What if CryEngine was used as Arma 3 future engine?

[tpw 2315]

Fair enough, I've never really noticed a problem tbh but then I could just be far too used to compensating. I've lobbed a nade into a building to clear it without taking damage in arma 2 many times... Incidentally in Crysis wars you can drop a frag through the skylight of a building and still get killed by it despite the thick concrete ceiling between the blast and you.

Admittedly there is only one way of chucking the grenade in Arma 2 and this does feel clumsy at times, if there was a way to roll it or underarm toss it would be quite good rather than having the frag bounce off the wall of the building and explode in your face. Adding this isn't necessarily an engine change though but some clever scripting could do it with the exsiting system.

ACE used to allow different styles of grenade tossing, including underarm. I think it's been removed now, but it at least proves the concept.

[maddogx 13]

That is all true, and is in fact a point that I often bring up myself :) the essential non-player-centric nature of the game engine. However, for SP gameplay purposes, a "bubble" of physics can be implemented I expect, centered on the player, like a sort of "view distance" but for physics. All events outside the bubble revert to old-style engine handling, there's probably no difference to the player experience to come across a situation that's happened outside the physics bubble long before the player got there.

In any case, it's just a musing on my behalf, I'd rather the current engine than a lesser but physics-enabled engine :)

This sounds like an idea I had been playing with, regarding ragdoll physics...

The main argument against ragdoll working in Arma2 is that the game world and potential battles are huge, so stuff like ragdoll physics would need huge amounts of processing power and - in multiplayer - bandwidth to work. But this is assuming that such physics would need to be handled all over the map, in real time and with full fidelity, while being 100% synchronized on all clients. I think that assumption is wrong.

If a bunch of AI are duking it out without any players being near enough to watch, why should full fidelity ragdoll physics be calculated for them? Wouldn't it be enough to simplify the animation, or as DMarkwick said, simply play the old static death anims? It would also be possible to delay distant ragdoll calculations, giving them a lower priority. Who cares if the death anim of an AI two miles away took ten seconds instead of one, and happened with very low fidelity? No one was watching anyway.

As for animation states being synchronized across the network, is it really necessary for it to be 100% precise?

Example: Say an AI local to a player is killed while standing on the edge of a building rooftop, while other players nearby are also watching. The main problem with ragdoll physics is getting all players across the network to see the AI fall in the same way, mainly to avoid situations where the AI is still lying on the roof for some players, but on the ground for others (since a body in Arma2 can be used as cover, and also has weapons and ammo that can be taken).

The first idea that comes to mind is: synchronize the ragdoll animation. This sounds simple, but is problematic due to bandwidth requirements. So, how about reducing the workload a little? Say, instead of synching the full animation frames, transmit only the position of the head, upper torso and lower torso while the ragdoll anim is playing, and let the other clients add in their own "details", for example where the legs and arms are concerned (hooray for inverse kinematics). The animations may not look exactly the same across all clients, but at least the body will end up at the same location and in a similar position.

The only real issue I see to implement something like this is development time/costs.

[Engioc 10]

Wow have you seen the new cryEngine 3 engine? This could suit a game like Arma!

Fully destructible enviorments, Open world . easy create to mods and can convert your work into the consoles in "real time" and last but not least, looks downright Sexy

http://www.joystiq.com/2009/10/14/cryengine-3-released-box-of-tissues-not-included/

Well I don't really like the idea of them going out and buying an engine. CryEngine 3 looks good but I'll explain my reasons why I don't like to see developers using some 3rd party product.

I was a long time fan of the old Might and Magic RPG series, now I know this is a very different kind of game but one of my memories of the last few episodes of MM was that they bought the LithTech engine and although without a doubt it was the best looking MM ever many fans including myself felt the game ended up feeling more like a FPS than a true RPG and much of that was down to the engine because it just didn't really suite the style of game it was being used for.

In the end the only way to get an engine that truly caters for the game your trying to make is to make the engine yourself and so I really hope BI stick with producing it themselves and I see no real reason why they cant particularly on the back of their successes with A2 and expansions. As others have pointed out although we have little real info there would appear to be a very good engine behind the CC game BI are working on and there is always a good chance that engine will also cater to their needs for a future ArmA game.

