Validity and realism of flight models

[Conure]

Hi guys,

 

Just curious about educated opinions really. People often talk about the realism of a flight model but how do they really rank up?

 

For example, I suppose on a sliding scale with Hawx/Battlefield at 1 with DCS/PMDG at 10, you'd rank RoF an 8 and Flaming Cliffs 2 (excluding SU25t) possible 7? Where does this leave the IL2 stuff?

 

I guess basically I'm interested in how realistic the many sims we fly are considered, though I would presume DCS and the latest PMDG stuff is considered the best.

 

What are your opinions?

[Frostiken]

DCS's flight model was signed off as realistic enough by actual A-10 pilots. ROF and IL2, in fact just about all WW2 'combat sims', are really just based on wikipedia and how the designers feel about it. They certainly didn't hire SMEs for every one of the 3 dozen aircraft you get to pick from, and it's one of the reasons I find WW2 flight sims a joke, because every aircraft is usually indistinguishable from another with a tiny handful of minor exceptions. So given the volume of aircraft available to you and the lack of real-life SMEs for them all, ROF / IL2 can vary between 1 and 10 just on sheer chance.

Edited August 14, 2011 by Frostiken

[Nate--IRL--]

What are your opinions?

 

There is no definitive answer - they are all, as you say, opinions.

 

Nate

[Yskonyn]

I think this has pogressed over the years;

For example in Dcs A10 wehave a fully cickable 3d cockpit instead of the 2d one of Falcon where the 3d cockpit only served as 'SA tool' so to speak.

IL2 lacks both. It does not have a clickable 2d or 3d cockpit, instead you control everything with keyboard commands or mapped joystick buttons.

This is an aspect of realism that has made big strides forward with dcs (and a10 especially), eventhough at the time we found it realistic in IL2 and Falcon.

Thats how Id like to illustrate my point.

 

Progress in hardware has made more processing power available over th years. Enabling devs to create more demanding engines to run their sims on hence upping realism in avionics and flightmodelling as a concequence but at the same time we were all amazed by the Janes sims of old back in the day, but They dont hold a candle to dcs. :-)

 

Realism is perhaps subjective (to knowing hardware limitations for one) but as long as the sim gives a believable sense of flying around in an aircraft that seems to perform according to known specifications, speeds, rollrate and what not; i think we can conclude a sim is realistic. At least at the time it was built. Perhaps years later we kinda chuckle about how on earth we could have ever found a sim realistic! The old Ocean sims come to mind. ;)

Edited August 14, 2011 by Yskonyn

[Conure]

Interesting replies, thank you guys :) I often wondered about the validity of older sims without the direct maths to work off. How about FC2 then, are they considered closer to the arcade than realistic?

[Frostiken]

Yskonyn, I actually think most "sim" developers stick to WW2 or WW1 mostly because it's really, really easy to make a sim for these old aircraft. For one, you have almost total freedom to make up what you want since only a scant handful of aircraft from that era survive, and there's probably fewer pilots even still alive. And I've honestly never found the clickable cockpit (or rather, the lack thereof) much of a technology thing, but more of an 'ease of implementation' thing.

 

To be fair though, cockpits weren't exactly complicated - the avionics suite practically consisted only of a pneumatic altimeter, an ADI, two fuel probes and maybe a radio.

[EtherealN]

Well, to be fair, there are flying examples of the key fighters from the WW2 era - the "survivors" - and there's also quite a few replicas flying around. But generally speaking I agree that it's harder to get SME's. Not that good SME's don't exist - there's still active pilots on those aircraft, and several operators of them. And of course, there's a certain operator of such aircraft that works with a well known simulator developer... >.>

 

But yeah, even though the potential SME resources are there, since they're few and far between it's really hard for developers to make use of them, and in the case of games like IL2 the target audience isn't such that recruiting and using such SME's would be that useful. All down to market segmentation, really.

[Headspace]

DCS's flight model was signed off as realistic enough by actual A-10 pilots.

What's your source? "Realistic enough" is pretty vague. Realistic enough for what?

[Yskonyn]

Well regarding the flightmodels of ww1 and 2 a/c we should not forget the progress and hence changes we have seen in the way aircraft were developed then and how they are now.

The whole certification process has been pretty much a trial and error procedure which has resulted in the system that is in place now for current aircraft, but was nowhere near as complicated back then.

So naturally there is less data to work with for older aircraft being put into a sim, but at the same time we might need less data because the whole aircraft was less complex.

Besides, ww1 and 2 aircraft were built and developed in wartime in an era where flight was still very much in its infant years.

There might not have been knowledge about certain (aerodynamic) side effects then that we do take intoaccount now.

 

So yes, ww1 and 2 sims are perhaps less work to code and might remain vague in their flightmodels but that is not neccessarily because of lack of data per se.

Like EtherealN mentions; there are a lot of airworthy vintage aircraft and they are just as much a source of data then for example build plans.

[Headspace]

And of course, there's a certain operator of such aircraft that works with a well known simulator developer...

