> If I understand what you said. Voxel data can be or is a subset of point
> cloud data.

Yes, voxel data could just be thought of as point cloud data where all 
the points are on a finite uniform grid (with no empty cells) within 
known limits. In which case it becomes more efficient to store the data 
as a list of values (colour, transparency, whatever) in some agreed 
order, rather than listing the coordinates of every point.

Point cloud data is totally arbitrary, you could have points 1mm apart 
in one area, but points meters apart in others. This is usual if you got 
the data from a laser scanner. The coordinates of each point could be 
anything (within the resolution of the number format you are using), so 
are not on a grid or equally spaced at all.

Think in 2D, voxel data is like a bitmap image, whereas point cloud data 
is like a list of 2D coordinates (perhaps with associated information 
like colour as clipka mentioned).

> So what is the problem with redefining coordinates in a df3 to points or
> small spheres instead of a cubical volume?

My guess is all the algorithms used to render df3's won't work with just 
an arbitrary list of point coordinates.

Post a reply to this message