Having said that I also don't really agree with comments about BI not developing a new engine (either a much updated version of their current version or something completely new which ever you prefer to call it). I really cant understand these kinds of comments simply based on the fact BI are a small time developer. Yeah sure they are but they are a GAME developer and if they want to remain in that business then at some point they will need a new engine and there is no way around that, either get a new engine or get out of gaming because no engine can last absolutely forever. Again I state I'm not suggesting they need to throw away all code and begin from scratch, I'm simply saying that it either needs a major overhaul before A3/A4 (I'm not saying all the changes need to occur in one version) or they will need to buy one because in 2, 3, or maybe even 4 years time when they might be looking to release an A3 the current engine will just make them look silly when its surrounded by other better looking games with newer engines. I'm also sure BI are already aware of that, these guys aren't stupid and I've got every faith they know what needs to be done. They probably already have a list of all the things they'd like to update/change about the current engine, prioritized and they'll work through them in a sensible manner and when ready they'll release the next game. That's what they do, its their job to produce games and the tools to continue producing them. If they listened to much of the attitude we read on these forums they would just give up and stop producing games because everyone seems convinced that they have zero chance of producing something new. What a load of BS, if they don't produce a new game, what is it you expect them to do with their time???? Do you really think that in 20 years time the current engine will make a salable game?

And don't anyone give me that Sim vs game garbage, its so often used as the argument as to why we don't need something better, is ok its a sim so we don't need anything new, forever. It's just a stupid, over used, excuse and no matter what you class the product as, it IS a game because it IS designed for our entertainment so drop the stupid labels and just realize it needs to compete with the other FPS on the game store shelf or BI will just go bust sooner or later. So really just pull your heads out of the sand and accept that like every other game out there A2 does need improvements in all aspects, thats what gaming companies are about, bigger, better.

Edited October 8, 2010 by Engioc

[gossamersolid 155]

cryEngine3 could NOT work in a game like ArmA... it would run like crap if you put a map the size of arma's in it.

People say Crysis' levels are large... they aren't though, there's loading points every 2 steps you take.

[Big Dawg KS 6]

Yeah sure they are but they are a GAME developer and if they want to remain in that business then at some point they will need a new engine and there is no way around that, either get a new engine or get out of gaming because no engine can last absolutely forever.

Care to explain your reasoning? Developers don't just switch to new engines for no reason, especially if the engine they've been developing was designed/optimized for the type of games they make. There's not one good argument in this entire thread to justify BI dropping RV; especially not after over 10 years of development. Yes, it will probably be necessary for them to make significant changes to RV to keep up with new technology. But that's not the same as just dumping RV and using different technology.

[froggyluv 2130]

That is all true, and is in fact a point that I often bring up myself :) the essential non-player-centric nature of the game engine. However, for SP gameplay purposes, a "bubble" of physics can be implemented I expect, centered on the player, like a sort of "view distance" but for physics. All events outside the bubble revert to old-style engine handling, there's probably no difference to the player experience to come across a situation that's happened outside the physics bubble long before the player got there.

In any case, it's just a musing on my behalf, I'd rather the current engine than a lesser but physics-enabled engine :)

This.

Similar to Mount and Blade, the player has an option for how many ragdolls he wants displayed in the options menu -from 10 to unlimited. In SP, I've used unlimted ragdolls for battles up to 900 AI strong without a problem. However, M&B battles all take place within a confined randomly generated battle terrain and bots don't use cover.

M&B multiplayer seems to handle all of the fast action ragdolling pretty well with servers up to 200 players as long as you have a ping under 100, your pretty good to go.

Even with MP kinks, would be a great nod to the lot of us who still prefer SP Arma :)

[Engioc 10]

Care to explain your reasoning? Developers don't just switch to new engines for no reason, especially if the engine they've been developing was designed/optimized for the type of games they make. There's not one good argument in this entire thread to justify BI dropping RV; especially not after over 10 years of development. Yes, it will probably be necessary for them to make significant changes to RV to keep up with new technology. But that's not the same as just dumping RV and using different technology.

