Depends on your firearm. a pistol bullet can travel anywhere from less than 1000 feet per second (subsonic 9mm pistol rounds) to 2750 FPS (5.56 nato rifle rounds) or more. a granularity of 1 calculation every 50 meters is, at roughly mach 1, 7 caluclations a second. Well within server tick rates.
On the issue of calulations times: if you're doing per-poly hit detection things might get a bit hairy and expensive, but considering the size of the hitboxes and the size of the maps (if you're using smallest source sizes and biggest maps this probably won't apply) you don't have much to fear. A difference of 1-2 world inches at 100 yards won't make much of a difference. You could make the weapon perfectly accurate in game, but the calculation error alone would make the weapon just as accurate as it is in real life. Sniper rifles could be higher-precision in calculation to account for increased RL accuracy. You'd want to limit the number of player-snipers, but I figure that with higher difficulty you'd have less people bothering to snipe. Yarin, you're really off on your figuring: 1 MOA (minute of angle, roughly 1 inch at 100 yds) is about as accurate as an average hunting rifle. 0.5 MOA is a worthy goal for dedicated sniper rifles, and 2 MOA is average for most AR15 and AK type automatic rifles. I'd also like you all to consider the following: why bother with tracelines and their associated calculation at all. Why can't you have a vphys bullet that operates like a snark? Damage doesn't happen until it touches something. You can even make it nonsolid and have penetration. just make the bullet put a traceline out behind it and when it flies through a wall, the traceline detects the opposite surface when it passes through and puts the decals, effects, and decrements it's 'penetration' counter? You might want to consider re-appropriating the SMG grenade's code: when it contacts a model, it does area damage and removes itself, right? remove the floor/wall contact code or replace it with penetration, and you basically have a very low impulse, very large bullet. If anyone wants to hear more about ballistics, which I can tell you quite a bit more about, you can contact me at [EMAIL PROTECTED] if enough people are interested in the nitty-gritty I'll send a more refined ballistics information e-mai to the list. On Tue, 4 Jan 2005 07:26:24 +0100, Yarin Kaul <[EMAIL PROTECTED]> wrote: > Hallo Jeffrey, > > Du schriebst: > > r00t 3:16 wrote: > >> > >> But suppose you wanted it so players who shoot weapons had to > >> compensate for distance, aiming a little ahead of a running player > >> etc. > >> > >> Is the source engine able to do this? > > > [...] > > The real issue is that things that move VERY fast through the world > > don't get their position updated often enough to follow the path of a > > true parabola. You get something like this (excuse the ASCII graphics)... > > > Game Ticks (StartFrame) at the 'v's > > > v v v v > > > x----------x > > / \ > > / \ > > / \ > > / \ > > / \ > > / \ > > / \ > > x x > > > Shouldn't that look more like this? > > \ / \ / > x---------------x---------- \ / > / \ / \ \--------x--- > / \ ------ > ------ \ / > ---x- > / \ > > I think the large steps which are emerging if the calculation is done > in real-time aren't such a serious problem in a game (it would be > serious if you are writing a software for simulation purposes, > indeed). > > If you fire a bullet over a distance of 100 meters, it may lose 1 > meter of height till it hits the target (Rubbish? I'm not an expert on > ballistics). The player wouldn't even notice that there are only a few > gradations. > > Got what I'm trying to say? :-) > > -- > Grüße, > Yarin <[EMAIL PROTECTED]> > > -----BEGIN GEEK CODE BLOCK----- > Version: 3.12 > GCS d-(++) s:+>: a--->-- c++(+++)>$ UL++>++++ P--(---) L+++(++)>++++$ > E--- W+++$ N+>++ o? K? w++(---) !O M-- !V Y(+) PGP--- t+++ 5-- X R- > tv++ b(+) DI@ D+ G++ e? h! r-- !y+ > ------END GEEK CODE BLOCK------ > > > _______________________________________________ > 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
