Concerned about KSP2 aerodynamics

[OHara]

On 10/29/2022 at 6:02 AM, eekee said:

if it's computing the wing area, why not instead make the more complex but realistic calculation of overall lift based on vehicle shape, as the FAR mod does?

FAR itself doesn't try to estimate the performance of wings based on their shape; it has numbers in the file FerramAerospaceResearch.cfg for each type of wing or control surface to describe its aerodynamics.  I looked through the code a couple years ago to try to understand why.  The analysis of the whole-body shape of the craft is too coarse to notice the critical details of a wing, such as the blunt forward edge and sharp trailing edge, that give it so much lift.  Wings and control surfaces are largely computed separately from the fuselage.  A control surface buried inside the fuselage will still steer the aircraft in FAR.

FAR doesn't do a low-level physics simulation, but uses heuristics developed by aeronautical engineering, so it does a good job on craft shaped like normal aircraft and rockets.

KSP1 does automatic computation of aerodynamic properties of each non-wing part, including parts from modders who provide no aerodynamic information, and caches the results in PhysicsDatabase.cfg.  Then has rules for cancelling some of the drag when you attach two parts.   So many people use stock KSP1 for things like challenges, that we ended up learning and sometimes gaming those rules.  I would rather if KSP encourages use to learn more real aerodynamics, than the rules of a video game.

In hindsight, I think it would have been better if KSP1 took the whole-craft approach of FAR (re-computing every time you eject a stage or un-dock something).   I hope KSP2 was able to take advantage of the benefits of hindsight.  It might not have been easy.  We'll see soon enough.

There are aspects of KSP1 stock aerodynamics that make the game easier than with FAR.  Lift and drag unrealistically increase at low speeds, allowing our spaceplanes to take off and land on a 2.2km runway without extreme measures.  The stall of wings is very soft and forgiving.  These adjustments affect curves defined in *.cfg files, so KSP1 players (using computers) can restore realistic low-speed flight with configuration-file changes.  I would think that if KSP2 was able to implement more realistic aerodynamics, they could adjust the lift-versus-speed curves in a similar way, to make a playable game.

[Dinlink]

This is a supersonic, probe-controlled airplane drone for simple science exploration and reconnaissance that was made by me using FAR tools to flight with FAR aero physics :

 

I dont know if you find it uncreative or unreasonably difficult to design. It flights just fine, without any SAS, just trimming. It is gliding at relatively low speeds. It was designed to be performant for supersonic, so maybe not the best example for gliding.

The first pictures shows the FAR overlay, that gives a lot of insight on how the plane is behaving, the most important parameter for me is the Lift-to-Drag ratio L/D, which for this glide is about 7... Real jet airliners and combat airplanes normally get L/D about 20... and specialized planes can even get to the range of 30... But Kerbals being Kerbals, we can assume they don't use airfoils, but flat sheets of metal... which could justify the crappy performance of their planes xD...

The second picture shows more of the details of the plane. I think a realistic aerodynamic model does not hinders the player experience, but enriches it.

I do agree that for KSP2, the user interface should be a lot better than the presented by FAR mod... and approachable tutorials and documentation should be included to introduce the player to the basics of aerodynamics and airplane design, like it probably will be with rocket design and orbital mechanics, which are concept about the same complexity or even of greater complexity than the aerodynamics knowledge needed to design airplanes in KSP with "realistic" aerodynamics physics engines like FAR.

Edited October 30, 2022 by Dinlink

[Guest]

11 hours ago, Bej Kerman said:

I've gotten the gist of it - KSP 1 aero works "well enough" and isn't worth changing.

That mindset has never led to anything improving ever.

Since scrolling up seems to be beyond your capabilities here:

18 hours ago, Master39 said:

I don't use FAR, but I want a more realistic approach.

FAR is a bad target for a game because is pure realism without any consideration for gameplay. KSP2 should aim for a better model than what is stock now, but still something way more forgiving than FAR while also keeping in mind that they're erasing a decade of intuition in building planes in KSP while changing model and that requires a ton of in-game tutorials and explanation to get people up to speed as fast as possible on the differences.

And on top of that there's my old argument that realism and the educational nature of the game should take the backseat when they get in the way of gameplay.

For the aero model I think that while a more realistic approach would improve the game in low speed scenarios there's plenty of unintuitive stuff happening at supersonic speeds and KSP scale start being a problem at hypersonic speeds. That's especially where playability and gameplay should take the lead.

