I am pretty amazed by SU11, especially the new 20 KM CFD that surrounds the plane (CFD stands for Computational Fluid Dynamics). The new 20 KM CFD simulates air movement around trees, hills, mountains, buildings, etc, around the plane, as well as thermals too. From my understanding, the 20 KM CFD is a first for a desktop flight simulator and no other desktop flight simulator has anything like it:
So after the work the MSFS team has done on the flight model over the last 2 years, and now with the new 20 KM CFD that surrounds the plane, what is your assessment of the MSFS flight model? Do you think MSFS has one of the most advanced flight models for a desktop flight simulator?
It really depends who you ask: A casual who doesnāt know better might say āyeah this is just like irlā but a veteran simmer or real world pilot might go āeehhhhhh no this is so not accurate.ā
Maybe not now after 11 sim updates. Thereās still better flight models in other games. But who knows where weāll be after another 11 sim updates? The potential for the most sophisticated and advanced flight model is there, but itās not gonna be for a while.
Weāre on the right track i think. Others will disagree. They can go play X-Crash or Pā¦
Thanks for the response! Iāve read a lot of opinions and responses from real life pilots. The responses do vary, but I donāt see any consistency in the responses from real life pilots that are saying the flight model for MSFS is worse than the competition.
Some say itās worse, some say itās better. One thing I do notice in the responses from real life pilots is that many say it depends on a plane to plane basis. For example, the Milviz 310 has a very good flight model, according to many real life pilots. Then I read other opinions of other planes in MSFS where real life pilots say the flight model is not as good. So it appears that it seems plane dependent.
Having said that though, the new 20 KM CFD from SU11 does really appear to be ground breaking technology for a desktop flight simulator.
you hit the nail on the head with this one. An easy comparison would be comparing the 737 by PMDG and a knockoff you found on Flightsim.to. The odds that the knockoff you found on flightsim.to is just a model and the thing flies likeā¦idk an Extra 300 are pretty astronomical. The Boeing Extra 73007-800 might be reacting to the atmosphere based on how the developer did the flight model on the plane so itās probably correct in the sense itās flying the way itās designed, even though itās not how that plane should be flying.
Developers can make whatever flight model they want really. It can be accurate or it might not. It seems to me that the average amount of time for at least an accurately flying airframe to be developed is around two years. Just based on the high fidelity planes we see available now
I feel like the MSFS flight model has improved significantly since release and is on par to X-Plane. Iām sure they have pros and cons of each.
As to each plane it may be a case of āGarbage in, Garbage outā. If an individual plane has a poor flight model data it does not matter how good the sim flight model is. The sim flight model is only going to be as good as the aircraft data it gets.
Itās certainly one of the most realistic, and I donāt think we can determine if itās the most realistic. The differences between current simulators are negligible, or at least small enough for most players not to notice. MSFS is certainly the closest most of us will ever get to flying a real plane.
Iām not a pilot in RL, but for me itās the first time that flying with small GA airplanes feels believable. I love how wind moves the plane around and would like even more turbulence effects! Graphics have been degraded to oblivion on Xbox but flight model for me is the BEST and most FUN it has ever been. Thank you Sebastian and team for this progress in aerodynamic modeling.
What I dream about now is real dangerous weather, turbulence so high that it breaks planes apart, just like in real life. Asobo is on the right path here and I think they can give us this extreme weather modeling in the future.
The sim flight model (atmosphere) is a completely different thing than the aircraft model it self. The sim flight model (atmosphere) needs to work with every single aircraft. A small paper plane should be able to fly in the sim atmosphere but because it has less mass it is affected more of the turbulence. Then if the turbulence is accurate modeled iām 100% sure is not. Will never be in a flight simulator. Can it be improved? For sure it can be improved. CFD is the way to go to have it simulated correct. Especially if they using LES-model in the future. But LES-model requires much CPU-power especially to run in real time that it needs to in a flight simulator.
In my opinion the atmosphere/air behaviour has improved much since release. The atmosphere feels much more alive. Continue the work on improving the atmosphere Asobo
Asobo said in latest q&a that they working on a full CFD simulation. If you turn on aircraft CFD then you see what they working on.
I agree the differences between the flight models of the top civilian simulators is very close. Having said that, I donāt think any of the other civilian simulators have anything like a 20 KM CFD which simulates the airflow in the atmosphere 20 KM out (or something comparable to that distance) from the plane. I think this is a first for MSFS, and MSFS is way ahead of the competition by just having this 20 KM CFD atmospheric simulation.
As you say, they definitely do not have such a big CFD (or any). But they do have a much more accurate flight model, that is inevitable, as the data comes directly from the manufacturer.
Still, even the big sims donāt get everything right, and getting into the real aircraft definitely feels different again. And certainly the scenery on MSFS is much more pretty, although thereās some gains there in the big ones too, but very low poly and low quality textures still.
The flight models we have now are undoubtedly getting closer and closer as time goes on, and 3rd party mods are also helping to bridge that gap even more. MSFSā limitations now lies in the limits it imposes on itself about the flight model (singular square wing, in conventional configuration is the only one you can have).
