The Physics Model

[sluflyer06 32]

39 minutes ago, Rob_Ainscough said:

We don't have enough computational power to perform real-time flight physics to any degree of accuracy under a wide environmental condition set and aircraft designs.  All is a compromise.

If P3D is one of your flight environments, then I'd recommend looking for aircraft that leverage P3D's ability to allow external computation of "flight physics" (there are many aircraft that do this in P3D such as MJC Q400, PMDG, A2A, FSLabs, Milviz).

If Xplane is one of your flight environments, then there are aircraft vendors (FF A320) that will extract the best out of it's flight physics but just like P3D it requires some adjustments ... the Flight Factor developers went into detail on some of the Pro's and Con's of XPlane's supported flight physics.

As far as MSFS, I've not much accuracy "yet" regardless of "sensitivity" settings ... a flight physics model should NOT be reliant on input device sensitivity and should operate equally within the devices calibrated range.  I'm still puzzled why after the latest MSFS update the SDK download link takes me back to v 0.10.0.0 of the SDK rather than the 0.11.0.0 I already have from the prior MSFS update??  I'm sticking with the 0.11.0.0 SDK under the assumption Asobo/MS just made a mistake with links??

Going forward, I hope MSFS embrace the idea to open the flight physics up to external vendors as they such complexity is very aircraft specific and IMHO should be managed by the aircraft vendor ... but this does increase the cost of development time by vendor (DLC provider) which they will need to recover.

Cheers, Rob.

Dead wrong on computational power for a real-time flight physics engine, I've been working in Flight Simulation for 11 years for a large DoD contractor, I've designed, upgraded, tested and delivered primary flight trainers for multiple fighter platforms (large dome, full cockpit sims with control loading, motion seats, gsuits, you name it). We run 50 microsecond realtime gurateed host system for the flight model on dual 6 core xeons that that use a whopping 2% CPU load and there's even a couple other runtime models on that PC (custom realtime variation of red hat). You do NOT need massive processing power for a real-time flight model running wind tunnel generated physics, etc.  The other runtimes we run need FAR more processing power than the flight model.

[sluflyer06 32]

7 hours ago, mtaxp said:

You really need to get into facts rather thab quoting other and then trying to twist the bottom line.

Long story short; x-plane is used in a pro manner to develop REAL aircrafts, test how they will fly, analyze performance, or in other words it is reliable enough for cessna, the us military, NASA...so if somedbody looking for aerodynamics research, x-plane is the best answer currently.

11 years in professional defense flight simulation and I can tell you we'd would never in a million years consider using any of these commercial products for any use case, nor have I ever heard p3d, xplane, or fsx mentioned by a single engineer. They don't come close to stacking up to our needs for fidelity, however I think MSFS is great, and xp11 is installed on my machine as well, I do expect xp11 and p3d will be massively obsoleted by MSFS within 12-24 months as it matures since the bones of it are more modern and you have the air mass, etc.

[jcomm 5,923]

The more I used flight simulators, specialy those that had models of stuff I fly IRL, the more I think they do their job acceptably provided I kept within what "ELITE IFT" used to designate "the normal IFR envelope..."

Either because we do not own fancy controllers, or because the flight modelling used in our desktop flightsims has it's limitations, flaws, bugs, I have learned that I should really never pretend to feel like in the real thing, or I was going to keep uninstalling each and every flightsim I used 🙂

Uninstalling has been my major occupation since I use fightsims, and it's a loss of time due to the fact that I used to put way too much expectations on a program that is better used if we really want to work our skills in terms of procedural trainning - not flight / hands & feet ... Forget it !

If I use a simulator to practice navigation, then I really don't have to be exigent about  it's post stall characteristics, or even if sideslip or forward slip maneuvers feel realistic.

If  I use a simulator for soaring, then the best thing I can do is to use it for decision making in terms of how to progress in a soaring task. Forget abou how good / close to real you feel the controls and sometimes even the modelled aircraft respond to your input... It's something none of the sims I ever tried even comes close to reproduce....

