Ciao, as no pilot I’d like to have a comment from real pilots about MSFS flight dynamics. Is it realistic? What do you think are the main differences between real fly and simulator? Reading around I think I understood it depends from the plane you fly too, is it true?
Thank you all
Hi @Looo5610
Your topic has been moved to the Aviate, Navigate, Communicate category of the MSFS forums. Please ensure you look for the correct category when creating a new thread in the future.
Thanks,
The MSFS Team
Whoa that’s some question! Where to start. At the risk of turning this into an Op-Ed, let’s take some different angles.
In the days of FSX, developers were able to precisely define a flight model through a vast number of look-up tables. A ton of work, but highly accurate and versatile.
MSFS has (mostly) left that scheme and moved towards the ‘real world’.
In the real world, there is no such thing as a flight model. An aircraft’s behavior follows directly from the interplay of geometry, weight and engine with the air. There is no ‘flight model file’ (Exceptions are from aircraft that are heavily controlled through fly-by-wire). In order to predict the aicraft’s behavior, aircraft manufacturers like Boeing, Airbus use supercomputer-based CFD (Computational Fluid Dynamics). Even then, they need to resort to wind tunnels and flight testing to measure and finetune the details.
Problem is, that kind of calculations is WAY too heavy for realtime PC-based work.
And thus, MSFS seems to have adopted a compromise. The flight model is based on the geometry and MSFS translates that to a ‘normalized’ aircraft, and allows for tweaking and modification. Unfortunately, it is unclear to most flight model developers how it works exactly, and a flight model for MSFS has thus become a major trial & error, tweaking (through ‘scalars’) and testing effort.
And here comes the issue…
On the one hand, developers boast largely the same: performance ‘by the numbers’, tested by real pilots, in accordance with the manual, etcetera.
On the other hand, whereas many pilots voted flight model as their #1 priority ([POLL] Your most important things when deciding to BUY a NEW AIRCRAFT? - #80 by MatronlyApollo8), I suspect many don’t have a clue to what it implies. Many reviewers are explicit on this. They say ‘it feels fine, but I’ve never flown one for real’, and that’s it.
Many pilots fly airliners. And yes, these are mostly fine because the maneuvers are mild, the flight envelope used is very narrow etc. The difficulty is much more severe in fighters, aerobatics, high-altitude, supersonic aircraft and such.
You may be surprised to learn that MSFS does not model flaps differently from slats, although they are vastly different in terms of effect. MSFS does not allow a wing profile to vary across the wingspan. MSFS does not allow an asymmetric aircraft (except perhaps an engine placement), MSFS does not allow custom lift and drag for the fuselage, MSFS does not allow multi-position spoilers/airbrakes. And I’m not even starting to talk about engines and propellers. And, Asobo has not adhered to the established aerospace models, definitions and equations, making life difficult for developers (e.g. wing camber in degrees instead of percentage, a left-handed coordinate system instead of right-handed).
Most pilots are probably totally unaware of these complications.
The benefit for MSFS is that it has become a lot easier to start developing flight models, because there is a lot less to define. And thus we see a lot of payware aircraft of, let’s say, marginal flight model quality at best.
For developers, flight modelling is close to a nightmare. The flight_model.cfg features tens of obsolete parameters. But some are described as ‘legacy’, but are still required and may not be omitted (e.g. “pitch_moment_delta_elevator”).
As a result, developers like A2Asimulations seem to have totally bypassed the Asobo code, and use a prevailing fully custom code. Which is perfectly fine, except that the development time and cost will go through the roof.
Some other noteworthy solutions are what is being done by DCS and Il-2. They feature a very sophisticated and proprietary flight model. As a result, it is much harder for developers to come up with additional payware. Il-2 apparently does all their flight modelling in-house and does not even feature add-on aircraft as far as I know.
So, looking forward to MSFS2024, MS/Asobo have four options, in my opinion.
Honestly, I’m not sure what to expect or hope for….
BTW… I hold an M.Sc. in Aerospace Engineering, specializing in flight performance. I have been flightsimming since 1985 (FS2 on a Commodore 64). I have held a private pilots licence. I am also a contractor Flight Model developer, although I must admit I have not done any modelling for FSX.