You fail to take note of the fact I said they don't need to throw away the current engine. All I've said is that as with all game engines and any other program it will need to be updated and at times have major updates done to it. So the reason I've not put up any argument for dropping RV is because I'm not suggesting they should, update does NOT equal drop RV and start from scratch. It means continual development on the engine they have, or if they feel for some reason its no longer going to meet their needs or it works out cheaper/better to buy a 3rd party product then do that instead (but my belief is RV is just fine and there should be no reason it cant be updated).

Look at any other game and engine around, they do NOT remain stagnant despite people here who seem to suggest they do, they are continually updated between each major release in a series. Sometimes these updates are small, sometimes they're significant changes to the code. I also said I don't expect or suggest that any changes they make need to be done before A3, they could and should be spread out over a reasonable amount of time as with any other game engine being developed. It's really people like you who put up no real argument why BI should stick with one engine forever as if it will make great games forever from one engine without even a single update. Current version of RV is going to look pretty dated in a few years from now and if BI dont find a way to update it then they might as well stop calling themselves game developers because they'll have no salable game to develop.

Can you put up an example of any game that has never had at least some update to its engine since its original development? All the engines people continually point to like Unreal etc and saying how old they are, well they aren't old, sure the original one is but what they're using right now has had plenty of changes since, some of those changes are pretty significant and if they weren't we'd all be laughing about how dated it looked/worked. Current system wont last forever, simple fact.

BI produce games, in order to do that they need to have an engine, and it needs to remain current so it competes with the rest of the games on the shelf. Can it do that without any updates? How?

How does an engine remain current without updates? How does anything in life work forever and never change??

What's more is A2 is proof that updates are required and BI know it. The current engine in A2 is not the same as the original OFP, its had plenty of changes. So as I stated once before on this forum I think some people have trouble with the concept that at some point you've made so many changes to an engine that regardless of its name and whatever original code does remain it is effectively a new engine. Slap a new name on it, stick a larger number on it, call it a new engine, or call it an update of current engine, it equals the same thing, you've made some significant changes to it. You can expect over the next 10 years BI will make plenty more changes.

If your simply looking at my statement which is they will need a new engine or stop making games, well maybe I should of made it clearer that new doesn't mean throw away RV. I consider a major update to RV to equal a new engine, ie Word 2010 is new. If your going to get picky about words:

Yes, it will probably be necessary for them to make significant changes to RV to keep up with new technology.

There is no "probably" about it, RV as it is now may be fine for A3 depending on when they make it, but its hardly going to be good enough for A4 and beyond. Updates WILL be required as new technologies develop.

Edited October 8, 2010 by Engioc

[dmarkwick 261]

The first idea that comes to mind is: synchronize the ragdoll animation. This sounds simple, but is problematic due to bandwidth requirements. So, how about reducing the workload a little? Say, instead of synching the full animation frames, transmit only the position of the head, upper torso and lower torso while the ragdoll anim is playing, and let the other clients add in their own "details", for example where the legs and arms are concerned (hooray for inverse kinematics). The animations may not look exactly the same across all clients, but at least the body will end up at the same location and in a similar position.

Hey - you've been reading my posts :D I've mentioned just this exact idea several times now. The ONLY thing that needs synchronising across clients is the torso position. That's all. So given that the torso position is being synched, it would be feasible for all client machines to generate arm leg & head movements via ragdoll to give visual dissonance to all bodies. It doesn't matter in the slightest if one client sees arms & legs in one position while another client sees them differently, if the torso itself is placed properly across both machines. It's an irrelevant detail, and probably comes into the sort of difference you'd see anyway with smoke effects etc.

Edited October 9, 2010 by DMarkwick

[Big Dawg KS 6]

Way huge post...

Yea, I didn't really read your entire post until I was about 4/5 done with mine and it was a rushed post (as is this). Still you seem to think that I implied RV needs no changes. That is totally inaccurate; I fully understand that BI needs to be continually developing RV. However, I really don't believe BI has the capacity to massively overhaul any significant part of RV just yet. Of course things will probably change now that BI has grown and encompassed more projects, so it's not unreasonable to think that they will in the near future.

