Game physics

[baddo 0]

If it's too costly to have realistically looking ragdoll physics in these kinds of games.

Then how about this.

Let animation system handle a death of a character up to the point it does currently in ArmA.

Then enable ragdoll physics mode and thus let the character follow the shape of the terrain that way.

One of the ugliest things I see in ArmA is the way the characters often have their body up in the air after death... because they stay in the position of the last animation frame. This could be solved by having at least a little bit of ragdoll physics in there.

[max power 21]

If it's too costly to have realistically looking ragdoll physics in these kinds of games.

Then how about this.

Let animation system handle a death of a character up to the point it does currently in ArmA.

Then enable ragdoll physics mode and thus let the character follow the shape of the terrain that way.

One of the ugliest things I see in ArmA is the way the characters often have their body up in the air after death... because they stay in the position of the last animation frame. This could be solved by having at least a little bit of ragdoll physics in there.

I think that that's how COD4 does it currently but I think that all the collision checking and angle solving with the ground is what is cpu intensive. The devs said that they were R&Ding ragdoll simulations for ArmA and said that no method performed well enough. Maybe this will change with the multi cpu support of ArmA2? Who knows.

[CyDoN 0]

Americas Army physics are nice. Old technologiacaly but realistic i think they should stick into that kind of physics (not source type...)

[Kirby 2]

I beleive CoD4 is semi-ragdoll, the best kind of ragdoll. Lol.

Looks pretty realistic...

Depends on the situation.

But a game at the scale of ArmAII probably wouldn't be good to put ragdoll with, atall.

[CyDoN 0]

Something must be done with Airplanes, yes we know that arma is not designed to support airplanes but the physics are not reallistic and i thing the same are used for choppers are also used for ariplanes. When u fly the A10 it feels like u are sliding in the air not flying...

[maddogx 13]

On the topic of ragdoll physics: Implementing a ragdoll effect for all soldiers in a game is simply not possible. The potential CPU load could be devastating in large games. The only real possibility to implement ragdoll would be to scale the ragdoll animation down over distance. Sort of an animation LOD, depending on scene complexity and distance, just like the current system.

So the highest LOD (close to the viewer) would be full ragdoll with good error compensation, the next lowest LOD would be simplified and so forth until the lowest LOD would be the simple animations we know now. This could reduce overall processor load and still give us a certain level of ragdoll effects.

But even that method would open a whole new can of worms - especially in multiplayer, since different players would be getting different ragdoll LODs, which would somehow need to be syncronized. Imagine a soldier is standing on a high roof and gets shot. One player is close by and the ragdoll effect makes the soldier fall off the roof. Another player is far away, so the static animation kicks in, making the soldiers body stay on the roof. What now? Does the machine with the highest LOD broadcast the final body state to all others?

And what happens if you come across a body that was previously hanging off the edge of a roof with a static animation? Does it just stay that way, or does the ragdoll effect kick in when you get close, making the body fall off?

It could be done, but it wouldn't be easy.

[frostwyrm333 1]

many older games used method of static death animation which slightly altered by terrain, so most anim physics will be moving hands and legs and maybe losing weapon. Jumping from towers would be rare (mass suicides in MP?) and still it doesnt have to be half life 2. Anims in arma 1 when corpse is hovering down the hill looks very bad.

And i think almost everything in arma is lodded, and there could be physics deactivation trigger for mp and sp, and some treshold for max rendered anims. So you want notice that half of the nuked squadmates wont get rag doll or very simplified one. May be depended on performence option.

[baddo 0]

Implementing a ragdoll effect for all soldiers in a game is simply not possible.

Well I think you exaggerate and generalize with that statement at least a bit don't you

Never say never.

What I actually want is not necessarily ragdoll physics. What I want is to get rid of the very ugly-looking "look I'm dead and 80 % of my body is up in the air!" poses the dead units have in ArmA. I don't care what technology is used to improve it. I see ragdoll physics as a possible tool to solve this problem, and I think it could be made to work like I said, only have ragdoll enabled at the very end of a unit's movement. You could even do it so that not all units would get it, but only the ones the player is going to see very soon. If a player never goes into an area where there are dead soldiers, then ragdoll death would not need to be enabled for those soldiers.

[max power 21]

Implementing a ragdoll effect for all soldiers in a game is simply not possible.

Well I think you exaggerate and generalize with that statement at least a bit don't you

Never say never.

