Browse Prior Art Database

Perspective View Truncation for Spatial Subdivision Modellers

IP.com Disclosure Number: IPCOM000121478D
Original Publication Date: 1991-Sep-01
Included in the Prior Art Database: 2005-Apr-03
Document File: 2 page(s) / 71K

Publishing Venue

IBM

Related People

Woodman, S: AUTHOR

Abstract

Disclosed is an efficient method for a Graphics Spatial Subdivision Modeller (such as the Winchester Solid Modeller (WINSOM90)) to subdivide a perspective view by a recursive binary chopping algorithm.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Perspective View Truncation for Spatial Subdivision Modellers

      Disclosed is an efficient method for a Graphics Spatial
Subdivision Modeller (such as the Winchester Solid Modeller
(WINSOM90)) to subdivide a perspective view by a recursive binary
chopping algorithm.

      The efficient truncation of a perspective pyramid is more
complex than a non-perspective pyramid in that instead of having to
subdivide a cubic universe it is required to subdivide one shaped
like a square- based, flat-topped pyramid.  The procedure is exactly
the same except for one slight problem; the universe can be very much
deeper than it is wide.  This is necessary to get a far-off horizon
feel in the rendered image but causes a problem because trying to
subdivide this type of shape will result in thin but very deep
voxels.  An established technique for speeding up classification of
primitives against these voxels is to surround the voxel with a
sphere and simply check whether the primitive intersects the sphere.
The speed of classification in this way more than makes up for the
loss of accuracy in the non-perspective case because the volumes
taken up by a cubic voxel and its surrounding sphere will be similar.
But this is not so with a narrow deep voxel.

      Proposed in this disclosure is that this deep universe is
truncated into several smaller flat-topped pyramids, each one of
which is rendered, in turn, from front to back.  The quadtree data is
kept between each invocation of the rendering code so that objects
which should be hidden do not get drawn in front of objects in a
closer pyramid.  The following algorithm is the way the universe may
be truncated:
      Truncate:
        compute ratio of voxel volume to surrounding sphere volume
        (voxel/sphere) if ratio is greater than ratio from previous
        leve...