 

9 hours ago, Dinlink said:

I do agree that for KSP2, the user interface should be a lot better than the presented by FAR mod... and approachable tutorials and documentation should be included to introduce the player to the basics of aerodynamics and airplane design, like it probably will be with rocket design and orbital mechanics

And that's the crux of it, mostly.

Even if we want to forget about gameplay and go for maximum realism using KSP educational purpose as an excuse the game should totally teach you those things.

A realistic aero model with a mysterious V*L/D number among 20 other similarly badly labeled values in a giant window taking half of the screen isn't teaching you anything. All the "educating" part happens off screen, off the game, while the games merely rewards you for knowledge gained somewhere else.

That's terrible whether you want to focus on gameplay or the educational aspect.

 

When you propose "The game should have FAR as stock" people don't think about the advantages of a better aero model, they just think about all their planes suddenly not working anymore and the screen being filled with numbers they don't understand and don't have the needed context to do the research on their own.

[Bej Kerman]

17 minutes ago, Master39 said:

For the aero model I think that while a more realistic approach would improve the game in low speed scenarios there's plenty of unintuitive stuff happening at supersonic speeds and KSP scale start being a problem at hypersonic speeds. That's especially where playability and gameplay should take the lead.

Orbital mechanics was unintuitive before playing KSP. That doesn't mean KSP should've ran in the opposite direction and made Star Wars esque space operas possible. Instead of running away from such "unintuitive" stuff, KSP should teach the player how to design their planes properly instead of mollycoddling them and removing challenges.

14 minutes ago, Master39 said:

realism and the educational nature of the game should take the backseat when they get in the way of gameplay.

They shouldn't. The gameplay is learning how to design rockets and planes in a realistic environment.

15 minutes ago, Master39 said:

people don't think about the advantages of a better aero model, they just think about all their planes suddenly not working anymore

And? They should build better planes instead. They can mourn their old borderline physically impossible abominations when they're done designing physically plausible aircraft.

16 minutes ago, Master39 said:

screen being filled with numbers

That's a UX/UI problem, not an aero problem.

[Guest]

54 minutes ago, Bej Kerman said:

That's a UX/UI problem, not an aero problem.

That's where you're missing the point.

When you say they should implement something like FAR, in that image you're giving people all the UX/UI and tutorial problems of FAR are included.

FAR doesn't teach you anything, just changes how thing works without giving you the documentation to understand what's going on.

That's terrible from both the educational and the gameplay purpose the mod can have.

 

The FAR experience for most people is "Your planes don't work anymore, close the game and get back after you studied elsewhere how aerodynamics works"

And I'm not talking about flying bricks or pianos, plenty of planes that look like planes IRL can stay in the air only because an insane amount of corrections from the onboard computer.

 

1 hour ago, Bej Kerman said:

The gameplay is learning how to design rockets and planes in a realistic environment.

Hardly given that all that learning happens outside of the game.

KSP is a game in which you design and build rockets using knowledge you already acquired somewhere else. If you don't have the required knowledge you have to close the game window and go search for that knowledge and and get back to the game when you've learned how things work.

The fact that KSP is a fun game make that studying part outside the game barely worth it.

[Luriss]

1 hour ago, Master39 said:

Hardly given that all that learning happens outside of the game.

Happened*

Considering KSP 2 is going to have animated tutorials this is bit of a moot point.

[Xelo]

Honestly really interested to watch KSP's animated tutorials on aerodynamics. I've always sucked at plane-looking planes even in stock, much less make a supersonic teapot. xD 

[Guest]

2 hours ago, Luriss said:

Happened*

Considering KSP 2 is going to have animated tutorials this is bit of a moot point.

If we're talking about the experience with FAR we have to take into account the experience people had with far up until today.

I'm all in favour of a more realistic approach to aerodynamic, but I would never use FAR as an argument for it.

To me FAR doesn't mean "realistic aerodynamic" but "your planes are not going to work anymore and nothing and no-one is going to tell you why".

And no, I'm not talking about making bricks fly, let's not kid ourselves and pretend that you can just make something barely resembling a plane and that would automatically be both stable and maneuverable at all speeds and altitudes.

Edited October 31, 2022 by Master39

[The Aziz]

Too bad you all want this game to be a lot less fun that its predecessor.

