is DCS a Ship Simulator and weapons systems Simulator only ?

[shagrat]

There was a very true comment on managing expectation. Because in the end there is a technical limit, to the "want this", "must have that" and "we need those".

A real life flight simulator for civilian pilot training cost more per day of operation, than the total of most of our combined PC, controls and periphery. It has a small Datacenter attached that powers the calculations and simulating the aircraft system, weather and effects on the aircraft. It does not calculate hundreds of units fighting, dozens of projectiles ballistically flying through the air, damage effects and AI for all the other planes and ground units... Despite easily 5-10 times the compute power of our average PC at home it does only(!) aircraft systems, Flight Model, weather and visuals.

DCS, a Combat flight simulator for a single home PC, handles aircraft systems (often quite deep, with avionics, electrical system, hydraulics, fuel etc.) in an impressive detail. A pretty good flight model for most modules, weapons, ballistics, damage to your own aircraft, other aircraft, ground units and environment, the AI and interaction of all of these AND on top some pretty good weather effects, including basic turbulence, wind, moisture, temperature... and pushes the limit of the hardware most of the time, already.

Of course it would be possible to improve the weather and atmospheric effects, but then we need to balance against something else. What could be removed? A combat flight simulator without combat? Or with a limit of 8 AI aircraft and 10 ground units? Or like in the early days, learn to live with single digit FPS?

The real challenge for a good PC combat flight simulation is to manage a good balance of all these aspects in a way that it still performs on most customer's hardware.

So the request to transform a combat flight simulator into a flight and weather simulator with thermals and up/downdrafts, I've only seen in a specific glider simulation, is maybe a case of excessive expectation on your side? Not that I think the weather/atmospheric effects simulation is perfect as it is. Of course not. There is a lot of room for improvement. But the current balance of all aspects relevant to DCS as a whole vs the compute power available is close to the best we can expect at the moment. And if your focus is on just flying and feeling the wind and weather like in a glider there are indeed other simulations that focus on these aspects and neglect the combat part.

Last but not least, despite a focus on realism and simulation of aircraft in incredible detail, an important part of DCS is the "game", be it PvE or PvP, scripted campaigns.

And at least for me DCS does a pretty good job at balancing all this into something I can enjoy at home, with a single PC.

[GOZR]

2 hours ago, shagrat said:

There was a very true comment on managing expectation. Because in the end there is a technical limit, to the "want this", "must have that" and "we need those".

A real life flight simulator for civilian pilot training cost more per day of operation, than the total of most of our combined PC, controls and periphery. It has a small Datacenter attached that powers the calculations and simulating the aircraft system, weather and effects on the aircraft. It does not calculate hundreds of units fighting, dozens of projectiles ballistically flying through the air, damage effects and AI for all the other planes and ground units... Despite easily 5-10 times the compute power of our average PC at home it does only(!) aircraft systems, Flight Model, weather and visuals.

DCS, a Combat flight simulator for a single home PC, handles aircraft systems (often quite deep, with avionics, electrical system, hydraulics, fuel etc.) in an impressive detail. A pretty good flight model for most modules, weapons, ballistics, damage to your own aircraft, other aircraft, ground units and environment, the AI and interaction of all of these AND on top some pretty good weather effects, including basic turbulence, wind, moisture, temperature... and pushes the limit of the hardware most of the time, already.

Of course it would be possible to improve the weather and atmospheric effects, but then we need to balance against something else. What could be removed? A combat flight simulator without combat? Or with a limit of 8 AI aircraft and 10 ground units? Or like in the early days, learn to live with single digit FPS?

The real challenge for a good PC combat flight simulation is to manage a good balance of all these aspects in a way that it still performs on most customer's hardware.

So the request to transform a combat flight simulator into a flight and weather simulator with thermals and up/downdrafts, I've only seen in a specific glider simulation, is maybe a case of excessive expectation on your side? Not that I think the weather/atmospheric effects simulation is perfect as it is. Of course not. There is a lot of room for improvement. But the current balance of all aspects relevant to DCS as a whole vs the compute power available is close to the best we can expect at the moment. And if your focus is on just flying and feeling the wind and weather like in a glider there are indeed other simulations that focus on these aspects and neglect the combat part.

