
wrong filter and may have come across a new bug in ParaView: load the
provided unstructured grid input.vtu and apply filter "Normal Glyphs".
Both ParaView 3.6.1 and 3.7.0 cvs build segfault (with 5, 50 and 5000
glyph max points). I couldn't find a bug report in Mantis about that.
File one?

Karl


> There is something wonky with whatever service you are using to post
> your files.  The first several times I tried accessing them, I just got
> some web pages with a bunch of advertisements in French (which I cannot
> read).  After several attempts, I finally got the files.
> 
> When I run the contour filter with the default isovalue for each input,
> I get roughly the same result.  The major difference is that normals are
> automatically generated for image inputs but not for unstructured grids.
>  Thus, the output for the unstructured grid resorts to flat shading,
> which makes it look like it has sharper edges even though the polygons
> are the same.  If you run the normals filter on the ugrid output
> (turning off splitting), you will see the result looks similar.  You can
> also try turning /off/ the Compute Normals feature for the /image/
> input.  With the normals gone, you will see that the result is the same.
> 
> By default, the contour filter does not compute or pass scalar values
> for image data.  If you click on the Compute Scalars checkbox, you
> should see them again.  The contour filter always computes scalar values
> for unstructured grids.  I’m not sure why the two are different.
> 
> -Ken
> 
> 
> On 9/8/09 1:05 PM, "fred" <fredmfp at gmail.com> wrote:
> 
>     fred a écrit :
>     > I forgot...
>     >
>     > Files are here:
>     >
>     > http://dl.free.fr/rtNWe2Uhg
>     >
>     > http://dl.free.fr/rUhCCAbyu
>     Hi,
> 
>     Nobody can help me on this issue?
> 
>     :-((
> 
> 
> 
>     Cheers,
> 
>     --
>     Fred