What I actually want is not necessarily ragdoll physics. What I want is to get rid of the very ugly-looking "look I'm dead and 80 % of my body is up in the air!" poses the dead units have in ArmA. I don't care what technology is used to improve it. I see ragdoll physics as a possible tool to solve this problem, and I think it could be made to work like I said, only have ragdoll enabled at the very end of a unit's movement. You could even do it so that not all units would get it, but only the ones the player is going to see very soon. If a player never goes into an area where there are dead soldiers, then ragdoll death would not need to be enabled for those soldiers.

I don't think he's exaggerating. Then he goes on to give a possible method for decreasing load... and then you go on to take the first sentence of his post and go on about how he's generalizing...

At best you come off like you didn't read his post at all.

[baddo 0]

Implementing a ragdoll effect for all soldiers in a game is simply not possible.

Well I think you exaggerate and generalize with that statement at least a bit don't you

Never say never.

What I actually want is not necessarily ragdoll physics. What I want is to get rid of the very ugly-looking "look I'm dead and 80 % of my body is up in the air!" poses the dead units have in ArmA. I don't care what technology is used to improve it. I see ragdoll physics as a possible tool to solve this problem, and I think it could be made to work like I said, only have ragdoll enabled at the very end of a unit's movement. You could even do it so that not all units would get it, but only the ones the player is going to see very soon. If a player never goes into an area where there are dead soldiers, then ragdoll death would not need to be enabled for those soldiers.

I don't think he's exaggerating.  Then he goes on to give a possible method for decreasing load...  and then you go on to take the first sentence of his post and go on about how he's generalizing...

At best you come off like you didn't read his post at all.

What?

Oh yes he was exaggerating. He said "Implementing a ragdoll effect for all soldiers in a game is simply not possible." So in what kind of game? See my point? He was generalizing and exaggerating his point quite a lot.

My point, why do people jump here and say "simply not possible" if they didn't try it by themselves? Did they try it? I don't know. But being overly pessimistic is surely not a way how technologies improve.

What he's describing is what I was talking about too, partly. The ragdoll effect wouldn't have to happen everywhere everytime. What the player sees is what matters. Multiplayer? You will have to leave unsynched other things too. Also, what I was explaining earlier about when the ragdoll physics would take over. It would be at the very last stage of a soldier's dying. It could be after the current ArmA death animation has played to the end. Now, the soldier's position in the gameworld would be fixed to this position (except the height maybe). Then only the soldier's limbs would be moved with the help of a little bit of ragdoll simulation to make it look like the soldier actually rests on the terrain and not in the air. You would not have to synchronize this in multiplayer. What difference to the gameplay would it make if a soldier has its left hand under his right armpit or in his crotch. But for the individual player, this would make things better, as the dead soldiers would look to be lying on the ground more naturally than they currently do in ArmA.

Anyways, for letting me know that I didn't read someone else's post.

[max power 21]

What?

Oh yes he was exaggerating. He said "Implementing a ragdoll effect for all soldiers in a game is simply not possible." So in what kind of game? See my point? He was generalizing and exaggerating his point quite a lot.

My point, why do people jump here and say "simply not possible" if they didn't try it by themselves? Did they try it? I don't know. But being overly pessimistic is surely not a way how technologies improve.

What he's describing is what I was talking about too, partly. The ragdoll effect wouldn't have to happen everywhere everytime. What the player sees is what matters. Multiplayer? You will have to leave unsynched other things too. Also, what I was explaining earlier about when the ragdoll physics would take over. It would be at the very last stage of a soldier's dying. It could be after the current ArmA death animation has played to the end. Now, the soldier's position in the gameworld would be fixed to this position (except the height maybe). Then only the soldier's limbs would be moved with the help of a little bit of ragdoll simulation to make it look like the soldier actually rests on the terrain and not in the air. You would not have to synchronize this in multiplayer. What difference to the gameplay would it make if a soldier has its left hand under his right armpit or in his crotch. But for the individual player, this would make things better, as the dead soldiers would look to be lying on the ground more naturally than they currently do in ArmA.

Anyways, for letting me know that I didn't read someone else's post.

He was obviously talking about having a complicated ragdoll simulation for every soldier on a complex mission in ArmA.  The post really isn't that difficult to understand.  Obviously it's possible to have every soldier have a ragdoll simulation in america's army, because every soldier does have one.  Moreover, later in his post he goes on to suggest ways to cheat it or to get around the cpu load. There must either be some kind of severe problem with english reading comprehension here or you're constructing a weak strawman argument for the sheer joy of debunking it.  Either way, the specifics of his post are clear, and they are not what you are arguing against.

[-Total- 0]

I will say that bodies laying suspended is an ambience killer in some minor ways, but not totally detrimental to the game.

