PDA

View Full Version : Recoil - let the experiments begin.


gal-z
10th Aug 2005, 04:35 AM
Recoil in INF seems very unrealistic to me. It may be way better than any other game out there, but needs a lot of work. The only weapons I have good experience with are M-16, car-15, M4A1 with trijicon 4x and a haris bipod and M-24. I have some minimal experience with IMI negev (almost the same as the minimi) and the FN mag.
Right now in game it is too difficult to hip shoot with the minimi imo and too easy to shoot rapidly or full-auto with 5.56 when standing, yet a BIT too hard when moving.
So I started experimentig with what I can, and got results for 2 positions:
Crouch, 4 rounds, almost as rapid as full auto (the officer thought I was using full auto and made me stop shooting). Prone, 5 rounds, full auto.
Note the differnece where in crouch it went mostly to the right and a bit up while in prone it went mostly to the left and very little up. Probably because in crouch your right side of the body can be pushed back while in prone the stock slightly slides to the right so the barrel points more to the left. Those reasons are just a theory though. I'm a right handed shooter and was focused mostly on aiming the first shot and barely focused on handling recoil (since it's the first time I was testing very raid/auto fire when actually aiming the sights). Targets were at 25m.
Main conclusion you can see here is that recoil affects way more to the left/right and can go completely off to the other side in a different position.
Also when zeroing my M4A1 with bipod and trijicon 4x a while ago, every shot made it go way to the left as well, so it's possible that in prone as a rule the recoil will make your weapon turn to the left.
With M-24 I can't really talk much about recoil because I didn't really cared about it when I fired, both because it's a bolt action, and that the recoil usually makes you blink because the scope gets kicked quite close to your eye and you're trained to reload immidiately after firing so you don't really get a chance to see the effect. I hope I'll have a chance to make experiments with M-24, but that probably won't happen. With M-16 and car-15 I should be able to experiment sometime in the future, and maybe even with a reflex sight (ITL MARS, Meprolight and or Elbit falcon reflex sight) to see how they work when firing rapidly (had a chance to play around with them, just not fired with them).
Another unrelated recoil problem I found in game is that the M4A1 "jumps" way less than the M16A2, which makes absolutely no sense.

I hope more people who can will make experiments to help make the reoil system in INF actually real and make people stop spraying full auto with minimal recoil.

Rostam
10th Aug 2005, 04:55 AM
Although I have no experience with firearms, I do think that recoil should be so much harder to control. If weapons would do more damage as well, then maybe covering fire would be more effective (right now covering fire means missing on purpose :p) and maybe there will be an actual use for semi - other than just being cool.
Glad to see you've put some effort in this, it was informative for me atleast.

zeep
10th Aug 2005, 07:02 AM
Interesting read. I have no real experience with those weapons myself but your findings make sense to me.

Psychomorph
10th Aug 2005, 03:28 PM
As I know many people say, that recoils are never really straight up and differs like your position is.

Im not sure too, but when you aim a sight and fire, isn't it the case that the frontsight actually doesn't move up, but it looks like the frontsight moves down, due to the rearsight beeing closer to the eye and having more movement, so when the rearsight is recoiling back and slightly up they is the effect like the frontsight is moving down.
Can you confirm this?

Rostam
10th Aug 2005, 03:37 PM
Psych, you mean dis-aligning the sights due to recoil? That would work really well I reckon.

Derelan
10th Aug 2005, 03:44 PM
Recoil could be done better, definately needs to be way higher on the weapon pack guns regardless of realism :P

Psychomorph's idea sounds pretty cool, its hard to keep a weapon perfectly aligned after firing.

Unfortunately, recoil has to be unidirectional so we can control it in-game, but it would be neat to see weapons realistically moving directions other than up because of recoil.

Psychomorph
10th Aug 2005, 04:14 PM
@Rostam & Derelan:
You're right, but I think I mean something a bit different. I mean the way the weapon recoils. In most games you see the frontsight going more up, but doe to the rearsight is closer to the eye it has a 'larger' movement scale, so the frontsight should disappear downside during the recoil, but should recover after this a bit.
Sure not 100%ly recovery, you need to realign it then.

I cheked INF again and indeed the single shot recoil does work as I explained. It just recovers to straight.

keihaswarrior
10th Aug 2005, 05:36 PM
Recoil definately needs to go more to the right and left. Right now, it is straight up which allows you to easily walk your fire up through the target while spraying on full auto.

We need a proper wound system before semi will be useful.

Derelan
10th Aug 2005, 07:06 PM
Recoil definately needs to go more to the right and left. Right now, it is straight up which allows you to easily walk your fire up through the target while spraying on full auto.

Well its not perfectly straight up, and most of us like to move the mouse down while firing, simulating the "hold your gun really tight"-anti-recoil.

gal-z
10th Aug 2005, 08:11 PM
It's quite perfectly straight up, or at least unnoticeably close. Which is far from the way it is IRL.
As for disaligning the sights, 5.56 recoil barely disalign the sights and it's natural to realign them. It is not natural, however, to aim left after the recoil offseted you to the right - you move it back as a reaction after you see where the recoil offseted you.
As for animation of gun moving backwards, it moves backwards then back to normal real fast you can't notice it IRL as well. The only reason you know it happens is that when you're new to using the weapon or when the recoil is strong (like M-24) it often would make you blink as a natural response for something getting close to your eye in high speed :P

Rostam
11th Aug 2005, 04:21 AM
Gal, I thought he meant exactly that the first time. After firing having the sights dis-aligned and re-aligned in a blink of an eye. Doesn't matter for semi auto, but it should make it a lot harder to put 30 bullets in a face while on full auto like some people do now :p

Psychomorph
11th Aug 2005, 04:53 AM
What I totally agree on is to give single shot more advantages (without really making fullauto burst that uneffective).
I like to use rifles on singleshot, it is done in real life (trained soldiers) and since weapons are lethal, fullauto is not that of an advantage it just suits more of a suppressing situation when you just spray at a undefined target.

