Flight Modeling vs. Flight Handling.
 

I'm not intending to open a can of worms with this, but we are having an interesting discussion of this on our forum, and wondered what all of you think about it.

We were discussing RoF actually, with one of the guys that has time in about a dozen or so different bi-planes, including a full scale replica SE5a and a three quarter scale Nieuport 17.

The question was do the planes in the simulation behave or "feel" like the real thing, or are they just generic?

This was his response as a real pilot and sim flyer...

To put it in a context that I can relate to being someone involved in motorsport of the 4 wheeled variety...

You can have two cars with identical performance numbers across the board, top speed, braking, acceleration, lateral Gs, etc... Yet one will have to be bullied into doing it and the other will behave as if connected directly to your brain.

His assertion is that (or seems to me to be that) just knowing the numbers and having the virtual aircraft meet them is not enough, you need the actual pilot's input on how the aircraft behaves in your hands, how it feels, how easy or hard it is to fly and operate.

This is what, to me, is missing in most simulations that I have experience with.

Discuss.

The 6DOF 'Flight Models' can be very accurate wrt performance values..

Note I said can be.. ;)

But to capture the 'quarks' and or 'personality' and or 'characteristics' of a plane.. What you call 'Flight Handling' can be very subjective sometimes..

Because the 'Flight Handling' can be affected by so many variables, one key one being the INPUTS from the pilot. Read different pilot different characteristics

The first PC game I remember making an attempt at simulating the 'Handling' was EAW..

Where the P51 flight manual stated to recover from a stall, you had to put the ailerons into the spin and the rudder opposite of the spin..

There are other types of characteristics.. For example take the Fw190 and P39's tendency to flip over onto its back in an accelerated stall.. Where some 190s pilots actually started to use it as an evation manuver.. Another example would be how some planes will buffet or give the pilot some sort of clue before it stalls.. Like a vibration in the seat or stick.. Where as other planes gave no warning what so ever.. Or take the classic 109 slats that sometimes opened up un-evenly can caused the plane to jerk around.. The list is endless!

So a lot of those kind of things could be added to the 6DOF flight model.. But they would be a 'wrapper' (think mission trigger like) to the 6DOF not really part of it..

Maybe in the near future when the average PC is capable of running a Fluid Dynamics version of a 'Flight Model' than and maybe only than will you get both 'Flight Modeling' and 'Flight Handling' characteristics

The good news is most of these characteristics have no affect on the performance.. So even though you may be able to recover from a stall in a P51 without following the procedure in the P51 flight manual does not mean the you can not still have a very realistic experience and apply realistic tactics..

To put it another way, most of those 'characteristics' go unnoticed because most of the time your not at that point in the envelope

Interesting topic.

:grin:

The way I understand it:

http://forum.1cpublishing.eu/attachm...7&d=1330307513

There are different paths - an infinite number of them, actually - that connect two or more fixed points (in the graph, points A and B). These fixed points are the performance numbers; those different paths, the distinct handling experiences, or distinct 'feels'.

It would be fortunate if for a certain WWII aircraft modelled the correct 'path' can be picked out by a real-life pilot of it, but in most cases the 100% genuine aircraft does not exist now.

--

Interesting thread.

In a related aspect, I've been flying RED online Hurricane Rotols and Spitfire 1a's. They handle very well for the majority of my flying except for one area: fine aiming. It seemed like I was "chasing the reticle" when attempting to place the gun sight pipper on a specific location of the target aircraft. With a suggestion from fellow pilot Catseye, I adjusted the joystick curve (X and Y axis) in my Warthog's TARGET software to change the linear sensitivity to an S-curve. I still get full deflection at the extreme travel of the joystick, but towards the center of both axis I can now make much finer adjustments -- it makes me feel less "ham-fisted". By the same token, for my CH Pedals I simply turned the sensitivity of the pedals from default "1" (= full sensitivity) down to "0" (= less sensitivity) in the ingame menu Options/Controls/Axis.


A side benefit is that ordinary maneuvring feels noticeably smoother overall, plus I'm now able to pour what feels to be a higher % of rounds on target now.

My apologies if this falls outside this thread topic or has been discussed at length in an earlier thread.

