Jump to content

Flight models in MSFX.. Just how good are they?


wehyam

Recommended Posts

I've just been on the X-plane Forum.. :?

And frequently a topic boils up agonising about the accuracy or otherwise of the flight models. Much is made of the fact that in X-plane the plane is 'modelled' as collection of surfaces with aerodynamic properties. lift drag etc.. and the actual flight characteristics are computed 'real time' from the 'theory' ... you 'build' your plane with a wing in the wrong place and the plane flies like its got a wing in the wrong place! you want to add another wing and it will compute the lift drag etcetteras and ..etc... this is ideal for a community that likes building planes...

Everynow and again.. MSFSX or 2004 comes up as the straw man..

:wink: They use look up tables .. they say.. :wink:

What I'd like to know is how are the flight models of MSFX planes 'modelled'? how are the differences (if any), between the Aerosoft Beaver with floats and the Aerosoft Beaver with wheels, actually modelled?

Suppose someone makes a plane to fly in MSFX.. is there a fixed format. Fill in the values of the following parameters weight wingspan engine HP ... etc etc and clunk... a piece of code falls out of the slot which flies a bit like a Beaver.. but which actually is completely independent of the actual shape of the animal you've skinned ... :lol:

Link to comment
Share on other sites

  • Aerosoft
I've just been on the X-plane Forum.. :?

And frequently a topic boils up agonising about the accuracy or otherwise of the flight models. Much is made of the fact that in X-plane the plane is 'modelled' as collection of surfaces with aerodynamic properties. lift drag etc.. and the actual flight characteristics are computed 'real time' from the 'theory' ... you 'build' your plane with a wing in the wrong place and the plane flies like its got a wing in the wrong place! you want to add another wing and it will compute the lift drag etcetteras and ..etc... this is ideal for a community that likes building planes...

Everynow and again.. MSFSX or 2004 comes up as the straw man..

:wink: They use look up tables .. they say.. :wink:

What I'd like to know is how are the flight models of MSFX planes 'modelled'? how are the differences (if any), between the Aerosoft Beaver with floats and the Aerosoft Beaver with wheels, actually modelled?

Suppose someone makes a plane to fly in MSFX.. is there a fixed format. Fill in the values of the following parameters weight wingspan engine HP ... etc etc and clunk... a piece of code falls out of the slot which flies a bit like a Beaver.. but which actually is completely independent of the actual shape of the animal you've skinned ... :lol:

Xplane uses a different system and it gives them far greater flexibility. But the proof of the pudding is in the eating and most flight models of FS can stand up to most of Xplane.

Not sure how it is done in Xplane but in FS you indeed give all the right parameters on all surfaces, engine parameters etc and you end up with a flight mode that should be close to the real thing. It will need tweaking to get the exact numbers and often a few changes to get the right 'feel' into it. It's not to hard to get the right performance numbers if you know what you are doing.

In the Beaver the difference the floats and the wheels are numerous, for starters the floats have more (and different kind of) drag, the weight is different and differently distributed (further from the CoG so pitch movement are dampened. All this can and is modeled. It's pretty accurate, when we have real pilots fly the model they will know for sure what the undercarriage is without looking.

Personally I would say that FSx is better inside the flight envelope of 'normal' aircraft. When you get to the ragged edge of the flight envelope or beyond (for example flying inverted in a 747) Xplane probably is better. For me that is not very important.

Link to comment
Share on other sites

If my earlier response appears .. sorry.. funger errirs...

Thanks for the answer Mathjis.. .. I suppose the X-plane advantage is that if you dont like something you can 'relatively easily' modify it.. and of course you can specify what airfoil you want (which is an aswer to the sort of question I dont ask!!)

BUT ... However.. this reminds me ... Mathjis are you (or Aerosoft) going to 'upgrade' the Supercub to SFX.. I remember you and David -Wi had something in the pot but it seems to have gone cold..

are you intending to do anything further or has it died??

regards ..

Link to comment
Share on other sites

  • Aerosoft
If my earlier response appears .. sorry.. funger errirs...