You know, even if they is a vest and the hit doesn't kill, the hit still affects the person. He recoils, fall, loose weapon control a bit all the variety of reactions on just a single hit.
I know they is a shock element too when a person is not that affected by the hit, but that should go only for less serious areas of the body.


Even if it sounds nasty, but beeing hit by a round should often create a situation where you are still alive, but weakened and disturbed so the guy just have to finish you. It's like you are alive, but as good as dead (with a slight chance to strike back of course).

When it is a firefight on distance the weakened person have a chance to survive, since the guy can't easily shoot you a second time (if not a totally free area without cover of course).


Sorry for slight off topic.

gal-z
11th Aug 2005, 09:43 AM
Realistic recoil should be the main thing that makes it impossible to do very good groupings with full auto.
As for injuries, it should be a different thread, but I know some1 who got shot and the bullet hit his NV sight he had in his back pouch and it didn't penetrate (didn't even reach the lvl3 vest). If I get a chance to talk to him I'll try to get more information.

Rostam
11th Aug 2005, 09:56 AM
Did it change direction and miss or did it stop all together? In any case, glad he didn't get hit :)

gal-z
11th Aug 2005, 10:07 AM
It was stopped. It was a pretty well-placed shot in the top-center of his back, where the pouch is and the NV sight is held during daytime.

Psychomorph
12th Aug 2005, 08:15 AM
gal-z how exactly an M4A1 behaves during fullauto (or other weapons like the tar, mitar)?
How does it sway? How the sights behave?

Eric220
12th Aug 2005, 08:55 AM
I too suggest the recoil on Auto/Burst weapons should be harder so noobs don't spray you. But I'm VERY amazed with the Robar effect with the sight aiming, that's beautiful.

gal-z
12th Aug 2005, 06:54 PM
M16A1 fullauto is the 2nd picture (and 1st picture is semi but very fast almost like fullauto). I didn't fire M4A1 in fullauto, but unless the different flash hider has a significant effect, it should be quite the same. If the effect IS noticeable, it will rise up less but left/right movement will be the same.

Eric220
12th Aug 2005, 07:07 PM
IMO if you waste a 30 round clip at the wall it SHOULD look something like this:

T

yurch
19th Aug 2005, 03:47 PM
Infiltration uses an instant displacement for it's recoil. It is for this reason I suspect that the recoil was made linear, as random displacements of any significant scale are quite disconcerting, especially at the edges of the freeaim extents.

As an example of nonlinear recoil, try firing 75 or more consecutive (without letting go of the trigger) shots through the SAW. Standing and walking around for extra fun. The left/right motions will make it difficult to even walk straight.

Of course, using just displacement for recoil is quite simple, too simple really. Why use algebra when calculus is more fun?

gal-z
20th Aug 2005, 10:03 AM
Firing 75 rounds without letting go of the trigger while running is reall hard IRL anyway :)

5eleven
20th Aug 2005, 01:43 PM
As I have said before, recoil IRL is subjective. It depends on the shooter's experience, control, stance, and practice practice practice. If a shooter knows how to control recoil, they can compensate appropriately.

Sort of like Infiltration. :D

yurch
20th Aug 2005, 08:01 PM
Firing 75 rounds without letting go of the trigger while running is reall hard IRL anyway :)
I only meant it as an example of the displacement method's effects.

Subjective or not, recoil should be a velocity or accelerational motion rather than an impossible instant relocation. It does spell a length of time where a single recoil action may disrupt aim, so it is an idea to consider for those who want to make things harder.

5eleven
20th Aug 2005, 08:50 PM
I understand that. But how does the learning curve fit in, ingame? You still learn to control recoil. As time is logged playing, you learn how to manipulate the environment, or the weapon, in order to put the rounds where you want, and to win.

I would doubt that implementation of "jumping" ingame immediately spurned forethought of "bunnyhopping".

You understand what I mean?

yurch
20th Aug 2005, 11:50 PM
What? I think we're talking about different things here. It's easier to 'learn' (and, master to a greater precision) the very set angluar distance infiltration uses than a sort of angular velocity representation that can be found in other games.

Recoil representations are used in games to specifically disrupt the player, generally abuses such as bunnyhopping don't arise from such additions.

5eleven
21st Aug 2005, 12:07 PM
Yes, we are. I re-read your post before mine, and I think I understand now what you meant. There should be a uniform formula, so to speak, and not randomness assigned to recoil. I was under the impression that the thread was addressing inaccuracy of recoil by stating that there was a specific, RL implementation dependent upon fire mode selection and some "control equality".

Jesus, I just confused myself.