I don't see what can be done about it short of actually flying RL versions of each modelled aircraft. And even then you are limited to a single instance which could differ markedly from other individuals. Pilot experience and style would also have a strong impact on what is essentially a subjective interpretation anyway.

In short, I don't see where this can go... Interesting topic though.

Fully agree with the gunsight handling comment! I am still struggling with this myself and tinkering with the response to try and get it 'just right'. Problem is that the sweet spot differs for each aircraft for me so I am finding myself changing it quite often. You certainly don't want ultra-responsive controls when flying the Blenheim for instance!

Interesting topic indeed!

I don't buy it though, or perhaps just not the car analogy.

If two cars have identical accelerations then they'll hit the same speed at the same time. Period. If the driver of one feels like he has to "bully" the car into doing it then that's his problem. Just a problem with his perception - because we already know the car is doing it just as well as the other car. If it didn't, then the accelerations would not be identical.

As far as this can relate to aircraft, well, I'm not sure what ElAurens is advocating exactly, but it does raise the issue of how to simulate what basically boils down to a pilot's opinion.

Being an engineer, my belief is that the nebulous qualities that we call "handling" or "feel" will show up when we consider things such as the aircraft's stability. For example the Spitfire was widely considered to be a forgiving aircraft but (test) pilots complained that the early versions "did not have enough elevator authority" and that it was difficult to control the aircraft, pitch-wise. When the issue was investigated it was found that the aircraft had a very narrow longitudinal stability margin.

I struggle to imagine a concept that can't be quantified, or related to some measurable quantity in some way.

Here we go http://www.cubpilot.com/Tspin/popcorn.gif.....

I think everyone that posted so far is correct in some way and touches upon all the main aspects of the topic.

I can certainly feel what ElAurens says, but that's through years of perception under controlled conditions (same control profiles, peripherals, software used, etc): in the previous IL2 series it didn't feel the same doing 500km/h in a 109 and doing the same in a 190, but it took quite some time to notice it.

In a similar fashion, when i first took up a 111 in CoD it felt much heavier and full of inertia than whatever bomber i had ever tried in IL2, but in a way that didn't feel uncanny. To the contrary, it felt just right and i had the feeling that the extra weight and size really showed.

I think the variables are too many and while it's true that i can't think of something that can't be quantified and measured (like Doggle says), maybe we don't have the technology to take advantage of it yet. Of course it's mostly fundamental newtonian mechanics at the speeds and masses we are talking about, but how many consecutive higher order derivatives of a certain function can a current PC compute per second and at how many instances and points across the aircraft's surface before melting? :-P

It's also difficult to quantify objectively because of the same reasons: few or no surviving aircraft, probably different in handling than when they came off the line after all these years (refurbishments, weight changes due to modern avionics installed per ATC rules, removal of guns, etc), a dwindling amount of veterans who all have their subjective opinion (depending on their flying habits) and a multitude of sim fliers with millions of combinations of different controllers and input curves.

It's a nightmare :grin:


I think that technology-wise Xplane is probably the closest one can get to having a wind tunnel emulator running on a PC (as long as the individual flyable is also done to a high standard to take advantage of the sim's engine in full), but sadly it will be some time before we see such technology in combat sims: it's so taxing that with full multi-core support in Xplane 10, a current PC can not run more than 4-6 AI aircraft with the same FM accuracy at the same time without noticeable performance loss. In fact, there are people who take Xplane flyables, simplify their FMs and reissue them as AI-only aircraft to populate the game world. :grin:

P.S. Very interesting topic ;)

I agree with this, but it comes back to perception/opinion. Maybe I feel the 109 "feels faster" and you have the opposite impression.

Is that even possible to quantify, let alone model?

My gut says no, because it's probably not an actual difference in the aircraft, just a perceived difference.

No two machines of any type are exactly the same, especially when they're of a more complex construction like a vehicle is, and especially after they've been subject to wear an tear.

Machines are built to operate within varying internal tolerances, and so while two vehicles may be superficially identical, there can be enough differences in all their variables to give each one a unique character or feel if one becomes familiar enough with them. Lubrication, hydraulics, different construction materials, different levels of wear and maintenance, and faults that aren't worth ironing out, can all vary and line up in such a way as to become noticeable in subtle or even obvious ways.

Simulations may never get down to that level of complexity and may always have to provide a single generalized representation of a machine, but who knows, one day all the variables might be able to be generated and result in genuinely individual virtual vehicles.