[Engioc 10]

Yea, I didn't really read your entire post until I was about 4/5 done with mine and it was a rushed post (as is this). Still you seem to think that I implied RV needs no changes. That is totally inaccurate; I fully understand that BI needs to be continually developing RV. However, I really don't believe BI has the capacity to massively overhaul any significant part of RV just yet. Of course things will probably change now that BI has grown and encompassed more projects, so it's not unreasonable to think that they will in the near future.

N problems Big Dawg, I should of read your post more clearly before i posted too. I think in the end we're both agreeing here that there's nothing really wrong with RV but it will need updates. I don't support BI going out and buying a 3rd party engine anyway because as I said I feel the only way to really get an engine that fully suites the game is for BI to make it themselves. The one game I've played where I knew it had a 3rd party engine being used , M&M, for me wasn't a success because although it did some of what the game needed, it wasn't 100% right for the game.

Also just point out I've got nothing against the current engine as it is in A2, OA, etc. It's a fine engine and game and I see no reason why BI cant continue releasing DLC and expansions to A2 without any major updates to the engine, everything they've done so far has been great and I'd happily pay for more. A2 is without a doubt one of my favorite games atm and despite the fact I play all kinds of games and there are quite a few I'm excited about atm (some of them from other small time developers like Nitro Cota) I'm very much looking forward to more from BI and ArmA, will be watching out for CC too.

Edited October 9, 2010 by Engioc

[noubernou 77]

BI produce games, in order to do that they need to have an engine, and it needs to remain current so it competes with the rest of the games on the shelf. Can it do that without any updates? How?

BI also produces simulators (VBS) and as far as their engine goes its pretty much the best in the field they are in. There really is not a lot of close competition there, and VBS2 looks a lot more like ArmA1 than ArmA2... and from what VBS2 costs and the licenses they have sold to a number of military organizations I can imagine they make quite a pretty penny from that side of the business as well.

[ray243 11]

All bi needs to do is add in physics and realistic damage system, and I'll be happy enough.

[Big Dawg KS 6]

All bi needs to do is add in physics and realistic damage system, and I'll be happy enough.

How could you miss the entire discussion? Physics are in the game. Simply saying "add in physics" is not even a valid suggestion.

[noubernou 77]

All bi needs to do is add in physics and realistic damage system, and I'll be happy enough.

Physics are handled in the game, but they have to be specified simulations of only certain subsets of physics properties. Each simulation is meant specifically for these objects, there are simulations for bullets that are specialized to handle bullet flight, shell simulation for larger rounds, rocket and missile simulations, then of course simulations for cars, helicopters, and airplanes.

Each one of these in the engine is made to do those specific calculations. Not many other games have the same level of physics modeled that the BIS engine does, and most skimp on simulating EVERY bullet, which the VR engine does simulate.

[ray243 11]

And, is VBS2 more superior than Arma 2's engine?

[Zipper5 74]

And, is VBS2 more superior than Arma 2's engine?

In terms of some of its features, yes, but remember that it's still running on RealVirtuality 2, ArmA's engine, whereas Arma 2 runs on RV3.

[froggyluv 2130]

Would be a treat if the Dev's could comment about ambitions and/or projects regarding the future of this series :)

[Engioc 10]

Physics are handled in the game, but they have to be specified simulations of only certain subsets of physics properties. Each simulation is meant specifically for these objects, there are simulations for bullets that are specialized to handle bullet flight, shell simulation for larger rounds, rocket and missile simulations, then of course simulations for cars, helicopters, and airplanes.

Each one of these in the engine is made to do those specific calculations. Not many other games have the same level of physics modeled that the BIS engine does, and most skimp on simulating EVERY bullet, which the VR engine does simulate.