Derelan
21st Aug 2005, 07:35 PM
Yeah yurch I like the idea of having recoil as a thrust towards a direction, rather than a movement to a new position. Recoil is a pushing force, not a teleporter :P

gal-z
25th Aug 2005, 10:09 AM
An accelerated motion in a very short timespan is almost the same as an "instant relocation", though I don't have any specific objection for making it that way. I'd just rather have an instant relocation representing recoil than not having realistic recoil at all, and I don't think it'll be bad at all. After all, right now the problem with recoil is not the way it's implemented, it's the amount and direction in which the weapon "jumps".

keihaswarrior
26th Aug 2005, 04:44 PM
Here's an idea. The recoil thrust direction is always directly away from the center of your freeaim. The closer you can keep the sight to the "sweet spot" in the center, the easier it will be to achieve tight groupings. This would simulate squaring up properly for the shot, also managing recoil would be in 2d (up,down,left,right) rather than 1 direction.

The best anaylogy I can think of to this method is like balancing a yardstick on your nose. If you can keep it right in the center, it is fairly controllable, but once it starts to move a certain direction it will accelerate rapidly unless corrected.

gal-z
27th Aug 2005, 07:24 AM
The problem is that kind of analogy is simply incorrect when it comes to shooting. :(

keihaswarrior
27th Aug 2005, 03:07 PM
The problem is that kind of analogy is simply incorrect when it comes to shooting. :(
Ah, ok. Can you please elaborate? You've got more experience shooting than I do.

Derelan
27th Aug 2005, 09:35 PM
But there aren't alot of peripherals on a computer to simulate the balancing act of shooting, keihas has a good idea.

ravens_hawk
28th Aug 2005, 03:08 AM
I'm telling you the next INF will need a joystick or a steering wheel at least. I mean really have you ever tried to pilot anything with just a mouse and keyboard, it's horrible. A clear compromise would be to use a gamepad but who would want to do that? Maybe an idea would be to move away from this whole keyboard/mouse thing and get into some freaky dual controller thing... you know mouse on one side freaky joystick with 27 buttons on the other.

Harrm
28th Aug 2005, 12:21 PM
I'm pretty nasty in a plane/helicopter/tank in BF42, and I use my mouse. Also, I control all of the vehicles in GTA:SA with nothing but my keyboard. As long as the force for each button push is reasonable, I'll never feel obligated to dust off my joystick for anything other than Freespace and Microsoft Combat Flight Simulator.

--Harrm

gal-z
30th Aug 2005, 12:07 AM
It's not that if you hold the rifle right your shots will group well. When you shoot, the rifle WILL get an offset, which is relatively fixed if you're not holding it in some wierd way (as in, if you're nolding it right in the shoulder in a resaonable shooting position). There may be a small difference between different people, but the INF soldier is the "average soldier +", isn't he? So it's not that if you're holding it right you get hardly any recoil and once you get some offset to one side you'll keep offsetting more and more in that direction. It's that whenever you shoot you get an offset in a direction depending on your posture (standing/crouch/prone) and if you don't bring the sights back to the target after the shot (before shooting a second shot), the rifle will keep offsetting more and more to that same direction.
Basically IMO the way it's done in INF right now IS good, it just needs a change of how much it recoils and to which direction.

Nukeproof
30th Aug 2005, 04:59 AM
Well so Keihas idea does make sense, doesn't it? (as far as I understood it)

When you do it like that:
I) The ironsight of the gun are perfectly alligned in the center of the screen (aka center of the freeaim box). Recoil is at a minimum at this center.

II) Aiming off center would result in a minor missalignment and result in increased, recoil depending on the direction of the aiming offset

The slight misssalingment of the gun would need to be visible on the screen (gun sligthly tilted sideways), so players intuitively have a clue what to do. The whole procedure (aiming at freeaim center) would be INF's interpretation of alligning ironsights and holding the gun propperly.

This way players in a guarding position would finally have a decent advantage over those walking, running or strafing around.

JFYI I've never shot a real gun, so I have no clue on realism but gameplaywise this sound very good to me.

Harrm
30th Aug 2005, 08:56 AM
That's not really accurate, is it? I mean, when you aim around your peripheral view with a gun in real life, it's not like your entire body stays stationary. You tilt your head instinctively and shift your weight. There would be an offset if you didn't tilt your head, but in real life view, you wouldn't even be able to see the irons anymore (since your eye is much closer to the irons than in INF).

And recoil isn't that hard to manage if you have the gun turned (as opposed to tilted, something I have never done or reccomend doing) to the side. As long as you've got the butt of the rifle propped up against your arm and your hand on the trigger, the recoil is the same pretty much no matter how you shoot it. I've seen guys firing weapons like they where paintball guns and nailing targets with relative ease.

Basically, what's wrong with the system we have now (besides all that "kick up and to the side" junk)? Yeah, some people get really good at it, but that's what happens when people play a game for extended periods of time. If you really want to balance it, just have the gun kick a small amount in a random direction. It's not realistic, but if you want a quick fix...

--Harrm

EDIT:
The problem is that kind of analogy is simply incorrect when it comes to shooting.

That's what I just tried to explain. In short: how you shoot on a screen is different than how you shoot in real life.

Demosthanese
30th Aug 2005, 08:42 PM
Look at it this way, your rifle is always going to start offset, as it is crooked int the meat of one of your shoulders. The only way to get it on center is to place the stock in the center of your chest, and that just won't work. Thus, the force is always to push back on the side it is held, causing the riffle to turn you in that direction. Holding it "off center" to the left while shooting right handed will still result in an impulse to the right side of your torso, causing you to rotate to the right.