[Bej Kerman]

4 hours ago, Master39 said:

When you say they should implement something like FAR, in that image you're giving people all the UX/UI and tutorial problems of FAR are included.

Irrelevant.

4 hours ago, Master39 said:

5 hours ago, Bej Kerman said:

The gameplay is learning how to design rockets and planes in a realistic environment.

Hardly given that all that learning happens outside of the game.

Talking about KSP 2, not KSP 1; Irrelevant.

56 minutes ago, Master39 said:

To me FAR doesn't mean "realistic aerodynamic" but "your planes are not going to work anymore and nothing and no-one is going to tell you why".

Given how terrible KSP 1 aero is, both mean the same thing.

4 hours ago, Master39 said:

plenty of planes that look like planes IRL can stay in the air only because an insane amount of corrections from the onboard computer.

Literally wrong. I['m not saying I'm disagreeing, I'm saying this literally isn't how planes work. All you need to keep a little glider or a Boeing airliner in the air is to hold the yoke back a bit so that the ailerons keep the plane from nosing down. Autopilot functions like holding a heading or pitch and maintaining a certain speed are usually only used for insanely tedious and simple tasks like keeping a plane straight on an intercontinental path.

57 minutes ago, Master39 said:

And no, I'm not talking about making bricks fly, let's not kid ourselves and pretend that you can just make something barely resembling a plane and that would automatically be both stable and maneuverable at all speeds and altitudes.

KSP 1 lets you do that. Can make a suborbital plane using two fairly subsonic looking wings and mostly circular fuselages, and it remains maneuverable at most altitudes. It's silly and daft, and it shows that KSP 1 used to be barely a mobile game before the devs tried patching a full simulator into it. KSP 2 is meant to be more than a mediocre game we all play anyway because there's nothing better to play (besides SR2 that's still fairly barebones in terms of game and community or something like Orbiter that's too hardcore).

25 minutes ago, The Aziz said:

Too bad you all want this game to be a lot less fun that its predecessor.

 Noted.

[Guest]

1 hour ago, Bej Kerman said:

Can make a suborbital plane using two fairly subsonic looking wings and mostly circular fuselages,

From saying that everyone not wanting FAR wants to "fly a literal grand piano" to this is quite the jump.

So we aren't talking about just making something that resembles a plane but also learning how differently shaped wings work at different speeds.

Still, tutorials first, if something is not explained in the game it should not be in the game. I don't care how realistic a feature is, if it's not 100% explained in the game it shouldn't be in the game.

[The Aziz]

So I won't be able to fly a train?

Because it doesn't match with very restrictive areodynamic rules?

 

Edited October 31, 2022 by The Aziz

[Bej Kerman]

33 minutes ago, Master39 said:

Still, tutorials first, if something is not explained in the game it should not be in the game. I don't care how realistic a feature is, if it's not 100% explained in the game it shouldn't be in the game.

Maybe just move the boulder instead of throwing away 200km of train tracks? Why is it that you consider "demolish the integrity of the game" to be a solution worth considering when compared to "just add a tutorial"?

24 minutes ago, The Aziz said:

So I won't be able to fly a train?

Because it doesn't match with very restrictive areodynamic rules?

To reiterate

Shouldn't be possible

[KSACheese]

I really think these concerns can be alleviated with the development of mods. Sure, it will probably take a while for a comprehensive atmospheric simulation mod to be developed, but it will eventually be released, that I am certain.

As much as I respect FAR for being a fantastic, challenging mod, one that I have used in the past, I would agree with some here that having a more realistic flight model (while, again, a great challenge) would be counter to the vision of accessibility that the KSP2 team seems to have.

Personally, I have been playing with the stock aerodynamics model for a little while now, and while yes, it is simplified, it is more fun for me to play with. I think it is more than acceptable to use a similar aero model in KSP2 as a baseline.

Also, the train bit is also quite funny, but brings up a point - a more advanced aerodynamics model would be prohibitive to the core KSP experience in a way, as, at least to me, KSP has always had a base of goofiness and a "lets see if this thing works" attitude. I think with a stock aero model like FAR, this wouldn't be (as) possible.

I do respect that people would want a more advanced aero model, since people have their preferences and their styles of play. I just think that starting at an accessible level and being able to mod it to a more advanced one would be a better way to go about it.

I don't think its really anything to be really concerned about, though. Advanced aero will come, one way or another.

