Novodex Physics Engine

Discuss about anything related to the Irrlicht Engine, or read announcements about any significant features or usage changes.
Masdus
Posts: 186
Joined: Tue Aug 26, 2003 1:13 pm
Location: Australia

Novodex Physics Engine

Post by Masdus »

Has Anyone looked at integrating the Novodex physics engine into irrlicht? It is free to use for non-commercial use. Check it out at http://www.novodex.com
Robomaniac
Posts: 602
Joined: Sat Aug 23, 2003 2:03 am
Location: Pottstown, PA
Contact:

Post by Robomaniac »

Novadex is one of the best physics engine out on the market today, stable, easy, etc.

Someone should look into implementing it
The Robomaniac
Project Head / Lead Programmer
Centaur Force
Masdus
Posts: 186
Joined: Tue Aug 26, 2003 1:13 pm
Location: Australia

Post by Masdus »

i'm adding it to my list of projects, but i doubt i'll have time to work on it anytime soon
coder2000
Posts: 40
Joined: Sat Oct 02, 2004 12:14 am
Contact:

Post by coder2000 »

Novodex looks easier than ode or newton but doesn't look like it has any ragdoll features.
Guest

Post by Guest »

Are you sure novodex is easier and more stable than Newton?
It is faster but not more stable. Everything in novodex is like a sponge just take a body and apply pressure down and See how everything squash like and sponge. Also the friction is not correct.
Roberto
coder2000
Posts: 40
Joined: Sat Oct 02, 2004 12:14 am
Contact:

Post by coder2000 »

I never said it was more stable. I just said it was easier. As well it depends on the type of friction you are using as well because there is no friction, isotropic friction and ansiotropic friction, each acting differently.
Guest

Post by Guest »

coder2000 wrote:As well it depends on the type of friction you are using as well because there is no friction, isotropic friction and ansiotropic friction, each acting differently.
What are isotropic and ansiotrpic friction? I know of two kinds of friction, static and kinetic.
Also, does anyone know what model Novodex uses for friction? Newton uses the Coulomb friction model, which takes friction into account cumulatively (harder to pull a box from the bottom of a stack than from near the top).

If Novodex is really being used by Unreal Engine, as I keep hearing, then I'd think it must be pretty darn good.
coder2000
Posts: 40
Joined: Sat Oct 02, 2004 12:14 am
Contact:

Post by coder2000 »

thats what novodex called them. I have no idea why.
brcolow
Posts: 56
Joined: Mon Jul 19, 2004 6:15 am
Location: Arizona

Post by brcolow »

It would be a good idea in theory..but the licsenses are totally different. Besides, what are the advantages of this VS ODE?
G'Day.
coder2000
Posts: 40
Joined: Sat Oct 02, 2004 12:14 am
Contact:

Post by coder2000 »

ease. although it doesn't look like it incorporates ragdoll it doesn't look like it would be that hard to implement. Coding with Novodex is as easy as coding with Irrlicht once you know what to do.
Guest

Post by Guest »

The problem with your statement is that you call easy as an absolute statement.
What you say is not different that:

Ogree is easier than Irrlicht
Direct3d is easier than OpenGl
WindowsXP is easier than max os X
A Camaro is better than a Mustang
And goes on and on.

For each one of those statements I am sure you will find 50-50 split. So Saying Novodex is easier than any other physics engine is your opinion. As some want once Said beauty is in the eye of the beholder, I add the same apply for any subjective property.

About Friction they can call it what ever they want the fact is that the friction is wrong in novodex, Also big problem with high mass ratios, if you place a big heavy object on top of a light one the light one compress like it was made out of rubber.

I am using the physics to simulate college level problems in mechanics. And the only physics engine that has given result that come close to the prediction is Newton. We were using ODE for a while and switched to Newton, we also tested Tokamak and Novodex. All these engines go berserk when you try to apply realistic values. Newton is the only one that holds on. They may be very fast and that makes then go for games, but not for good physics simulation.

It will be interesting to see what you come up with that mean there will more options for the community
Roberto
coder2000
Posts: 40
Joined: Sat Oct 02, 2004 12:14 am
Contact:

Post by coder2000 »

I was never saying that Novodex was better than Newton for friction. If you wish to turn this into a flame could you do it somewhere else please as this is about how to integrate novodex into irrlicht now how it compares to other physics engines.
Guest

Post by Guest »

It is not flame, in fact I said it will be very cool to have more options. But for the record
You were the first who brought the comparisons when you say this “Novodex looks easier than ode or newton but doesn't look like it has any ragdoll features.”
I just pointed out that, that was a very subjective opinion. But if it bothers you I will go about my business and not post in this thread anymore.
Masdus
Posts: 186
Joined: Tue Aug 26, 2003 1:13 pm
Location: Australia

Post by Masdus »

bit unfair to rate a games physics engine on its capability to perform college physics. These system were designed to create a resonable approximation of physics in a fast game environment. As such they only implement certain subset of features and rely more on fast stable method then extremely accurate one. Further it is often necessary to extend they system to cover situations that were not explicited coded into the library.

The advantage of Novodex as i see it is that it is a commerical quality engine, has a resonable sized community and has a dedicated team improving it.
Guest

Post by Guest »

accuracy = stability with physics engines. fast (linear time solvers) + stable isn't easily possible i think. novodex looks very cool but the question is how much parameter tweaking was necessary to make their demos stable enough. that's something you only can find out by working with it...
Post Reply