1 hour ago, robert young said:

I'm not going to argue with you, if you mean the potential. But this "flying on rails" cliche really needs debunking.

 

Oh my gosh yes!.... so tired of seeing people write about FSX rails vs XP's.  I'm a current pilot of a small Beechcraft, and I've also used FSX/P3D and XP11 extensively.  In either sim, the aircraft flight model can be done extremely well, or done very poorly (or somewhere in the middle).  I've had a wonderul "feeling" of flight in both FSX and XP11.  It HIGHLY depends on the developer's skill at creating that feeling of flight.  The plane I fly irl, trims out so nicely, I can take my hands and feet off the controls and it won't change heading or altitude, unless some turbulence appears.  I strongly feel people who continually mention this "rails" item, have never flown a small aircraft!

 

Edited December 6, 20205 yr by ryanbatcund

17 minutes ago, leprechaunlive said:

So why did you buy MSFS?

Probably the same reason why I did. The 3d aircraft models and textures are outstanding for default offerings. The scenery is beautiful. The lighting is great. The special effects are impressive. The user interface is not perfect but very well laid out. The weather is likewise beautiful to look at despite the complaints, and the potential is fantastic.

Other features are not so good - especially the important one relating to actual flight. In time it will no doubt improve, but some basic things are taking rather a long time to sort out. 

6 hours ago, MattNischan said:

I don't believe this is true either, to my knowledge. The border conditions (like stall behavior and such), are emergent properties of different calc segments hitting critical angles of attack at different times, due to their relative positions in the airstream.

-Matt

Speaking as someone who has never tuned a flight model for FSX or MSFS, but years ago took 4 years of classes on aircraft design (and then promptly went to work on mechanical stuff instead).....I think this is one challenge with this type of flight model - if the "emergent properties" are not correct, then correcting things can become a nightmare of interacting correction factors that are potentially harder to turn than the older table-base flight models where the interactions were more clear because the flight model was assumed to be based on tables of factors. 

Which is consistent with what Robert has been seeing in his efforts, and I think these are useful observations.

If you define an aircraft with the correct properties for all of the aero surfaces, and the cruise speed is correct at the correct engine output, climb rates are correct, but the stall speed is 10kts too high, what then?  What do you change that fixes one parameter without degrading the others? You'd need the wing to have a different CL vs Alpha curve to change the stall speed, without impacting cruise, so you tweak the airfoil definition to something other than the real airfoil to try to fix stall?

I would have been much happier if the sim had launched with ~9 aircraft instead of 28 (+2 steam gauge variants of those):

• A slow VFR-only taildragger like the Savage Cub
• A standard 2-seat trainer
• A standard 4-seat GA aircraft
• A high performance single piston GA aircraft
• A twin piston aircraft
• A single turboprop aircraft
• A twin turboprop aircraft
• A business jet
• An airliner
• A fighter jet trainer

And Asobo had spent a lot of time on the aerodynamics and engine performance of just those aircraft, with real-world test pilots, to bring the flight models (not the entire systems model of the aircraft, leave that for others) to nearly study level.

This would have proven out the "emergent properties" of the new flight model across the entire spectrum of expected aircraft that would use it in the future.  That would ensure that when other aircraft are added later, the model is on a solid footing.

Instead we got ~30 aircraft none of which are modeled particularly well (in my opinion).

I don't doubt that there is more power in the new method if implemented properly, I just don't think Asobo have done it properly.  The baked in instability/motion of the aircraft to make them feel more "alive" that has nothing to do with any aerodynamic effects on the aircraft being the biggest indicator of this.  X-Plane's blade element model was originated by someone coming at it from a different area of expertise/experience.

The issue now is that doing anything to try to update the behavior that impacts all aircraft has huge consequences, so it will probably be "correction factors" from here on out to avoid breaking existing aircraft.  Or they find a way to extend the flight model API to have multiple 'versions' so new aircraft can use a newer version of the flight model without breaking old aircraft tuned against the original version. 