Someone, in the midst of all the ragdoll discussion, mentioned the physics of the air vehicles.

The behavior of the helicopters is actually better than other games I have seen, but still not realistic. The fact that the throttle doesn't just give speed, but also affects the altitude was a great effort on BIS's part as well as the need to drop and pop the throttle while flaring the helo in order to hover or land.

If anyone has even sat in a military helicopter while in flight, they'd notice the pilot is moving all four limbs at any given point. A chopper has to be controlled the entire time.

If programmed to react to engine torque and all the other variables, the coppers would be the hardest part of the game. And personally, that would suit me just fine

[baddo 0]

What?...

He was obviously talking about having a complicated ragdoll simulation for every soldier on a complex mission in ArmA.  The post really isn't that difficult to understand.  Obviously it's possible to have every soldier have a ragdoll simulation in america's army, because every soldier does have one.  Moreover, later in his post he goes on to suggest ways to cheat it or to get around the cpu load.  There must either be some kind of severe problem with english reading comprehension here or you're constructing a weak strawman argument for the sheer joy of debunking it.  Either way, the specifics of his post are clear, and they are not what you are arguing against.

Why do you discuss me and not the topic? I certainly don't like the way you talk to me.

I pointed out that saying "impossible" in the topic that has been discussed here recently, is exaggaration and generalization, and I truly believe it is. The rest of his post makes sense. If you read my post you should notice that I wasn't arguying against his points, except the "impossible" part of his post. If you read my posts again, you will notice I also discussed the problems he mentioned and how to possibly prevent the problems from being actual problems for gameplay.

I suggest you go to a boxing gym and start to throw your insults around there and not here. I'm not interested to fight with you.

[max power 21]

What?...

He was obviously talking about having a complicated ragdoll simulation for every soldier on a complex mission in ArmA.  The post really isn't that difficult to understand.  Obviously it's possible to have every soldier have a ragdoll simulation in america's army, because every soldier does have one.  Moreover, later in his post he goes on to suggest ways to cheat it or to get around the cpu load.  There must either be some kind of severe problem with english reading comprehension here or you're constructing a weak strawman argument for the sheer joy of debunking it.  Either way, the specifics of his post are clear, and they are not what you are arguing against.

Why do you discuss me and not the topic? I certainly don't like the way you talk to me.

I pointed out that saying "impossible" in the topic that has been discussed here recently, is exaggaration and generalization, and I truly believe it is. The rest of his post makes sense. If you read my post you should notice that I wasn't arguying against his points, except the "impossible" part of his post. If you read my posts again, you will notice I also discussed the problems he mentioned and how to possibly prevent the problems from being actual problems for gameplay.

I suggest you go to a boxing gym and start to throw your insults around there and not here. I'm not interested to fight with you.

I'm not fighting, man. Nor did I insult you. I'm maintaining that you latched on to a part of his post and aren't understanding the gestalt of it. He said very clearly under what conditions it was impossible, and under what conditions it may be possible. I don't believe you are fully comprehending the post, and I will leave it at that.

[baddo 0]

What?...

He was obviously talking about having a complicated ragdoll simulation for every soldier on a complex mission in ArmA.  The post really isn't that difficult to understand.  Obviously it's possible to have every soldier have a ragdoll simulation in america's army, because every soldier does have one.  Moreover, later in his post he goes on to suggest ways to cheat it or to get around the cpu load.  There must either be some kind of severe problem with english reading comprehension here or you're constructing a weak strawman argument for the sheer joy of debunking it.  Either way, the specifics of his post are clear, and they are not what you are arguing against.

Why do you discuss me and not the topic? I certainly don't like the way you talk to me.

I pointed out that saying "impossible" in the topic that has been discussed here recently, is exaggaration and generalization, and I truly believe it is. The rest of his post makes sense. If you read my post you should notice that I wasn't arguying against his points, except the "impossible" part of his post. If you read my posts again, you will notice I also discussed the problems he mentioned and how to possibly prevent the problems from being actual problems for gameplay.

I suggest you go to a boxing gym and start to throw your insults around there and not here. I'm not interested to fight with you.

I'm not fighting, man.  Nor did I insult you.  I'm maintaining that you latched on to a part of his post and aren't understanding the gestalt of it.  He said very clearly under what conditions it was impossible, and under what conditions it may be possible.  I don't believe you are fully comprehending the post, and I will leave it at that.

I don't know why you keep insisting that.

You came into this discussion only to tell me that I have misunderstood MadDogX, post. That is your input to the topic at hand.