Haarm, recoil in a random direction is almost as bad as conefire. I'd say set the recoil to what one would experience when not trying to control the effect. Thus, the player moving the mouse would be the equivilant of the soldier tensing and flexing his muscles. It will become a natural response for the player, as it does for the soldier, and just as a soldier's muscles would tire, a players mouse would run out of desk space, and both would have to quit firing to get back on target.

yurch
31st Aug 2005, 12:54 AM
Well so Keihas idea does make sense, doesn't it? (as far as I understood it)No, he's using freeaim instead of misalignment. In the past I've implemented misalignment-directional recoil like yours (probably in some HL2 experiment) and I don't remember liking it too much. The recoil tends to 'pick up speed' (if, the recoil effects misalignment) at an alarming rate.

keihaswarrior
31st Aug 2005, 02:34 AM
...whenever you shoot you get an offset in a direction depending on your posture (standing/crouch/prone) and if you don't bring the sights back to the target after the shot (before shooting a second shot), the rifle will keep offsetting more and more to that same direction...
The way you put it here, it seems to agree with my idea.

The recoil based offset would occur in a direction away from the center of the freeaim. Then, you would have to correct (or the rifle will keep offsetting more and more in that direction), pulling your mouse towards the center of the freeaim. If you over correct, then the rifle will start recoiling in a different direction resulting in the balancing act I was referring to in the analogy of the yardstick on the nose.

Misalignment would be another thing to add that I wasn't really talking about specifically. As yurch said, if the rifle is offset AND misaligned it would really throw you off, so a little probably goes a long way.

Demosthanese
31st Aug 2005, 03:10 AM
KW, it's basic physics; if an impulse is applied off of the center of mass, the object will rotate in the direction of the impulse. In the case of INF Guy, it is a sharp impulse to the right shoulder, above and right of center mass. The center of mass will move backwards, away from the impulse with the right side moving away faster than the center, with the left side moving toward the impulse. As the center of mass moves back, INF Guy also pivots at the waste and leans back slightly. The result is the recoil will go up and to the right. As the traditional fighting stance has INF Guy leaning slightly forward, the upwardmovement is minimalised.

Nukeproof
31st Aug 2005, 04:45 AM
No, he's using freeaim instead of misalignment. In the past I've implemented misalignment-directional recoil like yours (probably in some HL2 experiment) and I don't remember liking it too much. The recoil tends to 'pick up speed' (if, the recoil effects misalignment) at an alarming rate.

I understand, some things just don't turn out to be what they looked in theory. I fully trust you - you'll do an awesome implementation anyways :)

keihaswarrior
31st Aug 2005, 02:04 PM
KW, it's basic physics; if an impulse is applied off of the center of mass, the object will rotate in the direction of the impulse. In the case of INF Guy, it is a sharp impulse to the right shoulder, above and right of center mass. The center of mass will move backwards, away from the impulse with the right side moving away faster than the center, with the left side moving toward the impulse. As the center of mass moves back, INF Guy also pivots at the waste and leans back slightly. The result is the recoil will go up and to the right. As the traditional fighting stance has INF Guy leaning slightly forward, the upwardmovement is minimalised.
That would be the case if a soldier was a rag doll or mannequin. But the soldier can expect this to happen and therefore compensate for the recoil trying to twist his torso.

I would think that if a soldier saw his shots always going up and right, then he would correct this by bracing against it. Therefore, the recoil wouldn't always go up and to the right.

Demosthanese
31st Aug 2005, 05:49 PM
In INF, that would be simulated by the player moving the mouse.

gal-z
8th Sep 2005, 02:03 AM
Went shooting yesterday with my car-15, and noticed once again that in standing and crouch it will ALWAYS recoil right+up.
The attached images show about where it recoiled to (where the 2nd shot was fired to, sort of), in standing and crouch positions with my car-15. Notice that with crouch it goes a bit less up, probably since the elbow is leaning on the knee. We've seen it in prone with the M16A1 in full auto that it barely goes up since both elbows are on the ground. Also we notice greater recoil with the car-15, and I can guess that the reason for it is the lighter weight of the weapon (about 2/3 of an M16A1).

Psychomorph
8th Sep 2005, 05:07 AM
Nice, go on with your tests :D.

It is known that full sized rifles have less recoil than their carabine counterparts. Due to a shorter and mostly lighter barrel (which creates a total weight loss) thw weapon recoils more. It is the case with all weapons, G36vsG36K, M16A2/3/4vsM4A1, SIG550vsSIG551, AKMvsAKMSU, MP5A2vsMP5K, Mk23vsUSP45. :D

Thats why I would like to have most rifles in both (or more) variations. I would love to use the G36 with more accuraccy on larg maps and a G36K with less weight and size in CQB.

When you run forward (+walk) how the recoil behaves (could be a problem for you to test.. just risk it :p ).

