[ODE] continuous forces w/ constant step size

Tyler Streeter tylerstreeter at yahoo.com
Fri Sep 17 12:48:32 MST 2004


--- John Miles <jmiles at pop.net> wrote:

> > I don't see how that would do anything different
> from
> > the pseudocode I posted.  In your pseudocode, the
> > input and continuous forces are still being
> applied
> > only once per visual frame, not once per time
> step.
> > Input will then be applied more often on a fast
> > machine than on a slow machine.
> 
> I didn't see your corrected post, but you still have
> a problem, in that
> you're throwing away the fractional part of dT that
> remains when the final
> "while dT >= STEP_SIZE" test fails.  This *will*
> cause ODE's notion of time
> to diverge from your app's.  Fix that first and see
> if it magically feels
> better.  I'll bet it will.

Actually I wasn't throwing away that fraction; I am
saving it until the next frame, then adding the new dt
to it like you suggest.  Sorry I forgot to mention
that earlier.

> It's OK to apply input forces less often on a slow
> machine.  They won't be
> as "valid," but your player doesn't have much choice
> since s/he is reacting
> to a slow and jerky portrayal of the game state. 
> That's the problem with
> owning a slow machine. :)
> 

That's a good point.  Essentially, you still need a
decent machine to get good results.  ...but I think
there are some things I can do to make the situation a
little better (namely, storing force events and
applying them every time step until they expire).

Tyler


		
_______________________________
Do you Yahoo!?
Declare Yourself - Register online to vote today!
http://vote.yahoo.com


More information about the ODE mailing list