Thanks for the answer Mathjis.. .. I suppose the X-plane advantage is that if you dont like something you can 'relatively easily' modify it.. and of course you can specify what airfoil you want (which is an aswer to the sort of question I dont ask!!)

BUT ... However.. this reminds me ... Mathjis are you (or Aerosoft) going to 'upgrade' the Supercub to SFX.. I remember you and David -Wi had something in the pot but it seems to have gone cold..

are you intending to do anything further or has it died??

regards ..

FSX also requires you to specify the exact airfoils shape, so not so much difference there. Saying the flight models of FS are only table based is a myth.

Not sure about the Cub. It's on the list but I need to know for sure that it will be worthwhile. It would probably mean doing a whole new 3d model, expensive stuff.

Link to comment
Share on other sites

As with FS9, FSX flight modelling seems to get better the higher the maintained fps value. Which is a problem in the later sim whcih struggles to maintain double-digit frame rates in some situations. One might make an interesting comparison with new hardware bought for FSX to actually go back and fly aircraft in FS9. The additional frames per second in FS9 transforms the fluidity and effectiveness of even the mundane flight models in FS9. Try it with a model you think you know well, and prepare to be greatly surprised...

One could also make an interesting case for arguing that the flight models in FSX are actually LESS sophisticated than in FS9. And where is the SDK?

Sadly, MS bought in the original .air file structure and has never really showed that they know what to do with it. Them they simplified it without understanding what needed simplifying and what simply needed explaining.

And sorry, but the idea of crunching the numbers and achieving flight model perfection is simply a misnomer. Compare a Beckwith 1% flight model with a Metzger, tuned-for-FS-amd-bugger-the-numbers flight model and `see` the difference. The fact is that some of the best ever flight models in the FS world do NOT obey the Book values, and fly all the better for it. It remains an art, not a science, although understanding the science only helps the art.

Link to comment
Share on other sites

Sorry Snave,

Metzger I got, Beckwith..?

I dont know wherewhat 1% ?

I think I'm half way agreeing with your disparagement of flight model purity per se.. 8)

I suppose in Sim-world a key thing is whether the plane is sufficiently (in principle) controllable that it is worth the investment (time and $) to learn the handling foibles and the systems. I suppose the heavier the metal the relatively more important the systems become where as for light GA stuff the key 'enjoyment' is whether the plane can be controlled and its worth learning how to fly it so you can put your Super Cub down on a 200ft sand bar (and get it off again)!

IMO (though I know nothing) The easy bit is to build the systems...am I right?

I was most surprised when I first flew a real plane (a C152) how much easier and more controllable it was compared with similar sim planes (which seemed to have little no inertia and drift and bob arround..) There is of course a confounding here with the 'controls'.. I now live off a MS-SidewinderFFB2 because I dont think the cheap (!) yokes work well enough.

I think all would agree that different planes should be different and have different feels and they should somehow behave 'right' at the user end..

The purist ideal would have 'perfect' physical modelling.. everything right.. like PHYSICS.

I think what you 'suggest' is that.. who cares what laws are broken in the implemenation as long as the plane flies right..

I also think you said that if you work with MS-stuff you have really only got that option? :?

I think X-plane would claim that their way is the way that, in principle (they may not be there yet), flight sims will become increasingly more realistic.. without having to rely on special case twiddling...

The argument may have more force in the 'design a bomber' application BUT the entire specs of the super cub are available and so the claim is that if you build it to spec (and the PHYSICS are right) you get a super cub.

Please Mathjis make me one...

Just as a matter of interest... I moved the rudder on a C150 into/onto the luggage compartment.. (you can do that in X-plane very easily).. It flies quite nicely..

NOW an advantage of X-plane is that I will be able to check out how this affects the flying characteristics.. if at all, because X-plane provides the measurements and timeseries data flight black box... everything..

This isn't meant to be a Nahnahnah comparison of the two.....

BUT it would be really interesting to see (how long it would take?) the

Aerosoft Beaver specs and twiddles put into X-plane and see how it flies.. (I'm not talking about the 'cockpits and instruments.. paint job and rivets..) just the plane.

enough already :D

Link to comment
Share on other sites

You would still need a supercomputer to simulate the Laws of Physics as they relate to flight model efficacy and accuracy.