Conglomera
8th Sep 2005, 08:51 AM
You may need to check this http://forums.beyondunreal.com/showthread.php?p=1892691

keihaswarrior
8th Sep 2005, 04:01 PM
Went shooting yesterday with my car-15, and noticed once again that in standing and crouch it will ALWAYS recoil right+up.
The attached images show about where it recoiled to (where the 2nd shot was fired to, sort of), in standing and crouch positions with my car-15. Notice that with crouch it goes a bit less up, probably since the elbow is leaning on the knee. We've seen it in prone with the M16A1 in full auto that it barely goes up since both elbows are on the ground. Also we notice greater recoil with the car-15, and I can guess that the reason for it is the lighter weight of the weapon (about 2/3 of an M16A1).
What range is that? What is the size of the target? and why didn't you fire the second shot to make sure where it would hit? Keep it up gal-z, RL experiments should be a big help to the INF devs both on the team and in the community.

gal-z
8th Sep 2005, 06:46 PM
range is 25m, target is a man 35cm wide I think.

MP_Lord_Kee
9th Sep 2005, 12:13 AM
The pics you posted gal-z are very consistant with my own experience of recoil. On full auto with Rk 62 (http://world.guns.ru/assault/as43-e.htm) from aprox. 25m range, only first couple of shots hit the target and due to recoil it goes upwards and to the right. Easier to control recoil if hipped compared to shouldered but less accurate unless you got tracers. First shot is always off target hipped unless you are very lucky. Interestingly with tracers in the mix, single shots hipped becomes surprisingly easy, 2-3 shots to hit a 10 cm diameter stationary target from approx 25m...
On the other hand, if you don't hit and kill with your first shot it becomes a moot point if the next rounds fired are of any big importance...

//K

gal-z
14th Sep 2005, 10:47 AM
Under ~15-25m it isn't a must to hit with your first shot. SF units in the IDF are taught to fire rapidly without using sights (with weapon shouldered) and have very good groupings - at 8-12m they hit with most of their shots inside of a circle with a radius of only a few CMs, and ALL shots hit the A4 page the circle is printed on.

Harrm
14th Sep 2005, 05:00 PM
I dont think this up-and-to-the-right applies to pistols, does it? My .22 sets in a pretty tight circle when fired (the offset is rarely more than an inch). The .38 too. The .44...well...let's just say there's never any need for a second shot.

As far as rifles go, my experience definately agrees with up-and-to-the-right, but I'm a pretty poor shot with a rifle anyhow, so that's not saying much. But I noticed my friends' father (who is left-handed and fires left sided...somehow) had a tendancy to let his shots drift almost straight up. Is this indicative of all left handed shooters?

--Harrm

-Freshmeat
16th Sep 2005, 04:23 AM
If anybody are able to get a digital video of them shooting, I would be VERY happy to analyze the videos to get an idea of the angular displacement per shot. It can probably be done with any digital camera that has opportunity for digital video, as long as it shoots at least 20 frames per second.

If anybody are interested, the videos should be shot with the camera on a stand, zooming to show the weapon only, so I can get as many pixels as possible to work with. A measure af the particular weapons overall length would be necessary as well to get a reference.

-Freshmeat (Teaching high school physics for a living)

keihaswarrior
16th Sep 2005, 05:17 AM
You already know the angular displacement. He told us the range and target size. That's all you need. For a high school physics teacher, your trig isn't so hot :p

Now we just need the same thing for all the other weapons from every possible posture.

*looks @ gal-z......... well what are you waiting for get crackin! :D :p


Edit: Gal-z, what is the height of the target? Both the torso box and head. Or maybe you could just measure the vertical and horizontal distance from the first and second red dots for us. :)

jayhova
16th Sep 2005, 08:29 AM
I have something of a question here. I envision recoil having a number of effects on the shooter. First and formost is the tendency to push the shooter back this has the added effect of rotating the shooter to the right and upward. However this motion is somewhat limited by the position and stability of the shooter. One of the other effects of recoil is not at all simulated in INF is the violent change in the position of the shooters head. one of the reasons it is so hard to judge where to aim your weapon after the first shot is that you head is being jarred by the repeated buffeting of a weapon against your shoulder. This would be like trying to aim while someone is skaking you by the shoulder. Of course the fact that the weapon tends to buck around as well couldn't help.

On a completely different tack, one of the things that has always disturbed me about INF is that fact that the things that should move (as you move do to walking breathing etc.) don't and the things that shouldn't move do. When holding a rifle the most stable part of the weapon is the stock. Why is it then that this is the part that moves the most in INF? The other part that bounces around is your head. I don't care how tired you are or how much your pack weighs peoples heads don't bob like ballons in a breeze. If you are breathing hard the thing that will most move is your chest arms and shoulders and anything attached to them. I so hate seeing everything near me bobbing around like I am standing on a boat.

-Freshmeat
16th Sep 2005, 02:02 PM
@KW: Sure, the displacement of Gal-Z's second shot could be calculated. It is a good place to start. However, it does not take into account all adjustments Gal-Z might have made more or less unconsciously. A movie would be a lot more reliable information. Further, it would be nice if other persons would join in as well.

-Freshmeat

Psychomorph
16th Sep 2005, 02:37 PM
Yeah, the community should work for their favorite mod.

Nukeproof
16th Sep 2005, 04:51 PM
On a completely different tack, one of the things that has always disturbed me about INF is that fact that the things that should move (as you move do to walking breathing etc.) don't and the things that shouldn't move do. When holding a rifle the most stable part of the weapon is the stock. Why is it then that this is the part that moves the most in INF? The other part that bounces around is your head. I don't care how tired you are or how much your pack weighs peoples heads don't bob like ballons in a breeze. If you are breathing hard the thing that will most move is your chest arms and shoulders and anything attached to them. I so hate seeing everything near me bobbing around like I am standing on a boat.