[Bej Kerman]

1 minute ago, KSACheese said:

As much as I respect FAR for being a fantastic, challenging mod, one that I have used in the past, I would agree with some here that having a more realistic flight model (while, again, a great challenge) would be counter to the vision of accessibility that the KSP2 team seems to have.

Accessibility does not have to go against realism.

May I ask that we see less black and white fallacy here? 

2 minutes ago, KSACheese said:

Also, the train bit is also quite funny, but brings up a point - a more advanced aerodynamics model would be prohibitive to the core KSP experience in a way, as, at least to me, KSP has always had a base of goofiness and a "lets see if this thing works" attitude. I think with a stock aero model like FAR, this wouldn't be (as) possible.

Goofiness takes a backseat to realism.

Edited October 31, 2022 by Bej Kerman

[t_v]

34 minutes ago, Bej Kerman said:

59 minutes ago, The Aziz said:

So I won't be able to fly a train?

Because it doesn't match with very restrictive areodynamic rules?

To reiterate

Shouldn't be possible

This is the most Kerbal thing I have seen in a while, and removing the ability to make stupendously silly things is just removing the core of the game. 

35 minutes ago, Bej Kerman said:

Why is it that you consider "demolish the integrity of the game" to be a solution worth considering

Anyways...

18 hours ago, Dinlink said:

I dont know if you find it uncreative or unreasonably difficult to design. It flights just fine

Sorry for not responding to this earlier. That does indeed look like quite a nice plane. I didn't mean to say that FAR's model ruins the creativity of builds, just constrains it to work in a smaller space of possibilities. Your point about the data table is an example of how KSP 2's aero model should be designed in a way that completely new players don't need such things to understand what is happening to their planes. I think that requires some adjustment of the parameters and physics, but maybe as you said, it is a matter of UI. 

Just a question: Would you be able to make the same place with forwards-sweeping wings? A lot of the more interesting-looking plane designs in KSP 1 use wings at strange sweep angles. And, if you are interested in an even harder challenge, could you try to approximate @The Aziz's train in FAR? It doesn't have to be stable, it just has to be able to barely control its flight trajectory. 

4 minutes ago, KSACheese said:

Also, the train bit is also quite funny, but brings up a point - a more advanced aerodynamics model would be prohibitive to the core KSP experience in a way, as, at least to me, KSP has always had a base of goofiness and a "lets see if this thing works" attitude. I think with a stock aero model like FAR, this wouldn't be (as) possible.

This is what I think too - KSP is at its core a physics simulator that has tolerances to let you do crazy things. The realism can led to the craziness, but it shouldn't be there to reduce it. I'm hoping that FAR still allows for this (I've been playing with FAR the last few months and so far, I haven't gotten a train to fly), but taking away the goofiness from less experienced players would just hurt the game more than the realism helps. 

[Alastor the Kerbal Demon]

Personally I think a balance between realism and gameplay is important for KSP 2, it is very crucial that realism doesn't cause someone to grow angry towards the game because whatever their designing isn't working and they can't find any digestible information for a specific problem they may have, but realistic enough that more serious users are able to accurately simulate their own designs in a more realistic manner.

Like what KSACheese was talking about, specifically the fun aspect of KSP, keeping the game engaging for people of all ages is just as important as a realistic game, and it would defeat the whole point if we now need a mod for a less complicated aero model.

Though this is just my own take on this question, I feel like having the game appeal to a wider audience would provide a greater impact, and also give a fun and enjoyable experience for everyone, I personally was inspired to play KSP after watching Thor LP's Insane Rockets Division series (If anyone still remembers that??)

 

[Bej Kerman]

24 minutes ago, t_v said:

This is the most Kerbal thing I have seen in a while, and removing the ability to make stupendously silly things is just removing the core of the game. 

No it isn't.

24 minutes ago, t_v said:

Anyways...

Noted and disregarded.

[Bej Kerman]

12 minutes ago, Alastor the Kerbal Demon said:

it is very crucial that realism doesn't cause someone to grow angry towards the game because whatever their designing isn't working and they can't find any digestible information for a specific problem they may have

Perceived problem: the game is too realistic

ACTUAL problem: no digestible information on troubleshooting designs

[KSACheese]

30 minutes ago, Bej Kerman said:

Goofiness takes a backseat to realism.

