New Racing Physics Engine

I added the zip file of the last compiled executables on the download link of the site.
If someone download it please give me any feedback.


Racing Physics Engine

I downloaded the file but the car doesn''t move. All keyboard/mouse inputs work except UP/DOWN/LEFT/RIGHT.

Did you press on p to toggle pause off ?

too slow...
it took me 1 minute to move the wheels abit...
and the FPS is only 30, not full screen.

Ok there must be a bug, on my PC it is running at 300 fps.
Is it fluid when you rotate the camera with the mouse ?
What is your configuration please ?
3d Graphics? Processor (intel/AMD)?
Thanks

Also you should try to turn off by default in OpenGL mode the Vertical Synchro. I dont do it in the code...

I get 20fps and it takes ages to turn the wheel and move. Weird. It looks interesting and promising though.

The left wheels roll forwards, while the right ones roll back.

For the wheel rolling behavior, it is normal i have not programmed yet the wheel rolling dependence with the physics.

But for the 20 fps it is not normal.
I am investigating on this strange behavior

Can you toggle the wired mode (default) off with "L".
As the wired rendering mode is antialiased it can considerably slow down the performances.

Thanks

Works fine on my system. 500+ fps with texturing on, else 400+ with it off. The sliding with turns seems slightly off sometimes -almost like there''s not enough friction -them again maybe it is just a visual effect from the wheels spinning the way they do-it just makes things look odd.

Yes i have to tune the car physics parameters...
I will do a version with configurable physics parameters soon.

Very nice first shot. Can you give some brief explanation on your implementation. and If you will manage to solve oscilation and friction problems, tell us how you did this.

All these people having problems with timing, you should really make it time independant so these guys running 30fps don't have this problem. It appears you set the timing up to work @ 300fps that your machine is running, rather than setting it to run at a constant speed irregardless of the frame rate.


--- Edit ---
By the way, I get around 200fps on my box in regular mode, and about 8fps in wire frame! It seems pretty good, a bit of work and I think it'll be ready to be placed on a real track

[edited by - Ready4Dis on November 28, 2003 10:25:15 AM]

DirectXXX:
I hope i will solve friction and oscillation problems by only changing the forces magnitudes to good values as i don''t see what kind of force i could add to the car physics

But i am not sure yet i will not add more forces to solve the problem if i am not able solve it with the first solution.

Ready4Dis:
I am not sure problem is not that i am not managing well the time as if you check at the fTimeElapsed value it should give you the time elapsed in seconds since the launch of the application.
It "fTimeElapsed" is not giving a good approximation of the time there is a bug. But i hope there is no.

The REAL problem in fact is that i have not optimized yet the physics engine. The problem can be explain this way, my main loop
perform the DISPLAY and X time a physics engine UPDATE( fDeltaTime / X) where X depends on the time elapsed since the last frame rendered. the optimal is when X = 1, but as soon as the rendering is slowering down the X increase. And as the UPDATE function is not yet optimized -> it is very bad.

So it is now my first priority to optimize the physics engine

I put a new version with better physics parameters.

Also i fixed something in the time management for the physics, and it seems that even with low frame rates the physics are reacting well.

Thanks for your comments

Racing Physics Engine

Looking good!

It would be nice if the car would fall back onto the table once
it falls off..

Hum... Good idea.
I think i will do it when pressing on a keyboard key.

Done. Now pressing ''R'' reset the car to its original position.
Also fixed a bug.
Download the new zip file.

can you give me some implementation details

did you used a physics lib.
What type of integration.
Methods for collisions and response.
and some other helpful info