I think that's were RAv3 looked a lot better. I'm really keen on testing INF-HL2...

gal-z
17th Sep 2005, 02:24 PM
Well my pictures aren't too accurate. After all they're from memory and the second shot was fast so it could've been a bit off of where I thought it was aiming, but definately not much. The target I used is actually a full body, I just cut it since I only aimed at the upper body part, but you can still *estimate* the offset.
Making accurate measurements or videoing myself isn't a practical option for me, at least ATM. It's hard enough to convince certain people to let me shoot :P

Gnam
18th Sep 2005, 03:08 AM
Infiltration uses an instant displacement for it's recoil. It is for this reason I suspect that the recoil was made linear, as random displacements of any significant scale are quite disconcerting, especially at the edges of the freeaim extents.

As an example of nonlinear recoil, try firing 75 or more consecutive (without letting go of the trigger) shots through the SAW. Standing and walking around for extra fun. The left/right motions will make it difficult to even walk straight.

Of course, using just displacement for recoil is quite simple, too simple really. Why use algebra when calculus is more fun?

I only meant it as an example of the displacement method's effects.

Subjective or not, recoil should be a velocity or accelerational motion rather than an impossible instant relocation. It does spell a length of time where a single recoil action may disrupt aim, so it is an idea to consider for those who want to make things harder.

Aaah. OK, I was wondering about that. I was experimenting with recoil in Halo a few weeks ago, and got the exact same effect as Inf; point of aim is istantly "teleported" straight up to the amount specified. By far, I think the most aparent effect of this discrency from real life is that a weapon will land the same auto grouping regardless of the the rate of fire. In other words, in real life, faster ROF's tend to produce tighter grouping on short bursts, because the 2nd and 3rd shot fire before the recoil of the 1st and 2nd round have ran their full course (the AN-94 is supposed to have such a high ROF in burst mode that it can potentially land 2 shots in the same hole at long range). The off shoot is that the recoil does catch up to you eventually, causing greater innacuracy at sustained bursts, and potentially delaying the time it would take to line up a 2nd burst (compaired to weapons with lower ROFs). Lower ROF weapons generally shoot bursts with wider groupings, because the successive shots fire after the recoil of previous shots had taken full effect on your point of aim. However, lower ROF's are easier to control in sustained fire.

In the Halo mod I rigged up, and, I assume in Infiltration, burst/auto grouping will be the same regardless of ROF because the effect on your point of aim is instantaneous. There is no way the shots could out-pace the effect of recoil on at any rate of fire. Weapons like the FAMAS with an ROF of aproximately 1000rpm should shoot tighter 3-round-bursts than other 5.56mm weapons of the same weight that shoot at 700-800rpm, but the difference is probably not existant in Infiltration, unless the programmers acheived it artificially (by lowering the overall recoil impulse).

yurch
18th Sep 2005, 11:56 AM
Correct. Inf has a 'kickmod' that increases for sucessive shots, but it usually isn't distinctly notable until the entire mag is involved. A three round burst and three single shots with no compensation will usually group the same.

BF2's SVD sniper rifle has a lesson we can learn from, as well. The recoil motion on this rifle follows a v-shape, recoiling up and then is (automatically) pulled down by the shooter, and usually ends up in a slightly different orientation.

The result of this is a half-second of disruption that nearly EVERYONE has to deal with before fring again - usually you only find people firing at maximum rate in close combat.

Of course, I think this particular recoil pattern would be more fitting to pistols.

gal-z
21st Sep 2005, 07:30 AM
Never fired a pistol so I can't comment about that.
About the 2nd shot firing before the recoil has taken a full effect - sounds a bit wierd to me. After all, in order to fire a second shot, the bolt has to move all the way backwards and forward before the shot is fired. The recoil effect is (according to my logic at least) should be fully done until then.

Nukeproof
21st Sep 2005, 07:52 AM
I agree - that's in fact what the G11 tried to overcome.

keihaswarrior
22nd Sep 2005, 01:41 AM
Never fired a pistol so I can't comment about that.
About the 2nd shot firing before the recoil has taken a full effect - sounds a bit wierd to me. After all, in order to fire a second shot, the bolt has to move all the way backwards and forward before the shot is fired. The recoil effect is (according to my logic at least) should be fully done until then.
The same is true for almost all automatic pistols. The slide must recoil back, then come forward chambering a new round. So I am not sure what gnam is talking about. While the recoil isn't instant, it is very fast.

jayhova
22nd Sep 2005, 03:12 PM
It's important to remember that recoil has a cumalative effect. Recoil in a rifle is buffered by the action. That is to say that until the bolt has moved entirely rearward the full impact of the recoil has not been felt. In addition to this the energy of that rcoil then has to transfer from the firearm into the shooter. the shooter will then compress and then expand in the direction of the recoil as the shockwave moves through him. Finally the shooters orientation will change as the upper right quadrent of his body moves in the direction of the recoil. All of this takes time. Now given that the initial direction of recoil is straight back it is possible to load a new round before the weapon has changed orientation. In the case of most weapons there will be at least some change before the second shot is fired.

