What if CryEngine was used as Arma 3 future engine?

[carlostex 38]

Wow. Arma 2 gold edition freeware. People can be so gullible

[birtuma 28]

irony about the DX11 remark is that

we have Ambient Occlusions already in A2: OA on DX9 engine for quite some time ... :D

also Specular Lighting is possible in DX9 too (don't we have it too ;) )

And still no reflections in the game... For example in the mirrors.

[avengerzx 10]

What a huge disappointment on how CryEngine 3 runs on PC (I had Crysis 2)

Those textures were downscaled 1/2 to fulfill the memory needs for the console and Crytek says DirectX 11 will be coming in a month or so.

Whats more exciting is, there is no Sandbox 3 included (TOTAL DISAPPOINTMENT)

And still no reflections in the game... For example in the mirrors.

Regarding this, it is called Render To Texture (Rendering a reflection image to a texture) nothing to do with Specular Lighting.

Something like this Render to Texture in VBS2

And don't get to much comparison on the engines, it is a little bit different (as mentioned before many many many times)

[wolfbite 8]

I declare this thread perma zombie.... Its dead and has nothing new ever... But it keeps comming back

[birtuma 28]

Regarding this, it is called Render To Texture (Rendering a reflection image to a texture) nothing to do with Specular Lighting.

Something like this Render to Texture in VBS2

I never wrote that it has something to do with Specular Lighting. I just wanted to say that if "they" boast of some techniques they have without using DirectX 11 or 10, we should mention some negative issues which are almost standard things in games nowdays.

And don't get to much comparison on the engines, it is a little bit different (as mentioned before many many many times)

I didn't...

[carlostex 38]

i'm fully aware of availability of DMM v2 API via AMD sponsorship for free to Bullet engine users ...

I'm starting to lose hope. I actually love what AMD is doing. Supporting Open Physics is the right way.

All of the Bullet Physics implementations described above can be run on any OpenCL- or DirectCompute-capable platform. On AMD platforms, ATI Stream technology is used to drive the enhanced game experience. As a further enhancement, AMD has developed new parallel GPU accelerated implementations of Bullet Physics’ Smoothed Particle Hydrodynamics (SPH) Fluids and Soft Bodies/Cloth. The new code written in OpenCL and Direct Compute will be contributed as open source.

This is an opportunity. I don't understand why the Spanel bros say nothing about this. Arma games could sell potentially 10x more with proper physics implementation. And the best way to do it is OPEN SOURCE OpenCL.

Alot of people i know didn't like the game not because it was hard to learn or graphics or bad performance, but because of silly physics like ping-pong grenades, horrible collision detection, tanks going up in the sky, fake debris...

I believe that actually, with OpenCL physics performance would improve a lot.

Arma 2 is a prety much driven CPU game. If you have good shader power, a lot of that power goes to waste. A Radeon 6850, has 960 shader processors. 80 of those would be enough for accelerating OpenCL based physics on RV3. While this is achieved, your CPU is free of RV3 older physics and you have more processing power available for whatever the engine needs. You could be playing with more than ever units on the map, without massive lag.

Forget Havok, or other engines. If there is a way... I think this is it. If this is not it, then i'm sure nothing will. And Arma series interest will fade away eventually. I remember when Arma came out. I was massively disappointed to wait so many time for a OFP follow up with decent physics. And physics on Arma wouldn't serve eye candy purpose. But to allow better and more realistic simulation.

I believe Dwarden has a vision on this subject. Too bad Spanel brothers have no plans for this.

I'm actually curious to see how take on helicopters is gonna be like. Let's see what kind of breakthrough they achieve with it. Good luck. Bu for me, there's no love like the first.

[MissionCreep 12]

I'm actually curious to see how take on helicopters is gonna be like. Let's see what kind of breakthrough they achieve with it. Good luck. Bu for me, there's no love like the first.

It was an April Fools Joke ... but if you are still interested, a Zombie Apocalypse DLC for Take on Helicopters has been announced.

[froggyluv 2136]

It was an April Fools Joke ... but if you are still interested, a Zombie Apocalypse DLC for Take on Helicopters has been announced.

Wait what?!? That was no joke - as there is a full sub forum thread down below. The Pre LC Zombie thing was a joke but the game is real.

+1 CarlosTex tho I doubt we'll live to see the day.

Edited April 4, 2011 by froggyluv

[PuFu 4600]

I believe Dwarden has a vision on this subject. Too bad Spanel brothers have no plans for this..

How would you know that...? FPDR

[metalcraze 290]

Have you seen the recent advertisement video for CryEngine's simulation side?

