The implementation went faster and with with less code than expected
( 2 days / 120 lines for the C++ terrain code ).
Abonnieren
Kommentare zum Post (Atom)
skip to main |
skip to sidebar
Sparse Voxel Octree Raycasting based Game Development.
Links
- About me
- Raycast.org - Games & Graphics News
- Download CUDA Voxel Engine Jan.2013
- Download Skeletal Animation Tutorial
- Download Terrain Engine Source Code
- Publications
- HVox Terrain Engine
- christmas radio
- Source Islands
- DanielStrayer's SVO Raytracer
- MakeMake Voxel Engine
- DEFTWARE's DEVBLOG
- Transvoxel based planet renderer
- NVIDIA's Efficient-SVO Raycaster
- John Whigham's SVO Raytracer
- Jeroen Baert's SVO Raytracer
- Cotter's SVO Raytracer
- Attila T. Voxel Research
- Raytracey's blog
- Graphics Runner VolVis
- Genova Terrain Library
- Branislav Siles Atomontage Engine
- Cyril Crassin Site
- Inigo Quilez
- Ompf Forum
- Humus3D
Follower
Über mich
Gesamtzahl der Seitenaufrufe
---
It looks interesting but I don't quite understand what I'm seeing. In the top image I see lines but no triangles... so I guess you are raycasting rather than rasterizing? But does the concept of clipmaps really apply in this case? And have you tried going truly 3D, rather using just a 2D heightmap (i.e. how does it work with caves and overhangs?)?
AntwortenLöschenInteresting stuff anyway.
The terrain is conventional triangle strips with VBO, so no raycasting. The plan is to combine voxel raycasting with triangle based terrain rendering in 2 passes. Since the voxel raycasting is already quite demanding, there is few time left for the terrain. The idea is therefore to first support heightmaps and perlin noise voxel based terrain in the future. I expect that the pure geometry clipmap terrain can run at 250-500fps on current hardware.
AntwortenLöschen