Last but not least, despite a focus on realism and simulation of aircraft in incredible detail, an important part of DCS is the "game", be it PvE or PvP, scripted campaigns.

And at least for me DCS does a pretty good job at balancing all this into something I can enjoy at home, with a single PC.

Going to answer this one since I made the limits of the axis while doing the mod decade ago.. all this while cruising or basics flights and combats yes it's fine .. but while in competitions it is another storry and this is what disturb me a lot..  

[Baldrick33]

19 hours ago, trevoC said:

Sims that try to limit this movement are worse IMO as there are so many different types of headsets and heads. I can't stand when they try to limit the movement as what will happen is you move your head to the side and all of a sudden your vision stops moving (as your vr head has hit their limit) but you are still moving your head. This doesn't seem right when in VR IMO and is worse than allowing a little clipping.

This is very noticeable as your head might hit the limit long before your sight or eyes start clipping outside the cockpit.

That is my experience too. As a long time VR user, the first adoptions of VR in race sims lacked positional tracking and many found this both immersion breaking and motion sickness inducing. Once it was added the issues went away. It is even more of a necessity in flight sims where you move your head much more. Also in the early days of VR tracking was more prone to breaking, which again was horrid as you moved your head and the car or aircraft would move with you.

The alternative for clipping is to temporarily break positional tracking. Many find that worst possible solution, myself included. It has been discussed many times as a possible option but often the request is for it to be mandated as some perceived quest for realism.

[Dragon1-1]

18 hours ago, GOZR said:

 but while in competitions it is another storry and this is what disturb me a lot..  

I don't think VR is particularly competitive for other reasons. TIR is actually more efficient, from pure competitiveness standpoint, because you can turn your head at unrealistic speeds and easily reach extreme angles. The ability of the VR users to see outside the canopy is minor. I disabled that restriction in the other WWII sim, mostly because I felt I didn't have my head close enough to bump into the glass (plus, the kind of leather cap pilots there wear would compress a little in such situation, anyway). I just try to avoid sticking my head through the glass.

[GOZR]

Going to do a video.. 

 

I have been....

[GOZR]

https://fb.watch/hbGvXE0r0y/

 

I try my best to get into the immersion but I am really rememded each time i fly DCS how thecsim need to add this feeling ..

Not offensing anyone here but it's the worse of all simulations..   all are becoming great but that missing component which is one of the most important is missing.. and kills everything..  And i think if you really try tonunderstand you'll have the same thought.. now the ones who do not see that its because they live games and complex systems and feel ok with..  but for the ones who are looking for a flight combat simulation with real8sm flight physics at list effects DCS is not it... yet..

 

Try to recreate this like the video above with DCS.. you ll have much more chance to do it in the other sims..   I am going to be banned or bashed but what I say is the truth.. in DCS if you want some feeling you will have to push turbulence very high or to the max to have some feeling but it will feel like you are on a boat..  and this is why i wrote that  DCS was a Ship simulator.. maybe the first engine was.. i do not know.. it is time to require it ..  now that that we succeeded to ask for a multicore support.. ( .. i bet it will be only two cores )

We need to really demand this..  and yes it will be real good.. but not how it is now.. 

Edited December 4, 2022 by GOZR

[GOZR]

 

Edited December 4, 2022 by GOZR

[GOZR]

For now I am investing on the other good ww2 ww1 sim.. dcs store is in pause for me.  With the new up coming jet sim in the horizon DCS now working on a dual processor mode.. not impressed.. dcs need to react fast and make this possible

[Art-J]

^ Noone's going to bash or ban you, mate, but I think you're really overestimating the importance of weather simulation in this platform. People don't come to DCS for realistic weather, but primarily for combat and weapon / nav systems modelling. I'm willing to bet that majority of users here don't even turn any tubulence on in their missions, or even add any crosswind over their home air bases, 'cause that's not what they're here for. In DCS physical weather aspects have always been "nice to have" rather than "must have", most customers are OK with that and thus so are the developers. You want more advanced weather you go P3D/XP/MSFS, that's always how it's been and it's OK.

Ironically, I say that as a guy, who uses both DCS and Il-2 pretty much as a cheap alternatives for MSFS, mostly for non-combat flying around and navigation. I know I'm in minority, though, so I don't bicker about things this platform is just not meant for. IF ED manages to improve this aspect after multithreading is implemented somewhere in 2024 (yes, I know what newsletter said, but I also know how mumbo-jumbo estimates work here ) then great. If not, so be it. 