What you get in any FS product is a simplification of the multifarious and vastly complex inter-relationships inherent in cause/effect simulations and the interactivity with a dynamic environment.

So yes, I am saying that `feeling right` is more important than being right. Jerry Beckwith helped create the `1%` flight model concept which defined the accuracy of the flight model by the data that is used to create it - the suggestion is that, as long as the (for example) Cn_r(alpha) Yaw Damping Coeff vs Body AoA is accurate the aircraft is accurate. This a misnomer, as one may not imply from proof of `A` that `B` is also correct without knowing all the parameters that impact on the accuracy of `A`, as these may be subject to alteration later on in the file code... This was the lesson the late, great Ron Freimuth taught us. So you can plug in as many accurate `numbers` as you like, but faced with an incomplete inderstanding of the correlation of that data to, say, the environment flight model in the sim, you cannot actually claim 1%, 10% or even 0.1%. A perfect example is the `stickiness` of the water friction in FSX. It still - and quite literally - sucks!

Is the `air` in Flight Sim heavy, or light?

But while the accuracy may not be truly accurate, you may certainly claim that you are on the path to true flight model fidelity when the majority of the numbers ring true. And it is at this point the `science` departs, and the `art` takes over. And this is where the artists like Metzger and Rob Young come in. They do NOT rely on accurate data, but instead can manipulate that data to give a reasonable, or even fantastic, approximation of the real thing, within the limits of the flight simulation product itself.

The problem is, and remains that system fidelity actually requires more computing power than the current desktops possess:

http://fltsim.group.shef.ac.uk/

Not only that, there is still an imprecise and incomplete knowledge of the various components of the .air file structure, and in the absence of the SDK, an incomplete understanding of the complex inter-relationships between the thousands of lines of code. Huge advances have been made in decompiling and understanding the file structure of the .air file, but to this day it remains an inexact science. From this, the logician concludes that the 1% flight model is doomed to fail to achieve the +/- 1% accuracy that is its aim, and title...

Read more here:

http://avhistory.org/communityserver/default.aspx

That is not to claim the 1% flight model is a failure, it most certainly improves drastically on the somewhat ham-fisted attempts in default flight models, and can still come very close to getting things nearly right in many simulated situations. But as anyone who has flown a real aircraft will confirm, real airplanes have a sense of mass, a sense of fluid response to their environment and, most importantly, a feedback loop that is completely absent in Flight Sim. Which is why real flying is easier than sim flying!

Then on top of all that is the difference in the individuals controllers, systems and practices and procedures. Not to mention their implied, but most often absent, knowledge of the aircraft in question. I recall a subject in a forum about the `excessive` climb performance of a very sophisticated airliner addon. Lots of snorting about 12,000ft/min climb rates being totally unrealistic! As it happened, my Brother in Law was a senior captain and check pilot on this aircraft, and was able to inform me that the aircraft in question, far from being `Boring`, was actually able to outclimb most sixties generation fighters when not hamstrung by huge fuel loads, self-loading cargo and full drinks trolleys, and that climb rates of 16,000ft/min were actually readily obtainable, if one wanted to fly a twin-aisle like a Tomcat! And this man flew Vulcans and can confirm that, even today, NO fighter currently made would out-turn a Vulcan above 50,000ft, not even the latest generation of FBW marvels. Try that in the sim and have all the F16 wannabes puffing and blowing about rubbish flight models when they can't stay with a fifties bomber in a tight turn at high altitudes...

I can also tell you that several of the very best flight model developers DON'T actually release the most accurate models they could, because they are fed up with the idiotic comments of pubescents who think they know how something should fly, because they've seen pictures of it - it's perceived as being better for business to `dumb down` the flight model because if the accurate one was released, no-one would believe it... This I have witnessed, and contributed to, first-hand.

So there are many factors contributing to `accuracy` above and beyond the mere application of the correct numbers to the sixth decimal place.

