The OgreOde Thread
-
- Gnoblar
- Posts: 3
- Joined: Mon Jan 05, 2004 1:16 am
- Location: WA, USA
-
- OGRE Expert User
- Posts: 286
- Joined: Fri Nov 01, 2002 3:54 pm
- Location: Berlin & Nuremberg, Germany
- x 1
-
- OGRE Retired Team Member
- Posts: 19269
- Joined: Sun Oct 06, 2002 11:19 pm
- Location: Guernsey, Channel Islands
- x 66
-
- OGRE Community Helper
- Posts: 1098
- Joined: Mon Sep 22, 2003 2:40 am
- Location: Melbourne, Australia
It's a compound geom; six boxes (body, 2 wings, tail "boom", tail "flipper" and tail fin) and three spheres for the wheels. I'm yet to be convinced that moving trimeshes are a good idea in a game environment.Is the helicopter's body a trimesh or just a box
At startup (of that demo) I get about 40fps with shadows on 70 with them off. Once I've smashed into the wall and the boxes are more spread out I get more like 65 with shadows on and 140+ with them off. That seems to be the limit of my card, since the other scenes (which are, physically speaking, much simpler) max out at about the same level. These are obviously very rough estimates, and that's on; P4 3.2GHz, 1Gb RAM, GeForceFX5600 256Mb, Ogre's running windowed at 800x600 32-bit no anti-aliasing.What sort of framerates are you getting, especially for that particular shot
Perfectly. It even gives a reasonable version of arcade-style car physics (powerslides, oversteer, understeer, wheelspin) straight out of the box. Without any fancy tyre modelling.how is ode working with your buggy
Seeing as ODE has been developeed over many years by a number of extremely clever people, including Russ, the lead developer, who previously used to develop physics code professionally; if you can come up with anything better I would be very surprised.i hate it and im writing my own physics code
Well sure, it can. As can sqrtf (or glBegin, or fprintf, etc, etc) if you call it enough times per frame.ode is bogging down the computer
Why am I not surprised!3rd eye brainworm posted 4 messages after this, all of which were devoid of anything remotely resembling content and were thus removed.

-
- Ogre Magi
- Posts: 1266
- Joined: Tue Aug 12, 2003 1:53 am
- Location: Melbourne, Australia
- x 1
howdy Monster, you are Ed Jones from the Ode list right? I love your remakes
that 2d Joint really should be in the core ODE. 2d/3d games rule
ps I see you're in Melbourne too. I've decided that Yake with lua scripting is the only way I'll ever actually finish a game, but I also do music if you need some

ps I see you're in Melbourne too. I've decided that Yake with lua scripting is the only way I'll ever actually finish a game, but I also do music if you need some

-
- OGRE Community Helper
- Posts: 1098
- Joined: Mon Sep 22, 2003 2:40 am
- Location: Melbourne, Australia
Shhh! You'll spoil my mysteriousness if you reveal my real name!
If only I could take credit for the 2D joint, but I just nabbed it off someone else! Glad you like my remakes, hopefully there'll be another one on the way soon. And thank's for the music offer.
The lesson for today is; don't try to destroy scene nodes to which particle systems have been attached at the start of a frame, do it at the end. You've no idea how many hours that's just cost me!
And here's today's screenshot;

[edit]picture updated; nicer looking particles[/edit]
If only I could take credit for the 2D joint, but I just nabbed it off someone else! Glad you like my remakes, hopefully there'll be another one on the way soon. And thank's for the music offer.
The lesson for today is; don't try to destroy scene nodes to which particle systems have been attached at the start of a frame, do it at the end. You've no idea how many hours that's just cost me!
And here's today's screenshot;

[edit]picture updated; nicer looking particles[/edit]
Last edited by monster on Wed Jun 09, 2004 3:11 pm, edited 1 time in total.
-
- Kobold
- Posts: 31
- Joined: Tue May 11, 2004 8:09 am
- Location: Tunisia
-
- OGRE Retired Moderator
- Posts: 20570
- Joined: Thu Jan 22, 2004 10:13 am
- Location: Denmark
- x 179
Wee-Hee !
I am almost speech-less ...

I am almost speech-less ...

/* 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.
-
- OGRE Community Helper
- Posts: 1098
- Joined: Mon Sep 22, 2003 2:40 am
- Location: Melbourne, Australia
Yeah, but they're not travelling that fast, and the physics is running at a consistent 200Hz. Better would be to attach a little Ray collision geometry object to the front of the missle to check if it's going to collide with anything within the next timestep. But that, as they say, is beyond the scope of this demo!Don't you have problems with the collision of the missiles and the
boxes since I suppose they're travelling at a high speed?
-
- Kobold
- Posts: 28
- Joined: Sat Jan 17, 2004 7:45 pm
Why not 4 more boxes for the prop? Just curious.monster wrote:It's a compound geom; six boxes (body, 2 wings, tail "boom", tail "flipper" and tail fin) and three spheres for the wheels. I'm yet to be convinced that moving trimeshes are a good idea in a game environment.Is the helicopter's body a trimesh or just a box

-
- OGRE Retired Team Member
- Posts: 19269
- Joined: Sun Oct 06, 2002 11:19 pm
- Location: Guernsey, Channel Islands
- x 66
I agree with monster's approach - if you can use a combination of simple collision proxies, for god sakes do, they're a lot less expensive than full tri mesh collision and you'll be hard pressed to notice a difference if you pick the proxies well. Too many people go immediately for the trimesh solution unnecessarily - ok it's the only option in some cases like world collision, but in those cases the local tesselation level is not high so it's justified.
-
- OGRE Community Helper
- Posts: 1098
- Joined: Mon Sep 22, 2003 2:40 am
- Location: Melbourne, Australia
Because it's just a demo! I've gotta leave something as a exercise for the reader!Why not 4 more boxes for the prop? Just curious.

