gdalgorithms-list

Re: [Algorithms] Curves...

[Ralf S. <rap...@ra...>]

|> Bezier or Nurb:s?? or maybe something else...
| If you want smooth looking terrains, both are good.
| If not, then you will have to do some bump or displacement
| mapping to get nice pits, valleys, and such.

a simple but effective trick i used in an experimental isometric-terrain voxel engine was this:

the terain itself was generated by a spectral-synthese heightfield generation programm written by Klaus Hartmann.
the data(512x512) was stored as a height-map and during runtime converted to 2d-splines.
these splines were rendered to the screen with a resulting size of 65536x65536(the entire "new" "base"-heightfield).

as Dave remarked everything was round and smooth...

to add artificall structure i "added" the initial heightmap with the size of 512x512 to the resulting heightfield(65536x65536) using tiling.
i think i have to remark: 
one nice feature of heightfields generated by spectrale-synthese is: they "tile" perfectly!!

well, there was structure! but unfortunately now the "__repeating__" heightfield "on top" of the base heightfield was visible.

so, as a last step i was computing the roughness at a given position at the base-heightfield and then scaling the detail-heightfield linear to the corresponding roughness.

here are some results:

http://www.rappongi.com/h1.jpg
http://www.rappongi.com/h2.jpg
http://www.rappongi.com/h3.jpg

last remarks:

the roughness was computed by computing the deltas-height of surounding heights.
... the roughness values between four samples on the base-heightfiled were computed using bi-linear interpolation.

finalHeight = heightFromSplines(x, y) + heightFromHeightfield(x & 511, y & 511) * roughness(x, y)


maybe this small trick will be usefull for someone else.

- ralf

RE: [Algorithms] Curves...

[Lipson, P. <Pet...@br...>]

I have a little energy for a debate...
  For my application, the key feature of patches is that it's a tremendous
compression technique. On PSX-2, at least, handling data is very expensive,
while computations aren't. There are a few cunning tricks you can use to
overcome a lot of the problems with patches.  Other problems can be avoided
by selecting content or techniques that don't stress the implementation.

Peter Lipson
Mattel Interactive

> -----Original Message-----
> From: Akbar A. [mailto:sye...@ea...]
> Sent: Friday, September 01, 2000 2:59 PM
> To: gda...@li...
> Subject: RE: [Algorithms] Curves...
> 
> 
> >Bezier patches are easy to render.
> opinion:
> imho patches are horrible, after talking with a whole lot of 
> people (i used
> to like patches) writing some code...
> 
> there are just to many problems with patches, it's not worth 
> the effort.
> they get split back to triangles, artists can do more with a 
> higher mesh of
> triangles.
> there are "so" many problems with patches that people don't see when
> starting out :|
> someone should really right an article on the downsides of 
> patch rendering.
> 
> sigh, this will probably start a debate why patches are good, 
> if only i had
> the energy to discuss more about this. i think i might have 
> already written
> a few points about this in a previous e-mail.
> 
>