And, secondly, just because they have data (what data?) from a manufacturer (why doesnāt MS/Asobo have the same access on the same planes?), how does that get related to a flight model improvement in a meaningful way when the flight model constituents in the other sims are MUCH MUCH more coarse than those found in MSFS? The last I heard the difference was over a thousand surfaces (MSFS, each interacting with each other and now off the surrounding air) vs on the order of tens of surfaces (not interacting with each other last I knew, i.e. surfaces donāt shadow airflow from other surfaces like they do in MSFS).
Iām asking, as you are a developer and I imagine have developed for both simulators.
As another example, the flight model for the C-172 in MSFS was developed from instrumented data of a real aircraft, as opposed to maybe wind tunnel data from a manufacturer, which is at best an approximation of data in a couple of regimes. Hence Iām wondering what data they have that makes their āflight modelā special? (which really isnāt a definition of flight model. It might only mean the initial conditions are closer than those used in one vs the other, which has a meaningful effect, but is not relevent to judging one flight model tools / methodology vs another per se. Hence, my confusion.).
Granted, weāre talking PCās here, so, whatās really inevitable is itās an approximation for all of them with much simplified calculationsā¦ Secondly, itās a little early to really judge the answer to the OPās question at this point as developers havenāt had a chance to exploit the advantages of the new SDK flight model tools yet if comparing current planes to the same in the competition.
As a layman, albeit one with an aeronautical background, my opinion is:
MSFS has the best atmospheric model - I would go as far as to say that.
MSFS has good flight modelling - out-of-the-box as it were - but perhaps not the best (yet).
To that end, the atmospheric model actually adds a much bigger layer of immersion than a precise-to-the-numbers flight model (for me), so Iām happy with it overall.
I feel that low-speed, high AoA flight feels too forgiving, and thatās a regime which can be common in the landing pattern, but Iām mainly informed by previous sims and my own expectations rather than any deep real-life experience to draw upon, so I am happy to be corrected if the data supports it
You also have to take into consideration that not a lot of aircraft fine tune their aircraft for CFD. PMDG for example uses their own custom flight model rather than the one MSFS provides, while it is amazing, it is not perfect and PMDG still have a bit of work to do to get the 737 right from the way the aircraft behaves in the air to the way it handles on the ground. Taxing for example feels so unrealistic and literally like you are taxing an aircraft that turns instantly regardless of if you are taxing slow or fast. Flaring on approach also feels really lightweight as if you are flying a Cessna.
The only aircraft I know that utilized the full CFD system is some of Asobo default aircraft like the C172(G1000) and such. Only third-party aircraft I know that use CFD are FlybyWire and Fenix. Let me tell you, the way these aircraft fly feel so much better than any other 3rd party aircraft out there. If you use these aircraft, you will notice how well the aircraft handles so much better on the ground than PMDG and how the Fenix feels on approach when coming into a flare. It actually feels like you are flying an airliner on how it feels in terms of weight.
It is all dependent on how well the developers use the tools that they have.
For that comment I was talking about professional D-level simulators for real life flight training.
Simply put, whilst for example, Airbus for example, shares some data with Asobo and Fenix for their performance, they donāt share everything. I would guess in fact that much data is still withheld from home use flight simulators.
That is made up for by us users, and the developers themselves, but āadjusting thingsā to make it feel correct, or to meet certain performance targets.
The fixed base sim (big real sims) will have much more accurate data, and be required to meet certain performance specifications. They are complex machines, but a large part of their cost will also be coming up with that high-fidelity flight model and meeting the manufacturerās stringent requirements. Requirements that a home based sim such as MSFS will never have to meet, nor probably even have access to.
So even though MSFS may have 1000 surfaces, those surfaces have to absolutely have to be correct to ensure those interactions are correct. Without 100% the manufacturer data, and the requirement to strictly meet their tolerances, home sims can never compete with the big professional flight simulators. But they get closer every iteration!
Another point though that I was trying to make, is that MSFS basic model that all aircraft work with, is a straight wing, conventional (single symmetrical (left/right) main wing in the middle of the aircraft, tailplane at the back) aircraft, where as other simulators may be able to simulate the interaction against the atmosphere in a much more accurate way. Even if they have less interaction points, they may be able to model swept wings for example, or canards as another.
Ok, now it makes TOTAL sense! Thank you for the contextā¦ I was so confused
P.S. MSFS does support canard configuration. I think youāre correct however that they still donāt model biplanes? Which seems awfully silly to me. One of the most basic configurations. I donāt know about swept wing configuration either, that also seems to me to be a very basic one that should be handledā¦ Delta configuration? I wonder how Carenado handled the v-tail of the V-35?
Oh interesting! I wonder if 3 lifting surface is also enabled?
Yeah some of the limitations are interesting to say the least. Especially with high speed or supersonic aircraft. Biplanes have a whole suite of interesting aerodynamics and interferences. And the V-Tails must just be approximated.
Thatās one thing that the other sim can do, is to allow aerodynamic experimentation.
Maybe one day weāll have a few extra build options.
As I understand it, according to Seb the eventual atmospheric model will be in place that devs will then only have to get their physical modelling and performance right.