I wouldn't say Il-2 is THAT much better in this regard anyway. Both sims have the same, simple wind direction and turbulence settings which throw the planes around a bit and that's that. DCS adds some extras like different wind speeds at altitudes plus dynamic weather (oh wait, I've heard it got somewhat broken in 2.8), but that's more for a navigation challenge - a different aspect altogether.

Edited December 4, 2022 by Art-J

[BitMaster]

...don't I have this 256-core CPU somewhere in a box, next to the 12TB/2PB RAM-Disk combo thing Steve gave me, maybe just good enough for weather in a 30km circle and 10km high. 

naahh..not there, it's got to be in that box with the JFK evidence stuff...I will try it out after feeding the unicorns

stay tuned.

 

Not saying it wasn't nice to have it ALL, who wouldn't, but you have to measure your wishes and expectations against real hard core numbers or end up feeding unicorns.

Once DCS has really become multi-core competent and general CPU's have dozens of cores and even more RAM there "might" be enough juice left to think about "weather".

 

Look at a tree with all his leaves in the wind, stop the frame !, manipulate ONE leave...and calculate the change 12h later.....  forget it !

It's an extreme example but still, one tree, one leave, half a day later.

*From my Physics teacher in the 80's, just to show how complex thermodynamics are.

 

[Ala13_ManOWar]

3 hours ago, GOZR said:

For now I am investing on the other good ww2 ww1 sim..

You mean that awfully sucking arcade?? You sure you're a pilot? 

[Hawkeye60]

If you think VR is fun, Get out  of the cockpit while auto pilot is on, have a chair nearby, sit on the wing edge, and look down!

(Have barf bag handy!)

 

For real fun, set a mission with a few bombers attacking a airfield.

Spawn on the airfield, in VR, get out, and wait.....

Watch bombs fall!

Kiss butt goodby!

[GOZR]

11 hours ago, Ala13_ManOWar said:

You mean that awfully sucking arcade?? You sure you're a pilot? 

Arcade in what explain .. again this is not a personal attack but a way of awareness to make a change .. like now the Multicores

6 hours ago, Hawkeye60 said:

If you think VR is fun, Get out  of the cockpit while auto pilot is on, have a chair nearby, sit on the wing edge, and look down!

(Have barf bag handy!)

 

For real fun, set a mission with a few bombers attacking a airfield.

Spawn on the airfield, in VR, get out, and wait.....

Watch bombs fall!

Kiss butt goodby!

LOL  yes !!

What I do sometime is go into a busy server get into a plane and get out or seat in cockpit and watch teh aircraft take off landing or display .. that looks quite fantastic and teh sound as well.. well done.. and fun..  Gives me flash backs lol

[Ala13_ManOWar]

4 hours ago, GOZR said:

Arcade in what explain

Sorry, just a joke, we aren't supposed to discuss any other products in here, but I thought you already knew by now… just, in everything…

[GOZR]

5 hours ago, Ala13_ManOWar said:

Sorry, just a joke, we aren't supposed to discuss any other products in here, but I thought you already knew by now… just, in everything…

LoL yes. Interesting to compare with other sims with out saying names..  but DCS is great far better than others regarding the systems implementation .. and this is not in question.. but it is the most arcade alike of all of them in the atmospheric effects onto the the plane and that should be better.. 

Now i am going to retest one ww2 sim with a North African name that actually has both.. atmospheric and systems within.   This is a good topic

[Dragon1-1]

The only thing that one currently lacks is VR support. I'm waiting on that to check it out, the devs are supposedly working on this.

EDIT: It looks like it's in closed beta since early this year, so it should come out sooner or later.

Edited December 5, 2022 by Dragon1-1

[Ala13_ManOWar]

3 minutes ago, GOZR said:

LoL yes. Interesting to compare with other sims with out saying names..  but DCS is great far better than others regarding the systems implementation .. and this is not in question.. but it is the most arcade alike of all of them in the atmospheric effects onto the the plane and that should be better.. 