I would be satisfied with basic stability and control behaviour of the simulated aircraft, acceptable speeds / engine parameters if it's a motorized aircraft, systems detail and possibly even the simulation of failures, and all of that within normal flight situations / envelopes.

ATC is another great potential for these flight simulation platforms, and I regret never having explored it in what it's probably the most proficuous option - VATSIM / IVAO / PilotEdge...

I could never find an acceptable ATC robot 😕

If & when MFS can provide such an acceptable basis, I'll be glad to use it, just as I was when although rather limited visually, I used ELITE IFT or Aerowinx PSX . These were great sims to "train" procedures and I could take the aircraft POH / FCOM and obtain consistent performance from my inputs. That's all I should require from a desktop flightsim.

If it offers nice / great out-of windshield graphics, weather rendering and effects ( although weather effects can bring problems to the flight dynamics model that I would rather prefer not to experience ... ), then even better !

MFS offers great visuals, so that's a ( + ) for it... But the other platforms also have great graphics for my level of exigence. I am perfectly satisfied with all of them graphically.

For a while ( way to long a while, I must admit... ) I decided to divert into combat flightsims, because i assumed they would do a better job in the "feel of flight" area - as if that was possible using my cheap controllers 😕 What a big big waste of time ... and $$$.

Sorry for the long post, but I really thought I needed to share these thoughts ...

Edited March 28, 2021 by jcomm

[Biology 1,347]

10 minutes ago, jcomm said:

The more I used flight simulators, specialy those that had models of stuff I fly IRL, the more I think they do their job acceptably provided I kept within what "ELITE IFT" used to designate "the normal IFR envelope..."

Either because we do not own fancy controllers, or because the flight modelling used in our desktop flightsims has it's limitations, flaws, bugs, I have learned that I should really never pretend to feel like in the real thing, or I was going to keep uninstalling each and every flightsim I used 🙂

Uninstalling has been my major occupation since I use fightsims, and it's a loss of time due to the fact that I used to put way too much expectations on a program that is better used if we really want to work our skills in terms of procedural trainning - not flight / hands & feet ... Forget it !

If I use a simulator to practice navigation, then I really don't have to be exigent about  it's post stall characteristics, or even if sideslip or forward slip maneuvers feel realistic.

If  I use a simulator for soaring, then the best thing I can do is to use it for decision making in terms of how to progress in a soaring task. Forget abou how good / close to real you feel the controls and sometimes even the modelled aircraft respond to your input... It's something none of the sims I ever tried even comes close to reproduce....

I would be satisfied with basic stability and control behaviour of the simulated aircraft, acceptable speeds / engine parameters if it's a motorized aircraft, systems detail and possibly even the simulation of failures, and all of that within normal flight situations / envelopes.

If & when MFS can provide such an acceptable basis, I'll be glad to use it, just as I was when although rather limited visually, I used ELITE IFT or Aerowinx PSX . These were great sims to "train" procedures and I could take the aircraft POH / FCOM and obtain consistent performance from my inputs. That's all I should require from a desktop flightsim.

If it offers nice / great out-of windshield graphics, weather rendering and effects ( although weather effects can bring problems to the flight dynamics model that I would rather prefer not to experience ... ), then even better !

MFS offers great visuals, so that's a ( + ) for it... But the other platforms also have great graphics for my level of exigence. I am perfectly satisfied with all of them graphically.

For a while ( way to long a while, I must admit... ) I decided to divert into combat flightsims, because i assumed they would do a better job in the "feel of flight" area - as if that was possible using my cheap controllers 😕 What a big big waste of time ... and $$$.

Sorry for the long post, but I really thought I needed to share these thoughts ...

"Uninstalling flightsims is a temptation I can never resist..."

