Sign in to follow this  

Newton Game Dynamics - 3 platforms supported!

This topic is 4594 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

Yes! I'd recommend Newton to anybody interested in a physics engine, it was a deam to use and integrate, with a simple API, powerful options, new Custom User Joints for flexibility, and WAY easier to use than ODE (and Novodex, but honestly I didn't look into Novodex deep enough. Newton Was "it"...)

Go Newton!

Share this post


Link to post
Share on other sites
I liked Newton a lot from the 3 completely free engines...but they've got to speed up the solver somehow. It still can't compete with the rest in terms of actually getting things done quickly.

Share this post


Link to post
Share on other sites
Quote:
Original post by Promit
I liked Newton a lot from the 3 completely free engines...but they've got to speed up the solver somehow. It still can't compete with the rest in terms of actually getting things done quickly.


In fact, there are 3 solver modes now in latest patch, so you can choose speed and accuracy you want. But I figured that it's not solver thing - it's collision detection part that slowing things the most. Newton has very reliable CD but it's quite slow comparing to other engines. But there are many tricks that allows to speed it up.

A a small proof, I encourage you to download following videos and see Newton in action:
http://www.tmreality.com/destruction.avi
http://www.tmreality.com/seaside.avi

It's complete game made with Newton and the only speed issue was rendering performance.

Tomasz Zielinski
TMReality

Share this post


Link to post
Share on other sites
Guest Anonymous Poster
I agree, it looks like there is more that just having a fast solver to physics.
Here is a small game with more rigid bodies density I found on the ogre forum
[url]http://www.ogre3d.org/phpBB2/viewtopic.php?t=9724&postdays=0&postorder=asc&start=0[/url]

Share this post


Link to post
Share on other sites
There are more games that use Newton.

But if it's still too slow then custom solver or collision library could be used.
That mix should be interesting as Newton is powerful library and swapping only one r two components while using the rest should make explosive mix

Share this post


Link to post
Share on other sites
Guest Anonymous Poster
http://newtondynamics.com/forum/viewtopic.php?t=1408&start=15

there was a bug with newton's mesh optimizer. maybe that caused the low collision performance?

Share this post


Link to post
Share on other sites
Here's something that may be of interest to all of you; over at the OGRE forums, a member of the community over there, monster, has been working on a wrapper that nicely packages several different physics engines together.

The engines can't react with each other, of course, but the cool thing is, you can call any library you want and check out how each engine handles the given situation, without having to figure out different syntax and whatnot for each individual engine to find the same results. Once you write the code to do it once, the only thing needed beyond that is to select the engine you want to use. Here's the thread.

According to monster's estimates, it should be available sometime around the end of the month, which isn't long. This project also isn't specific to OGRE, monster says in the thread that he's simply using OGRE as the graphical output, but the whole thing can be dropped into pretty much any project [smile].

Share this post


Link to post
Share on other sites
http://www.adrianboeing.com/pal/index.html

But no code as far as I can see.

This chap emailed me a while back with a executable demo showing various scenes running with different physics engines (my jiglib, novodex, newton, tokamak and ode) - was very interesting.

Share this post


Link to post
Share on other sites
What was the result. Your listing subject that jiglib and novodex are the more powerful engines, but I did not see jiblib listed in that page.

How can I get the API, I am looking for freeware physics engine that is robust accurate, and stable.
It needs to be able to simulate rigid bodies without auto-sleeping, and also has the ability to create user defined joints, because it is for a technical project in AI not for a game.

Ah I forgot it has to model coulomb friction fairly accurate and give me the feedback, because the experiment depends of that.
This is for training a neural net for automatic robot guidance system, so the measurement must be reliable.

Convex hull collision or even better mesh to mesh collision would be desirable.

I have been using Newton with very good success but if your project is faster and more or equally stable, I still have time for more evaluations.

Share this post


Link to post
Share on other sites
Quote:
Original post by jovani
What was the result. Your listing subject that jiglib and novodex are the more powerful engines, but I did not see jiblib listed in that page.


No, that was definitely not the implication - the list was not in any order.

Quote:

I have been using Newton with very good success but if your project is faster and more or equally stable, I still have time for more evaluations.


From what I can tell it's about the same speed as Newton (which means far slower than Novodex), and about as stable, but probably less accurate. Definitely less finished, so stay with Newton!

Share this post


Link to post
Share on other sites
I downloaded your test demo, to check it out, but it did not seem very stable.
When I touch the wall or the genga tower with the car the entire structure start to swaying out of control until it collapses into the ground, but if move out of the way it became stable again. So it seems that is been kept stable by artificial means.
Also if I hit the f key, then all the structures start swaying until they all collapse in about 15 second. And the boxes keep jittering on the floor. Perhaps is not as stable as the others but is it very fast.
I think you are correct is not ready for prime time yet. Good start thought.

Share this post


Link to post
Share on other sites

This topic is 4594 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this