Wobbly Rockets with David 'Trigger' Tregoning - KSP2 Dev Chat

[Rudolf Meier]

1 hour ago, The Aziz said:

... while on the subject of said video, there was a mention of a real rocket that has radial attachments that don't flex - true, but are they connected to the stack by a single attachment point, like in KSP(1 and 2)? Because I don't think so. Any more than one should prevent any flex, but because of the way the construction works in-game, the only solution for multiple attachments are struts. And honestly, one strut per radially attached part is enough.

I always read "1 attachment point" and people complain about it (not you, it's just that it crossed my mind when I read what you wrote). But, it is not 1 attachment point. It is 1 joint. That's a difference! A joint is what the software uses to calculate and predict the relation between two parts. That's all. If you stack and bolt 2 tanks on top of each other, you may use 100 bolts in reality and you still have just 1 joint in the game. That's fine, because this 1 joint simulates the behavior of all those 100 bolts which form "the connection".

So, my first point is, that it is irrelevant how many "attachment points" we have. But just how good those simulate reality. That's also why it is perfectly fine to use chains of Rigidbodys connected via ConfigurableJoints to simulate a rocket. The only problem is the flaw or design choice (maybe it is not even a flaw, we could debate this... but for now it is what we have to work with) which makes joints enormously weak, if the weights of the connected parts are not equal enough.

The game is built on top of that system. That's why the game needs to handle this (in the background in my opinion). You don't want that you have to design your ships in a special manner, just because of the engine which does not simulate the stuff correctly. Therefor we need a fix in the game, which does this, preferably without the player remarking anything.

And by the way... I often hear stuff like "we don't want workarounds" and stuff like that. But did you know that in KSP 1 parts were sometimes non physical parts and then turned into physical parts when it was needed? ... you could say, that this is also a "workaround". It could be that this idea could also be used to solve the wobbly rocket problems in KSP 2. But... I'm not sure if that's an easier solution to implement...

[Infinite Aerospace]

OK, I feel I have to share an experience. I decided to recreate my Skylon build from KSP1 in KSP2 and I have a few takeaways from it, some good, some bad and some downright shocking:

 

Procedural wings are one of the best additions to the game, without a doubt. Making wings with an aesthetic is easier than ever. I love the recolouring options, so now I can have a black Skylon as Alan Bond intended. Result!

 

Now, onto the bad. There must be some incredible amounts of drag from some parts, that I'm not aware of as the thing has ten R.A.P.I.E.R engines and struggles to get up-to 300 m/s in level flight. The thing is pointy as hell as well, there's not a single flat edge, everything is streamlined.

 

Now, the downright shocking. What the hell is wrong with the joint system in the game? This is one of the biggest aircraft I've attempted in KSP2 to date and it just has no structure at all. Doesn't matter if it's strutted, or not strutted. The entire fuselage just wobbles around and looks like an accordion, with gaps all over the place. Can we please do something about this? What is the point in things like cargo bays of they have absolutely no ability to hold their structure!?

 

Keep in mind, it worked in KSP1 and flies right on up-to orbit with a decent payload and now the thing can barely fly due to how spaghetti it is, the constant flapping of the control surfaces doesn't help either. Come on developers, what we doing here? Let us just 'weld' parts together to get rid of the hopeless joint system.

[The Aziz]

15 minutes ago, Infinite Aerospace said:

get rid of the hopeless joint system.

Can't wait to lose the entire vehicle because a solar panel broke off.

[Royalswissarmyknife]

1 hour ago, The Aziz said:

Can't wait to lose the entire vehicle because a solar panel broke off.

Why Wait?

Ksp-2 kind of already does that when you lose something.

[Pthigrivi]

All this discussion again makes me grateful that they’re really thoroughly considering this problem and its implications and not just diving head first into the most obvious and terrible solutions. 

[Infinite Aerospace]

On 10/8/2023 at 11:41 AM, The Aziz said:

It's manageable when there's no change in diameter. Go watch the latest @ShadowZone's video to see what it's all about.

... while on the subject of said video, there was a mention of a real rocket that has radial attachments that don't flex - true, but are they connected to the stack by a single attachment point, like in KSP(1 and 2)? Because I don't think so. Any more than one should prevent any flex, but because of the way the construction works in-game, the only solution for multiple attachments are struts. And honestly, one strut per radially attached part is enough.

The thing is, the wobble is manageable if all the parts are stacked in a vertical manner. However, this isn't applicable to aircraft and space planes, as these seem to have very little rigidity, it's not so much a wobble as it is 'sag'.

[NH4Cl Enthusiast]

15 hours ago, Infinite Aerospace said:

Now, the downright shocking. What the hell is wrong with the joint system in the game? This is one of the biggest aircraft I've attempted in KSP2 to date and it just has no structure at all. Doesn't matter if it's strutted, or not strutted. The entire fuselage just wobbles around and looks like an accordion, with gaps all over the place. Can we please do something about this? What is the point in things like cargo bays of they have absolutely no ability to hold their structure!?

Reminds me of a time in KSP1 at some version 0.something.something when  tried to build a Concorde replica. The fuselage did the accordion thing so badly it touched the ground and no amount of struts could help it. 

It just highlights how the joint wobble doesn't simulate anything worthwhile in the real world but is just a function of the game engine. In real world we would have a huge amount of tools and options to fix issues with structural integrity. In the game we have only struts.

It doesn't punish bad design, it punishes anyone using the in-game building system. We don't have procedural tanks or fuselage parts or welding, so we stack the small parts together with the only option we have - joints. This isn't bad design, it's the only design available and in many instances not even fixable with struts.

Adding a number of ugly, massless ghost parts to fix a problem caused by the wobble isn't some master engineering skill, it's also largely a trivial thing to do and adds very little to gameplay compared to what you'd get just by increasing the joint stability in KJR-style.

At the same time we completely miss other real life structural issues when parts that should bend, like wings, just don't. So we have irrelevant-to-real-life issues with the wobble, unrealistic parts to partially circumvent some of them but no tools to fix the ones that in reality are probably easiest to fix, we can have massive parts like wings which are unbreakable and to top it all off massive performance issues caused directly by increasing part count...

I'm all in favor of structural integrity being a part of the game, and things like joint strength and aerodynamic forces need to be there, but the silly rubber noodling is more detrimental than fun and is an incredibly poor "simulation" of said design concerns. 

 

5 hours ago, Pthigrivi said:

All this discussion again makes me grateful that they’re really thoroughly considering this problem and its implications and not just diving head first into the most obvious and terrible solutions. 

Fair, but is this really relevant right now? More important than adding re-entry heating, science, colonies, interstellar, multiplayer, resources, fixing performance and so on?

I can't understand why they would waste time on this. They could just slap on a passable interim solution in the time it took to make the dev video and move on to something worthwhile. It's like their house is on fire but Nate is making a video on how to furnish the living room. 

They already have solutions available that would be simple and fast to implement that the player base is familiar with and would probably be willing to accept as interim solution that can be revisited in some distant future if necessary. Just copy KSP1 or KJR. It's not perfect but passable for now.

Make a game worth playing first, then contemplate on the nuances of wobble.

[Infinite Aerospace]

14 minutes ago, NH4Cl Enthusiast said:

Reminds me of a time in KSP1 at some version 0.something.something when  tried to build a Concorde replica. The fuselage did the accordion thing so badly it touched the ground and no amount of struts could help it. 

It just highlights how the joint wobble doesn't simulate anything worthwhile in the real world but is just a function of the game engine. In real world we would have a huge amount of tools and options to fix issues with structural integrity. In the game we have only struts.

It doesn't punish bad design, it punishes anyone using the in-game building system. We don't have procedural tanks or fuselage parts or welding, so we stack the small parts together with the only option we have - joints. This isn't bad design, it's the only design available and in many instances not even fixable with struts.

Adding a number of ugly, massless ghost parts to fix a problem caused by the wobble isn't some master engineering skill, it's also largely a trivial thing to do and adds very little to gameplay compared to what you'd get just by increasing the joint stability in KJR-style.

At the same time we completely miss other real life structural issues when parts that should bend, like wings, just don't. So we have irrelevant-to-real-life issues with the wobble, unrealistic parts to partially circumvent some of them but no tools to fix the ones that in reality are probably easiest to fix, we can have massive parts like wings which are unbreakable and to top it all off massive performance issues caused directly by increasing part count...

I'm all in favor of structural integrity being a part of the game, and things like joint strength and aerodynamic forces need to be there, but the silly rubber noodling is more detrimental than fun and is an incredibly poor "simulation" of said design concerns. 

 

Fair, but is this really relevant right now? More important than adding re-entry heating, science, colonies, interstellar, multiplayer, resources, fixing performance and so on?

I can't understand why they would waste time on this. They could just slap on a passable interim solution in the time it took to make the dev video and move on to something worthwhile. It's like their house is on fire but Nate is making a video on how to furnish the living room. 

They already have solutions available that would be simple and fast to implement that the player base is familiar with and would probably be willing to accept as interim solution that can be revisited in some distant future if necessary. Just copy KSP1 or KJR. It's not perfect but passable for now.

Make a game worth playing first, then contemplate on the nuances of wobble.

That sums it up quite well, don't get me wrong it wasn't that floppy but it didn't have any real rigidity. It wasn't a single fuselage, it was a collection of parts held poorly together through some mechanism. The issue is, in my opinion that the game treats the 'middle' of a part as the joint, so any turning moment occurs around that point. The edges the cargo bays for example should be treated as solid, but they aren't as evident how they clip into eachother as they flex. Right now though, I cannot make it work, nothing I do stops the deformation so it's been shelved for now.

[Periple]

21 minutes ago, NH4Cl Enthusiast said:

They already have solutions available that would be simple and fast to implement that the player base is familiar with and would probably be willing to accept as interim solution that can be revisited in some distant future if necessary. Just copy KSP1 or KJR. It's not perfect but passable for now.

Make a game worth playing first, then contemplate on the nuances of wobble.

Maybe they're trying to avoid having an interim solution that will break lots of craft when they roll out the proper one? Performance considerations aside, simple quick and dirty solutions might have this side effect as very likely they would make some joints stronger than they ought to be. 

[Stoup]

34 minutes ago, Periple said:

Maybe they're trying to avoid having an interim solution that will break lots of craft when they roll out the proper one? Performance considerations aside, simple quick and dirty solutions might have this side effect as very likely they would make some joints stronger than they ought to be. 

I just feel like refusing to address the issues as they are currently, even if with a temporary solution, is sending a bad message.

That it doesn't matter if the game's less fun or enjoyable right now, since they want to make a 'perfect solution' down the line.

And besides that, we're so early in early access, I hope people aren't expecting complete continuity with craft files through the entire EA process. EA IS the time to break such things

[NH4Cl Enthusiast]

22 minutes ago, Periple said:

Maybe they're trying to avoid having an interim solution that will break lots of craft when they roll out the proper one? Performance considerations aside, simple quick and dirty solutions might have this side effect as very likely they would make some joints stronger than they ought to be. 

Again, fair point, but is it really relevant right now? When there is a massive discontent among players, nothing on the roadmap implement after almost 8 months and very basic features like heating still awol. And this is what they choose to focus on? Also save game compatibility is not an issue for a long time and I'd argue that KSP players are also not overly concerned about having to go back and fix issues with old crafts. I personally would even enjoy the challenge.

At this point it's also ok to play the EA card. Anything after 1.0 should be somewhat compatible unless there's a very good reason to break crafts and saves but it's totally fair to expect your 0.99 save won't work perfectly in 1.0. 

Unless there is a fundamental reason to go and change the joint physics in some way which profoundly affects the implementation of heat, resources, interstellar or something else important in the long run, tinkering with something like this just seems like a red herring to distract people and/or really poor project management.

[gluckez]

2 hours ago, NH4Cl Enthusiast said:

It just highlights how the joint wobble doesn't simulate anything worthwhile in the real world but is just a function of the game engine. In real world we would have a huge amount of tools and options to fix issues with structural integrity. In the game we have only struts.

This is true, and got me thinking and doing some digging. I went through the code myself and noticed that the current implementation is only using configurable joints. There's definitely good arguments for that, probably the best one being the connected anchor and primary and secondary axis that you can set, as well as limiting rotation and motion between parts to some degree, but it's also prone to buggyness, and thus wobble. 
One option would be, that in case of vertically stacked parts, you could use the fixed joint. But the problem there is then that there's no wobble whatsoever. I think most people wouldn't mind that, for obvious reasons, but then you lose certain aspects that make the games physics simulation so fun. For example, you would expect a massive rocket, consisting only of small fuel thanks to eventually collapse in on itself, because that's just bad design, and fixing it all in place with fixed joints would cause it to be really stable, whereas reality wouldn't exactly agree with that. In that case, it's not the joints between the parts that are flexing, but the parts themselves, and replacing all parts with a soft-body (to a certain degree) probably isn't the way to go either.