I feel you - It has been 12+ months and I still couldn't decide on my main sim. At this point I'm thinking of making my own using Unreal Engine...

Edited March 28, 2021 by BiologicalNanobot

[marsman2020 1,310]

3 hours ago, BiologicalNanobot said:

Some geometrical information is better than none, which is the case in FSX/P3D. What matters most is if end result matches global coefficients, which seems to be the case for Cl and Cd and less so for Cm (and other stability related ones)

Otherwise P3D wouldn't allow a true biplane flight model either.

There is way more to the flight model then the global coefficients, if you want behaviors to be correct at the edge of the flight envelope (like stalling the inside wing during the base to final turn).

[TrafficPilot 121]

7 hours ago, BiologicalNanobot said:

It doesn't even come to a conclusion, it just says that MSFS uses a similar approach to X-Plane (using local coefficients and integrating surface element forces), I can't see why this caused you to think the post says X-Plane's flight model is inferior.

Some flying schools still use FSX and P3D for ground school, theory and navigation training. Both are useful training aids and "within the confines of a non-motion environment" they simulate actual flight pretty well IMO.

Have you checked out the weather engine in MSFS? I've uploaded a short vid of how the new MSFS weather "gusts" affects the default Cessna 172.

I've called it "MSFS TWERK" because that's what it reminds me of. If you think it looks realistic please remind me never to fly with you 😄

http://hhuk.net/MSFSTWERK.mov

Graphically I think MSFS is ground breaking. But they can't claim it's a "flight simulator" anymore. It's a real pity.

 

[LRBS 947]

Pretty interesting statements here and some totally outrageous.

 

As a real airline pilot, I can share with you some of my experiences in million dollars simulators, some at the factories where I did some type rides, and a few airlines.

 

I will refer to only 2 major manufacturers that use state of art simulators with all the bells and whistles and visuals for B748, A380, and A350.

 

Outstanding companies with the unlimited support of maintaining their simulators.

 

Their flight simulators are only within 85% or so accuracy of the real airplane.

 

So, it is impossible for them to create 100% aircraft accuracy and behavior.

 

It is total nonsense to have some developer claim that their flight model and behavior is within 5% of the manufacturer when a factory simulator cannot achieve that standard.

 

They have fixed base (no motion) and full base simulators (full motion) and we do training in both. I mentioned this because there are so many "experts" here implying as to be a problem that there is no motion.

 

Yes, we do also training on a fixed base, do we like it? no, is it real? no, all that it matters is how the airplane (simulator) reacts to the control inputs, how it will react to winds, how will react to inertia, etc.

 

Coming back to the MSFS "physics model" there is a total disconnect and not even close to a realistic aircraft behavior, the airplanes are not so twitching up and down, so unstable at only 2 KTS wind changes with such an unrealistic air mass reaction, bad ground effects in many aspects....

 

Then we have people making statements like "Having never flown a real plane in real life I can tell you the physics are 101% correct" or fighting all over that the airplanes are flying "beautiful".

 

After such a long time since the launch took place, I expected to see more concern and changes to the flight model, ultimately, they call it a flight simulator. Obviously, it is not of any concern to them regarding the flight model.

 

To close this, there is no real-world 100% accurate flight simulator, it will be impossible to have one 100% for the PC but MSFS aircraft flight models at the present time are "first" from the bottom.

 

It's only up to the customer to request for a fix or just pretend that is OK.

Edited March 29, 2021 by killthespam

[Doug47 574]

1 hour ago, killthespam said:

Pretty interesting statements here and some totally outrageous.

 

As a real airline pilot, I can share with you some of my experiences in million dollars simulators, some at the factories where I did some type rides, and a few airlines.

 

^^^
This should be a sticky in every sim platform sub section and on every flight simulator forum. And given a AVSIM Gold Award for the post of the century. 

 

 

Edited March 29, 2021 by Doug47

[DrumsArt 270]

3 hours ago, killthespam said:

Pretty interesting statements here and some totally outrageous.

 

As a real airline pilot, I can share with you some of my experiences in million dollars simulators, some at the factories where I did some type rides, and a few airlines.

 

I will refer to only 2 major manufacturers that use state of art simulators with all the bells and whistles and visuals for B748, A380, and A350.

 

Outstanding companies with the unlimited support of maintaining their simulators.

 

Their flight simulators are only within 85% or so accuracy of the real airplane.

 

So, it is impossible for them to create 100% aircraft accuracy and behavior.

 

It is total nonsense to have some developer claim that their flight model and behavior is within 5% of the manufacturer when a factory simulator cannot achieve that standard.

 

They have fixed base (no motion) and full base simulators (full motion) and we do training in both. I mentioned this because there are so many "experts" here implying as to be a problem that there is no motion.

 

Yes, we do also training on a fixed base, do we like it? no, is it real? no, all that it matters is how the airplane (simulator) reacts to the control inputs, how it will react to winds, how will react to inertia, etc.

 

Coming back to the MSFS "physics model" there is a total disconnect and not even close to a realistic aircraft behavior, the airplanes are not so twitching up and down, so unstable at only 2 KTS wind changes with such an unrealistic air mass reaction, bad ground effects in many aspects....

 

Then we have people making statements like "Having never flown a real plane in real life I can tell you the physics are 101% correct" or fighting all over that the airplanes are flying "beautiful".

 

After such a long time since the launch took place, I expected to see more concern and changes to the flight model, ultimately, they call it a flight simulator. Obviously, it is not of any concern to them regarding the flight model.

 

To close this, there is no real-world 100% accurate flight simulator, it will be impossible to have one 100% for the PC but MSFS aircraft flight models at the present time are "first" from the bottom.

 

It's only up to the customer to request for a fix or just pretend that is OK.

Thank you very much. It puts the ideas in place and things in perspective...

This post should be pinned 😎

Richard.

[WestAir 2,522]

4 hours ago, killthespam said:

They have fixed base (no motion) and full base simulators (full motion) and we do training in both. I mentioned this because there are so many "experts" here implying as to be a problem that there is no motion.

Hi Captain. I am that "expert" I think you're referring to, but I think you misunderstood my point. (Which means I didn't make my point clear enough, so I apologize). I only mentioned motion in relation to force feedback. If you're climbing at 90 degrees with afterburner, it takes more effort to make a nose down correction with the G force acceleration / added weight fighting against both your arm and the yoke; even if you could program a simulation that was 1:1 in realism, it would fall short with a $40 joystick sitting on a desk, and some pilot somewhere will accurately conclude that the feeling is wrong, even though the numbers are right.

I'm sure we've all had moments where a bump or whatever forces us to make inputs we didn't intend to make.

[mrueedi 702]

5 hours ago, killthespam said:

Their flight simulators are only within 85% or so accuracy of the real airplane.

Imho this number has not much use because just the "accuracy of a flight simulator" is a far too broad term. So the question is, the accuracy of what exactly is only 85%?

The accuracy of the flight performance surely matches the real aircraft within a couple of percent points (considering that also the performance of real aircraft can differ by that from aircraft to aircraft). Also the simulation accuracy of the avionics systems would be a joke, if it would not be better than 85%. What I can imagine, to miss the 100% mark, is the simulation of the flight handling. And what is known to be a rather poor compromise even in full motion simulators is the simulation of the motion:

Do we really need full-motion flight simulators? (linkedin.com)

The physics model does not need to be so bad. Calculating the forces and the reactions correctly, so the motion on the screen would be accurate, is not impossible. Looking at the MSFS flight model the ingrediences for a real good flight model are certainly there.

5 hours ago, killthespam said:

Coming back to the MSFS "physics model" there is a total disconnect and not even close to a realistic aircraft behavior, the airplanes are not so twitching up and down, so unstable at only 2 KTS wind changes with such an unrealistic air mass reaction, bad ground effects in many aspects....

