Manta Interactive Ray Tracer Development Mailing List

Text archives Help


Re: [MANTA] Message from Dylan


Chronological Thread 
  • From: James Bigler <bigler@sci.utah.edu>
  • Cc: manta@sci.utah.edu
  • Subject: Re: [MANTA] Message from Dylan
  • Date: Thu, 15 Jun 2006 09:57:06 -0600

Hi folks,

As an experiment, I modified the objviewer scene to use a GriddedGroup
instead of a BVH.  I got it to build, but I see huge stairstep artifacts
along one side of the test object (a cube), while the other side looks
good.  Any thoughts?

http://www.cs.utah.edu/~lacewell/broken_cube.png

The artifacts look like ray packet problems. To be honest I haven't used/testing the GriddedGroup much, so I'm sure there are bound to be problems.

I'll take a look and see if there is anything obvious.

James

The grid prints out the correct bounding box; just to check, here it is
with the "-bounds" option enabled:

http://www.cs.utah.edu/~lacewell/bounds_cube.png

Input obj file:
http://www.cs.utah.edu/~lacewell/cube.obj

Session:

% bin/dm_demo -scene "lib/libscene_gridobjviewer.so( -file
/home/wally/models/cube.obj -cellfactor 1.0 -lambertian)"

Using 1 workers plus one gui thread.
Min: -0.5 -0.5 -0.5  Max: 0.5 0.5 0.5  translate: 0 0 0
Total triangles added: 12
Grid: Bounding box -0.5 -0.5 -0.5 : 0.5 0.5 0.5
1. Initializing grid
2. Count objects
3. Count cells
4. Allocating 108 list entries
5. Filling in lists
6. verifying grid
GriddedGroup: 12 objects, targeted cells: 12, actual cells: 27, 3x3x3,
built in 0.000258 seconds
Warning resize is broken.

Thanks,
Dylan






Archive powered by MHonArc 2.6.16.

Top of page