Gnam
24th Sep 2005, 01:37 AM
Yeah, I'm not saying that the round fires before any recoil is felt, but the rounds fire earlier in the "recoil cycle" than lower ROF weapons. Granted, I don't know exactly how fast a rate of fire you need or exactly how much time it takes for the recoil of a single shot to have full effect on your body (probably varies with the weapon and the shooter) but this effect does exist, and Inf does not simulate it. If the AN-94 were in Inf, it would not fire any tighter bursts than the AK-MSU.

gal-z
24th Sep 2005, 09:34 PM
Before putting something like this in game, you'll have to experiment with it a lot, and feel exactly how significant it is if at all, which would be very very hard. Personally I hadn't noticed such an effect, and I'm not surprised, as it either does not practically exist or is very small, at least with my weapon :P

yurch
24th Sep 2005, 11:13 PM
Well, the problem in field 'testing' something like this is that you've basically got a free-floating weapon - the task of keeping the weapon on target will have so many factors, recoil's role will be impossible to distinguish.

I don't doubt this burst recoil effect, as I know a person is often still moving after the bolt has completely cycled, so there's still that potential time window. Too bad you can't bench-test people.

Psychomorph
25th Sep 2005, 02:58 AM
Im sure it's not a big problem for some people out there to test some guns, but hard to tell if they have such an eye for detail as gal-z.

Hmm gal, you have to go to a shooting range and test some various guns (just an idea).

Gnam
27th Sep 2005, 01:38 AM
Yeah, I think the only way to tell for sure is to compair various compairable automatic weapons of identical caliber, and close to the same length, weight, ergonomics, etc, but vary the fire rate.

I would bet a M3A1 'Grease Gun' shoots larger burst groups compaired to Thompson's and other .45 cal SMG's of that era (the fire rate of the M3 was close to 350 rpm). Another good compairson would be the Soviet PPS-43 to the PPsh-41. With the highest fire rate in a 5.56mm assault rifle that I know of, the Fa-mas would be a good compairson to almost any other 20" 5.56mm rifle, though preferably the later would have as low a fire rate as possible (AFAIk no 5.56 AR's have a fire rate under 750rpm, so it may be tough). Of course, there would still be other factors to consider which could cause one weapon to shoot tighter bursts than another, but it might give you some clue as to whether this is true. If every higher-ROF weapon shot tighter bursts, then clearly you'd be on to something.

jayhova
29th Sep 2005, 06:42 AM
Bear in mind that recoil works 2 ways. Since the shooter is expending energy to hold the weapon in position the energy of the recoil is returned as well. So a slow rate of fire weapon may give more time for the weapon to return to the original position. This is especially true of light weapons firing a heavy round.

gal-z
29th Sep 2005, 02:24 PM
If we take putting the weapon back into account and not leave it to the player with the mouse (which is an idea though I doubt it won't take too much control away from the player), only then will low rof weapons will spread less because they can be put back on target faster. If we leave it to the player, it will just be easier with lower rof weapons (assuming same recoil of course) but will still depend a lot on the player.
So we need to decide if recoil compensation is computer or player controlled. Right now I think player controled will be a better idea because making the game try to put the weapon back to where it was will take too much away from the player and won't be completely realistic anyway.

As for firing with different weapons, it's really not so easy for me. Non-pistol civilian shooting ranges are non-existant in Israel AFAIK, and in the pistol ranges you can only fire what they have (probably not too much variety there) or your own pistol, and I'll have to pay for the ammo as well :( and it's still just pistols. In the army we get to go to shooting ranges once in a while, usually happens every 2 weeks but we probably won't have time for it in the next month due to the holidays. And even then I can only fire my own weapon (everyone has the same weapon in the HQ, though there's a low chance I can get my hands on an M4A1 for shooting). Constantly checking the targets is impossible as we're always in a hurry, and for the same reason it's hard to convince the officer to let me shoot all the ammo that I manage to gather (and I gather a LOT of ammo that gets "neglected" :P).

Harrm
29th Sep 2005, 06:17 PM
Or maybe there should be a more significant weapon shake that holds true while firing the weapon?

--Harrm

Psychomorph
30th Sep 2005, 03:55 AM
@gal-z:
As I know the recoil causes always a kickback-to-front after the actual recoil kickback-back. So the weapon returns a bit, but since it is a human fireing, the weapon never returns to exact the same position where it was before it was fired (it is pointed slightly more up and sideward after the shot). That means you still have to reposition the weapon after the shot.
That means useing fullauto the screen doesn't move up, not really the weapon, but the bullet spread will be more, but if you control the weapons with the mouse you can control the recoil better and minimize the bullet spread.

That's why I think both should be the case, an automatical compensation and the need to reposition the weapon manually with mouse movement.


One I can say, no fireing till the screen looks up! I would suggest, that when kicking fullauto with a highrecoiled weapon, the screen should move only very slightly up (like shaking), but the weapon only start to move slightly up after few shots (when no time to be compensated automatical [like when using bursts]).
So you look still forward, but the weapon moves and fires more up and sways slightly around a bit every time it recoils.

When you stop fireing, the weapon moves back till it's pointed forward again (if you do not make it faster with the mouse).

gal-z
30th Sep 2005, 12:48 PM
I know that we're usually noy allowed to fire fullauto because if you're really really not careful you can hit the person standing next to you, or at least be somewhat dangerously close to doing it. So lots of firing with no compensation SHOULD cause extreme turning.

Domox
10th Oct 2005, 10:28 AM
In all the trainingg videos etc I have seen on full auto weapons, even full auto is done in short bursts (6-10) rounds max because the body torques slighty to whatever side you hold the gun, even if the shooter is particularly solid and knows what to expect from the hardware.