I think though that such a level of complexity is only part of what would add up to a more accurate 'feel' when flying in a game or sim, with most of it being dependent upon an equally accurate depiction of the physical interaction of the plane with the atmosphere around it.

That space the plane's fly around in isn't empty. It's a constantly changing set of influences that are mind-bogglingly complex and variable, and while a lot of them can be depicted, visually and even via force-feedback, I'm not sure how complex the depiction of the weather and atmosphere would have to be before it allowed one to get a real sense or feel of actual flight. Some would say it's already been done in some ways, and some would say it will never be possible without using a hardware system that somehow allows you to experience the changes in g-force and air pressure that can occur in flight.

With computer simulations we're very dependent upon audio and visual cues/clues, and somewhat dependent upon force-feedback effects to give us an idea of what the plane is doing and being subjected to from outside, and so, as complex as that can ever be, it's always going to be somewhat limited in what it can convey. But I'm of the opinion that even without the added sensory input that you can get in real-life, there's still scope for a more realistic and complex simulation of the machines and their environments than what we've experienced so far.

The sensation of flight is something that results from such a complex arrangement of factors that there are bound to be lots of new breakthroughs, all giving a more accurate experience than the last, as the simulation technology itself becomes more complex and able to recreate things in finer detail.

The sights, the sounds, the control inputs, the machine and environment systems, all need to keep developing and improving in unison, as indeed they have been over the years. But people need to keep supporting the flight-sim genre of games and simulations to keep those developments and improvements happening. I think for as long as people are using computers there will always be someone somewhere using them to simulate flight. I think it's as irrepressible an urge as the desire people have to fly for real, but if the rest of us want to enjoy the results of those works of passion, we need to make sure we give ourselves that opportunity by supporting those we know of who are currently pushing the technology further towards that goal of a more complete and accurate flight-sim experience, wherever they may be.

I saw this video my brother had shown me for Gran Turismo 5, they had this rig set up that plugged into a car's electronics and got data from its driving on the track. Then the programmers would pull the usb flash stick and upload the data into a computer that compiled the data.

My brother has driven the new BRZ / GT86 (FT86) test production car, and he also has GT 5 on a PS3 and he says the handling, acceleration, etc is similar. He is really excited about this new car and is amazed how GT 5 gets the GT86's characteristics down (rear wheel drive, drift / skid factor, braking / handling, sound, etc).

The only thing missing is the G and feeling of momentum and motion, wheel tension, pedal pressure etc the real thing has.


The technology to record the actual flight data / getting experience pilots to work with is available, as there are some examples of the warbirds. For planes that don't have flyable model, and there isn't a pilot that flew in them when they were available . . . they'd just have to get a working one as close and go from the records / pilot records of how it responded.

But the thing is GT 5 had a Modern Warfare esque budget, and 1C doesn't have that . . .

Then you have to have the fun factor / time, and that cuts out the realism.
Where as in real, missions can take a long time / black widow planes will have quirks that will crash newb and intermediate pilots . . . even experienced fliers unaware of the ship he is flying . . . guns will jam.

So you could have sudden things uber realistic but it just robs the fun factor and people don't play. Now I know the hardcore guys and gals will say "No make it as realistic." but there even then they will find certain aspects modeled realistically won't work in the game.

So there is a balance between realism and fun.

But IL-2 is the sim that gets close. And the devs show they are willing to work to get it as close as possible.

It wouldn't do any good to have someone with experience post their observations. There would be someone shouting them down because they're old and can't remember or the aircraft they're flying now is less weight that it was when used in combat, etc.

Nothing personal Doggles, but I suspect you have never driven on a race track (road course). Two vehicles of similar performance can behave very differently and one can indeed be easy to extract it's performance from while the other is a workout to drive.

In a sim there is none of this.

They are all easy. And if they are not (rare) then players tweak thier joysticks or put trim on a slider, or what have you, to make them easy.

No offense taken. You're somewhat correct, I've only been out on a track twice ever.

Similar performance, or identical performance? First of all that's different terminology than you were using before and secondly I'd agree that two similar vehicles can behave differently.

But two identical vehicles cannot behave differently. If they behaved differently then they are either not identical or are not receiving the same inputs.

