Please Add Realistic Aerodynamics (FAR) and Perhaps Procedural Nozzles for Optimizing at Different Pressures

[Missingno200]

On 7/17/2021 at 10:49 AM, AlmightyR said:

 

It's not impossible: Multiple other games have done it. (War Thunder, ...

I feel the unnecessary compulsion to point out that War Thunder does not in fact have multiple aerodynamics models applying themselves differently to every plane, they have 3 different sets of stats for each plane for each mode, that is slightly tweaked to give better balancing and closer to realistic flights. In other terms:It's a lot like sticking 3 extra pairs of wings in your fuselage for an "arcadic" experience, and just removing them if you wanted a "realistic" experience without FAR.

Edited July 21, 2021 by Missingno200

[Incarnation of Chaos]

9 hours ago, Lt_Duckweed said:

 

It's ok to have a high skill ceiling, but it needs to be based on clearly designed and laid out game mechanics, not strange arcane exploits that live only in the minds of vets and in secluded discord servers.

 

Gosh i wish i had more than one like, because this is a golden nugget for the archives.

[OHara]

On 7/19/2021 at 9:53 PM, TLTay said:

Would KSP2 need a different model for hydrodynamics? Or would it still be able to use the aerodynamic model for underwater work?

A single model would work for atmosphere and ocean, taking into account the different density of air, water, etc.

There is tricky physics at the air-water surface: wave drag and hull speed and planing.   KSP1 ignores all of that, except for buoyancy..

 Water is about 800× as dense as air, so at a given speed, the water forces on wings and control surfaces would be 800× as strong as aerodynamic forces.  More likely, the speeds through water lower by a factor of 30.  (KSP1 reduces the hydrodynamic forces, by the factor  buoyancyWaterLiftScalarEnd = 0.025 in physics.cfg, so water forces are only 20× as strong as air forces, at the same given speed.)

KSP1 also has several other adjustable parameters in physics.cfg.  Some are apparently for making splashdowns look believable.  There is an artificial extra drag that, as I remember, was added to more quickly bring the velocity of a capsule below 0.1 m/s so it could be recovered.  Better, I think, would be to let us recover 'splashed' craft while they are realistically drifting through the water.   FAR is also rather messy and arbitrary in its treatment of water.

I was frustrated trying to deal with these more complicated rules when building submarines  with KSP1.

Edit:  If buoyancy from hydrodynamics also applied to aerodynamics, that could be a natural way to simulate bases floating in Eve's dense atmosphere.

A FAR-like evaluation of the shape of the entire craft,  might also evaluate the buoyant volume.   (Finding the centre of buoyancy at each depth, is similar to the evaluation of sections that FAR does to evaluate area ruling.)
Players could then clip parts to express the tight packing of a dense submarine, or sparsely fill a fairing to make an Eve airship.

Edited July 21, 2021 by OHara
more thoughts on buoyancy

[AlmightyR]

On 7/19/2021 at 11:30 AM, Deddly said:

1) KSP should have FAR and only FAR from day one

2) KSP should have the regular aero model and FAR should be added later as a toggle

3) KSP should have the regular aero model only, and FAR should remain a mod as it is today.

Agreed.

But there is a forth alternative some have argued for or painted others as arguing for:

4) Regular aero model, AND a FAR-like model as a toggle, at launch rather than later.
(overkill, IMO)

On 7/19/2021 at 11:30 AM, Deddly said:

recognise that each alternative is just as valid as yours

On 7/19/2021 at 11:30 AM, Deddly said:

so I would be over the moon with either suggestion 1 or 2

I disagree that each of the alternatives are just as valid as any other, and have largely already pointed out and explained their differences, advantages and disadvantages, in my opinion.

I disagree particularly with #1, as, to me, that is by far the worst option. The only advantage I can see with it is that it's the best for people who like a FAR-like model (and to be clear, I'm one of them; FAR is a must-have on my KSP1; but that's KSP1; a game I also played without FAR for years); but then it's cancelled out by being the worst for everyone else. And at that point, I see no good side to it.

But I thank you for at least presenting your point in a nuanced, respectful manner, without misconstruing anyone's argument.

And I'm along side you in that I would be ecstatic with happiness if a FAR-like model was implemented a secondary model with a toggle on the official game. Exactly as you said: Probably not going to happen, but we can dream.

[AlmightyR]

On 7/19/2021 at 3:45 PM, Lt_Duckweed said:

2 wildly different aero models supported at the same time has a pretty much 0% chance of happening.

Almost zero is not zero. Even the "almost zero" part is debatable: It really depends on the community and the community pressure for it.

On 7/19/2021 at 3:45 PM, Lt_Duckweed said:

1. Developing 2 different models costs dev time (and therefore money)

Agreed. But that is not a problem as long as the devs expect to profit more out of implementing than they expect the cost to be. Again, community pressure for it is a big factor there.

On 7/19/2021 at 3:45 PM, Lt_Duckweed said:

2. 2 different methods have to have all their interactions regression tested with every release (costs money)

Technically true, but functionally false, IMO. Automated testing is an integrated part of development nowadays and doesn't need to be reimplemented from scratch for every release. It has a very small footprint overall, and a microscopic one between releases, to be honest.

On 7/19/2021 at 3:45 PM, Lt_Duckweed said:

3. 2 different methods means twice as much surface area for bugs. More bugs to fix, harder to diagnose (costs money).

Again, technically true, but functionally false. And if you put it into perspective, since a large portion of KSP1 players use FAR, a FAR-like secondary model is pretty much guaranteed to be implemented to KSP2 through mods if it's not done by the official team, and so you're still dealing with the same problem here, because the "surface area for bugs" is equal at best, and greater at worse; as is the difficulty of diagnosis; through an external 3rd-party implementation. This is basically hot-potato'ing the problem to a different set of hands, and then claiming "not my problem". Not a realistic  criticism of the idea, IMO.

On 7/19/2021 at 5:53 PM, TLTay said:

Would KSP2 need a different model for hydrodynamics? Or would it still be able to use the aerodynamic model for underwater work?

As @OHara already pointed out in a different way, most of the interactions are dealing with fluid dynamics (name is misleading; it applies to gasses too); and so the same model applies well for interactions with both water and air for the most part, with some caveats at special points like buoyancy and density.

[GKSP]

My (somewhat unorganized) two cents on the topic.

Whatever Take Two decides to implement, there should only be one stock aero model.  On the dev team, and on the players, that would make everything easier to facilitating development, creation, usability, and sharing of crafts.  That being said, the aero model should not be what's implemented in KSP1.  I've only recently dabbled in FAR, but one of the things that stood out to me was the drag model.  At a minimum, KSP2 should have a more realistic drag model that gets rid of weird exploits and doesn't force you to think about drag in a non-intuitive way.

An example of what I mean about the drag model:
When making replicas, there is often the need to clip parts everywhere to get a shape just right or for whatever reason.  Or it could be a customized cargo bay for a spaceship, where parts that should be shielded from drag are experiencing drag.  This causes things like people creating rather OP engines (BDA Saturn) when in reality, instead of extra thrust, there needs to be less drag.  As a side note, I'm pretty sure this is what the node attachment drag thing was meant to fix, but obviously people surface attach things too.

On the OP engines such as the Saturn from BDA, by countering excess drag (that shouldn't even be there) with thrust, the thrust makes other aspects of the craft which rely on TWR very unbalanced.  I, for one, tend to believe that more realistic = more balanced.  Not only that, but enhancing realism should make some things intuitive for new players (attaching parts to a backwards node and flipping them around is one example of something new people would not think of, and is only done as a workaround/cheat of the physics system).

With all I've said about realism, I don't think that the FAR-level realism is necessary.  Sure transonic flows and whatnot may be an interesting read, but most people will be interested in going to space.  They'll probably experience it for a couple seconds when their rocket with 50 boosters first launches on a Mun speedrun challenge.

Worst comes to worst, the aero model is like KSP1 and a FAR type mod will come out for KSP2.

[kspnerd122]

Adding FAR would probs just ruin the game for people who dont like FAR, also, it means u have to fairing everything, etc, it feels worse to play with FAR

[mcwaffles2003]

On 8/4/2021 at 10:08 PM, kspnerd122 said:

it means u have to fairing everything

How is this bad?...

Wind forces should be able to tear a craft apart if it has a bunch of stuff dangling off of it

[kspnerd122]

Well, for people who dont use FAR already, I could see this being INFURIATING

the heatshield wing thing is literally only because that was NOT the intended use of that

Why make everyone have to learn FAR in order to play KSP, I get it, you like FAR, but FAR should stay a MOD not a feature in the basegame, since then at least its optional, as opposed to being forced on for all players, regardless of how many people dont want it

[Bej Kerman]

9 minutes ago, kspnerd122 said:

Well, for people who dont use FAR already, I could see this being INFURIATING

the heatshield wing thing is literally only because that was NOT the intended use of that

Why make everyone have to learn FAR in order to play KSP, I get it, you like FAR, but FAR should stay a MOD not a feature in the basegame, since then at least its optional, as opposed to being forced on for all players, regardless of how many people dont want it

FAR does not massively spike the difficulty of the game, you simply have to get used to having to put more effort into ridiculous stunts. Besides, you'll have more things to be concerned about, like three-body orbits around Rask and Rusk, and interstellar flight.

[kspnerd122]

Interstelar is fine, I do that with KSPIE every 3 days or so.

[cfds]

The only excuse for KSP1 stock aerodynamics is that the game was developed by amateurs. Since KSP2 is developed by a professional team, I expect a professional aerodynamics model.

The "worst" consequence of a FAR-like stock model is that rockets look like rockets and supersonic planes look like supersonic planes...

[KerbalSofaProgram]

On 8/18/2021 at 8:19 AM, cfds said:

The only excuse for KSP1 stock aerodynamics is that the game was developed by amateurs. Since KSP2 is developed by a professional team, I expect a professional aerodynamics model.

The "worst" consequence of a FAR-like stock model is that rockets look like rockets and supersonic planes look like supersonic planes...

This is the correct answer! No really though, I just read through this whole thread... it wasn't fun, but I'm pretty invested in the implementation of aerodynamics in KSP2. I think it's important that they ditch the "drag cubes" (?), and get on with a more accurate model. 

I'm quoting you here just to let you know I'm addressing you.

On 7/21/2021 at 8:19 PM, AlmightyR said:

Almost zero is not zero.

Tell me if I'm wrong but your main argument for KSP1 aerodynamics has been crowd appeal and accessibility, right? I'm going to assume you said "yes" and move on. Now I'm not always great at articulating my thoughts, but I'll try... First off, I'm going to shelve the discussion about two flight models, I think it's a bad idea, or not even a bad idea, just not a "worth it" idea, at least. I'll get back to it later if I still feel like typing. Anyways!!! I want to establish that realistic, does not mean difficult. Sure, in stock you can build things you can't in FAR, but it doesn't really... Make sense? Why things sometimes act the way they do in stock? On that note FAR has always felt "easier" for me, because it's a little more realistic, you know, so it's a little closer to reality, which I'd say I'm familiar with. What I'm saying is stock is easier because it's more forgiving sometimes, but harder because it just doesn't make logical sense... And with something like that, you can't just say, "well it's what people are used to so let's keep it". No if they can I believe they absolutely need to find a model more like FAR. Doesn't have to be the same thing, but it has to be as close to reality as they can make it, without making it frustrating or thinking they need to make a sci-fi style super simulation.

Ok on to the two flight models one game... it's just not going to happen. No matter if I agree with you on this or not. It's not going to happen. I suspect we all know it's just not going to happen... So let's all have a productive conversation. If a developer happens to read this we ought to make it easy to see our consensus on the topic... So let's try to get to one. I'm going to do that in another post so it looks better.

EDIT: the forum won't let me make two separate posts. Probably a good idea. But I want to bring attention to the last part so let me put some in bold. And underline it. There.

1. Is there anyone who's happy with KSP1 aerodynamics? Not just content, but it's perfect and shouldn't be changed AT ALL. Like a shark...

2. Who wants to see a more true-to-life flight model, so long as it's well explained and doesn't detract from a fun experience?

3. Who wants as much tricky realism as possible? What this option might include I don't really know... I'm not a professional.

So I imagine most of you thought #2 sounded pretty reasonable, right? If you picked #3 you're going to have to tell me what sort of things you have in mind... cuz i dunno fluid dinamiks. If you picked #1, I suggest you forget about KSP2 and stick to KSP1, and I take your choice personally and my day is ruined. But really why wouldn't you want something better? Anyways so I'm imagining most people would say #2 or #3, and I'm hoping the developers are going for #2... Going for #2. Funny. But #2 is the most reasonable choice and I doubt anyone disagrees. How far #2 goes towards #3 is a little bit harder to figure out, and lends some credence to AlmightyR's argument. Though I think two whole different flight models is the wrong way to go about it. Either way I don't think KSP1 cuts it... I honestly think if everyone had played KSP1 for the first time with a modern base game + a modern FAR nobody would have any complaints, but to switch over after you've gotten used to stock is different. And I just don't think being used to something is a reason to keep it when you have something better. So if I could pick I'd DEFINITELY pick a professional version of FAR over a redone KSP1 version of aerodynamics.

Edited September 17, 2021 by KerbalSofaProgram

[moar ssto]

I don't think there is much issue with stock aero model, except for the need of more informantion that allow users to better understand their craft global behaviour, and more stability(removing load/computer caused inconsistencies, like fairings not occluding a a part during one loading, while occluding the same part on the same craft when loading the same save next time again).  Stock aero model,being not realistics and intuitive, encourages you to do more science in the game, to summerise behaviours and interactions, and encounrages looking at data, like drag cubes, float curves, lift coefficients ect, numerical wing loading, rather than intuitive things like how pointy/streamlined it looks, and whether wings look big or small., though I won't disagree that for starters, this feels more blackbox and frustuating than a more shape dependent system.  Stock aero model also encourages you to use some level of matheatical models,  complicated enough to let you ponder for a while sometimes, while simple enough that you don't need intense extra game computational resources to do optimisations.  A FAR like shape dependent system could preserve some of these, but I think the complexity involved, for those that are interested in deep optimisation, is just way too much, directly working with shapes, will involve a lot more variables, and the math propabbaly won't be as easy as something you can do on the back of few sheet of paper and a calculator, but would probably involve decent amount of programming, and possibly modding to extract detailed craft shap from the game. 

    As for aero exploits, and peculiarities, as long as being fully aware of them, a player can simply choose to view them as pathways for optimisation, and use them, or bugs that hurt gameplay/realism, and avoid them. These things, as long as the player being aware of them, is easy to control to whether include or not include in a craft. 

    There is, however,  a much bigger problem with ksp's physcis, is it's inconsitency and errors, which would be far more manecing than now if when we introduce interstellar distances, and celestial bodies with more extreme physical characteristics. (orbital drifting, loading under terrain/mid air, terrain bugging out for distant bodies, structure of craft degrading over times you load it ext.)  If you have played with some real scale exoplanet packs, for which we are talking about distace measured in lightyears, you will see, how annoying, and sometimes deadly, these issues are.

Edited September 24, 2021 by moar ssto

[pandaman]

What I want from the 'aero model' is that it behaves in a believably realistic way. Exactly HOW they achieve that doesn't really concern me greatly, beyond just being quite interested in what method they use.  They could use a perfect full voxel type model and still cock it up with the implementation.

KSP1s method 'works' ok, it is obviously not overly realistic, but it does the job and is adequately useable.  Whether they use FAR, a refined version of the KSP1 model, or something else entirely isn't the issue for me, it is how it plays.

[Bej Kerman]

On 9/17/2021 at 10:52 PM, KerbalSofaProgram said:

So if I could pick I'd DEFINITELY pick a professional version of FAR over a redone KSP1 version of aerodynamics. 

Completely agree. Using the stock aero in KSP 2 would be like using complicated SOI arrangements for Rask and Rusk when there's a far better system (three-body) that can be used. There's a certain finess FAR has that's lost on the stock model - why go back to vessels not updating outside of staging events and a soupy atmosphere when we could use a model that just feels good and isn't as prone to issues? FAR isn't difficult beyond having to get used to an aero model that isn't all shambles, I've tried it myself and it clicked. Aircraft at low speeds no longer felt like they were on rails and supersonic craft didn't feel as unnaturally tough. I'm not sure what others see in the stock model other than not having to put effort into building atmospheric vessels. Again, why use something that just works when there's stuff that works far better and is a lot more dynamic? 

Edited September 24, 2021 by Bej Kerman

[KerbalSofaProgram]

On 9/24/2021 at 1:36 PM, pandaman said:

What I want from the 'aero model' is that it behaves in a believably realistic way. Exactly HOW they achieve that doesn't really concern me greatly, beyond just being quite interested in what method they use.  They could use a perfect full voxel type model and still cock it up with the implementation.

KSP1s method 'works' ok, it is obviously not overly realistic, but it does the job and is adequately useable.  Whether they use FAR, a refined version of the KSP1 model, or something else entirely isn't the issue for me, it is how it plays.

That's my general sentiment. I kind of had issues referring to FAR as if what I'm trying to say is, "hey take that mod and just do the same thing", but since it's the topic of the thread I did specifically refer to FAR. The main point I'd like to push is I hope they give aerodynamics some serious love. Everything you put in space has to get there through the atmosphere... I know that's apparently not actually going to be true in 2, but still. Atmosphere is important! And learning about it, and building around it, modeled in a way that's as realistic as possible while still remaining enjoyable, is 100% in the spirit of what the developers seem to going for. I just want to push this message as hard as I can because I worry there will be shortcuts in this area, and I have some secret hope someone with some say-so reads this and says, "you know what? I agree".

[Ninemeister]

Has anyone heard any new information on the aero model with KSP2?

I've seen all the recent media releases showing airfoil adjustability and whatnot, but still haven't seen anything mentioning the physics (or any change thereof).

[mattinoz]

On 7/14/2021 at 12:45 AM, The Physics Engineer said:

It won't and shouldn't be implemented in the core game, that's what mods are for. Making it part of the game would ruin gameplay for many players as orbiting stations and satellites would start drifting from their original orbits and would require to be adjusted constantly or could end up being accidentally slingshotted away of the orbiting body, which could be frustrating. In a mod this becomes an optional challenge.

 

I haven't used FAR, but from what has been said here, I agree it should stay as a mod too. 

Mods or DLC or both or a JV between Mod makers and the team like content creator packs (which lets the mods run console).

Many options to push the game in a certain direction test the market potential to amp up the challenge factor. If the game runs 10+ years making money to justify it the better for everyone. 

[MarcAbaddon]

I haven't really played enough FAR to have a really founded suggestion. I do know that atmospheric flight is important to me, but secondary to the orbital mechanics part of the game. So I would not really appreciate an aerodynamic model that makes getting through out of the atmosphere or landing a shuttle type vessel much harder than it is in KSP 1.

This being said, however it will be implemented, I would like to see obvious weaknesses with the KSP 1 model eliminated, most importantly that hidden or shielded parts can still add drag or that some parts that should produce a little bit of lift are missing it.

Speaking of the topic, in terms of simplifications, I would also not mind having some 'nicer' thrust curves as well. Especially with RAPIERs where currently it is hard to get above MACH 1 and then they suddenly kick-in like crazy.