Well I do believe its physics could be better in some areas but I do take note that what physics are there go way beyond what anyone else bothers to do. Don't get me wrong the physics as is are amazing level of detail with every bullet etc as you said, but I do feel sometimes the vehicles dont react how you might expect. People seem to bounce off vehicles in a very unrealistic way, you can drive full throttle at someone and they sort of slide along in front and dont go under or get killed (I'm not saying they never get killed, but there are plenty of times they don't die when you would of expected they would). Seen this happen many times in ArmA so it could still be improved. I know the aim of the game is not to run people over so you could argue its a minor problem but then again there are times I've decided to simply ram foot soldiers who were close enough, saves ammo :).

Would be a treat if the Dev's could comment about ambitions and/or projects regarding the future of this series :)

That would be nice but at the same time it would be a bad idea to make too many comments because it could lead to a lot of disappointments when things get left out at the end as is always the case with any game. One thing I hate these days is how so many games are hyped up and people know so much about the game before its even released that there is nothing left to discover or be amazed at once you finally get to play it. Personally I often try to stay away from forums and fan sites in the lead up to many game to avoid all the hype, its probably one reason I dont end up with disappointments because I play without any pre conceived ideas about what the game should have other than the basics you expect in a series. Eg I'll be buying ACB and TDU2 once released but I know very little about either atm because I dont want it spoiled, and that extends to movies too. It's like knowing who won the football game before you finished watching it, its just not the same experience when you already know the result.

Edited October 11, 2010 by Engioc

[=wfl= sgt bilko 10]

I have no experience with VBS2. But from the vids I've seen it looks like it's more used for smaller group/platoon size scenarios (enter village, run into ambush etc).

Doesn't look like you run 100s of AI (Air, Armor, Infantry) or any multiscale pvp on maps the size of Chernarus in VBS2 scenarios.

In that perspective I don't understand the comparison between A2 and VBS2.

They seem to cater for two different audiences. The things we ask for in A2 might simply not be common in VBS2.

As said I have no experience with VBS2 so I might be wrong.

[noubernou 77]

I have no experience with VBS2. But from the vids I've seen it looks like it's more used for smaller group/platoon size scenarios (enter village, run into ambush etc).

Doesn't look like you run 100s of AI (Air, Armor, Infantry) or any multiscale pvp on maps the size of Chernarus in VBS2 scenarios.

In that perspective I don't understand the comparison between A2 and VBS2.

They seem to cater for two different audiences. The things we ask for in A2 might simply not be common in VBS2.

As said I have no experience with VBS2 so I might be wrong.

From my experience with VBS2, in a lot of ways its able to support a lot more players at once. Thats mainly because you can do some pretty intense things like offload AI to other machines (though how often this is done i do not know) and for the most part the scenarios are run on local networks so latency is not really an issue.

Also PvP runs smoother in either A2 or VBS2 simply because there is less AI that needs to be calculated. AI is the real killer in A2 in terms of performance.

[nodunit 397]

don't forget VBS2 generally uses lower res textures and none of the fancy shaders such as normal maps, specular maps, reflection and who knows what else. uh oh did I just necro

[KCIV 10]

I think putting arma on Cry would be amazing! I understand the whole "it wouldnt be large scale and we couldnt have 50000000000 ai doing stuff" but come on really.... who does that??? If you could take arma scale and put it on the cry engine it would be a total win all around. it would finally allow SO many things that arma's engine does not.(Good quality CQB, Texture to render-so many options there-, destructible environment, more realistic flight physics) no offense to anyone but seriously there are just some things that just need to added to the existing engine if its to stay around for another game. Its a great engine no lie there. Its also Old. And even the Best die and get replaced. there are just some things that need to either be added or put into a new engine.

speaking of VBS lol VBS has many features that arma does not. Features that shouldnt be left out in arma. you think they would at least take RV2 and make RV3 the same but better. instead of taking alot of features out......

