RE: [Plib-devel] Nested LOD selectors almost works...

[Vallevand, M. K <Mar...@UN...>]

That's it!

Yes, all of my tiles share the same origin.

A simple tree looks like this:
root
 |-LOD Selector
    |-branch
    |  |-LOD Selector
    |  |  |-branch
    |  |  |  |-1x1 tile at (0,0)
    |  |  |  |-1x1 tile at (0,1)
    |  |  |  |-1x1 tile at (1,0)
    |  |  |  |-1x1 tile at (1,1)
    |  |  |
    |  |  |-2x2 tile at (0,0)
    |  |
    |  |-LOD Selector
    |  |  |-branch
    |  |  |  |-1x1 tile at (0,2)
    |  |  |  |-1x1 tile at (0,3)
    |  |  |  |-1x1 tile at (1,2)
    |  |  |  |-1x1 tile at (1,3)
    |  |  |
    |  |  |-2x2 tile at (0,2)
    |  |
    |  |-LOD Selector
    |  |  |-branch
    |  |  |  |-1x1 tile at (2,0)
    |  |  |  |-1x1 tile at (2,1)
    |  |  |  |-1x1 tile at (3,0)
    |  |  |  |-1x1 tile at (3,1)
    |  |  |
    |  |  |-2x2 tile at (2,0)
    |  |
    |  |-LOD Selector
    |     |-branch
    |     |  |-1x1 tile at (2,2)
    |     |  |-1x1 tile at (2,3)
    |     |  |-1x1 tile at (3,2)
    |     |  |-1x1 tile at (3,3)
    |     |
    |     |-2x2 tile at (2,2)
    |
    |-4x4 tile at (0,0)

However, the bounding sphere for each tile is in the
correct location.

The solution is what?  Insert a transform branch above the
LOD Selector?  A transform branch in place of the regular
branch?

Regards.
Mark K Vallevand   ma...@rs...
Outside of a dog, a book is man's best friend.
 Inside of a dog, its too dark to read.  - Groucho


> -----Original Message-----
> From: Steve Baker [mailto:sjb...@ai...]
> Sent: Friday, April 28, 2000 1:37 PM
> To: pli...@li...
> Subject: Re: [Plib-devel] Nested LOD selectors almost works...
> 
> 
> "Vallevand, Mark K" wrote:
> > 
> > Nested LOD selectors almost works.
> > 
> > In my program, I recursively partition a tile into sub-tiles at
> > at higher level of detail.  If a tile can be partitioned, I create
> > a LOD selector with 2 kids - one for the lower level of detail
> > tile and a second one that is a branch containing the higher
> > level of detail sub-tiles.  Every thing works together really
> > nice.  The tree seems to be correct, all the leaf nodes at all
> > levels of details are correct, and their bounding spheres
> > have the correct center and radius.  Lots of brute force
> > debugging and print statements makes me believe things
> > are correct.
> > 
> > But, the same debugging and print statements show me
> > that the range calculated in the cull of each LOD selector
> > is the same for each selector.  As I move the camera
> > around the terrain, the range changes and I get different
> > culling, but it is funny because the range calculated seems
> > to be wrong.
> 
> Hmmm - the LOD range is measured to the local origin of the
> model...so could it be that your terrain tiles all share
> a common origin?
> 
> > So, my questions...
> > How is the range calculated when culling to the viewing fustrum?
> 
> I don't calculate the range.  I transform the bounding sphere
> center using the current modelview matrix and compare it to the
> plane equations of the six clipping planes that make up the
> view frustum.
> 
> -- 
> Steve Baker                  http://web2.airmail.net/sjbaker1
> sjb...@ai... (home)  http://www.woodsoup.org/~sbaker
> sj...@ht...      (work)
> 
> _______________________________________________
> plib-devel mailing list
> pli...@li...
> http://lists.sourceforge.net/mailman/listinfo/plib-devel
>