Gameplay? No it didn't feature gameplay. It featured some dude, walking like a bro gangster from da block talking about how awesome this simulation is

It's laughable, they use engine to boast about their mocap or something instead of showing some gameplay, weapon modelling, environment scale and AI

Here it is:

http://www.gametech.ru/cgi-bin/show.pl?option=news&id=19678

I mean compare it to any VBS2 video

Edited April 4, 2011 by metalcraze

[froggyluv 2136]

I think the stuff by RealTime Immersive looks very interesting. Doubt they could match Arma2 in terms of scale but looks like they could make a pretty good tac-sim or even mil-sim of a smaller scale -with bigger physics :p

[WackyIraqi 2]

I think all of this Cryengine non-sense would go away if BI gave the physics system the major overhaul it needs. Thats about all it has over RV3.

[carlostex 38]

How would you know that...? FPDR

Bad choice of english... But this engine has now 10 years. Physics were never a priority, and now that there's a very good technology to give RV3 what has kept Arma series from being the hit OFP was. Newcomers see the silly physics events and immediately go away without exploring the other good stuff in the game. Collision detection, makes urban combat very disappointing.

I just don't think the Spanels want to implement a good physics engine on RV3. From OFP physics improved yeah, but not to the point that would help the simulation.

[Mr. Charles 22]

I declare this thread perma zombie.... Its dead and has nothing new ever... But it keeps comming back

QFT +1

:j: just let it die peeps

[Inkompetent 0]

Physics were never a priority, and now that there's a very good technology to give RV3 what has kept Arma series from being the hit OFP was. Newcomers see the silly physics events and immediately go away without exploring the other good stuff in the game.

You mean like the flying BMP-1s, and M113s bouncing like rubber balls on mine fields, and those awesome physics? :p

[metalcraze 290]

Newcomers see the silly physics events and immediately go away without exploring the other good stuff in the game

I don't see how this is a bad thing. If one dislikes the game just by a few physics quirks (and aren't physics nerfed down hard in Crysis MP?) one shouldn't be playing it.

And you aren't correct either. In fact these people do stay in AA2, only they complain about how it doesn't have physics and CryEngine has forgetting that in CryEngine everything spawns around the player and everything waits for the player to come. There's nothing like AI tank driving 2 kms away and ruining some wall to little bits which obviously needs to be calculated (and if it's MP - sent over the network) which will produce the opposite kind of complaints:

"why ArmA doesn't have 10 cartoony shaders per square pixel from Crysis and gives bad FPS on my system?!"

I mean surely I want to see cool physics stuff in games (one of the reasons why I prefer GeForce cards) but in ArmA2 it will be way too taxing if it isn't the cosmetic stuff like crashing glass that only you need to see but physics of vehicles, buildings crashing to little bits etc

Edited April 6, 2011 by metalcraze

[Big Dawg KS 6]

Collision detection, makes urban combat very disappointing.

I'm not sure why some people talk about the collision detection in RV3 like it's fundamentally flawed or needs to be "replaced". As far as I can tell, all the basic stuff you'd expect in any standard physics engine is there. The only issue I would say it has is that the collision detection isn't adequately optimized/prioritized (perhaps too infrequent update interval, or the update interval is not appropriately fixed). For what reasons nobody can say (except perhaps the devs), and while I could take a few guesses, I really doubt there'd be any benefit from using completely new middleware.

[ziiip 1]

I dont know why people are still debating about this absolute nonsense, but I suppose that many of the community addons could be used in Cryengine 3 this summer.

[pathetic_berserker 4]

I declare this thread perma zombie.... Its dead and has nothing new ever... But it keeps comming back

+ many

Topic is interesting from an addon makers point of view and lover of sandboxes but belongs in off topic/general and has run its course.

[flattermann 12]

it's the crybabyengine, never stops rollin'

[carlostex 38]

You mean like the flying BMP-1s, and M113s bouncing like rubber balls on mine fields, and those awesome physics? :p

Yes. Those are awesome (irony). That is why i advocate for physics that allow better simulation not eye candy fancy physics.

---------- Post added at 09:41 AM ---------- Previous post was at 09:34 AM ----------

I don't see how this is a bad thing. If one dislikes the game just by a few physics quirks (and aren't physics nerfed down hard in Crysis MP?) one shouldn't be playing it.

And you aren't correct either. In fact these people do stay in AA2, only they complain about how it doesn't have physics and CryEngine has forgetting that in CryEngine everything spawns around the player and everything waits for the player to come. There's nothing like AI tank driving 2 kms away and ruining some wall to little bits which obviously needs to be calculated (and if it's MP - sent over the network) which will produce the opposite kind of complaints:

"why ArmA doesn't have 10 cartoony shaders per square pixel from Crysis and gives bad FPS on my system?!"

I mean surely I want to see cool physics stuff in games (one of the reasons why I prefer GeForce cards) but in ArmA2 it will be way too taxing if it isn't the cosmetic stuff like crashing glass that only you need to see but physics of vehicles, buildings crashing to little bits etc

I never said that CryEngine was the way to go. In fact i think it is absolute nonsense.