You mean like someone who would have a "collection" of these aircraft? A fighter collection of sorts?

[airea]

Condor soaring simulator has the most realistic flight model I have ever seen...

 

It is tested and actually played by real life glider pilots...

[EtherealN]

Ouf, it's good that you included the modifier of FM you've seen, since that's more observational than anything else...

 

I used condor as part of my training a couple years back when I was taking my glider certification, and it was nice in that it actually included a model of the plane I was going to use for my first single-seater certification. And sure, it was fairly close, but I wouldn't say it's the most realistic. The important thing to remember with Condor is that it's not really used to learn how to fly, it's used to learn principles and in part as a rehearsal tool - for example, before the spin exercise we would take Condor for a spin (pun intended) and thus be able to get a feel for the sequence of events in the upcoming exericse.

 

As an example, "real pilots" using a simulator isn't necessarily meaning that the flight model is spot on - the DTS that was developed for the USAF/ANG actually has a less capable and less realistic flight model than what you have with the DCS product! This is caused by the simple fact that the real pilots don't use the desktop simulator to learn how to fly, they use it to learn proceedure and potentially to do rehearsals ahead of an exercise or training lesson.

 

With Condor, the same thing - you use it to learn the relevant principles affecting glider flight, the procedures involved in thermal soaring, ridging, waving and so on, but you don't strictly speaking learn how to fly with it.

 

I'd personally say the Condor Soaring flight models is somewhere between the SFM FC2 planes and the AFM Su-25T. It's a bit hard to compare directly though since the flight envelopes are so different.

[Yskonyn]

^ Very clear explanation, there.

Makes me think about my 'sims' in flight academy; one was a wooden 'bathtub' with a replica cockpit of some imaginary IFR suited aircraft.

It was almost steam driven! ;) Still it, had a great flight model; you would stall when airspeed dropped below Vs, now can you believe that?

Seriously though, it was a perfect procedural trainer indeed.

 

Same goes for FNPT2 sims. They are static sims with visual modules pretty much resembling graphics you get in FSX. They model a certain aircraft and they have flightmodels approved of those modelled aircraft, but still you do not learn how to fly in an FNPT2.

You learn how to procedurally operate that aircraft. That's different.

 

Level-D sims are different, but they are approved for training students to fly on that specific aircraft. In these you actually do learn how to fly that particular aircraft. 'Seat of the pants feeling' of handling said aircraft is reasonably accurate (except for taxiing).

This 'feeling' is very important for learning how to fly. Flying is very much a 'seat of the pants' affair. Feeling when an aircraft is dipping below its glidepath even before you actually see it on your instruments, for example, is how you know you've learned to fly for one thing. This is something you will never get in a sim, not even in a Level-D sim, because in the end this is a computer. And if you keep the numbers aligned it will stay that way. In an actual aircraft we have mother nature contending as well. And even if you fly on the numbers you can still drift off if you dont correct it. Just a small example.

Edited August 15, 2011 by Yskonyn

[NineLine]

I think this has pogressed over the years;

For example in Dcs A10 wehave a fully cickable 3d cockpit instead of the 2d one of Falcon where the 3d cockpit only served as 'SA tool' so to speak.

IL2 lacks both.

 

Just FYI, the new IL2: Cliffs of Dover does have alot of clickable stuff in their cockpits, with a more advanced engine management system. I still wont say one way or another about how real it is as I am not a pilot, I only play one in my computer room, but least there is a more immersive start up procedure and a little more to flying the plane than other games before it.

 

Yes, there are many issues with IL2: COD as well, but I dont wanna mare this thread with that either ;) Just wanted to say, IL2 is progressing towards a more 3D clickable cockpit...

[Yskonyn]

^I was referring to IL-2 Forgotten Battles and its expansions. ;) But I could've been more clear about that.

[Yo-Yo]

The discussion like this are fed with the lack of information about real principles the different sims are based on.

There are several main types of physical modelling. Any movement of a rigid body can be decomposed to its CG movement along a trajectory (long-period movement or LPM) and rotating movement around the CG (short-period movemnt, SPM). The oldest and simpliest is the model where CG movement is calculated using standard equations. The model uses L/D polars and thrust tables. SPM is half-scripted, half-calculated using the pre-calculated solution of differential equation (exponent or damped sinewave, for example). This is the way SFM works. It can provide very low discrepancies with RL prototype, but SPM is really poor.

 

MSFS model uses different way to model SPM using the zoo of coefficients for differential equations describing SPM. Very hard to tune especially in post-stall area, but SPM is more realistic than SFM. MSFS also provides realistic CG movement effects on trim and balance.

 

Il-2.... it uses the almost same LPM model as LO SFM so we could suggest that trajectory parameters of the model can be relatively accurate but the SPM is realised in very strange way, so, I am afraid, it is very far from RL though the model is rigid body model. The most obvious fact is that CG movement is not kept in account in the model....

