Hi Bob,
I don't think your email made it to the list - rejected by our
majordom as too long because of the attachments.
The problem with trasnparency is that it only works right when
the polygons are sorted in back to front order, but the sort
order depends on how the user rotates the view. The Java3D
team has plans to add dynamic sorting, but its a very hard
problem and won't appear in Java3D for a while.
This same sorting problem is the reason the volume rendering
needs six rather than three different sets of planes. But the
sort is more complex for iso-surfaces because they are irregular.
You might be able to do something with:
display.getGraphicsModeControl.setTransparencyMode(
javax.media.j3d.TransparencyAttributes.SCREEN_DOOR);
It didn't work in an early Java3D implementation, but I
haven't tried it recently. Screen door transparency
avoids the sorting problems (we use it in Vis5D). Let me
know what success you have - I'll help if I can.
Cheers,
Bill
> We are trying to visualize an isosurface with transparency. However, the
> transparency does not appear to be 100% consistent in that the level of
> transparency does not appear to be the same for all parts of the surface.
> It looks as if the surface has been partitioned down the middle
> (so-to-speak) and 2 or more halves (or quadrants?) have different levels
> of transparency, and that the transparency changes as the volume is
> rotated.
>
> We have extended the Gridded3DSet with a class of our own that provides a
> modified make_isosurface() method. It looks as if an earlier mailing
> relates to this issue:
>
> > Actually, VisAD volume rendering is more complex than the
> > Delaunay.main() logic. Volume rendering from Linear3DSets
> > creates six different sets of transparent faces and switches
> > between them as the volume rotates. There is no good way to
> > do this with Irregular3DSets or even Gridded3DSets.
>
> We've attached 2 snap shots of what we are seeing. Is there a simple way
> to remove the visual effects of this partitioning when viewing our
> transparent isosurface?
----------------------
Bill Hibbard, SSEC, 1225 W. Dayton St., Madison, WI 53706
hibbard@xxxxxxxxxxxxxxxxx 608-263-4427 fax: 608-263-6738
http://www.ssec.wisc.edu/~billh/vis.html
