Project OgrePlanet - Slowly coming back to life
- Lee04
- Minaton
- Posts: 945
- Joined: Mon Jul 05, 2004 4:06 pm
- Location: Sweden
- x 1
Whats bug is that Hexidave that you couldn't reproduce?
Whats bug is that Hexidave that you couldn't reproduce?
Something on some other computer system that you don't have access to or some bug on your machine if so what?
Something on some other computer system that you don't have access to or some bug on your machine if so what?
Ph.D. student in game development
-
- Kobold
- Posts: 31
- Joined: Sat Jul 07, 2007 2:30 pm
Ogre 1.5 CVS
Did anyone made it work with Ogre 1.5 CVS , I manage to compile it but there`s no planet .It works fine with 1.4.
- HexiDave
- OGRE Expert User
- Posts: 1538
- Joined: Sat Jan 14, 2006 8:00 pm
- x 1
I'm currently rebuilding the entire system from the ground up - I have contract work on the side (plus 'real life' stuff). I'm at the bed-rock level and it'll remain that way until I'm satisfied it's not going to be a massive failure like some of my other plans. Basically, I'm just laying out test after test of systems and sub-systems - I've got a lot of perfectly functioning things (quad-trees that can attach to other quad-trees in any configuration with proper neighborhoods, large world coordinate systems, etc), but I need to make it all function together.
This is essentially going to replace SimplePagedTerrain as I'll have multiple basic world configurations to work with (flat, spherical, toroidal, etc). As I've noted in the SPT thread, I promise to show progress when I get something to show that isn't just numbers in a DOS window.
It's been a quite a long time since I originally started working with terrain and 3D programming - OgrePlanet was basically my way of getting into C++ in a big way, and it was a colossal failure/success. On one hand I created a system that could handle procedural planet generation and on another I completely missed the point of how STL works... So, in the end, OgrePlanet V1 sucked and is dead.
OgrePlanet V2 is going to knock your socks off.
This is essentially going to replace SimplePagedTerrain as I'll have multiple basic world configurations to work with (flat, spherical, toroidal, etc). As I've noted in the SPT thread, I promise to show progress when I get something to show that isn't just numbers in a DOS window.
It's been a quite a long time since I originally started working with terrain and 3D programming - OgrePlanet was basically my way of getting into C++ in a big way, and it was a colossal failure/success. On one hand I created a system that could handle procedural planet generation and on another I completely missed the point of how STL works... So, in the end, OgrePlanet V1 sucked and is dead.
OgrePlanet V2 is going to knock your socks off.
- Alexander
- Gremlin
- Posts: 175
- Joined: Sat Aug 05, 2006 3:55 am
You mentioned texturing earlier, so I thought you'd like to read this:
http://www.gamedev.net/community/forums ... ?jn=263350
http://www.gamedev.net/community/forums ... ?jn=263350
- HexiDave
- OGRE Expert User
- Posts: 1538
- Joined: Sat Jan 14, 2006 8:00 pm
- x 1
That is a good find - I generally peek at his dev journal now and again as I am doing things similarly to the way he's doing them, but I try to solve the problems on my own as I'll understand it better in the long run if I struggle with it. I had come across that problem some time back and nearly got as far as he did with it, so it's good to see there IS a solution to that.
- HexiDave
- OGRE Expert User
- Posts: 1538
- Joined: Sat Jan 14, 2006 8:00 pm
- x 1
Floating point precision runs out fairly fast for a good sized planet, but overall the large-space coordinate system is simple. I'm pretty sure I have a decent explanation of Floating-Origin use over in the SPT thread in Showcase. There are other things, like using 64-bit integers to do all the math until it's time to calculate geometry, then switch over to doubles, then cast them to floats when it's time to pack them into buffers.
My hardest issue so far has been getting the geometry and the management systems to cooperate in an elegant fashion. I've gone through a few dozen versions and I keep thinking up better ways to do it, but some of the earliest were very hacky. I use a quad-tree that handles neighborhoods and the geometry uses GeoMipMapping, so the quad-tree has to be extra smart in the case of a planet, as you've got 6 quad-trees hooked together (the geometry creation phase takes planar coordinates and turns them into 3D coordinates that match the curvature of a sphere). The biggest problem arrives when you're splitting/unsplitting the tree nodes - you don't want to destroy geometry as soon as you split as you need new geometry to replace the old (i.e. one chunk of terrain splits into 4 new ones), so you need to make sure things stay in place without fouling up the quad-tree split/unsplit mechanics and GeoMipMap stitching.
A lot of the simple tricks that make 2D terrains so nice just do not translate to planets, so you have to think of a lot of new tricks and get creative with making the most with what you've got.