In my time on the pistol lines the recoil follows a basic V form if things are somewhat loose, however i have seen guys who really know what they were doing eat a hole the size of a golf ball in a target at 10m... pistol recoil is a toughy because _alot_ of it is dependent on the ammo/grip/length of the reciever etc. A desert eagle mk XIX for example shooting action express will smack you in the forehead if you dont watch it, while the same gun with .357 chambered is a whole nother story, its very easy to control imho...

what would be nice is an RPG element where the experience of the player can be transformed into various points that stabilize the gun easier or decrease recovery time after shots.... the only problem is how would you prevent people from hacking this element... :) I will tell you it would be hard if not dang near impossible....

Demosthanese
10th Oct 2005, 03:53 PM
In all the trainingg videos etc I have seen on full auto weapons, even full auto is done in short bursts (6-10) rounds max because the body torques slighty to whatever side you hold the gun, even if the shooter is particularly solid and knows what to expect from the hardware.
If they are standing, that is true to some extent, if not, then it is for the simple reason that the barrel gets hot and melts during a sustained firing. Even mounted weapons, such as the cannons on fighters, fire in bursts for this reason.



what would be nice is an RPG element where the experience of the player can be transformed into various points that stabilize the gun easier or decrease recovery time after shots.... the only problem is how would you prevent people from hacking this element... :) I will tell you it would be hard if not dang near impossible....
It's easy to prevent people from hacking it. Simply don't impliment such a silly idea.

Domox
13th Oct 2005, 07:50 PM
If you look at how you adapt to your situations and how you learn in RL you would realize the idea isnt so 'silly'. ;)

Crowze
14th Oct 2005, 06:15 AM
All learning and 'role playing' should be done by the player, not the game. If you learn how to snipe then you will probably be a good sniper. The only issue with that is there is precious little documentation on how to be a good sniper in Inf, but you can always ask :).

Nukeproof
14th Oct 2005, 06:58 AM
Actually, mild roleplaying stuff like BF2 does with bonus weapons, is at least good for longterm motivation.

DEFkon
20th Nov 2005, 02:13 AM
The problem with trying to simulate recoil is that your not really trying to simulate the weapon so much as you are trying to simulate a person, which means even with "extensive expirementation" with real life weaponary the results are going to be biased.

Personally i always thought that it would be intresting to let the player input some personal information that would be reflected in the avatar. Age, Sex, Height, weight, build glove size.
and then some questions to determin strength, endurance, and other physical traits. How much can you bench, your fastest mile, ect ect.

That way the players avatar could more closely mimic their actual physical traits. Note that nowhere did i mention any type of "skill" based system that would effect aiming or reloading times. Just how far can you run, how much can you carry, how much recoil will you "feel".

Of course people could lie, and say that they were olympic athletes, which is just fine, everyone's free to mess around and see what it's like. maybe a positive benefit would be for players to actually want to exercise in real life.

Harrm
20th Nov 2005, 10:13 AM
Personally i always thought that it would be intresting to let the player input some personal information that would be reflected in the avatar. Age, Sex, Height, weight, build glove size.
and then some questions to determin strength, endurance, and other physical traits. How much can you bench, your fastest mile, ect ect.

That idea actually sounds awesome. Makes you wish it was feasible.

I gotta' get back into airsoft...

--Harrm

gal-z
26th Jan 2006, 10:09 PM
I would say the fat geeks that can't move at all (that have an equal right to play this game) would either hate this idea or cheat it :D
My suggestion is to solve this by simply deciding on standard size, strength, stamina, training etc that the INF soldier has and use these to simulate everything - as you said, it's not just simulation weapons, it's simulating a person. I still think everyone should be equal though and play an equally skilled/trained/whatever soldier.

Psychomorph
27th Jan 2006, 01:17 PM
Yeah sure, different body heights, strengh and training levels. Now who's realism nitpicky, huh?
Why not making someone hearing better and another beeing red-green blind? One is ill, another is hungry.


Ahhh feels good, I had to do it.

Craetech
2nd May 2006, 01:03 PM
Here's my 2 cents..

1. Ironsights alignment by player
One of you guy's idea made me think of something else. The use of freeaim to manually align the front and rear posts of the ironsight. When the weapon is shouldered, there will be a slight misalignment based on the position of the weapon relative to the "sweet spot" or center of freeaim. If weapon is above the center, the front post will be above the rear post, etc. This misalignment at most is good for a ~15m shot. But you will need to move the weapon to the center of freeaim to align the sights for perfect aim, and hold RMB to "lock" the alignment (as tho there's no freeaim). Reloading, going prone, jumping, falling, getting hit, jogging and sprinting will force to unlock so the player needs to re-align after. This is more realistic (in a simulation sense), requires players to really practice to be good, and ensures difficulty for players to twitch-aim and hit someone from over 150m.

2. Recoil
I support the implementation of recoil setting the aim diagonally up-right (or up-left if there's left-handedness in future engines) because it proven realistic and also for gameplay. It is way too easy to compensate an upward-only movement (I can burst the M16 and place all 3 shots at a target 100m away) as compared with a diagonal one as the player needs to handle x- and y-axis instead of just a single axis. It may be a good idea for a different angle of diagonal based on the weight, barrel length, stance, etc.