I respectfully disagree to this sentiment. KSP, in its base form, has never been about strict realism. Personally, I think that most of the realist side of things has come from the modding community, with mods like RSS/RO, FAR, Principia, Kerbalism, and so on. I mean, look at the Kerbals. If they aren't a physical embodiment of the goofiness that KSP has in all aspects of the game, I don't know what is.

30 minutes ago, Bej Kerman said:

Accessibility does not have to go against realism.

May I ask that we see less black and white fallacy here?

I see your point, but I would say that realism, in some aspects, hinders accessibility. I've struggled quite a lot with FAR as an experienced KSP player myself, so being tossed into a realistic aero experience as a new player may be overwhelming. This isn't black and white, and I thought I made that apparent in my original post.

To be clear, I think that advanced aero has a place in KSP. There is a huge portion of the community that plays KSP for realism, and I appreciate that aspect of the community. I just don't think it should be something that is present in the base game. As it is a job for the modders now, I think it should be in KSP2.

Edited October 31, 2022 by KSACheese
Clarification

[The Aziz]

1 hour ago, Bej Kerman said:

To reiterate

Shouldn't be possible

You just made the game straight up boring for thousands of players. If the only other possibilities are perfectly streamlined rockets and aircraft, a huge chunk of what was considered "fun" by many, is gone. This is not a hardcore simulation, never was, and never should be. You could keep arguing that the Kerbal universe should be scaled up to human proportions because that's realistic. Goofiness should come in pair with relative realism to keep the "game" part in the game.

[Bej Kerman]

Just now, The Aziz said:

You just made the game straight up boring for thousands of players. If the only other possibilities are perfectly streamlined rockets and aircraft, a huge chunk of what was considered "fun" by many, is gone. This is not a hardcore simulation, never was, and never should be. You could keep arguing that the Kerbal universe should be scaled up to human proportions because that's realistic. Goofiness should come in pair with relative realism to keep the "game" part in the game.

Smaller universe to make trips shorter =/= letting you turn off gravity with wibbly devices so you can make your ships look goofier with no dV expense

[t_v]

25 minutes ago, Bej Kerman said:

Smaller universe to make trips shorter =/= letting you turn off gravity with wibbly devices so you can make your ships look goofier with no dV expense

The length of times doesn't make a difference with time warp. The reason the universe is smaller is because it lowers delta-v numbers and makes the game easier. Unrealistically so? Yes, completely. The type of rockets that can get to space in KSP 1 are ludicrous from a real science standpoint, because they are tiny and have next to no delta-v. You can do a whole Duna mission with less delta-v than it actually takes to get to orbit. The point is that realism is put aside so that people can actually experience the game. I've played with RSS, and I liked the experience, but if I were to ask my friends who don't even know that you have to go sideways to orbit to play RSS, they would never experience the game. The difference between FAR and stock isn't the difference between gravity and no gravity, it is the difference between RSS and stock scale. 

Edited October 31, 2022 by t_v

[Alastor the Kerbal Demon]

1 hour ago, Bej Kerman said:

Perceived problem: the game is too realistic

ACTUAL problem: no digestible information on troubleshooting designs

The problem is condensing pages of information regarding the specifics on air resistance, displacement, etc. Is very very difficult, as these require calculations that people without a high school education (Or college) would not have the skills necessary for algebra, not everyone that plays KSP is going to be a wizz at algebraic formulas.

Most people want to know why, and how this small design change could effect aerodynamics, and just having a tutorial only going "If Problem A occurs, do this" would not answer these questions, but then having these being answered would clutter the tutorial up with multiple subjects

[regex]

The current crop of realism-askers really need to step up their game, how can you have realistic aerodynamics without a realistic planet?

In terms of gameplay I think the simulation in KSP1 worked quite adequately. People could mess around but it felt okay in general. For anyone who needed more there was FAR. Contrary to anything I've said in past years, part of the charm of KSP is being able to mess around, to build flying trains.

Now that I think about it the biggest thing KSP could borrow from FAR is the voxel-based drag model, that would help with obscured parts and custom builds, batteries tucked between parts, things like that (is that even an issue? I haven't played in years). But I also recall Ferram musing (again, a long time ago) about how to decide what was actually a wing for the purposes of FAR, including wacky part orientations and things like that. I think certain things are just going to be unsolvable and, quite frankly, this game isn't a piece of hardcore aircraft design software.

Anyway, the stock KSP1 aero simulation is generally fine for a light physics sim and I don't see a need for more with KSP2. Get FAR if you want to go further.

Edited October 31, 2022 by regex