Now i am going to retest one ww2 sim with a North African name that actually has both.. atmospheric and systems within.   This is a good topic

Yep, interesting how different people see also quite different things. DCS is by no means perfect, that's obvious, but from my standpoint atmospheric implementation is at least quite good, far better than anything previously seen in simulation (at least some years ago for sure). Wind is horribly modelled in that other arcade game for my taste, and compared to the real life I can know. I get sometimes people might like funny things, half my squad flies it indeed and so did I a bunch of years ago (since there were no alternative ), but I couldn't stand any more that lousy modelling after I got my license around 2011 and P-51D (released in 2012) was just perfect at that time with those PCs available and anything alike I had used before, so we made a separated squad just for DCS together with another pilot mate inside my squad. Yes, now it's been a long way and things have happened and everything, but I still believe DCS implementation is the best in the available software, even better than newer titles in many areas. Since we know they're working in so many aspects of the core sim, and those atmospheric effects you mean are also a part of the work, I'm fine enjoying it as it is now while I wait for those changes. It's no problem for me, I'm just not that impatient despite some sort of eternal deadlines at times, but I also know the quality they strive for and I just look forward to it.

[Aquorys]

On 12/4/2022 at 6:42 PM, BitMaster said:

Once DCS has really become multi-core competent and general CPU's have dozens of cores and even more RAM there "might" be enough juice left to think about "weather".

I think people greatly overestimate the performance gains that the use of multiple cores will likely produce. For things that need to be done every frame, you get about 16 milliseconds to get everything done for each frame if you're doing 60 fps. Multithreading is most efficient when you have highly independent, long-running tasks. Games are usually the opposite of that: very interdependent, extremely short-running tasks.

Now you need to delegate to multiple threads from the main thread, then it takes time until all those threads start running, modifying data that multiple threads need to access now requires acquiring and releasing synchronization objects, also implying that the CPUs would often have to write through their fast local CPU cache to slower caches or the even slower RAM so that the data is kept in sync for all CPUs, and then you need to wait for all the threads to complete, and you don't know when that will be, and the chance that one of the threads will not complete in time is probably higher too. Add in that none of the general purpose operating system kernels - be it NT (Windows), XNU (macOS), Linux or any of the various BSDs - have hard-realtime capable schedulers, and you will slowly begin to see the almost infinite multitude of problems and performance-adverse conditions that you will typically run into whenever you try to make use of multiple processors under such conditions.

[Dragon1-1]

1 hour ago, Aquorys said:

I think people greatly overestimate the performance gains that the use of multiple cores will likely produce.

With the approach ED adopted, I agree, but that's what you get with only two threads. That said, done properly, multicore could be the gamechanger it's made out to be. Most things don't need to run per frame, and there are many largely independent tasks (for example, each aircraft's FM is basically an isolated physical system, as is each AI "brain") that can be split over multiple cores. Flight sims do have a lot of room for scalability, particularly mission-based ones, where mission scripting and AI can sometimes hog the CPU, in addition to physics and graphics.

[Aquorys]

Things like strategic or tactical AI decisions could possibly run asynchronously, mission scripting maybe as well to some degree, physics not so much. You need the physics pretty much on each frame to get somewhat accurate and especially repeatable/predictable behavior, which in turn is important for the exact position of all objects and as a result, weapon guidance, collision detection, etc.

 

[Mars Exulte]

21 hours ago, Aquorys said:

I think people greatly overestimate the performance gains that the use of multiple cores will likely produce.

They do. It will help with consistency so highs and lows of FPS are closer together. On low end missions, it will make little to no difference, on heavily populated or script intensive missions it will help stabilise FPS. It will not magically double your framerate.

Similar with Vulkan, although as it's GPU related and involves streamlining the flow to/from, it likely will yield SOME improvement, especially in combination with the CPU overhaul, but will also help bring the lows and highs closer together and make them more stable.

It's going to improve QoL with some performance increases. It's not going to make a potato VR capable, nor is it going to resolve the fact VR is itself a demanding (and still newborn) technology and somewhat of a luxury item, especially for software that was not designed with it in mind. People keep ignoring that last part in their concern threads about ED's work ethics and competency, VR hasn't been around long, and how easy it was to attach it to pre-existing software has varied widely.

[schmiefel]

I think some run into a great mis-understanding what a - at least fixed base - simulator can simulate at all. Plus never mix what cameras record in a moving and shaking environment with what a human beeing 'feels' beeing in the same place: the human system - eyes, vestibular system and seat of one's pants - compensates for a lot of different aspects and mixes them in the right way so that one at least trained or used to person would normally rarely really see that shaking but just feel it mostly.

