Ok. I think i got the idea.. ".Since a tick is no larger than 30msec, a bullet will only fall sv_gravity * 0.03 (18 inches for HL2) in a single tick while it will move forward somewhere near 1000 inches"
Sorry if i am saying bullshit since i haven't done this kind of thing on HL2... YET! So, if the bullet didnt hit anything in this tick on these 1000 inches, you could save the bullet data and recalculate the trace on the next tick. (maybe starting from where the last tick left off or maybe starting from where the last one started and calculating it for 2000 inches) and so on until it hits something or exceeds an maximum distance travelled. Am i right? On Wed, 2 Feb 2005 17:17:49 -0800, Jay Stelly <[EMAIL PROTECTED]> wrote: > There are three issues here: > 1) Collision representation for the bullet > In HL2, they are point sized. You can also make them out of > boxes or fully general vcollides. > 2) Velocity of the bullet > In HL2 the velocity is infinite - they travel instantaneously > until they hit something (simulated with a single traceline). If you > want to simulate gravity you'll need some notion of finite speed for the > bullet. > 3) How to approximate the parabolic motion of the bullet in flight > In HL2 there is no parabolic motion, so it's not a problem. But > you really have two options here: Use vphysics, or use multiple > tracelines/boxes with subdivision. What I propsed earlier was just > using one traceline per tick which means that the bullet's curve is > approximately a line each tick. Since a tick is no larger than 30msec, > a bullet will only fall sv_gravity * 0.03 (18 inches for HL2) in a > single tick while it will move forward somewhere near 1000 inches > (obviously depending on velocity) in the same tick. So there's not tons > of error there, but if you need more precision just subdivide dt and > iterate. If you use vphysics it will do it's own approximation for the > motion which you cannot configure. > > Jay > > > -----Original Message----- > > From: [EMAIL PROTECTED] > > [mailto:[EMAIL PROTECTED] On Behalf Of > > Hasan Aljudy > > Sent: Wednesday, February 02, 2005 5:03 PM > > To: hlcoders@list.valvesoftware.com > > Subject: Re: [hlcoders] Weapon ballastics > > > > but bullets in HL2 and HL1 already are tracelines .. like > > everyone have been saying, they are instantaneous, and if you > > think that's bad, then this is why people want to make it > > more bullet-like. > > > > r00t: > > "Husan, I was thinking about doing this actually. > > Not sure of the performace hit doing a traceline everyframe > > on a bullet but something to test out" > > > > well, bullets are already tracelines, and I think the engine > > does tons of traces every frame (I would guess every moving > > object uses traceboxes for collision detection) etc. > > > > On Wed, 02 Feb 2005 18:27:14 -0600, Adam Mason > > <[EMAIL PROTECTED]> wrote: > > > > Iwas thinking .. why don't you just pre-calculate the > > spot where the > > > > bullet will land if it were to fall acording to gravity .. > > > > do a traceline and see how far your target is, then > > figure out how > > > > much it would fall ... etc, you get the idea. > > > > > > BUT... if the bullet is traveling slow and far enough you'd have to > > > lead a moving target, then your traceline will not actually > > be to the target. > > > > > > Adam "Tobias" Mason > > > Back East Design > > > > > > _______________________________________________ > > > To unsubscribe, edit your list preferences, or view the > > list archives, please visit: > > > http://list.valvesoftware.com/mailman/listinfo/hlcoders > > > > > > > > > > _______________________________________________ > > To unsubscribe, edit your list preferences, or view the list > > archives, please visit: > > http://list.valvesoftware.com/mailman/listinfo/hlcoders > > > > > > _______________________________________________ > To unsubscribe, edit your list preferences, or view the list archives, please > visit: > http://list.valvesoftware.com/mailman/listinfo/hlcoders > > _______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders
