Bombs are too slow!

[max power 21]

Vilas I doubt any of us here have the equipment to take an accurate measurement of a real bullet.

There are all kinds of ballistic charts online for popular bullets. All we need is a measurement of a 556, then we can look it up on ar15.org.

edit:

This thread has ballistic measurements with a test mission. It seems to confirm that the bullets are "1/3 too slow".

"To spoil it a bit I can tell you that it takes about 0.8 seconds for an m16 bullet to travel 450 meters." The testing methodology for this statement was based on a mission available on page 5 or so.

Well, I was looking at the initial velocity numbers and comparing them with the muzzle velocities of their real life counter parts, and with the exception of the sd ammo which seems to initialize a little fast, they all seem very close.

Apparently, I've done a bunch of work on this problem that I don't remember.

A good summary of the findings are available in this Devheaven ticket.

The devheaven ticket has been closed. The ballistics problem seems to have been solved in OA or something.

Edited December 16, 2010 by Max Power

[madrussian 347]

Haven't read through the entire thread, but for those who think that ArmA2 gravity is off, have you considered:

Gravity vs. Terminal Velocity

My guess is bombs hit their terminal velocity well before the equivalent "free-fall in a vacuum" comparison. Anyway, ArmA2 gravity feels right to me. A lot of other games speed up the physics, hence another reason BIS games are superior. Other than being ancient and not processing collisions very well, Open Dynamics Engine (ODE) will model gravity just fine as long as you don't go tweaking it (i.e. speeding it up).

[max power 21]

Well, regardless of our own personal theories, if we are able to provide some data like the ballistics thread and make a ticket, we can probably get this bomb simulation looked at.

[Opticalsnare 12]

Haven't read through the entire thread, but for those who think that ArmA2 gravity is off, have you considered:

Gravity vs. Terminal Velocity

My guess is bombs hit their terminal velocity well before the equivalent "free-fall in a vacuum" comparison. Anyway, ArmA2 gravity feels right to me. A lot of other games speed up the physics, hence another reason BIS games are superior. Other than being ancient and not processing collisions very well, Open Dynamics Engine (ODE) will model gravity just fine as long as you don't go tweaking it (i.e. speeding it up).

The gravity in arma2 isnt really correct imo, look at how vehicles can sometimes float about on small hills and have a sorta weightless feel to them. Either their mass isnt right or the gravity is just weird. Imo its almost like the gravity ingame is 1/2 of what it should be.

[noubernou 77]

The gravity in arma2 isnt really correct imo, look at how vehicles can sometimes float about on small hills and have a sorta weightless feel to them. Either their mass isnt right or the gravity is just weird. Imo its almost like the gravity ingame is 1/2 of what it should be.

For vehicles yea.

Gravity is different I am assuming in all the different simulation classes. The only one where it is right is the shell simulation, where it is standard gravity (9.80665m/s).

[madrussian 347]

Well, OK if you guys have really looked into this and proven that gravity is actually inconsistent across different game assets, then yeah that should be addressed. Of course terminal velocity should factor as applicable. Although bombs are heavy, they do sport a pretty good surface area that must be taken into account.

Regarding floating over hills while driving, and things feeling "weightless", etc, yeah imho BIS needs to implement a new physics engine badly. I would personally love to see PhysX (with optional HW acceleration). But keep it all true to real life, without monkeying with the new physics engine's standard overall gravity value, or standard time step, etc. In other words, never speed anything up beyond real life values. You can tell when other games do it, and it really cheapens the experience.

Anyhow regarding the status quo, keep in mind that things could always be a lot worse! ;)

[blakeace 11]

I'm using a Eulers method to try and predict the impact point, factoring in a constant air friction coefficient.

Similar to how I was calculating the artillery shells with my excel calculator.

When I use 9.8 for gravity and 0 air friction the bombs land way long. The closest approximation I have achieved so far are the earlier values I posted.

Though with values that impact correctly at lower altitudes result in long shots at higher altitudes. Looking at the results of dropping bombs from a stationary platform (thanks Maturin). Which from above looks similar to how those base jumpers with the wing suits, maybe there is a small lift calculation that is translated into a forward thrust? To get impacts at higher altitudes correct currently involves reducing the air friction coefficient, which I imagine would be similar to a higher coefficient, that has an opposing force that is increasing its forward velocity.