You have ignored everything else I have said about the topic than my comment about "impossible". Why you do that is hard to explain, what I can think of is that you are here to have a fight and not to discuss the topic. Why else would you ignore ~99.9 % of what I have said.

You are simply put wrong about what you are saying. You can stop repeating it now.

Edit: I have to add that maybe you have something to say relating to the topic? Riding on "I think you have not comprehended" doesn't in my opinion make a good discussion. Especially when that is the only thing you can say. So please, do keep discussing on this topic but if all you can tell us is that I haven't comprehended something, then you should stay out of this topic.

[max power 21]

I don't know why you keep insisting that.

You came into this discussion only to tell me that I have misunderstood MadDogX, post. That is your input to the topic at hand.

You have ignored everything else I have said about the topic than my comment about "impossible". Why you do that is hard to explain, what I can think of is that you are here to have a fight and not to discuss the topic. Why else would you ignore ~99.9 % of what I have said.

You are simply put wrong about what you are saying. You can stop repeating it now.

Edit: I have to add that maybe you have something to say relating to the topic? Riding on "I think you have not comprehended" doesn't in my opinion make a good discussion. Especially when that is the only thing you can say. So please, do keep discussing on this topic but if all you can tell us is that I haven't comprehended something, then you should stay out of this topic.

Comprehension is quite important to discussion, actually.

Baddo, your posts are absolutely surreal. It's like you're substituting an image of yourself on someone else and then taking them to task on your own issues. You generalize a whole post from a single sentence, and accuse them of making a generalization. You repeat yourself and accuse me of repeating. You get offended and start getting combative and accuse me of wanting to 'fight' you. I suggest that you're missing the totality of Maddog's most, and you come back with that I'm ignoring the point of your argument. Your kaleidoscopic posts are puzzling to say the least.

I kept telling you that you weren't taking the full post in context because you kept on defending your very wrong interpretation of MadDog's post. Maddog was not generalizing. He was actually being *very* specific. I sincerely think that we have a failure to communicate here due to either a poor understanding of the issues or a poor grasp of english, which seems to not be your first language.

I was taking the time to focus on this issue because I wanted to be able to move forward in the discussion with everyone understanding the complex issues. It's useful to be able to actually concentrate on an issue and discuss it.

....I find it again very interesting that you accuse me of not contributing anything else to the topic despite the fact that I have meaningfully contributed to quite a few pages of that topic... with several of my posts on the very same page as the start of this argument!

If you wish to discuss this issue further, I can be reached in PM.

[baddo 0]

You generalize a whole post from a single sentence

I did not do that. But someone here did.

You are actually telling me what did I think about MadDogX's post.

Continuing this discussion is obviously totally worthless. It has been totally worthless from the very moment when you initially commented my words by taking that one sentence from my post and totally ignoring the rest.

[maddogx 13]

Whoa guys, chill! No need to argue.

I think Baddo may have slightly misunderstood me. When I said it would be impossible to implement ragdoll for all soldiers in a game, I wasn't talking about Americas Army or something. I was talking about big mutha ArmA missions - say with >100 soldiers.

Not only is this more than you will find in any other multiplayer FPS at any one time, but also there is something many people forget about ArmA. Something which sets it apart from most other games:

Corpses can be used as cover.

This makes their position very important in some circumstances,

so in a multiplayer session, corpses would HAVE to be syncronized. It's not about their cold dead hands dangling over their crotch, it's about the position of the body and how it can be used.

I believe this is one of the main problems the devs faced with ragdoll in ArmA. I personally don't blame them for not implementing it.

[Johnnie_Walker 0]

I would like to know the forum nick (if any) of the new physics programmer hired by bis

[stakex 0]

Whoa guys, chill! No need to argue.  

I think Baddo may have slightly misunderstood me. When I said it would be impossible to implement ragdoll for all soldiers in a game, I wasn't talking about Americas Army or something. I was talking about big mutha ArmA missions - say with >100 soldiers.

Not only is this more than you will find in any other multiplayer FPS at any one time, but also there is something many people forget about ArmA. Something which sets it apart from most other games:

Corpses can be used as cover.

This makes their position very important in some circumstances,

so in a multiplayer session, corpses would HAVE to be syncronized. It's not about their cold dead hands dangling over their crotch, it's about the position of the body and how it can be used.

I believe this is one of the main problems the devs faced with ragdoll in ArmA. I personally don't blame them for not implementing it.

I wouldn't say its impossible to do ragdoll on a very large scale such as ArmA as far as lag is concerned... but its not viable. Im sure it can probly be pulled off on very high end PC's. But it would not be possible on current gen consol, and making a game that needs a very high end PC to run is not a good way to sell games.

