Game physics

[Alex72 1]

Unreal Tournament, at least the old one which I remember, uses a special pose when the player is typing. So maybe a new "hand signal", where player takes hand to ear, while a player is typing (or chatting?) would be useful?

Always wanted an animation that throws hand up and like "YO!" so others could see who you are. :) Sorry not physics though. :o

[Wolfstriked 11]

Quirky feel is what kinda ruins Arma for me.Stuff like the weapon behaving radically at corners.If you walk into a wall the game does nothing.If you walk into a buildings corner the game decides which direction it wants to point you in and just does it.Walking thru doorways is another.If the game is not gonna model actual weapon in 3D space then remove all the other quirky weapon in 3D behaviors so its way more smooth.

I think a way to give a little better feel would be to have a slight delay in the weapon compared to the actual viewpoint.If you go to iron sight and press freelook you'll get an idea what I mean.Move the view up and the weapon takes a second to follow.BEing to connected to a weapon the way it is in A2 causes the stiffness.Give a little bit by freeing up the weapon and voila.I do not mean aiming zone as that just moves the weapon without moving the actual view point.I have that turned off since it feels wrong for me.

Another area is stuff like dragging a body.Why can't the game allow you to drag a body and when you press release you just drop your hold.Instead I am pointing in one direction trying to clear myself and this body around a corner and when I press release I am thrown around 180deg and moved forward in new direction....the direction the enemy is shooting from.

The ladders is horrendous and while I feel its not a ladder game it could be made more intuitive.Press V to climb over anything but only when you near an object and have the V do same thing when you near a ladder.Maybe turn off headbob while climbing?

[deathil93 10]

Hi,

I have been playing BI' games since '04 and I must say, I loved them all.

But I still remain unsatisfied with the ArmA generation games, its not the graphics or features, but its the technology.

So I have a couple of questions:

1) Why didn't BI integrate some sort of a ragdoll physics system in-game?

2) Why aren't the animations more detailed? I mean like to the extenet that you open the tank's hatch and get in or actually openning an ammo crate and taking out some mags or a rifle?

Hopefully someone has the answers so I can go all next time I read the thread :)

[Big Dawg KS 6]

1) Why didn't BI integrate some sort of a ragdoll physics system in-game?

2) Why aren't the animations more detailed? I mean like to the extenet that you open the tank's hatch and get in or actually openning an ammo crate and taking out some mags or a rifle?

I'll respond to both of these with another question:

Is it really worth it?

[madrussian 347]

IMO, moving beyond ODE to a real physics engine would be well worth it to BIS. Think of all the mod creators that would make the move!

Aside from the dire physics, the OFP/ArmA engine allows for ultimate possibility. Why not add physics too?!? Seems like a good business decision to me.

[NeMeSiS 11]

Aside from the dire physics, the OFP/ArmA engine allows for ultimate possibility. Why not add physics too?!? Seems like a good business decision to me.

Cause most people are already crying about the performance.

[madrussian 347]

From my experience programming in ODE something like a whopping 8 years ago, that physics engine is actually far from optimized. Using a modern physics engine would actually boost ArmA2 performance!

Making rag-dolling etc optional would give a mighty performance boost for folks who only want physics applied to the things it's currently applied to (and who would disable rag-dolls). In addition to performance boost, the physics on already applied things like vehicles would run better and provide a more realistic experience. (i.e. No more flying tanks or things embedded briefly into the ground.)

Using a modern physics engine does not have to kill game performance or break the bank in terms of implementation. Physics engines run in parallel to the game so removing the old-outdated-busted ODE and replacing with any modern physics engine, while it will incur some cost, should more than make up for the expense in terms of new players and modders (i.e. paying customers) into the fold seeking a modern optimized engine to play with.

To summarize we the players / modders win in three ways:

1. For those who don't care about rag-dolls, they can turn them off and still see better general physics -and- a boost in performance.