By not taking a few aircraft across the range of configurations to near study level flight models pre-release, it can really make a model like this that depends on emergent properties a pain to work with.

Edited December 6, 20205 yr by marsman2020

36 minutes ago, robert young said:

Probably the same reason why I did. The 3d aircraft models and textures are outstanding for default offerings. The scenery is beautiful. The lighting is great. The special effects are impressive. The user interface is not perfect but very well laid out. The weather is likewise beautiful to look at despite the complaints, and the potential is fantastic.

Other features are not so good - especially the important one relating to actual flight. In time it will no doubt improve, but some basic things are taking rather a long time to sort out. 

allright fair enough 😄

1 hour ago, marsman2020 said:

Speaking as someone who has never tuned a flight model for FSX or MSFS, but years ago took 4 years of classes on aircraft design (and then promptly went to work on mechanical stuff instead).....I think this is one challenge with this type of flight model - if the "emergent properties" are not correct, then correcting things can become a nightmare of interacting correction factors that are potentially harder to turn than the older table-base flight models where the interactions were more clear because the flight model was assumed to be based on tables of factors. 

Which is consistent with what Robert has been seeing in his efforts, and I think these are useful observations.

If you define an aircraft with the correct properties for all of the aero surfaces, and the cruise speed is correct at the correct engine output, climb rates are correct, but the stall speed is 10kts too high, what then?  What do you change that fixes one parameter without degrading the others? You'd need the wing to have a different CL vs Alpha curve to change the stall speed, without impacting cruise, so you tweak the airfoil definition to something other than the real airfoil to try to fix stall

I don't doubt that there is more power in the new method if implemented properly, I just don't think Asobo have done it properly.  The baked in instability/motion of the aircraft to make them feel more "alive" that has nothing to do with any aerodynamic effects on the aircraft being the biggest indicator of this.  X-Plane's blade element model was originated by someone coming at it from a different area of expertise/experience.

 

I agree with all of this. I think there is a lot of focus on what system should be used and its relative sophistication, rather than whether the emerging work is convincing. There is a lot of talk about the marvelous potential etc. When I look back, I see that usually there are just a few key parameters that need to be addressed almost every time and the rest is often surplus to the task at hand. Just juggling a small number of variables can throw up hundreds of permutations, which is another reason why keeping things simple (and I mean in the simple-sophistication sense, not simplistic) is often the best approach.

In a sense nearly all aircraft fly in broadly the same way. And making them do so is not really a process of convoluted systems, but an aptitude for observation and tuning fairly straightforward variables that contribute to authenticity.  And the majority of sim pilots just want an aircraft that responds believably to input or weather - and of course a reliable autopilot.

 

 

55 minutes ago, leprechaunlive said:

allright fair enough 😄

also, not to beat a dead horse: but the preview videos mentioned the advanced aerodynamic model and the infamous 1000 points of bla bla... I will stop short of saying it was "false advertisement" because I'm not knowledgeable enough to make such a claim, but as someone with some C172/C182/C152/glider time, the end results are far from what I was hoping for. The default C152 and Rob's Bonanza Turbo mod are the closest at the moment (to my subjective opinion).  I also agree with @ryanbatcund : the "flies on rails" thing is overrated.

 

Edited December 6, 20205 yr by evaamo

3 hours ago, robert young said:

I'm not buying that Matt! You are implying here quite clearly that a few old fashioned farts with a luddite view of the world and well past their sell-by-date somehow just don't "get" how 1000 lift points is the future. They are SO yesterday!

I'm not here to make enemies or to make people upset, and that's not what I'm implying at all here. My intent is always to be as positive as possible and as forward looking as possible.

The reality is, that while there may indeed need to be more parameters that need to be exposed at present (and I agree there do absolutely need to be) the reason that something like real wings snapping in a stall is because the air will never move over each wing precisely like the other and one end of the wing will almost invariably start stalling before the other end. The new flight model is light years closer to that type of reality than a table that looks up values and acts on the center of the plane, and way closer to the fidelity of XP, which we all praise for flight model. I don't that's controversial to suggest.