MP is of course another issue, and you make a great point about it. Ragdoll is not needed in ArmA, but the devs should really work on the death animations. They were better in ArmA, but still need work.

[baddo 0]

A possible solution for the synchronizing problem in multiplayer with the ragdolls could be this:

1) Make ragdoll death simulation individually on each client and on the server.

2) Interpolate the pose of the dead unit on clients relatively slowly to that of the server. This could happen over a period of like 30 seconds or more, maybe over several minutes. The server would have sent the correct pose 'matrix' at an appropriate time, and the client would then make interpolation between its current pose and the one from the server. The speed of the interpolation could be adjusted to fit the situation.

Units which are out of the player's sight would not have ragdoll simulation at that player's computer at all, it would just receive the info from the server at a time appropriate for the situation. And delaying the synchronization all the way to the end of the mission could be possible; if the player doesn't come close to that part of the island, then why bother synchronizing at all.

If the player is in an airplane, the synchronization of ground unit ragdoll poses could also be 'relaxed' because it doesn't happen so that the player could reasonably notice exactly what happened for one dead unit on the ground.

Instant synchronization of the pose would be a performance problem. So why not do it slowly. Most of the time the players would not notice it, they are concentrating on other things.

Also, the synchronization could be delayed until the player goes far-enough from the unit to be synchronized: then no visual interpolation is needed as the player isn't watching.

Also, consider if the player has a battle at South Sahrani, kills some bad guys, then goes to North Sahrani, kills some bad guys, then goes back to the location of the first battle in South Sahrani: would the player remember anymore in which kind of poses the corpses originally were? Not very likely.

Surely there would be odd situations every now and then, when the player would notice that the server didn't register a bullet hit to another player who is hiding behind a corpse, but when the player in fact saw on his machine that there should have been a direct hit. But how often this kind of thing would happen? I think it would be very rare, thus it would be insignificant to the gameplay. I'm making an assumption here about ArmA's internals that the server is the one who decides if a hit happened or not.

As an effort to summarize what I'm saying, my point is to delay the synchronization of ragdolls or not doing it at all if possible. When the synchronization happens, it could be slowly interpolated depending on how near the player is. All this is based on an assumption that the server is authoritative over bullet hit calculations and over other important gameplay calculations. The server would always have the 'correct' ragdoll poses.

I would be surprized though, if the BIS devs didn't already think about this kind of method.

[maddogx 13]

All this is based on an assumption that the server is authoritative over bullet hit calculations and over other important gameplay calculations.

I'm pretty sure that assumption is not correct.

Bullet hit calculations appear to be executed on the machine local to the shooter and then broadcasted. I draw this conclusion from the fact that players can cheat by increasing the damage values of weapons on their machine, and these effects will work in MP. If the server was authoritative for hit calculations, this would not work.

[baddo 0]

All this is based on an assumption that the server is authoritative over bullet hit calculations and over other important gameplay calculations.

I'm pretty sure that assumption is not correct.

Bullet hit calculations appear to be executed on the machine local to the shooter and then broadcasted. I draw this conclusion from the fact that players can cheat by increasing the damage values of weapons on their machine, and these effects will work in MP. If the server was authoritative for hit calculations, this would not work.

In my opinion it would not be right to check if a hit has happened on clients. Notice I do not speak of what kind of weapon a player has, only if a hit happened or not.

It is hard to believe that ArmA would work so that it let's the clients decide if a hit has happened. Due to lag you can have significant error in the location of units shown on your screen. I don't see any other way to reasonably solve this problem other than to do the hit-checking on server.

I remember from the OFP times that I thought some times that I had scored a hit in multiplayer, but in fact I hadn't. The reason why I thought that I had hit another player was that the game had too much lag and the other unit was displayed in the wrong place on my screen. If the hit-checking was done locally on my computer, I would have had a hit. The same thing happens in ArmA in my opinion.

[dentist guba 0]

i think the physics need improvement, especially the vehicle handling, this game will be competing with OFP2 which has the same engine as DIRT.

as for the movement of people i think it needs to be much smoother, hopefully the 360 version at least will have proper analogue speed control.

[KAOSInc 0]

As above...vehicle handling! These programmers have never driven a car from the road onto grass I'm sure of it! Cars just don't decrease their speed considerably when driving from the tarmac to off road but they do decrease in the steering. Sandy areas....yes, the speed does drop greatly...but not on the grass!

And motorbikes don't do wheelies going up hills all the time. How about adding leaning forward and backward to keep the front wheel down?