2. For those that love physics, we can play / mod to our hearts content with the new system.

3. Perhaps most importantly, all of us will experience an influx of new players / modders moving onto the ArmA2 (3?) scene. More modders = a greater amount of content and more innovative groundbreaking content. Give the modders a full set of physics commands and they create primitives and connect them together with joint of all types, and then apply torque to the resulting objects and joints. We will see discoveries and innovations we cannot yet imagine, all usable in this wonderful open-world editor.

What does BIS win? Well more income from all the new players of course! :)

Then we the community win again because with all this new capital BIS can move forward and create wonderful new open-world sandbox products with surprising new and innovative features, and the cycle continues.

There is indeed a business case for BIS to move forward with a desperately needed cost-effective physics engine replacement. ;)

Edited May 19, 2010 by MadRussian

[deathil93 10]

From my experience programming in ODE something like a whopping 8 years ago, that physics engine is actually far from optimized. Using a modern physics engine would actually boost ArmA2 performance!

Making rag-dolling etc optional would give a mighty performance boost for folks who only want physics applied to the things it's currently applied to (and who would disable rag-dolls). In addition to performance boost, the physics on already applied things like vehicles would run better and provide a more realistic experience. (i.e. No more flying tanks or things embedded briefly into the ground.)

Using a modern physics engine does not have to kill game performance or break the bank in terms of implementation. Physics engines run in parallel to the game so removing the old-outdated-busted ODE and replacing with any modern physics engine, while it will incur some cost, should more than make up for the expense in terms of new players and modders (i.e. paying customers) into the fold seeking a modern optimized engine to play with.

To summarize we the players / modders win in three ways:

1. For those who don't care about rag-dolls, they can turn them off and still see better general physics -and- a boost in performance.

2. For those that love physics, we can play / mod to our hearts content with the new system.

3. Perhaps most importantly, all of us will experience an influx of new players / modders moving onto the ArmA2 (3?) scene. More modders = a greater amount of content and more innovative groundbreaking content. Give the modders a full set of physics commands and they create primitives and connect them together with joint of all types, and then apply torque to the resulting objects and joints. We will see discoveries and innovations we cannot yet imagine, all usable in this wonderful open-world editor.

What does BIS win? Well more income from all the new players of course! :)

Then we the community win again because with all this new capital BIS can move forward and create wonderful new open-world sandbox products with surprising new and innovative features, and the cycle continues.

There is indeed a business case for BIS to move forward with a desperately needed cost-effective physics engine replacement. ;)

Couldn't have said it better :)

I wouldn't mind BI intergrating the Euphoria physics engine in ArmA3 hehe

So why BI is waiting so long?

[Flash Thunder 10]

I agree, as of now I have 2 cores on my CPU that are not even being fully productive.

More Motion Capture would be nice, for things like opening up doors, kicking open doors, vaulting through windows, reloading weapon systems etc.

Aslong as theres fewer scripts to run to simulate the effect you should get better performance as stated.

Edited May 19, 2010 by Flash Thunder
Wrong wording

[Deadfast 43]

I agree, as of now I have 2 entire CORES on my CPU not even being touched.

There was no AI present on the map, simple 15km view distance test.

[Flash Thunder 10]

:o Ooops worded that wrong, meant to say that 2 of my cores are not being used as much as the others.

its about 40 40, 80, 80,

with about 30 AI

and 5KM viewdistance with me 150% 3D resolution

But I do have a pretty crappy clockrate on my CPU

I have an AMD Phenom 9750 quad core @2.4 ghz If I wasnt so Stupid/ less paranoid of destroying my system, I would overclock it to around 3.2ghz.

Been reading alot how Hardware accelerated physics are actually boosting CPU performance in most games, sucks that it makes no difference when I enable Physx on Arma 2.

