 | | | Fine displacement tips very welcome | Fine displacement tips very welcome 2004-06-30 - By Morten Bartholdy
Back
The docs state that view dependant fine displacement tesselates the geometry
relative to camera raster space, and I thought this meant that regardless which
type of geometry it would tesselate away, creating subpixel triangles in sets,
resulting in very fine detail, with little memory penalty.
I noticed that I could not get my NURBS grid to subdivide fine enough this way,
but Polys do fine, and I guess it has to do with the Max subdivision limit.
A simple test with a default poly and a default NURBS grid with identical
fractals applied for displacement, and identical Geometry Approximation
displacement settings show quite different behavior. Polys get very detailed
and NURBS do not, until i up the surface parametric subdiv and the displacement
max subdiv limit.
In short, it seems I can't render my NURBS grid detailed enough because it
crashes when I up the Max subdivision limit enough (with optimized BSP) so I am
about to abandon my NURBS for an ocean surface and switch to polys.
Does anybody here with experience in the behavior of fine view dependant
displacement, care to share some insight on the dos and dont's when using polys
for this and making it render fine?
I need to move from quite far away and up through the surface, so view
dependant it is, and my NURBS just started to look faceted when I got close.
Final render will be 2K so memory and renderspeed are issues.
TIA!
--
Morten Bartholdy | morten@(protected)
Denmark
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=iso-8859 (See http://iso-8859.ora-code.com)-1">
<META content="MSHTML 6.00.2800.1400" name=GENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=#ffffff>
<DIV><FONT face=Arial size=2>The docs state that view dependant fine
displacement tesselates the geometry relative to camera raster space, and I
thought this meant that regardless which type of geometry it would tesselate
away, creating subpixel triangles in sets, resulting in very fine detail, with
little memory penalty.</FONT></DIV>
<DIV><FONT face=Arial size=2></FONT> </DIV>
<DIV><FONT face=Arial size=2>I noticed that I could not get my NURBS grid to
subdivide fine enough this way, but Polys do fine, and I guess it has to do
with
the Max subdivision limit.</FONT></DIV>
<DIV><FONT face=Arial size=2></FONT> </DIV>
<DIV><FONT face=Arial size=2>A simple test with a default poly and a default
NURBS grid with identical fractals applied for displacement, and identical
Geometry Approximation displacement settings show quite different behavior.
Polys get very detailed and NURBS do not, until i up the surface
parametric
subdiv and the displacement max subdiv limit.</FONT></DIV>
<DIV><FONT face=Arial size=2></FONT> </DIV>
<DIV><FONT face=Arial size=2>In short, it seems I can't render my NURBS grid
detailed enough because it crashes when I up the Max subdivision limit enough
(with optimized BSP) so I am about to abandon my NURBS for an ocean surface and
switch to polys.</FONT></DIV>
<DIV><FONT face=Arial size=2></FONT> </DIV>
<DIV><FONT face=Arial size=2>Does anybody here with experience in the behavior
of fine view dependant displacement, care to share some insight on the dos and
dont's when using polys for this and making it render fine?</FONT></DIV>
<DIV><FONT face=Arial size=2></FONT> </DIV>
<DIV><FONT face=Arial size=2>I need to move from quite far away and up through
the surface, so view dependant it is, and my NURBS just started to look faceted
when I got close. Final render will be 2K so memory and renderspeed are
issues.</FONT></DIV>
<DIV><FONT face=Arial size=2></FONT> </DIV>
<DIV><FONT face=Arial size=2>TIA!</FONT></DIV>
<DIV><FONT face=Arial size=2></FONT> </DIV>
<DIV><FONT face=Arial size=2></FONT> </DIV>
<DIV><FONT face=Arial size=2>--</FONT></DIV>
<DIV> </DIV>
<DIV><FONT face=Arial size=2>Morten Bartholdy | <A
href="mailto:morten@(protected)">morten@(protected)</A><BR>Denmark<
/FONT></DIV></BODY></HTML>
|
|
|