Obvously you don't understand the implementation of physics i advocate. Performance with OpenCL GPU acceleration wouldn't be affected. As a matter of fact, it could increase. Your CPU would never have to do those calculations again. I'll say again: CPU's are great for general processing, but not for massive parallel workload. GPU shader units will deliver awesome performance on that.

About MP: I always mentioned that MP is the only problem with this because of client sync, unless we have 10Gb connections and much lower pings this will be almost impossible. That's why i said for the time being BIS could still use their older physics for MP.

---------- Post added at 09:46 AM ---------- Previous post was at 09:41 AM ----------

I'm not sure why some people talk about the collision detection in RV3 like it's fundamentally flawed or needs to be "replaced". As far as I can tell, all the basic stuff you'd expect in any standard physics engine is there. The only issue I would say it has is that the collision detection isn't adequately optimized/prioritized (perhaps too infrequent update interval, or the update interval is not appropriately fixed). For what reasons nobody can say (except perhaps the devs), and while I could take a few guesses, I really doubt there'd be any benefit from using completely new middleware.

Collision detection is not that good nobody can deny. When you're inside a bulding how much difficult is to fight inside of it without your gun being stuck in doors or turning your character around and sticking to walls like hell? Have you have climbed up in a tower and fell down to your death because of ghost railings? Collision detection is very very important for a game of this type.

[DeclaredEvol 10]

irony about the DX11 remark is that

we have Ambient Occlusions already in A2: OA on DX9 engine for quite some time ... :D

also Specular Lighting is possible in DX9 too (don't we have it too ;) )

http://msdn.microsoft.com/en-us/library/bb147399(v=vs.85).aspx

anyway i'm not saying DX11 is not better at all... it shines in completely different areas

also You dont know how epic's licensing work because each major v needs new license ...

so if you buy v2 you dont get v3 for free same goes for v3 to v4 etc.

also You get for free only engine updates in minor builds

if there is major jump e.g. 2.0 to 2.5

and you released prior the 2.5 release you will need to pay upgrade fee

ofcourse UDK licencing is different then UE itself ... as it's aimed on indie and small budgets ...

anyway why we discussing UnrealEngine or CryEgine etc. is beyond my understanding ...

Well, the API and SDK is already a masterpiece... and by the way things are going... Cryengine is also working on something with Realtime Immersive Industries, that is sort of dedicating its own engine towards some what your goal. It isn't ultimately as "simulative" but its pretty damn good for the ArmA 3 Battlefield, anyways... I really want to see a big change in this next game!!! People will continue to support your next games, I still say that a transition from just essentially realistic game play to essential both realistic game play and effects because it would grab more peoples attention... especially if the AI and Voice Acting was a little more up to earth. Most people want to feel more of a extreme sense of fear when playing, HiFi Sound System is definitely important in plucking some hairs off of peoples necks lol

Edited April 8, 2011 by DeclaredEvol

[Big Dawg KS 6]

Collision detection is not that good nobody can deny. When you're inside a bulding how much difficult is to fight inside of it without your gun being stuck in doors or turning your character around and sticking to walls like hell? Have you have climbed up in a tower and fell down to your death because of ghost railings? Collision detection is very very important for a game of this type.

Like I said, the collision detection is there and it does work, it's just not very precise at times. It's probably an optimization issue. Getting stuck indoors is more of an animation/controls issue than collision detection (I mean, the collision detection is actually working in those situations). But these are tweaks, not at all indications that a new or overhauled physics engine is needed. If you put the same models, animations, and control system into another engine, you'd still have the same issues.

[metalcraze 290]

Collision detection is not that good nobody can deny. When you're inside a bulding how much difficult is to fight inside of it without your gun being stuck in doors or turning your character around and sticking to walls like hell?

I dunno. Ask it to your 1m long rifle?

Why does every topic with complaints I read is about "baawww I don't want to learn to play, please dumb it down" and never, never about "make this feature more complex and realistic"?

If ArmA was just as dumbed down and primitive pop-corn as Crysis people wouldn't be playing it for years, dropping it after a few weeks like it happens to 95% of arcade shooters that come out.

Edited April 9, 2011 by metalcraze

[Pulverizer 1]

Guns sticking to walls is a silly feature in the name of realism because it's so unrealistic that your guy is rigid like a robot and has the rifle bolted onto his wooden arms. IRL, you can quickly turn around in a tight hallway with a rifle, without looking like a retard aiming to the sky or at your legs, or slowly putting your rifle to the rest position and back. So without another feature to complement it, it's broken.

If guns stick to walls, then there needs to be animations for letting the gun come back quickly when it bumps into a wall, like Infiltration mod or Red Orchestra 2.

Edited April 9, 2011 by Pulverizer