[ODE] simulation steps in a game

[Bruce Williams]

Animation movement is not really needed at near frame rate. If several
animated objects move independently at frame rate, you don't have a "all
jump now" effect and the eye is happy. Back in the prehistoric days of 8bit
processors, that was one method used at both Williams and Atari where I
worked. Classic cell animation used this also, moving different cells each
camera shot.

They had a "processor bandwidth problem" also that limited the "cell rate",
the animators wrist could move the pen only so fast....

Has anyone ran ODE on a timer to provide the minimum "good" simulation? You
get "beat problems", but a slower animation rate actually minimizes this
problem. I have used a steady dynamically calculated "intelligent" movement
and quick per/frame tweened ( simple linear interpolation ) values to avoid
beats.

What I am saying is there are system level solutions that avoid the need for
acceptable brute force ability.

Bruce Williams
"Two is not equal to three, even for large values of two"


> -----Original Message-----
> From: ode-admin@q12.org [mailto:ode-admin@q12.org]On Behalf Of Nate
> Waddoups
> Sent: Thursday, April 11, 2002 5:07 PM
> To: ode@q12.org
> Subject: Re: [ODE] simulation steps in a game
>
>
> On Fri, 12 Apr 2002 henry.maddocks@met.co.nz wrote:
>
> > Here's my render loop... [simplified for brevity]
> >
> > void render( float gameTime) // milliseconds
> > {
> > 		while( physicsTime < gameTime)
> > 		{
> > 			odeWorld->step( ODE_WORLD_STEP);
> > 			physicsTime += ODE_WORLD_STEP;
> > 		}
> >
> > 	scene->render( gameTime);
> > }
>
> But what if the processing for a single odeWorld->step takes
> significantly
> longer than gameTime?  That's the condition in which my app
> slows down,
> which as someone pointed out the other day is going to be a problem in
> a multiuser situation.
>
> Does anyone have any ideas for handling that condition?  The
> first thing
> that comes to mind is "don't let the world get that complex" but in my
> application that's not an option.
>
> --
>
> Nate Waddoups
> Redmond WA USA
> http://www.natew.com
>
>
> _______________________________________________
> ODE mailing list
> ODE@q12.org
> http://q12.org/mailman/listinfo/ode
>