Yes, could you bend over backwards if you were the best of the best and make a flight model that exhibited some quality boundary condition flying in the table-sim model? Absolutely. But that's the thing, Robert, you are the best of the best. Most other even great flight models in FSX flew like you were driving a car on an imaginary road in the sky, totally uninvolved in the air around them. Even someone who is only a quarter as talented can at least get some good feeling stall behavior now, as opposed to before, when you had to have a collected trove of clever table values to insert at the boundaries collectively or individually learned over literal decades of messing with the sim. We're nowhere near that level of experience in the new sim yet.

The only thing I'm suggesting here is that the sim community keep a much more open mind than it has or tends to. If that's a crime, then I am absolutely guilty.

-Matt

 

4 hours ago, robert young said:

flying the Marchetti Sf260 which I co-designed back in 2004 for FS9 and updated in 2007

I bought this plane. THANK YOU. What a great plane. 

41 minutes ago, MattNischan said:

I'm not here to make enemies or to make people upset, and that's not what I'm implying at all here. My intent is always to be as positive as possible and as forward looking as possible.

The reality is, that while there may indeed need to be more parameters that need to be exposed at present (and I agree there do absolutely need to be) the reason that something like real wings snapping in a stall is because the air will never move over each wing precisely like the other and one end of the wing will almost invariably start stalling before the other end. The new flight model is light years closer to that type of reality than a table that looks up values and acts on the center of the plane, and way closer to the fidelity of XP, which we all praise for flight model. I don't that's controversial to suggest.

Yes, could you bend over backwards if you were the best of the best and make a flight model that exhibited some quality boundary condition flying in the table-sim model? Absolutely. But that's the thing, Robert, you are the best of the best. Most other even great flight models in FSX flew like you were driving a car on an imaginary road in the sky, totally uninvolved in the air around them. Even someone who is only a quarter as talented can at least get some good feeling stall behavior now, as opposed to before, when you had to have a collected trove of clever table values to insert at the boundaries collectively or individually learned over literal decades of messing with the sim. We're nowhere near that level of experience in the new sim yet.

The only thing I'm suggesting here is that the sim community keep a much more open mind than it has or tends to. If that's a crime, then I am absolutely guilty.

-Matt

 

OK. No worries. I am open minded and want nothing more than for the current system to work in practice. The current methodology for creating a wing drop due to assymetrical lift has the current system trigger a roll in a random direction. It could be left or right, but not because of any assymmetry but because it is random. So at the moment there appears to be no implementation of lift differences, only a generalised tendency to roll instability near or at the stall, which is sort of correct, but slight rudder should create an assymmetry. It will eventually but only if you apply absolutely mega amounts of rudder.

You can even apply rudder to the left which at the stall should encourage a left wing drop, owing to the inner wing losing lift, but it doesn't. It's still random. This is very similar to the No 1538 table in FSX which is CL Alpha Roll moment vs Yaw rate (in other words a wing drop - but it does actually tend to drop on the correct side). The two parameters are very similar, but at the moment the FSX one works better. And that isn't bias either, or being old fashioned. Or being stuck in the past. It actually does work better!

The other thing that FSX does better is that you can see very clearly the variables because you can present them in graph form rather than a row of rather cold text. The availability of a graphical representation speeds up flight model creation by a huge factor.

Maybe in time things like this will be implemented properly. I'm hoping they will be. If they are I will be the first to embrace the new aerodynamics.

Edited December 7, 20205 yr by robert young

4 hours ago, robert young said:

Well, forgive me but I'll link you to a video of me flying the Marchetti Sf260 which I co-designed back in 2004 for FS9 and updated in 2007.

Great flying and a wonderful video! Wish I had that plane for P3D4.5!

Al

1 minute ago, ark said:

Great flying and a wonderful video! Wish I had that plane for P3D4.5!