I'm sure that driving sim players are tweaking their steering response curves too.

This would require the same input hardware too.
See? It's not going to happen.

For those that are convinced that two aircraft performing the same are behaving in the same manner, pls take a look at january issue of Air Force Monthly. There is an interview of an ex F-16 pilot flying now F-15C in Hawaï...

You'll change your mind in a single page of reading ;)

What we lack in CoD is inertia (mainly in rotation) and external perturbation of flows around our aircraft.

It's not just handling either. When flying a real aeroplane your senses are assaulted by noise, constantly changing g-forces, constantly changing visual stimuli over your whole visual field (as opposed to a screen in front of you), and so on. It's a far more intense experience compared with flying a PC sim, especially aerobatics. Even the weight of a helmet takes some getting used to. And the controls of anything from a light twin up can be physically far heavier than PC controls, requiring a bit of muscle to move.

Of course. But let's play it simple ;)

Well since I got my CH yoke, the Ju88 and especially the Heinkel feel totally different, much heavier and more 'real'. For me the difference between flying a 'heavy' with a light,quick moving Fighterstick and the slower moving, more physical actions of the yoke has made a surprising difference.
So, how much are our peripherals playing a part in the equation.

Exactly if you include these handling models you have to add pilot exhaustion, from what i have read you had to be quite strong to pull off tight turns in a 109...

This is why I stay far far away from this topic there are just far too many variables to deal with

Such as aircraft from one production facility being deemed poorer that others etc... HUGE can of worms.

For what it's worth, a long time ago I was taking flying lessons on Piper Cherokees. There were half a dozen of those and I flew all of them.

While the big numbers are very certainly similar, the feel was different. They each had their own personalities: sitting higher or lower, tightness in the controls, etc...

So there is indeed a lot of subjective variables here...

Doggles, two different cars can have identical (as close as possible in the real world) measured performance, but can handle and feel very differently.

I suspect it will take another 5 to 10 years of computer development to enable the kind of nuance I'm getting at here.

I flew a few different Pipers that were in the 140 stable during training time. I have to say that I did not notice a difference between them. I'm sure though that they existed, but I never noticed any...too young and nervous I guess:-P

I don't understand why X-Plane is so "hyped" - or have I missed to un-tick some default "arcade" setting?

Look at the clip below where I do my regular test of the FM in X-Plane (latest version 10). Some barrel rolls right after take off.

http://www.youtube.com/watch?v=ciJCTBW7zGc&hd=1

I'm sorry for being a bit evil in my comments ;)

Do this with a 88 in cod and you're a smokin crater. :D

This can be experienced now in the more advanced car-racing sims. With all the aero, tyre, suspension, gearing and other parameters you can adjust (sometimes with the assistance of telemetry systems like those used in real-life), two drivers can get the same performance (with lap times identical down to a fraction of a second) out of two differently set up cars of exactly the same make and model. Swap those two drivers cars around and with practice they might be able to adjust to each others set ups and still achieve the same lap times, but not without a struggle.

The thing with civilian flight sims is that a lot depends on the actual flyable and how it's modelled. They are very modular in nature and not all flyable aircraft use all of the code.

The same holds true for FSX too, i've seen aircraft that make CoD's CEM seem like a stroll in the park and i've seen others where keeping within the engine operating limits is just done for immersion and there's no consequence if you go over them.

This is pretty much the case with Xplane too, the core sim engine can calculate a bunch of stuff, but does it only if the flyable you are using requests that information. For example, i've downloaded the Xplane 10 demo and while their stock aircraft handle in a believable fashion (eg, crosswinds on light GA aircraft will lift your wing and throw you off course, needing aileron into the wind to keep staight, etc), some havn't included cowl flap operation and the CEM is simplified.

Maybe. But those cars will differ significantly in some quantifiable manner, which is all I'm saying. There's no magic.

It might not be the "headline" numbers like top speed or horsepower that are different, but some number, somewhere is causing the difference that the driver feels.

I agree with you that a lot of these numbers are probably not modelled in games these days, but I feel like that's a budgetary consideration, not a system-requirements consideration.

While I'm no engineer, I did work on the engine production line of Toyota for 4 years, the last 2 were on the engine testing bays.