That is something that can hardly if at all be simulated and there is no chance to simulate it right for the most use cases without a proper motion rig and then, if you don't get it right, even experienced pilots would get motion sick very easily because it doesn't fit to the real experience. In rather every clip I have seen, where real pilots 'tested' and experienced DCS they told that the aircraft systems are simulated very properly, but it lacks of course the feeling of the real thing because there are no G-forces, no moving parts etc. So even if the simulation would be able to push all those environmental data and physics through the PCs processing power  onto the screen you wouldn't think its right, because your are not connected to the simulation like in the real thing. So besides the technical restrictions modern hardware and its limits provide, it makes only sense to a certain degree to simulate that all like nature works and even sometimes you have to do something 'wrong' to trick the human experience.

Just my 2 cents. Pushing the boundaries is always a good thing but you have to know the limits that make sense.

[Dragon1-1]

2 hours ago, Mars Exulte said:

Similar with Vulkan, although as it's GPU related and involves streamlining the flow to/from, it likely will yield SOME improvement, especially in combination with the CPU overhaul, but will also help bring the lows and highs closer together and make them more stable.

In context of VR, it should greatly help overall performance, given that it only renders one scene from two different angles, while DX11 renders the whole thing twice. If implemented well, it should bring DCS closer to being inline with other high end VR titles, and that's what everyone is asking for, really.

It won't run on a potato, but that's not what we want. We want it to run smoothly on something less overpriced than a 4090. 4K VR came out way back when the 1080ti was the king, and so that GPU (mind you, it's still on the beefy side, even if it's not on top anymore), paired with a good CPU, should be able to drive a G2 at 100% resolution and mid-high eyecandy settings. I believe that would do it for most people, and it's certainly within the realm of possibility. You can already come close if you optimize the heck out of your setup, but we're not quite there yet. 

3 hours ago, Aquorys said:

Things like strategic or tactical AI decisions could possibly run asynchronously, mission scripting maybe as well to some degree, physics not so much. You need the physics pretty much on each frame to get somewhat accurate and especially repeatable/predictable behavior, which in turn is important for the exact position of all objects and as a result, weapon guidance, collision detection, etc.

That still doesn't mean all physics needs to run serialized on a single thread. There are ways to parallelize physics calculations, PhysX has no problems being multithreaded, for example, and in fact takes advantage of the fact GPUs are optimized for doing many things in parallel. In fact, this particular problem should be well explored because physics is computationally expensive, so spreading that load is a big win.

[GOZR]

Dragon1. Agreed

Edited December 7, 2022 by GOZR

[GOZR]

10 hours ago, schmiefel said:

I think some run into a great mis-understanding what a - at least fixed base - simulator can simulate at all. Plus never mix what cameras record in a moving and shaking environment with what a human beeing 'feels' beeing in the same place: the human system - eyes, vestibular system and seat of one's pants - compensates for a lot of different aspects and mixes them in the right way so that one at least trained or used to person would normally rarely really see that shaking but just feel it mostly.

That is something that can hardly if at all be simulated and there is no chance to simulate it right for the most use cases without a proper motion rig and then, if you don't get it right, even experienced pilots would get motion sick very easily because it doesn't fit to the real experience. In rather every clip I have seen, where real pilots 'tested' and experienced DCS they told that the aircraft systems are simulated very properly, but it lacks of course the feeling of the real thing because there are no G-forces, no moving parts etc. So even if the simulation would be able to push all those environmental data and physics through the PCs processing power  onto the screen you wouldn't think its right, because your are not connected to the simulation like in the real thing. So besides the technical restrictions modern hardware and its limits provide, it makes only sense to a certain degree to simulate that all like nature works and even sometimes you have to do something 'wrong' to trick the human experience.

Just my 2 cents. Pushing the boundaries is always a good thing but you have to know the limits that make sense.

I understand and yes.. " aircraft systems are simulated very properly,"  yes but this is not the problem..

Take the MiG21 or F14  pull the stick on a heavy turn .. you'll see air friction buffets  retalted to a visual effect corresponding to a turn etc. .. well this can be applied to turbulence as well but they are not at the moment.