My hardest issue so far has been getting the geometry and the management systems to cooperate in an elegant fashion. I've gone through a few dozen versions and I keep thinking up better ways to do it, but some of the earliest were very hacky. I use a quad-tree that handles neighborhoods and the geometry uses GeoMipMapping, so the quad-tree has to be extra smart in the case of a planet, as you've got 6 quad-trees hooked together (the geometry creation phase takes planar coordinates and turns them into 3D coordinates that match the curvature of a sphere). The biggest problem arrives when you're splitting/unsplitting the tree nodes - you don't want to destroy geometry as soon as you split as you need new geometry to replace the old (i.e. one chunk of terrain splits into 4 new ones), so you need to make sure things stay in place without fouling up the quad-tree split/unsplit mechanics and GeoMipMap stitching.
A lot of the simple tricks that make 2D terrains so nice just do not translate to planets, so you have to think of a lot of new tricks and get creative with making the most with what you've got.
- HexiDave
- OGRE Expert User
- Posts: 1538
- Joined: Sat Jan 14, 2006 8:00 pm
- x 1
Just to prove I'm not always slacking, here's some shots showing GeoMipMapping with normals on a full 3D spherical world:
(Note, this is in Debug mode)
I locked the updating so you could see what it looks like zoomed out. The normals are per-vertex there, but I've done it with textures as well - I've got a better way of handling that (although, I'll have per-vertex options as well). You can see that I'm getting closer to the surface and there's still a lot of geometry detail that's very close together.
Texturing and tangent-space normal mapping works as well, but I didn't have any good screenshots of that yet.
You can't see it there, but the other 5 faces of the cube I use are there and are acting properly with the stitching. I'll show that off once I get the quad-tree edge heightmap synthesis fixed (the edge data I store in the heightmap to avoid sampling neighboring heightmaps means that the very edge of the quad-tree screws the edge data up - simple fix, just time-consuming).
(Note, this is in Debug mode)
I locked the updating so you could see what it looks like zoomed out. The normals are per-vertex there, but I've done it with textures as well - I've got a better way of handling that (although, I'll have per-vertex options as well). You can see that I'm getting closer to the surface and there's still a lot of geometry detail that's very close together.
Texturing and tangent-space normal mapping works as well, but I didn't have any good screenshots of that yet.
You can't see it there, but the other 5 faces of the cube I use are there and are acting properly with the stitching. I'll show that off once I get the quad-tree edge heightmap synthesis fixed (the edge data I store in the heightmap to avoid sampling neighboring heightmaps means that the very edge of the quad-tree screws the edge data up - simple fix, just time-consuming).
- trilobite
- Silver Sponsor
- Posts: 135
- Joined: Thu Jul 13, 2006 9:16 pm
- Location: San Juan Capistrano, CA, USA
- x 1
- tdev
- Silver Sponsor
- Posts: 244
- Joined: Thu Apr 12, 2007 9:21 pm
- Location: Germany
- x 14
do you intend to release the code under LGPL?
can we help you with this plugin anyhow (coding, testing, ...)?
can we help you with this plugin anyhow (coding, testing, ...)?
working on: http://rigsofrods.com | OgreProcedual (project) | Caelum (project) | Paged Geometry (project) | OISB (project) | OgreAngelscript | OgreMeshTool
- Sovaka
- Greenskin
- Posts: 109
- Joined: Sat Dec 20, 2008 6:26 am
- Location: QLD, Australia
- x 3
- Contact:
-
- Gnome
- Posts: 334
- Joined: Thu Jun 28, 2007 2:12 pm
- Location: Brazil
- x 5
- Contact:
Re: Project OgrePlanet - Slowly coming back to life
where i can download this?
links are broken
links are broken
- Lord Alexion
- Kobold
- Posts: 29
- Joined: Wed Oct 19, 2005 2:46 pm
- Location: São Paulo, Brazil
- Contact:
Re: Project OgrePlanet - Slowly coming back to life
Just bumping to check if the project is still alive...
sorry if I'm being inconvenient ^^o
sorry if I'm being inconvenient ^^o
- jacmoe
- OGRE Retired Moderator
- Posts: 20570
- Joined: Thu Jan 22, 2004 10:13 am
- Location: Denmark
- x 179
- Contact:
Re: Project OgrePlanet - Slowly coming back to life
It's been well over a year since HexiDave posted anything new, so I guess the project slowly went back to hiatus.
In plain English: not alive.
In plain English: not alive.
/* Less noise. More signal. */
Ogitor Scenebuilder - powered by Ogre, presented by Qt, fueled by Passion.
OgreAddons - the Ogre code suppository.
Ogitor Scenebuilder - powered by Ogre, presented by Qt, fueled by Passion.
OgreAddons - the Ogre code suppository.
- Sovaka
- Greenskin
- Posts: 109
- Joined: Sat Dec 20, 2008 6:26 am
- Location: QLD, Australia
- x 3
- Contact:
Re: Project OgrePlanet - Slowly coming back to life
I'm not so sure
I have done some searching and it appears HexiDave has uploaded some recent videos to YouTube
He is still working on it to be sure and if he is still discussing it, he's just not discussing it here.
How goes the hard slug HexiDave?
I have done some searching and it appears HexiDave has uploaded some recent videos to YouTube
He is still working on it to be sure and if he is still discussing it, he's just not discussing it here.
How goes the hard slug HexiDave?