Well said by Matronlyapollo8. Very thorough. The only thing I’ll add is that sitting in a desk chair gives none of the feedback a real aircraft will give you. ‘Feel’ is a big part of piloting. But as far as consumer-grade simulators go, MSFS does duplicate many (not all) real world dynamics.
As someone who did his PPL just last year, after decades of simming, I’d venture that it’s not horrible. It’s not great, but it’s not horrible. As another member of this community (@CharlieFox00) once said, the aircraft in the game don’t necessarily fly like THE airplane they represent but they do fly like AN airplane.
I’d also offer that, as part of my PPL, I did some hours in a certified RedBird simulator (one needs at least 5 hours of instrument time for a PPL in Canada and a couple of those hours can be done in a simulator - the rest have to be done under the hood in a real airplane). In my opinion, the RedBird’s C172 flight model S.U.C.K. sucked in comparison to MSFS’ C172 flight model. (It really was bad, especially in the yaw axis.)
So yeah. MSFS’ flight model isn’t perfect but it isn’t garbage. It won’t teach you to fly, but it will let you practice certain aspects of flying once you know how. And it’s fun and lets you explore the world.
Edit: I should add that, while learning to fly, flying the C172 in MSFS felt nothing like flying the real thing. But as I got used to the real thing I could see the similarities again. It was an unexpected transition.
Edit 2: I like my answers to be as accurate as possible, so I’m going to revise what I said above about flying the C172 in MSFS feeling nothing like flying the real thing when I was first doing my training… That’s an overstatement. There were things that felt very similar and things that felt different but it’s hard for me to put my finger on what the cause of that is. There’s a certain ‘je ne sais quoi’ about flying the real thing that’s missing in the sim, and I really felt it at the beginning of flight training. Maybe it’s the lack of “life or death” pressure in the sim, or the lack of the same types of feedback, or the flight model, or myriad other things? I can’t say. But I did find that as I got more comfortable in the real thing, I could see more similarities with the sim again. And towards the end of my training, I often used the sim to prep for and to review after my lessons, with good effect.
@MatronlyApollo8 - Very well said. I feel option 3 (more CFD) is the most probable future - they have said as much for MSFS2024, but they have promised 3rd party devs much more ‘control’ than the current SDK allows. We can only hope so.
The flight dynamics are frustrating - there are aspects that are great and aspects that are really very poor, even for a home entertainment flight simulator (e.g. the difficulty in obtaining correct stall speeds when using CFD or the atrocious weathervane effect).
I would add for anyone reading, that when discussing flight dynamics and flight models, it is important to keep in mind that they are two different things - the flight dynamics are the core physics and are (to some extent) a black box that aircraft devs have to live with.
Flight models are however down to the dev, and there are tools available in the SDK to make the aircraft behave and perform closer to the IRL (although only by this ridiculous, almost ‘artisanal’ approach of tweaking and tuning). Some devs do a much better job than others. I guess some figure that many simmers don’t really care that much (and maybe they are right - it leaves me stumped why some of the flight dynamics/physics bug reports/wishlist items attract as few votes as they have).
And this is why some of the comparisons made on YouTube between MSFS and other sims are almost completely meaningless - all they are usually comparing is how a stock C172 in one flies compared to a stock C172 another. Great- if you are only interested in finding the best stock C172 simulator!
So when some folk cry that ‘The MSFS flight model is no good!’ in some aircraft thread, do they really mean to say ‘this dev did not do a good job with their aircraft flight model’?
Good points @LKFJP and @AlpineB4652. So indeed, the default C172 is quite good, and the WBSIM mod made it even better. I have not explored it throughout the flight envelope, but the ‘normal’ use is good indeed. And the same goes for some other aircraft. For the C172 it is important to understand that there is a lot of real-life data and information available. Try to find something for older aircraft, and you’re in the dark.
Good split between the SDK and the Core physics, and indeed comparing aircraft among each other as opposed to comparing to real-life. Comparing between each other is useful if there is more than one offering. And to what extend a developer does a right job within the limitations is another good point.
Thanks for your inputs!
From Todays FAQ on MSF2024:
But what is ‘improved’?!
Thank you all guys, what you wrote explaine well the situation. As I said I’m not a pilot, but an engineer and I’m always very interesting in this kind of simulation. At this point let’s hope MS will go in the best direction with FS2024!