You yourself cite a good example. Your FF stick absolutely sucks at representing the real forces in a real aircraft, but it's an entertainment product for an entertainment market. When I taxi a Cessna I don't feel the bumps on the tarmac through the yoke. Not ever. But those bumps are important to me in representing an aircraft travelling over a realistic surface in the sim and are a vital component in my enjoyment of the sim.

X-Plane suffers the same self-belief that it's very few parameters, input accurately, represent a true flight model. Not really. It's accurate enough in a different way, and it airfoil modelling does have an advantage when you crunch the numbers, but X-Planes environmental dynamics are way behind FS, when you think of the achievement of weather specialist products such as Active Sky or FSMetar. So how accurate is a flight model, when the right wing planform of correctly defined size, drag and coefficient of friction meets an airstream that is less than 10% accurate of the real patterns of air in the sky?

Link to comment
Share on other sites

If you want to do a comparisson of Beavers then have a look at http://shadetreemicro.com/. One of the first things I did when I bought X Plane recently is buy this and see the difference. To start off with I didn't see that much. However, what Simon says about frame rate is very true. Once you get above 40 or so fps you have almost a completely different aircraft on the screen.

I've yet to get the fps in FSX above 15 on the Beaver so I can't do the comparisson the other way but if I add the bells and whistles to X Plane and slow it down the gap in handling shrinks accordingly.

If you look at http://www.x-plane.com/contact.html and scroll down to 'How Do Make My Sim Run Faster' there's some good advice on how unshackle the frame rate.

Link to comment
Share on other sites

:)

You raise some good points here Snafe

but first...

****

"You yourself cite a good example. Your FF stick absolutely sucks at representing the real forces in a real aircraft, but it's an entertainment product for an entertainment market. When I taxi a Cessna I don't feel the bumps on the tarmac through the yoke. Not ever. But those bumps are important to me in representing an aircraft travelling over a realistic surface in the sim and are a vital component in my enjoyment of the sim.

****

Oh you are right... :wink:

I turn all the forcefeedback boxes in FSX to OFF.. and use the MS sidewinder 'control' download just to give me some 'stiffness' and a 'reliable' return to centre..... .. if I'm flying a Hurricane in IL2 I use the FFB for the recoil !!

real airplanes have a sense of mass, a sense of fluid response to their environment and, most importantly, a feedback loop that is completely absent in Flight Sim.

I think this is the nail oh the head... BUT some sims seem to capture that a bit better.. I'm thinking of IL2.. the question is how is that 'feel' captured because that is the important bit..

The model doesnt have to capture the plane/weather-air dynamic relationship exactly.. because that is untestable... what is has to do is capture the feel of a 'C150 at 60knots 20deg flaps, a little turbulence and a 6knot crosswind and what I'm grasping for is to understand what it is that is actually varied to tweak it to acceptable... the number of available prameters must be huge... some must be clamped .. some must be adjustable... Is there a book.. describing this.. There are simple rules relating lift/drag to IAS these could all be fiddled but to make it 'look' real the ground speed and has to be about right.. and then the dials can fiddled!

This is fascinating subject.. I know nothing.. I am going to read the 'history' .. is there a 'bible' that I should know about..

:lol:

Link to comment
Share on other sites

If you want to do a comparisson of Beavers then have a look at http://shadetreemicro.com/. One of the first things I did when I bought X Plane recently is buy this and see the difference. To start off with I didn't see that much.

I've got the Shadetree beaver (and the FF & IP scenery packages).... Its not as nice candy as Aerosoft.. but at least its got a 2D panel so the Freda checks are easy!..

I've just been 'flying' the two.. I find them really quite difficult to compare ... marginally I prefer the FSX set up (Vancouver+ scenery) mainly because there seems to be less tree clutter so there are more 'fun' places to put the plane down... the major diffferenc between the planes is on the ground.. The ground handling of the FSX plane seems much 'weightier' with far less weather-vaning..