The above is just guessing mind you.

Below is a video of the script doing the calculations, though very much wip. You can see on the first run at a higher altitude the shot is a little long, though when performing a low level run it calculates quite accurately. Note bomb release was calculated by the script so that removes human timing on hitting the fire button.

Lj7YoKWquc8

Edit:

Ok I wrote and quick and dirty acceleration monitor for the bombs.

Then added it to a hind, climbed to a height and released a bomb while in auto hover mode.

Initial acceleration is straight down at @ 16m/s^2.

Vert accel then drops to basically zero, while a forward accel increases to @ 5m/s^2.

Once the forward accel stabilises at @5 the vert component then drops again to around 9-10 m/s^2.

These then slowly drop as I assume air resistance play on then.

I don't pretend to be a physics whiss, so here is the code I used to test for others to confirm.

This would imply I believe that the grav is actually around the 9.8 mark, and it is this other component that changes things in my calculations as I am not taking it into consideration.

It does seem to be some sort of lift that is converting into a forward component by the looks of it.

this addEventHandler ["fired", "Bomb1 = _this select 6"];

SR1 = [] spawn
{
waituntil {!isNull Bomb1};
_dt = 0.01;
_oldvel = [0,0,0];
while {true} do
{
_vel = velocity Bomb1;
_accx = ((_vel select 0) - (_oldvel select 0))/_dt;
_accy = ((_vel select 1) - (_oldvel select 1))/_dt;
_accz = ((_vel select 2) - (_oldvel select 2))/_dt;
hintsilent format ["%1 %2 %3" , floor _accx,floor _accy,floor _accz];

_oldvel set [0, _vel select 0];
_oldvel set [1, _vel select 1];
_oldvel set [2, _vel select 2];
sleep _dt;
};


};

Edited December 19, 2010 by blakeace

[maturin 12]

"To spoil it a bit I can tell you that it takes about 0.8 seconds for an m16 bullet to travel 450 meters." The testing methodology for this statement was based on a mission available on page 5 or so.

According to the Winchester Ballistics Calculator available online, travel times to 500 yards (60 degrees F, 1.5 inch sight height, 300 yard zeroing) are the following:

http://ballisticscalculator.winchester.com/

7.62x39mm takes 1 second

7.62x51mm take .7 seconds

5.56x45mm takes .6 seconds

I looked through that thread and found detailed test results.

In-game, 7.62x39mm takes roughly 1.2 seconds to travel 500m.

7.62x51mm take about .7 seconds

5.56x45mm takes .8 seconds

So certain rounds are too slow. Some are not. The behavior of pistol, submachinegun and subsonic ammo is extremely wonky. There isn't much evidence of a sweeping problem with drag and gravity, just faulty configs. Or who knows, maybe his framerate wasn't high enough and the speeds or measurements were interfered with. Performance is a factor in rate of fire and many other things.

If someone wants to test the drag and gravity theories, they should compare bullet drop to real life, not travel time, as this should be independent of the bullet's size and speed, assuming Newtonian physics are at work.

Edited December 19, 2010 by maturin

[noubernou 77]

Blake, try using the CBA perFrameHandler for the most accurate results, it runs every single frame guaranteed. You can calculate the delta based off of diag_tickTime as well.

In fact I would suggest that for everyone that is doing any sort of diagnostic experiments.

[vilas 477]

Myke;1813940']If this helps' date=' i've measured (roughly) the droptime from a hovering Mi-24, dropping FAB-25:

Height: 100.394m

Time: 8.209s

Height: 201.315m

Time: 11.377s

Height: 301.501m

Time: 13.426s

Height: 401.541m

Time: 14.963s

Height: 401.587m

Time: 15.122s

Probably not high science but i guess it did the job pretty well.[/quote']

thanx for your work !

looks like:

(if we use "free falling") g=2h/t^2

it is "moonish" 1/3 of earth, 3 m/s

---------- Post added at 11:56 AM ---------- Previous post was at 11:51 AM ----------

7.62x39mm takes 1 second

7.62x51mm take .7 seconds

5.56x45mm takes .6 seconds

I looked through that thread and found detailed test results.

In-game, 7.62x39mm takes roughly 1.2 seconds to travel 500m.