Arma needs physics engine and 64bit support though :p

[11aTony 0]

I guess it would slowly be the time to have some kind of PhysX in BIS games. While graphics detail evolved to a really nice level you pretty much play OFP. Now I dont need the character to open tanks hatch but things like how motorcycle behaves over bumps or if player walks of 1m high object. I am satisfied with what it is but it needs to keep up with other games as well. A lot more players would start playing it. Not necessarily 12 year old kids.

[Dwarden 1125]

this is discussed into depth and many times (aka repeating already written)

http://forums.bistudio.com/showthread.php?t=26335

merging ...

[deathil93 10]

this is discussed into depth and many times (aka repeating already written)

http://forums.bistudio.com/showthread.php?t=26335

merging ...

Sorry for that, the search came up empty :)

[snoops_213 75]

I guess it would slowly be the time to have some kind of PhysX in BIS games.

Shot me for saying it but they have PhysX in VBS2 now, according to the site info,

Limited support for the PhysX library from Nvidia is being deployed within VBS2. Hardware acceleration is not required at this time. It is anticipated however that hardware acceleration of the PhysX routines will only be available on Nvidia platform video cards.

Maybe we could get some of that at some point...... in 2 - 3 years :D

[mr burns 131]

Shot me for saying it but they have PhysX in VBS2 now, according to the site info,

Limited support for the PhysX library from Nvidia is being deployed within VBS2. Hardware acceleration is not required at this time. It is anticipated however that hardware acceleration of the PhysX routines will only be available on Nvidia platform video cards.

Maybe we could get some of that at some point...... in 2 - 3 years :D

Having tried the Physx humvee in VBS2 US Army lite i don´t think it´s worth to long for. It simply didn´t feel that good. When people hear Physx they think Mirrors Edge or Nvidia Demos, but what VBS2 offers (at least in the lite version) just isn´t that good.

Also US Army Lite isn´t licensed (?) for civilians, that´s why i deleted it rather quick before some speshul forces detachment dropped through my window :D

[deathil93 10]

PhysX? thats not a real solution for all players, because if the guys at BI decide to go for that, than the ATI users (like me) won't have the same experience as nVidia users.

I suggest that if BI wants to go for a commercialy available physics engine, they should either go for Euphoria (GTA IV physics were beautiful) or Havok (not as 'realistic' as Euphoria but also a good alternative). Those 2 are the most advanced (and famous) engines in the market AFAIK.

But, if BI wants to develop an in-house physics engine, it would be fine with me :)

Cause most people are already crying about the performance.

Dunno what about you, but I have 2 rigs and on my old rig with a P4, HD2400Pro and 1.5GB of RAM, ArmA1 and 2 ran just fine on medium-high settings (30-60FPS).

AFAIK, most people out there have way more advanced rigs than a P4.

But the main cause of preformance issues are the bugs, there is ofc improvement with each game released. But all that can be fixed with a driver update or a game patch.

Edited May 30, 2010 by deathil93

[Pyronick 21]

PhysX? thats not a real solution for all players, because if the guys at BI decide to go for that, than the ATI users (like me) won't have the same experience as nVidia users.

I suggest that if BI wants to go for a commercialy available physics engine, they should either go for Euphoria (GTA IV physics were beautiful) or Havok (not as 'realistic' as Euphoria but also a good alternative). Those 2 are the most advanced (and famous) engines in the market AFAIK.

Havok is very expensive middleware. There are cheaper (free) alternatives like Bullet physics.

They lack IDE's but I assume Bohemia Interactive develop their own toolsets.

Also Euphoria is not a physics engine, but rather a biomechanics simulator.

It uses physics engines (like Bullet in GTA IV/RDR) and AI to create animations on-the-fly.

But, if BI wants to develop an in-house physics engine, it would be fine with me :)

If so, I really hope they'll fork an existing engine...

[NeMeSiS 11]

Dunno what about you, but I have 2 rigs and on my old rig with a P4, HD2400Pro and 1.5GB of RAM, ArmA1 and 2 ran just fine on medium-high settings (30-60FPS).