One other aspect to consider is clipping. Configurable Joints in Unity allow you to give an offset on the rotation and position of an object, and the engine would then use provided spring and damping values to try and keep it there, unless other forces (drag, collisions, gravity, thrust...) act on it. Fixed joints can't solve that problem, the only alternative would be to parent the gameobjects, but then you have overlapping colliders, and external forces couldn't break these things apart in any natural way, it would have to be coded, which would again result in several different sets of the rules of physics. Another way would be to use procedural meshes for the colliders, but that would create a big performance hit, and would be a mess to sync up in multiplayer, as well as pose a bunch of challenges when it comes to texturing and normal mapping. 

I'm pretty sure the devs have already explored all these options, and are indeed working to make things more stable. From what I've seen, with a decent ascent profile, and a good amount of struts, you could launch the ISS in one go on a rocket and you would get there. But IRL that would be an aweful choice, there's a reason why it was assembled in orbit. For the time being, I think we're stuck with the wobble, at least until all the forces on the joints are properly configured, but that's part of optimisation and is expected to take some time in EA.

[Rudolf Meier]

2 hours ago, NH4Cl Enthusiast said:

Just copy ... or KJR. It's not perfect but passable for now.

... no, just kidding. But to be honest, I'd love to see people trying the new options of KJR Next with fully disabled autostruts. And then we could try to tweak the system and learn something about the possibilities. And not just repeat stuff that's no longer true today...

[NH4Cl Enthusiast]

Just now, Rudolf Meier said:

... no, just kidding. But to be honest, I'd love to see people trying the new options of KJR Next with fully disabled autostruts. And then we could try to tweak the system and learn something about the possibilities. And not just repeat stuff that's no longer true today...

Haha, sorry no offense intended! I just meant that a perfect solution would probably involve things like soft body physics while magically not impacting performance and so on. 

I haven't tried KJR Next since I'm not actively playing right now, but the concept seems solid. It just further goes to show that it's weird of the devs to focus on this. Your solution seems like a good option, whatever KSP1 did seems solid enough assuming it's translatable to KSP2 (I know very little about Unity physics engine), the community would probably be ok to use a mod for now anyway. So why spend time on this?

[Scarecrow71]

2 hours ago, NH4Cl Enthusiast said:

Again, fair point, but is it really relevant right now?

Considering the number of people here upset over wobbly rockets, coupled with the numerous discussions on the topic...yes, it is absolutely relevant right now.  And should be the second priority (after orbital decay).  Implementing new features won't mean much if you still have to worry about flying a noodle.

[NH4Cl Enthusiast]

17 minutes ago, Scarecrow71 said:

Considering the number of people here upset over wobbly rockets, coupled with the numerous discussions on the topic...yes, it is absolutely relevant right now.  And should be the second priority (after orbital decay).  Implementing new features won't mean much if you still have to worry about flying a noodle.

Well ok, you are right and I agree on what you said, the wobble itself is an issue. But my point is rather that since to me it seems like we already have relatively easy solutions available which take fairly little dev time, why go through the shenanigans of making an entire video on it, several tools to fine tune a solution and overall spend time and energy to dig deep into the issue when a quick duct tape solution would do.

I appreciate the sentiment of doing things well, but just seems odd to me that they spend so much effort on this when a quick patch note would suffice. 

But again, I'm not a developer so maybe it does make sense? It's just as a regular player I haven't been convinced and I would really rather hear about the design philosophy on the upcoming science mode or something than the metaphysical properties of rocket wobble.

[Scarecrow71]

2 hours ago, NH4Cl Enthusiast said:

Well ok, you are right and I agree on what you said, the wobble itself is an issue. But my point is rather that since to me it seems like we already have relatively easy solutions available which take fairly little dev time, why go through the shenanigans of making an entire video on it, several tools to fine tune a solution and overall spend time and energy to dig deep into the issue when a quick duct tape solution would do.

I appreciate the sentiment of doing things well, but just seems odd to me that they spend so much effort on this when a quick patch note would suffice. 

