Keep ending up inside K's
Moderator: Moderators for English X Forum
-
- Posts: 562
- Joined: Thu, 11. Nov 04, 22:25
Keep ending up inside K's
I have been trying the ship surface subcomponent destruction gameplay a bit and taking a small ship into proximity with K's and often end up inside the model of the K, from where it is quite hard to get out.
This is with collision avoidance on, which doesnt fix this.
It is not helped by some very odd collision physics whereby the small ship I am flying sometimes cannot maneuvure if any part of it comes into contact with the K's hull. All of a sudden it cannot roll pitch or yaw or move and struggling to get away then leads to it ending up inside the K model in a mad spin.
The problem is the effect of contact on the S ship steering. In a lever friendly physics environment then turning the S ship would cause it to displace its center of mass pushed by leverage from the point of contact, but there does not appear to be leverage applied in this model or if there is it is not working as one would expect. So the CoM stays where it is and the turning and several directions of movement are obstructed by the point of contact and local geometry of the K model.
Thats it really, I just don't think that should be happening, so it deserves thinking about.
This is with collision avoidance on, which doesnt fix this.
It is not helped by some very odd collision physics whereby the small ship I am flying sometimes cannot maneuvure if any part of it comes into contact with the K's hull. All of a sudden it cannot roll pitch or yaw or move and struggling to get away then leads to it ending up inside the K model in a mad spin.
The problem is the effect of contact on the S ship steering. In a lever friendly physics environment then turning the S ship would cause it to displace its center of mass pushed by leverage from the point of contact, but there does not appear to be leverage applied in this model or if there is it is not working as one would expect. So the CoM stays where it is and the turning and several directions of movement are obstructed by the point of contact and local geometry of the K model.
Thats it really, I just don't think that should be happening, so it deserves thinking about.
bloop
-
- Moderator (English)
- Posts: 31785
- Joined: Fri, 16. Apr 04, 19:21
Re: Keep ending up inside K's
This has been issue for some time now for smaller ships doing capital proximity manoeuvres. See here.
A dog has a master; a cat has domestic staff.
-
- Posts: 6680
- Joined: Mon, 5. May 08, 00:05
Re: Keep ending up inside K's
I'm periodicly hugging K's and I's with M gunboats (Peregrine, Prometheus, Katana) and never ended insided their geometry. One can clip through objects on very high speed collisions (e.g. when using traveldrive) when the game can't keep up with constant collision checks (as one is too fast). Make sure your rig can handle X4 on a high level fps-wise - 40-60 fps. Otherwise you're prone to slip through.
Also, it's a long time since release - these things are both fundamental and characteristic to the games engine, those won't change. This can be considered an optimisation and compromise in order to keep the universe simulation up. One simply just can't have both - high entity count and very precise collision detection as both are costly in terms of computation.
Best advice is to get used to it and / or update your hardware.
Also, it's a long time since release - these things are both fundamental and characteristic to the games engine, those won't change. This can be considered an optimisation and compromise in order to keep the universe simulation up. One simply just can't have both - high entity count and very precise collision detection as both are costly in terms of computation.
Best advice is to get used to it and / or update your hardware.

Spoiler
Show
BurnIt: Boron and leaks don't go well together...
Königinnenreich von Boron: Sprich mit deinem Flossenführer
Nila Ti: Folgt mir, ihr Kavalkade von neugierigen Kreaturen!
Tammancktall: Es ist eine Ehre für sie mich kennenzulernen...
CBJ: Thanks for the savegame. We will add it to our "crazy saves" collection [..]
Feature request: paint jobs on custom starts
Königinnenreich von Boron: Sprich mit deinem Flossenführer
Nila Ti: Folgt mir, ihr Kavalkade von neugierigen Kreaturen!
Tammancktall: Es ist eine Ehre für sie mich kennenzulernen...
CBJ: Thanks for the savegame. We will add it to our "crazy saves" collection [..]

-
- Posts: 562
- Joined: Thu, 11. Nov 04, 22:25
Re: Keep ending up inside K's
Interesting link Alan Phipps thanks, will read.
M ships are a lot bigger than S and this may effect model interactions due to the hull geometry around contact points. Even so in the linked thread the OP had the same kind of problem with an M. My experience is S ships with L, which can fit in the nooks which M ship's size keeps them out of and this causes a real problem a bit too often imho. So often you have to be constantly aware and work around it.