Just my pennies on the subject

Edited November 9, 2010 by KCIV

[noubernou 77]

I think putting arma on Cry would be amazing! I understand the whole "it wouldnt be large scale and we couldnt have 50000000000 ai doing stuff" but come on really.... who does that??? If you could take arma scale and put it on the cry engine it would be a total win all around. it would finally allow SO many things that arma's engine does not.(Good quality CQB, Texture to render-so many options there-, destructible environment, more realistic flight physics) no offense to anyone but seriously there are just some things that just need to added to the existing engine if its to stay around for another game. Its a great engine no lie there. Its also Old. And even the Best die and get replaced. there are just some things that need to either be added or put into a new engine.

speaking of VBS lol VBS has many features that arma does not. Features that shouldnt be left out in arma. you think they would at least take RV2 and make RV3 the same but better. instead of taking alot of features out......

Just my pennies on the subject

Lots of people use lots of AI, especially in co-op missions.

As far as VBS2 goes I can also say there is a lot of things VBS2 doesn't have that ArmA2 does and I wish it did... :p

[the_antipop 10]

I believe ArmA 3 would work with CryEngine 3. Although, I don't know purchasing such an elaborate engine is in BIS's budget. If ArmA 3 was to use CryEngine 3, a 'real' publisher deal is almost confirmed. Not only is the engine proven itself in Crysis, but it's not 'buggy'. And with a box with ArmA 3 andm "powered by CryEngine 3", on it, its bound to sell.

About the size issue. i.e CryEngine 3 wouldn't be able to handle the size of say Takistan or Chern... I disagree. The size of the island in Crysis is quite large. And it is an island, the game isn't linear, you can go over to the other side of the map. But also, there is an extreme amount of detail. Detail that isn't exactly needed in a full scale war game such as ArmA, therefore reduces some lag which makes room for more square km's of land.

Anyway. I'd happily see CryEngine 3 be used with ArmA 3. Although, I'd rather see Real Virtuality Engine 4 be released with ArmA 3, but this time, no bugs. BIS's workforce has been expanded, they now need to use it to their advantage to get the most out of it, interms of engine wise and game wise. Therefore I believe RV4 and ArmA 3 would coincide, but also should a console release. But, nothing rushed. A non-buggy BIS game would be a dream come true. And even better graphics than ArmA 2 would be sweet. :)

My 2 cents.

[nodunit 397]

I think putting arma on Cry would be amazing! I understand the whole "it wouldnt be large scale and we couldnt have 50000000000 ai doing stuff" but come on really.... who does that??? If you could take arma scale and put it on the cry engine it would be a total win all around. it would finally allow SO many things that arma's engine does not.(Good quality CQB, Texture to render-so many options there-, destructible environment, more realistic flight physics) no offense to anyone but seriously there are just some things that just need to added to the existing engine if its to stay around for another game. Its a great engine no lie there. Its also Old. And even the Best die and get replaced. there are just some things that need to either be added or put into a new engine.

speaking of VBS lol VBS has many features that arma does not. Features that shouldnt be left out in arma. you think they would at least take RV2 and make RV3 the same but better. instead of taking alot of features out......

Just my pennies on the subject

By comparison helicopter combat in crysis is AMAZINGLY dull, the startup is..nonexistant, hardly any waiting time, very little reason to worry of crashing since the helicopter 'hovers' OFP style to the ground, you can't perform any acrobatic maneuvers or sharp banks, hell maneuverability is very simplified and 'on rail', you can't look around and the instrument panel is essentially dead, nothing moves. What is better about that?

I've never seen render to texture and I've never seen fixed wing outside of a community addon, flowgraphs would become the way to animate vehicles and such but it would apply to a map per basis not the vehicle itself.

Moving to Crysis would give us CQB, a generic fps view unless freelook can be implimented along with free aim. We would get vehicles but to what degree? Sure the tanks and cars look beautiful but what do they do other than say fire a cannon and machine gun. I honestly don't know why people get so wet over imagining Arma in a cryengine, it would give us physics and 'great' CQB sure but we would lose far far more than what it's worth.

Then you look at community, learning to use new tools, getting used to new tools, overcomming troubleshooting, writing tutorials and references there is far more at stake and at least 60% of assets you see on crymod never get released, what does that tell you? The engine is great for graphics and physics but for complex simulation, not the right way to go

Edited November 9, 2010 by NodUnit