In reality (i.e. a "real" game) you'd probably only use boxes for the rotors when they're spinning really slowly. When they get up to anything like operating speed you'd be better off using a proper (as opposed to capped) thin cylinder or a sphere (and then disregarding collisions that aren't close to the "equator") since the boxes would probably be spinning too fast for sensible collision detection.
Plus, in a real simulation there's about a million other things to consider when modelling a helicopter. Almost all of which I don't understand. This demo just shows that you can fly stuff around by simply applying forces in the right places. It's the crappest helicopter simulation ever!
-
- OGRE Community Helper
- Posts: 1098
- Joined: Mon Sep 22, 2003 2:40 am
- Location: Melbourne, Australia
-
- Halfling
- Posts: 74
- Joined: Tue Dec 17, 2002 11:57 am
- Location: Somerset, England
-
- OGRE Community Helper
- Posts: 1098
- Joined: Mon Sep 22, 2003 2:40 am
- Location: Melbourne, Australia
Actually, ODE's new QuickStep is much faster than WorldStep or StepFast, AND more accurate.
No it's not. It's faster, for large systems, but it's less accurate. If you can come up with a solver that's very fast, very accurate and scalable then I daresay you could make a lot of money.
The ODE docs say;
"For large systems this is a lot faster than dWorldStep, but it is less accurate."
I don't know how it compares to step fast, that's never been in core so I've never used it. All these demos (actually it's just one demo, you can flip between the different scenes at will) use quickStep.
The realistic-ness of your sleeping depends on how you set your auto-sleep parameters. You can set up things to auto-sleep in really bizarre positions if you like! Auto-sleeping is new (ish), previously you could do it, but only manually. The test_crash demo still does manual sleeping I think.Boxes also sleep more realistically than before.
Oh ye of little faith!Heh, I was gonna ask if the copter was fully working or if you just plopped it in there and rigged up a box proxy and hooked up the keys to move it.

It uses a lot of damping to make it controllable, where it should use a feedback-based fly-by-wire type system.
-
- Halfling
- Posts: 64
- Joined: Mon Apr 05, 2004 6:25 am
- Location: North Vancouver, Canada
-
- Halfling
- Posts: 74
- Joined: Tue Dec 17, 2002 11:57 am
- Location: Somerset, England
-
- Kobold
- Posts: 28
- Joined: Sat Jan 17, 2004 7:45 pm
You're right of course. I was just assuming a non-moving rotor for this "simple" demo.monster wrote:In reality (i.e. a "real" game) you'd probably only use boxes for the rotors when they're spinning really slowly. When they get up to anything like operating speed you'd be better off using a proper (as opposed to capped) thin cylinder or a sphere (and then disregarding collisions that aren't close to the "equator") since the boxes would probably be spinning too fast for sensible collision detection.

-
- Kobold
- Posts: 35
- Joined: Mon May 03, 2004 5:03 pm
And your helicopter is just one mesh? How do you know exactly the position and size of the wheels, the flipper, the boom, etc. when you are creating the Geom?monster wrote:It's a compound geom; six boxes (body, 2 wings, tail "boom", tail "flipper" and tail fin) and three spheres for the wheels. I'm yet to be convinced that moving trimeshes are a good idea in a game environment.Is the helicopter's body a trimesh or just a box
Aren't you using any physics for your heli? Or your Geom as a Body to?
-
- Kobold
- Posts: 31
- Joined: Tue May 11, 2004 8:09 am
- Location: Tunisia
I can maybe reply in the place of monster (I'm massively using his Jeep's simulationtlaloc wrote: And your helicopter is just one mesh? How do you know exactly the position and size of the wheels, the flipper, the boom, etc. when you are creating the Geom?
Aren't you using any physics for your heli? Or your Geom as a Body to?

Every part is a single mesh. The size can be retrieved using the bounding box. However, I guess, the position of the wheels, for example, are hard-coded. Or did you find a solution for this monster?
-
- Halfling
- Posts: 74
- Joined: Tue Dec 17, 2002 11:57 am
- Location: Somerset, England
-
- Gnoblar
- Posts: 13
- Joined: Mon May 17, 2004 10:00 am
- Location: Taree, Australia

I can't find cylinders in the Ogre implementation, how would you do a cylinder (like a pound coin)?
Paul.
If in doubt, whip it out!
-
- Kobold
- Posts: 31
- Joined: Tue May 11, 2004 8:09 am
- Location: Tunisia
I think this must be done in the ode wrapper's in order to make things more elegants, and the demo is just there to show the usefulness and the simplicity of writing some code above this wrapper. isn't it?I wouldnt be surprised if they were trial and error hard coded values. This is just a demo after all...