But I noticed another thing.. they clutter of the world with giant trees (this was in x-plane). this gave an impression of speed that felt much slower than one would expect... I wonder how accurate the optical flow distortion patterns are .. and whether overscaled objects interact with the human visual system to make an illusion .. I should know the answer but my brain is broken.. :roll: have you ever noticed this 'slow' illusion? :?

I've not found a good source of planes light GA planes in FSX but... I think my problem is that I dont quite know where to position Sim world in my flying training.. maybe I should just leave it in the IFR space.. and for VFR just enjoy the sunsets and the various special 'sceneries' ... :wink: which would be just as good (maybe better) sitting in a bath :lol::lol:

Link to comment
Share on other sites

The ground handling of the FSX plane seems much 'weightier' with far less weather-vaning..

Got it in one! I've never done a 180 on take off before :shock: Brakes off and start rolling?? Surely not?

I do get the impression aircraft are hovering rather than moving. Then you get closer to the ground things speed up quite rapidly. I find approaches in FSX much easier. Now the question is which is more optically correct?

Kind Regards,

Cliff

Link to comment
Share on other sites

Talking about flight models.. has anyone tried the SF260 from the Realair stable... :wink:

I think it actually sideslips... and it can certainly spin.. sort of fun when its safe! :lol:

Link to comment
Share on other sites

  • Aerosoft
:)

This is fascinating subject.. I know nothing.. I am going to read the 'history' .. is there a 'bible' that I should know about..

:lol:

It is a highly interesting topic, one we always follow closely. We need to find a middle ground between realism and what the customer expects. The two are often rather far apart.

There a dozens of books, but the one that I seem to fall back on every week is "Understanding Flight" by Andersen and Eberhart (McGraw edition). it's simple enough a manager type like me can understand it, hahaha, cheap as well. http://www.amazon.com/Understanding-Flight...8616&sr=1-1

Link to comment
Share on other sites

Thanks Mathjis,

I already have that book..

I was thinking more of a book that tells you how you 'do' flight sims, and how they work...

How do you guys 'make' a plane...

:wink:

In X-plane there is a program.. Xplane maker which has a lot of boxes in which the builder types parameters ranging from Vne to the engine power, the coefficient of friction of the tyre, and the size of the water rudder etc etc..

Of course being X-plane the documentation is a bit sparse! and there are more than a few runes :roll:

Anyway...

I was wondering how it is done in FSX... and whether theres a manual -how to do it book or posting... or do you have to give M$ a small fortune for the key?

regards ....

:lol:

Link to comment
Share on other sites

  • Aerosoft
Thanks Mathjis,

I already have that book..

I was thinking more of a book that tells you how you 'do' flight sims, and how they work...

How do you guys 'make' a plane...

:wink:

In X-plane there is a program.. Xplane maker which has a lot of boxes in which the builder types parameters ranging from Vne to the engine power, the coefficient of friction of the tyre, and the size of the water rudder etc etc..

Of course being X-plane the documentation is a bit sparse! and there are more than a few runes :roll:

Anyway...

I was wondering how it is done in FSX... and whether theres a manual -how to do it book or posting... or do you have to give M$ a small fortune for the key?

regards ....

:lol:

If you have FSX professional you got all the tools we use (how cool is that), but basically there is just documentation on how to make the config files. Not an expert on this myself, I did learn a lot by just fiddling with the parameters in the aircraft.cfg file. The .air file is a bit more complex and you need a freeware tool like AirEd to edit it. So no real tricks, by far the most complex is gathering the data.

Link to comment
Share on other sites

Thanks for that.. M.Lion...

I read the article .. of course that's going to (already did) wind up the xplane supporters..

I'm looking to understand what is involved in 'tweaking' the planes parameters to give it the 'desired' behavious :roll: ..

the article tried to find a set of 'behaviours' which the plane (C172) should have and rated the sims against them.. I was interested to see that the RealAir came of well..

What would have been nicer would have been some comparison of the parameters which showed/caused the differences....

At the moment X-plane is undergoing a bit of a 'wriggle' trying to fix the radii of gyration calculations (among other things) to try and get the right performance but I think it applies mainly to heavies...

as for the 'religious war' I go with what snave said above...