Al

The Lancair Legacy is not quite so lively but it can do aerobatics nearly as well as the Marchetti. Thanks for the kind comments!

2 hours ago, robert young said:

 

In a sense nearly all aircraft fly in broadly the same way. And making them do so is not really a process of convoluted systems, but an aptitude for observation and tuning fairly straightforward variables that contribute to authenticity.  And the majority of sim pilots just want an aircraft that responds believably to input or weather - and of course a reliable autopilot.

 

 

 

+1 on this

Whilst there are exceptions to this, rotary wing, vector thrust, canards, transonic flight and compressibility etc -  none of them are relevant to the current crop of aircraft.   Furthermore I think people generally do not expect 100% realism at the edge of the envelope. You do not necessarily expect a 747 doing an Immelmann to behave like it would if you were insane enough to try it in real life.

It is however reasonable to expect planes to behave realistically when flown in typical real world scenarios such as circuits or in cruise climb or in flare or ground roll out.

 

Can we just be clear here?

Rob you are an amazing at what you do.

FlybyWire you guys are too.

But this 1000 lift points is a little bit silly. Firstly it's 1000 for the entire plane, this often gets compared to another sim (I shall not name) which has only 600. But what people don't say or know is that is 600 per part, a wing could have 3 parts maybe more so that is 1800 points just for 1 wing.

All this talk of MSFS flight model is the future... Err no it's not, it's already beaten by again I won't bother naming it but the unsaid SIM is also older. 

And I am sure the next iteration of the better SIM is only going to get better. It might be the best for an MS flight Simulator and maybe in time it could be become the best, but all this time? So what have I bought? What am I playing? What am I trying to simulate with all these planes which don't work at all as they should not even to default standards.

The real cold hard facts that everybody is so adverse to saying, because it's so un ethical to tell the truth these days.... Is none of this matters, none of it until ASOBO wake up, and start taking the dynamics seriously. Till they get the SDK in order, till stop and hire people who know what they are doing. So the flight model is light years ahead of the fsx... Ok so why isn't the flying experience light years ahead in terms of flight dynamics? Why is things like engine temps, fuel flows, ITT temps not all spot on? 

Why is my plane bouncing around like super ball on game night with no wind, no forces acting on it? Why is the A320 modded or not still dropping a wing when landing? 

We can't even step back and say wow... We really got stiffed with the in modded defaults, but hey this add-on that add-on etc really fly's well, because nobody can make one without a working SDK. I mean the Long-EZ great plane but the flight model can't support Cannard even though the update before the latest said it was available in the SIM? Nothing about this, so why can't a "Light years ahead" flight model do this out of the box? I'm asking not digging?

With regards to rails or on rails. Funny isn't it, and kinda goes back to the whole want change resist change...when your planes bouncing up and down everyone is complaining, when you have it all trimmed in, power set and she sails through the sky nice and level with only the bit of +/- pitch change it's all on rails.... You know just like those big old airlines cruising at 350 all level and on rails. 

Ultimately the SIM needs a massive overhaul or it's going to remain a very pretty, but very unrealistic flight simulator. 

 

 

Edited December 7, 20205 yr by a321

4 hours ago, marsman2020 said:

• A slow VFR-only taildragger like the Savage Cub
• A standard 2-seat trainer
• A standard 4-seat GA aircraft
• A high performance single piston GA aircraft
• A twin piston aircraft
• A single turboprop aircraft
• A twin turboprop aircraft
• A business jet
• An airliner
• A fighter jet trainer

And a glider and helicopter...

5 hours ago, robert young said:

Well, forgive me but I'll link you to a video of me flying the Marchetti Sf260 which I co-designed back in 2004 for FS9 and updated in 2007...

 

 

 

Man I used to LOVE doing aerobatics, hammerheads and precision inverted runway passes in that plane!

(you'll have to insert your own soundtrack)

 

Edited December 7, 20205 yr by Stoopy