 

AFM. The basic principles of AFM are well known. Rigid body, any element the airframe fragmented to produces it's own forces, regardless of their sources (aerodynamics, contacts, impacts, etc). Aerodynamics forces of each elements are stored in databanks.

SPM and LPM are no artificially separated. Su-25 in FC/FC2 uses this model.

 

As ROF model is grown from the AFM I can suggest that it uses the same principles. As far as I know, DM is integrated into FM so the main differences are in elements interaction.

 

AFM-M proceeds AFM principles further. The airframe FM got minor enhancement and the most significant changes were made in engines model. True gas dynamic engine model provides realistic engine behaviour regarding ambient air parametres. There are some new features - hinge moment of controls (turn into MRFCS and take a look at the ailerons while running off-airstrip), true landing gear kinematics (try to retract it performing a 1.5g turn in Ka-50), true hydrolics, fuel and electric systems. Prop and articulated helo rotor true models as well as slipstream interaction with the airframe were added for Ka-50 and future helos.

Edited August 16, 2011 by Yo-Yo

[EtherealN]

One of these days you'll have to make a sheet containing the mathematical models. (Hm. Unless they are regarded a trade secret.) No need for code or anything, I just want to put it up on one of my screens as math-porn; and it always impresses the boss when there's something complex-looking on the screen. :D

 

Thanks for the illumination Yo-Yo. :)

[Headspace]

Sounds like FEM anyway, which is super well-documented and used for all sorts of stuff (and has been for years). IIRC X-Plane utilizes some version of that for its modeling. Rise of flight's developers say they use it in their marketing slicks.

 

You have discrete points. Moments are numerically calculated based on those points on the body based on a variety of factors.

 

Finite Element Method

Edited August 16, 2011 by Headspace

[Yo-Yo]

One of these days you'll have to make a sheet containing the mathematical models. (Hm. Unless they are regarded a trade secret.) No need for code or anything, I just want to put it up on one of my screens as math-porn; and it always impresses the boss when there's something complex-looking on the screen. :D

 

Thanks for the illumination Yo-Yo. :)

 

Some of them are already on our site. For SFM of course and briefly for AFM and further.

Edited August 16, 2011 by Yo-Yo

[Booger]

I guess basically I'm interested in how realistic the many sims we fly are considered, though I would presume DCS and the latest PMDG stuff is considered the best.

 

What are your opinions?

 

I would have to agree that it's incredibly subjective. For some, "more realistic" is simply having the maximum amount of switchology possible.

 

For others, it's something as simple as FFB.

 

I too have had a soaring rating and I simply can't imagine any current desktop simulation package coming close to real-life experience of actually being in the pit, as I've stated before. Procedures, indeed...but that's another topic.

 

Slightly on topic - I think we all have read posts (or have heard someone) state how slow it feels in _____ sim...how the developers have borked up ____ (aircraft). What people don't get is the developers probably have it spot on (or darn near it) but since you're staring forward at a flat screen, you don't "feel the speed".

 

Anyway, DCS (mainly BS) seems to be realistic enough...though I'm not an actual helicopter pilot. Having had some feedback from an actual Ka-50 pilot, he seemed to be quite happy with it. It has almost all the elements it can provide to be "realistic" while not losing the entertainment value it needs to maintain for the greater target audience. Not many would like to sit on the apron for long periods of time as systems align/spool up.

 

Again, it's subjective.

 

I'm glad you mentioned CG Yo-Yo. Last night I tried a flight with just 6 Vikhrs in my loadout. That turned into an unexpected chore upon entering a hover & at low speeds. I loved how much I hated it.

[Yskonyn]

Thanks for the input there, Yo-Yo. It made me look into things a bit further.

It still remains a nice discussion though. In the end I think the dream of simulator devs would be to create one that is as much true to real life as possible.

 

And indeed, I for one, have no idea about the technical mumbo-jumbo concerning coding stuff. As a real life pilot, however, I do know what would feel realistic, at least. And I do know what basics need to be involved into a mathimatical model for it to represent realism.

Center of gravity is one of them, indeed.

Aerodynamics seems like the most daunting task to program, if you ask me. It's so complex and even today the big brains learn new things about aerodynamic effects. I imagine that would be hard to code.

Still, with some level of abstraction we can come very close to reality, or at least presenting a simulator that feels realistic enough.

I think this underlines the subjectiveness in this discussion about 'what is realistic' even more.

 

Details like fluid viscosity, temperature effects, torque in airframe, electric current, etc all make the simulator even more realistic when looked at purely. Still, these things are 'under the hood' for many end-users and might not add to the feeling of realism for them, as would go for a lot of 'under the hood' stuff.

Edited August 16, 2011 by Yskonyn

[Yo-Yo]

Sounds like FEM anyway, which is super well-documented and used for all sorts of stuff (and has been for years). IIRC X-Plane utilizes some version of that for its modeling. Rise of flight's developers say they use it in their marketing slicks.

 

You have discrete points. Moments are numerically calculated based on those points on the body based on a variety of factors.

 

Finite Element Method

 

Nothing new under the Moon... of course yes.