You should try FSX. It flies on rails.

Fixing that "flying on rails", is specifically something they tried with MSFS. At least in the basic flight model (the formulas, the design of the calculations, the considered effects) they succeeded imho. The issues you mention are coming from the parametrization, which might be tuned up a bit too much. Something which is btw. also done in full motion simulators.

Edited March 29, 2021 by mrueedi
more precise wording

[ray hughes 165]

19 hours ago, orchestra_nl said:

IMO there are only two types of people who can know this: those who work at Asobo and possibly real world pilots who have been using both simulators. Everyone else is just using their gut feeling, with a dose of prejudice, expectations, hate and hope thrown in there as well.

As a private pilot (long since retired from flying) - and a user of P3dv4/5 as well as MSFS2020 i can say neither quite add up to a real world feeling at the moment. The closest thing I have seen was a youtube clip from a fella flying a new bush aircraft in Xplane 11. The reactions movement and dynamics looked about right. I did have Xplane 10, but didn't go for Xplane 11 as I felt the controls and scenery were lacking -  or just too different from what i was used to. 

However, that hasn't stopped me enjoying MSFS 2020.  I am sure it is but a patch or two away from feeling dynamically like the real thing!!

[EvidencePlz 276]

17 hours ago, Chock said:

Most of that is down to the controllers we use and the fact that the sim is to some extent designed to use these too. If we take for example a Spitfire, the real thing's joystick has a pitch throw of about 14 inches or so, but your average joystick for a PC probably isn't even a quarter of that

Hmm...

https://forums.eagle.ru/topic/266444-so-you-want-to-fly-a-spitfire-watch-this-video-to-get-started/

https://youtu.be/Gs2m-xlyK7o

 

[Dominique_K 10,138]

7 minutes ago, mrueedi said:

 

Fixing that "flying on rails", is specifically something they tried with MSFS.

 Asobo started their work from FSX 2006 instead of the state of the art of the FS franchise which was P3D payware aircraft circa 2015 which dont give the flying on rail sensation. 

They wanted to render something convincing rather than precise, see Bossard’s interview below, not a bad goal on itself for a flight simulator.  The problem is that convincing is subjective, precise is not. In the EP video on aerodynamics #2 Wloch hinted at an update in this department, evidence that they are not totally happy with what they did. 
I read carefully what the pilots in this community say, including tin this thread. As their opinions vary widely, I surmise that the debate will not be closed anytime soon. As a pure sim user, the overall sensation is rather good, except maybe at takeoff (controller sensivity problem) . 

This is what Bossard said
The previous flight simulators were based on an engineer approach with a concern to have all the correct parameters and instruments but too disconnected from a physical experience. Now, with the technologies at our disposal, we can reproduce the entire Earth and therefore favor another type of flight: visual flight. It is also the first type of flight that we learn in pilot training: finding one's way in relation to the coast line, the river, etc. We really put the emphasis on the feeling of flight, on the fun, on the pleasure of feeling the masses of air, on the dynamic camera. 

 

For me, the main technical difficulty of this project was precisely to approach this overall sensation of flight so that it was not only mechanical. And to get there, we had to link all the simulation parameters to have experience precision at all levels. 

[tweekz 2,165]

6 hours ago, killthespam said:

After such a long time since the launch took place, I expected to see more concern and changes to the flight model, ultimately, they call it a flight simulator. Obviously, it is not of any concern to them regarding the flight model.

I agree with you. It's about time ASOBO brings serious changes to the flight model. It's not that the idea is fundamentally flawed IMO, it is just not done. There are many things to tweak or add before it will be fully satisfying.

If they can add world update after world update, they better put that effort in the flight model for once. I don't like the trend that sim updates have become bug fix colletions rather than adding new core content.

Edited March 29, 2021 by tweekz