AFAIK, most people out there have way more advanced rigs than a P4.

But the main cause of preformance issues are the bugs, there is ofc improvement with each game released. But all that can be fixed with a driver update or a game patch.

My game runs fine on a X2 4200+ 2GB RAM HD4870, but i dont see alot of people around here who are happy with their performance. There are several large threads about it.

[deathil93 10]

My game runs fine on a X2 4200+ 2GB RAM HD4870, but i dont see alot of people around here who are happy with their performance. There are several large threads about it.

They are probably the ones who OC all the components to get uber-FPS and are satisfyed only with 100FPS OR have a realy low-end computer.

Havok is very expensive middleware. There are cheaper (free) alternatives like Bullet physics.

They lack IDE's but I assume Bohemia Interactive develop their own toolsets.

Also Euphoria is not a physics engine, but rather a biomechanics simulator.

It uses physics engines (like Bullet in GTA IV/RDR) and AI to create animations on-the-fly.If so, I really hope they'll fork an existing engine...

Still, you gotta admit the physicis simulation in Euphoria is realistic and beautiful. With an engine like that, the game will be realy amazing and it would fill the technology gap :)

Also, I believe it is cheap enough for BI to use. My guess, it costs around 20-50,000 dollars. I think BI spends more than that on engine devlopment, so it could be an awesome and realistic choice.

[madrussian 347]

Shot me for saying it but they have PhysX in VBS2 now...

I think I just shit a brick. :bounce3:

That is by far the best news I've read of any kind in the last several months, and in case anyone is wondering, I am dead serious.

Hardware acceleration is not required at this time.

Not required, but oh yeah baby you can turn it on!!! Judging by the way the VBS stuff makes it's way down to ArmA series, maybe we will get real physics in BIS games within a mere year or two. Not to mention hardware accelerated at that!

Hold onto your hats, boys.

edit:

Still, you gotta admit the physicis simulation in Euphoria is realistic and beautiful. With an engine like that, the game will be realy amazing and it would fill the technology gap :)

Agreed, Ephoria is mind blowing. :)

Also, I believe it is cheap enough for BI to use. My guess, it costs around 20-50,000 dollars. I think BI spends more than that on engine devlopment, so it could be an awesome and realistic choice.

I don't think it's remotely cheap though. Considering only two game developers (last I checked) can afford it. When they don't list a price on the website, and ask you to contact them, and then mention they will send a team to assist you in implementing, well it ain't gonna be cheap. :D

Having said that, I would love to see Ephoria in ArmA and Carrier Command.

Edited June 5, 2010 by MadRussian

[PuFu 4599]

Still, you gotta admit the physicis simulation in Euphoria is realistic and beautiful. With an engine like that, the game will be realy amazing and it would fill the technology gap :)

Also Euphoria is not a physics engine, but rather a biomechanics simulator.

It uses physics engines (like Bullet in GTA IV/RDR) and AI to create animations on-the-fly.If so, I really hope they'll fork an existing engine...

What he said...

Also, I believe it is cheap enough for BI to use. My guess, it costs around 20-50,000 dollars. I think BI spends more than that on engine devlopment, so it could be an awesome and realistic choice.

I'd firstly try and understand how game engines works, before suggesting stuff i am not aware of...

[gossamersolid 155]

the physics in arma2 have improved quite a bit since ofp1 though.

So at least they are making progress.

[jeza 5416]

I too would 'like' some form of improvment, but would take an engine change i guess not sure, something not to ragdoll like i.e L4D. maybe something along the lines of CODWAW, that seemed pritty good to me. But for me ArmA/Arma2 gameplay overshadows any little bugs and blips :)

[Rayers12 19]

I can't recall any other popular modern game that doesn't have a physics engine. IF you want this game taken seriously, physics are a must.

@Gossamer: OFP1 came out in 2-freaking-001. A little bit doesn't count.