Hi,
Not directly what you want but I have had problems using
System.currentTimeMillis() in other applications. As you say the
resolution is very poor (about 10 millisecs). I changed to using the
J3DTimer. Atleast on NT I have found this to give a much finer and more
accurate picture of the amount of elapsed time. This may help.
Cheers,
Brad
Georg Rehfeld wrote:
>
> Hello group members,
>
> has anybody of you successfully extended ViewPlatformAWTBehavior
> to handle KeyEvents and is willing to share hints or code?
>
> I'm trying to implement a simple Quake style behavior, where the
> mouse controls the view direction and some keys (say CURSOR UP/DOWN
> and LEFT/RIGHT) move the ViewPlatform forward/reverse and left/right
> (called 'strafing' in Quake) with respect to the current view
> direction. Needless to say, that the horizon shall remain
> horizontally, ever :-).
>
> This works, in principal.
>
> The problem is: since I started to respect deltaTime (via
> System.currentTimeMillis() as in KeyNavigatorBehavior) my
> Behavior became really worse:
>
> most notably, instead of accelerating smoothly to the desired
> maximum speed ever, it shows 3 different real behaviors randomly:
>
> - sometimes it only moves the VP in much too small steps, not
> accelerating at all, while the accelerating key is held down.
> When it is released, the VP suddenly has the maximum speed
> (decelerating then, as wanted, immediately).
>
> - most times it almost accelerates just fine to the desired
> speed, but in all the motion, when the framerate drops (i.e.
> due to new objects coming into view) the motion TOO is affected.
> Which should not be the case at all, as I'm respecting
> deltaTime from the last frame and that should compensate by
> moving the VP farther, when deltaTime is high. Besides, my
> deltaTimes are almost ever 10, 20, 30 ... milli seconds, due
> to the bad PC time resolution, possibly I should integrate
> that, but then, how does the KeyNavigatorBehavior class
> deal with that, I couldn't find any special code?
>
> - sometimes the final speed seems to be about double of the
> desired speed, then the movement seems to be smooth.
>
> My Pseudo-Code is as follows:
>
> protected synchronized void
> processAWTEvents(final AWTEvent[] events)
> {
> motion = false;
> // set motion = true, if some acc/decelerating
> // or movement is in progress
> if ... motion = true;
> // handle all AWT events
> ...
> processKeyEvent(event)
> ...
> }
>
> protected void processKeyEvent(final KeyEvent evt) {
> // record key states into instance variable(s)
> // and set motion = true on any recognized key event
> ...
> }
>
> protected synchronized void integrateTransforms() {
> // Check if the transform has been changed by another behavior
> targetTG.getTransform(currentTransform) ;
> if (! targetTransform.equals(currentTransform)) {
> resetView();
> }
>
> calcSpeed();
>
> // modify the target transform, this simply works
> ...
> }
>
> private void calcSpeed() {
> long now = System.currentTimeMillis();
> long deltaTime = now - lastTime;
> lastTime = now;
>
> // check for too old time sample
> if (deltaTime > 2000) deltaTime = 1;
>
> // protect against 0 deltaTime, this happens sometimes
> if (deltaTime < 1) deltaTime = 1;
>
> float dt = (float)deltaTime / 1000f;
>
> // exploit key states and then
> // integrate acceleration/drag over dt
> ...
> // integrate velocity over dt
> ...
> // set instance members for integrateTransforms()
> ...
> }
>
> Any hints are VERY welcome
>
> thanks
>
> Georg
> ___ ___
> | + | |__ Georg Rehfeld Woltmanstr. 12 20097 Hamburg
> |_|_\ |___ [EMAIL PROTECTED] +49 (40) 23 53 27 10
>
> ===========================================================================
> To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
> of the message "signoff JAVA3D-INTEREST". For general help, send email to
> [EMAIL PROTECTED] and include in the body of the message "help".
----------------------------------------------------------------------------
This Email may contain confidential and/or privileged information and is
intended solely for the addressee(s) named. If you have received this
information in error, or are advised that you have been posted this Email by
accident, please notify the sender by return Email, do not redistribute it,
delete the Email and keep no copies.
----------------------------------------------------------------------------
===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff JAVA3D-INTEREST". For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".