"Its entertainment ... enjoy... "

:lol:

Link to comment
Share on other sites

If realistic FM and accurate physics are your thing, you cannot afford to miss "Condor", a soaring simulator. In these two categories it beats both FS9/X and X-plane hands down (and every other flightsim I ever have come across in the past 20 years). Strange thing is that almost nobody has ever heared of it. Condor-flyers are an extremely rare species. But most of them are real world-pilots, and back what I say about the physics.

Link to comment
Share on other sites

If realistic FM and accurate physics are your thing, you cannot afford to miss "Condor", a soaring simulator. In these two categories it beats both FS9/X and X-plane hands down (and every other flightsim I ever have come across in the past 20 years). Strange thing is that almost nobody has ever heared of it. Condor-flyers are an extremely rare species. But most of them are real world-pilots, and back what I say about the physics.

I already have Condor and 'used' it a couple of years ago as an first stage of learning to fly .. practicing 'coordinated turns and circuits'... for hours before moving 'real world'.. :D ...

My only real objection to it is the scenery (its really artificial) and the one I suppose you cant complain about... there are no GA planes in it except tow-planes which the AI fly!! :roll:

Also... my major enjoyment from sims is gained not from 'high tech and system flights.... although a bit of proper navex never hurts anyone.. :wink:

but its the takeoff and landing on 'demanding' strips and the occasional bimble through some really good scenery... eg Vancouver+ tongass etc..

However... it will be interesting to compare the Aerosoft gliders when ready with the condor experience.. regards :lol: I

Link to comment
Share on other sites

Hey ho Skybird4..

I read your 'article'.. I admire your patience with some of the responders..

reminds me of a quote... it went something like this

"with stupidity even the gods struggle in vain".. :lol:

However... I agree with much of what you say and I'm glad to see the real world pilots using it came to your rescue... In fact I've just updated my version.. I gave up gliding last year when I concentrated on trying to get my private pilots licence.. maybe time to start again.. I want to get a motor glider.. :roll: . !!

BTW In my limited gliding experience I found ridge flying is the best sort of flying..of any I've done.. low enough to see the detail.. trouble is that so much RW gliding is time waiting and launching others planes...

regards

Link to comment
Share on other sites

Yes, I agree that condor is the state of the art when it comes to GA/glider flight models (as is Lock on for fighter planes...). I expect you can currently not reach this level of realism and detail in flight physics with a simulator like FSX or Xplane which are involving a much larger variety of aircraft.

Regarding the sceneries in condor, there has been quite some development lately. Some of the new freeware sceneries are pretty good. And as the condor comunity is growing larger, the first payware photo sceneries for condor will soon hit the market (simfrance, postfrontal). If this pays off, maybe some other addon developers will start converting some FSX/FS9 photo sceneries for condor, too? We can only wait and see.

Link to comment
Share on other sites

  • Aerosoft
Yes, I agree that condor is the state of the art when it comes to GA/glider flight models (as is Lock on for fighter planes...). I expect you can currently not reach this level of realism and detail in flight physics with a simulator like FSX or Xplane which are involving a much larger variety of aircraft.

Regarding the sceneries in condor, there has been quite some development lately. Some of the new freeware sceneries are pretty good. And as the condor comunity is growing larger, the first payware photo sceneries for condor will soon hit the market (simfrance, postfrontal). If this pays off, maybe some other addon developers will start converting some FSX/FS9 photo sceneries for condor, too? We can only wait and see.

If a newer version would offer a bit more option for scenery it would certainly be something to consider. Right now the limitations will make it hard and the commercial market is very limited. If we got a good idea of the current market most of our addons have larger distribution. But believe me, I am keeping an eye on it.

Link to comment
Share on other sites

Many Thanks Skybird04,

Bought the sim yesterday and have had loads of fun with it. Very quick to start up and configure. I'll confess to cheating and having the sim show me where the lift is but I'll wean myself off that over time.

Thank you for bringing it to our attention :D

Kind Regards,

Cliff

Link to comment
Share on other sites

Guest
This topic is now closed to further replies.
×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. Privacy Policy & Terms of Use