But again, I'm not a developer so maybe it does make sense? It's just as a regular player I haven't been convinced and I would really rather hear about the design philosophy on the upcoming science mode or something than the metaphysical properties of rocket wobble.

Again, dealing with new features when the base game isn't up to snuff won't do anything but add problems.  You need to solve the base problems before you add stuff that could potentially add more problems.

[regex]

8 hours ago, Periple said:

Maybe they're trying to avoid having an interim solution that will break lots of craft when they roll out the proper one? Performance considerations aside, simple quick and dirty solutions might have this side effect as very likely they would make some joints stronger than they ought to be. 

I would hate to have a poorly thought out interim solution become the permanent one later down the road because the playerbase couldn't let go of it or we had moved so far forward in development that changing it would significantly affect established gameplay/require an extensive refactor.

[Alexoff]

3 hours ago, Scarecrow71 said:

Again, dealing with new features when the base game isn't up to snuff won't do anything but add problems.  You need to solve the base problems before you add stuff that could potentially add more problems.

Specifically, wobbling would not greatly hinder the addition of science to the game. At the very beginning of our careers, we use primitive small crafts, which are not greatly harmed by wobbling. And you/devs can bring wobbling to the level of KSP1 by adding a couple of zeros to the game parameters. But I think that if wobbling is corrected, fans will begin to demand science, which is significantly more difficult and longer to do, and such conversations allow devs to switch attention and buy time.

[regex]

39 minutes ago, Infinite Aerospace said:

But, in-action is just a giant middle finger to the community if I'm honest and I fully have the teams back at this point.

vOv There's no conversation to be had here if your idea of "inaction" is imvestigating what will be a good solution before implementing it. Enjoy your middle finger, I guess, Intercept can't work fast enough for you people.

[The Aziz]

8 hours ago, NH4Cl Enthusiast said:

a quick duct tape solution would do.

I hate half-assed solutions because they never work for long term and look ugly as hell. So it really wouldn't do.

44 minutes ago, Infinite Aerospace said:

That is shocking player numbers, even for an early access title

I can tell you another game that is early access, is an open world sandbox, is a sequel of a renowned game from few years back, and has been a huge success on day 1 despite little content, and also has very low player count. It's not really shocking.

[Vanamonde]

Posts not about rocket wobble have been moved here. 

 

[Kimera Industries]

On 9/29/2023 at 11:50 AM, Stewcooker said:

"You could turn off all the joints, single rigidbody,, but then you lose some of that novelty..." I feel like "wobbly rockets" aren't an interesting "novelty", they're inaccurate and immersion breaking. Yes KSP prides itself on being cartoony and accessible, something I feel like the dev team may have leaned a little too hard into at times, but at the end of the day this is a physics-based sandbox game trying to emulate real-life rocketry, and in real life rockets don't wobble or bend when they encounter shear forces, they break. Often catastrophically. At this point my vote is to echo what @regex says and implement node to node welding of parts of the same size or cross section. You can still calculate shear forces and aerodynamics per part, and if shear forces reach a certain threshold, *bam*, explode the part, decouple the nodes, and then you've got a realistic emulation of a rocket.

I completely agree, this feels like a good, useful solution- of course, in-game testing could say otherwise, but it seems like it would work well.

As far as docking port rigidity, I feel it should be the same as with the rigidity of parts attached in the VAB, if a fix like this is implemented.

After all, the ISS and other stations don't wobble, just like how irl rockets don't. Those big interstellar ships seen in the trailers, probably assembled in orbit with docking ports, are gonna suck if you fire your engine and get a bowl of spaghetti. Please don't do something like autostrut if but for the simple fact that it gives me anxiety to make sure I got every part.

[Datau03]

9 hours ago, Kimera Industries said:

 Those big interstellar ships seen in the trailers, probably assembled in orbit with docking ports, are gonna suck if you fire your engine and get a bowl of spaghetti. 

I agree with everything you said, just wanted to hopefully excite you by reminding you that we will be able to build craft directly in orbit via orbital colonies

Edited October 12, 2023 by Datau03

[Pthigrivi]

2 hours ago, Datau03 said:

I agree with everything you said, just wanted to hopefully excite you by reminding you that we will be able to build craft directly in orbit via orbital colonies

Which is very much welcome. The only place I ever encountered wobble in KSP1 was with really big orbitally assembled ships, so to be able to create continuous structures and properly strut any out-rigged modules will be a huge help.