No two engines were ever identical. In particular the power output and emissions they produced where all different. While they were within tolerances and differences were very marginal, there were differences non the less. Multiply this by adding other components to the cars (gearbox, drive train etc.) no two cars rolling off the production line were ever the same. You probably wouldn't notice the difference immediately but as the car was used over time, tolerances become wider as the car loosens up and quirks begin to appear.

Saying that two identical cars / aircraft would perform / feel the same is correct, but only on the drawing board. In real life it's very hard to get two mechanical devices to perform "exactly" the same. There will always be some variation in real world products, otherwise we wouldn't need to have warranties would we ;)

But I'm not saying two identical machines, I'm saying different machines with identical (or close enough for the real world) performance numbers.


I realize this is a somewhat cerebral discussion.

There have been posts in the past, going back all the way to the start of IL2, where there are those that say that if two aircraft have the same wing loading and power loading then they will perform the same.

This is a common thread among many sim pilots.

I'm saying this is not necessarily the case. One could be so difficult to fly that achieving it's max performance would be nearly impossible, whereas the other could have such beautifully harmonized controls that relatively inexperienced pilots could achieve the outside of the envelope with ease.

In the sim this difference would be so slight that it would make no difference.

It this making sense?

Since when did pilots fly on intellect alone...

I reasoned today that Pilot experiences where based upon:

Aircraft handling
Pilot Energy
Aircraft Energy

Thats it.

Argh, I should have quoted excerpts of the thread to direct my responses.

I was wanting to present my experiences to CaptianDoggles point about two identical machines performing the same (post #16).

As for your initial post El, I understand where you're coming from. From the many posts, videos and books I've read / viewed over the years I think it boils down to wear and tear. It's no secret that pilots found things "different" when jumping into a new plane from their old one.

I do recall during my Airforce Cadet days when taking a joy flight in a Macchi trainer at RAAF East Sale. The pilot I flew with had recently transferred from F-111's to be an instructor, and he was having a hard time adjusting to the Macchi because he was accustomed to having his own ride (F-111). At East Sale, he had use what ever jet that was available on the day. He mentioned that he found himself always checking the IAS gauge before going into an maneuver not just because of safety, but also because all the trainers "felt" different to him. He knew all the jets could all perform at the same specs, but to him they all had their own "feel".

Two identical machines (even down to the molecular level) can never perform truly identical unless they resided in their own foot print.

Basically, because of this:

Right.

And it's those differences that can account for differences in performance.

The point I was trying to make is that people should stop using the term "identical" when really they mean "similar". Because if they were identical, they wouldn't have differences.

Without splitting too many hairs..

Most of the attributes that set one plane (or car) apart from another can not be 'simulated' on a $1000+ desktop PC with $400 worth of joysticks, throttle, rudders

Why?

Because most of the attributes that set one plane (or car) apart from another is the way it moves..

And as we all know our lazy-boy ain't going nowhere when we fly our games.

A lot of people confuse flight simulation with flight simulators.. A modern PC flight simulation game can and in most cases does run a far more complex flight model than the air force F16 flight simulators of the 80s and 90s..

Mater of fact most military and commercial (read air liners) flight simulators care very little about how realistic the performance values are..

Why?

Well because most military and commercial flight simulators are more concerned with training the pilot on how to make use of all the systems on board..

That is to say most military and commercial flight simulators 'ASSume' the pilot already knows how to fly.

But there are things that can be done short of a million dollar motion platform to simulate motion..

How?

Easy, because the human senses are very Very VERY easy to trick!

For example, in a stationary jet simulator (F16 if I remember) simulated the sensation of the air breaks by simply attaching a small motor to the shoulder harness (seat belts for car drivers).. Thus when ever the pilot would apply the air breaks, these motors would simply tighten up (read pull back) on the harness, which to the pilot felt like he was being forced forward into the harness.

All the brain needs is the visual and the physical 'que' from the sensation of the harness and the brian will do the rest (read fill in the blanks)

That is just one example, there are tons of ways to trick the human senses into thinking the body is moving when in fact it is not moving at all.

I think ATAG Snapper is spot on, flight handing is all up to the local hardware settings, as such all controlled by us whiners, while flight modelling is all about the numbers put in there by Luthier and team.

Altering the axis curve can make wonders to the flying experience but despite that, I for one still can't hit a barn from the inside with my lovely Hurricane :(

All there is to it IMO.