7.62x51mm take about .7 seconds

5.56x45mm takes .8 seconds

there is mistake in BIS config, in real 7.62 bullet has smaller velocity, in game both have the same 900,

[max power 21]

Don't forget that the test results in that thread are antiquated. The devheaven ticket suggests that they have changed since then.

Updated by 50.cal 3 months ago

Seems solved!

Results at 450m:

MK16 -> 0.66s

M16 Acog -> 0.67s

Edited December 19, 2010 by Max Power

[galzohar 31]

thanx for your work !

looks like:

(if we use "free falling") g=2h/t^2

it is "moonish" 1/3 of earth, 3 m/s

Yes, about 3.4 if you actually plot a fit, however if you plot a fit you also see that the "free fall" function doesn't really fit, so air friction is probably not being negligible here (in game terms). So basically it needs to be fitted into a formula that takes air friction into account to get the gravity and air friction that the game is using, and of course to get good values you might need a few more data points.

[blakeace 11]

Blake, try using the CBA perFrameHandler for the most accurate results, it runs every single frame guaranteed. You can calculate the delta based off of diag_tickTime as well.

In fact I would suggest that for everyone that is doing any sort of diagnostic experiments.

Cheers will have a look at that.

[noubernou 77]

This ticket might be of interest to people who did some research in this thread: http://dev-heaven.net/issues/18809

[CyclonicTuna 87]

This is what a real Jdam (GBU-53) looks like. As you can see the bomb is clearly visible for a moment, but it comes in very fast. I did a little math and I'm guessing that if 1000 pound bomb is dropped from a seling of 30.000 feet it comes in at approxamemitly 300 miles an hour. Which is quite fast. But this is just a rought estimate and the real thing is based on release speed, wind, temperature etc. That doesn't take away the fact that its a lot faster then bombs are currently dropping in Arma.

Sooo...yeah. I think it needs to be fixed too ;) Edited April 29, 2011 by CyclonicTuna
video doesn't work

[maturin 12]

Please do not quote images / videos!

This is what a real Jdam (GBU-53) looks like. As you can see the bomb is clearly visible for a moment, but it comes in very fast. I did a little math and I'm guessing that if 1000 pound bomb is dropped from a seling of 30.000 feet it comes in at approxamemitly 300 miles an hour. Which is quite fast. But this is just a rought estimate and the real thing is based on release speed, wind, temperature etc. That doesn't take away the fact that its a lot faster then bombs are currently dropping in Arma.

Sooo...yeah. I think it needs to be fixed too ;)

You're overlooking the obvious.

Bombs in Arma 2 are never dropped at 30,000 ft. They are dropped at lower altitudes and therefore don't have time to accelerate.

Edited April 30, 2011 by W0lle

[galzohar 31]

Also air friction will greatly limit a bomb's speed, both IRL and in-game. Of course the game may or may not have this simulated properly.

[fincuan 0]

Just a small note on the bullet physics in Arma2:

After recording times of flight vs. distance for a bunch of rounds and calculating their ballistics coefficients from that data(using G1 drag model) I'm able to use a regular ballistics calculator(Strelok, can be found on Android market) on my phone for my Arma2 sniping needs. The result isn't strictly 100% accurate, but in practice enough for landing first shot hits on human sized targets up to 2000 meters and beyond. In my mind this shows that while BIS ballistics for bullets aren't entirely true to real life, they can still be very close with proper config values for the bullets and proper muzzle velocities for weapons(this is really defined for the magazine in Arma2).

Ballistic coefficients for a few western rounds in ACE2:

5.56x45 0.369

7.62x51_ball("normal" 7.62 NATO) 0.456

7.62x51_notracer("precision" 7.62 NATO) 0.534

0.50_ball 0.711

Muzzle velocities(m/s):

M4SPR 823.1

M24 772.3

M110 786.0

ScarH 786.0

M107 887.0

TAC-50 872.4

[b00ce 160]

Seems pretty fast to me. :rolleyes:

yYXy2EiZQOA

Mk.82 dropped at 3200 meters (10,000 feet) AGL.

Edited May 1, 2011 by b00ce

[[frl]myke 14]

@b00ce &CyclonicTuna

when using youtube tags, don't insert the complete url, just the video code yYXy2EiZQOA

[maturin 12]

Aaaaand end thread. At 3000m, it had plenty of time to reach terminal velocity. Although since it is a guided bomb with a non-ballistic flight path, it may bleed off some vertical speed by gliding.