On 29 Feb 2000 08:24:35 -0500, Nieminen Juha wrote:
>  Wouldn't displacement mapping require LOTS of triangles? And I mean _LOTS_.
>  Just think that you want to apply a 640x480 displacement map to each side
>of a cube. If I have understood correctly, you will need at least 2 triangles
>per map pixel (like in a heightfield), which would make a total of at least
>640x480x2x8 = 4915200 triangles.
>  Since this kind of triangle mesh can not (AFAIK) be optimized like a
>height field can be, it would take rather lot of memory.
>
>  Perhaps I'm just exaggerating.

There are ways to make it less memory-intensive.  One is to displacement-map 
adaptively.  This is particularly useful if you have information on the 
footprint of the ray, but if you're not careful it can lead to cracking.  
Another is to displacement-map on demand and cache the results, but 
raytracing can be markedly less coherent than other rendering methods, so a 
cache might not be as good a solution in a raytracer.  There have been a 
number of papers written on the subject.

-- 
These are my opinions.  I do NOT speak for the POV-Team.
The superpatch: http://www2.fwi.com/~parkerr/superpatch/
My other stuff: http://www2.fwi.com/~parkerr/traces.html

Post a reply to this message