It undoubtedly glitches through but I am not sure where the physics are calculated, are they on the GPU in X4? Is the physics sim really linked to the graphics rendering rate?
As for FPS, I play on an RTX 2070S at 1080p, with an i7 9700K / Z390 / 32GB quad DDR4 / 1kw Seasonic. In X4 it gets 90 FPS in open space and 57 in a crowded battle in an asteroid belt by a gate, on ultra, which should be enough to play the game. I doubt better hardware would fix this.
This is more about the physics model not behaving as expected, as I said, imho it deserves thinking about.
M ships are a lot bigger than S and this may effect model interactions due to the hull geometry around contact points. Even so in the linked thread the OP had the same kind of problem with an M. My experience is S ships with L, which can fit in the nooks which M ship's size keeps them out of and this causes a real problem a bit too often imho. So often you have to be constantly aware and work around it.
It undoubtedly glitches through but I am not sure where the physics are calculated, are they on the GPU in X4? Is the physics sim really linked to the graphics rendering rate?
As for FPS, I play on an RTX 2070S at 1080p, with an i7 9700K / Z390 / 32GB quad DDR4 / 1kw Seasonic. In X4 it gets 90 FPS in open space and 57 in a crowded battle in an asteroid belt by a gate, on ultra, which should be enough to play the game. I doubt better hardware would fix this.
This is more about the physics model not behaving as expected, as I said, imho it deserves thinking about.
bloop
-
- Moderator (English)
- Posts: 31785
- Joined: Fri, 16. Apr 04, 19:21
Re: Keep ending up inside K's
I think a part of the issue is that the collision detection is not always sure what is inside and what is outside for the capital or station model structure. As a result, instead of reliably rebounding away from the larger structure when a part of your ship collides with it, sometimes part or all of your ship is sucked in instead. Add to that the possibility for your ship getting into wild and rapid spinning and yawing on collision, and the situation can quickly become uncontrollable.
The same sort of thing can happen when your ship 'sucks in' a colliding drone, pod or deployable.
You can readily see this in action when you use your spacesuit from your ship held in the drag field of a moving capital; you may end up inside your own ship structure or collide and end up inside the nearby capital - and often involving uncontrolled spinning by the suit too. (This is when teleporting and/or insta-docking become essential!)
The same sort of thing can happen when your ship 'sucks in' a colliding drone, pod or deployable.
You can readily see this in action when you use your spacesuit from your ship held in the drag field of a moving capital; you may end up inside your own ship structure or collide and end up inside the nearby capital - and often involving uncontrolled spinning by the suit too. (This is when teleporting and/or insta-docking become essential!)
A dog has a master; a cat has domestic staff.
-
- Posts: 562
- Joined: Thu, 11. Nov 04, 22:25
Re: Keep ending up inside K's
Rapid spinning suggests to me the collision physics sim is applying supersized numbers and/or possibly has tiny divisors somewhere in the calc.
I wonder if it is possible to impose a limit on the kind of output that can be generated in that way. Either by limiting output or input or both e.g. prevent divisors below a certain size, i.e. round them up.
This may be something to do with the leverage again. It does not easily move the CoM but it does cause rotation. So if the CoM is close enough to the point of contact, e.g. when the ship model is transiting the boundary of another collision volume aka glitching, the distance between the point of contact and the CoM can get very small and thereby generate high rates of rotation when plugged into the standard collision sim.
You may be onto something with differentiating inside and outside the model. It does seem as if it fails to do that. I remember this kind of thing did not happen in Jumpgate, your ship just used to be destroyed by collisions! But even if you had a well shielded ship which could survive collision it would always bounce off and not cross the boundary into the internal space of the colliding object. Maybe distance to CoM could help, or an internal no go volume just inside or congruent with the collision volume.
They have had problems like this in Empyrion where the deformable voxel based terrain mesh does not always form cohesively and avatars can glitch through and fall to the center of the planet, which they solved by adding another boundary inside the mesh and if the player falls inside a planet they are transported back to the surface magically, which is preferrable to avatar death and an irretrievable backpack, so noone complains! But it is not ideal as the fall is usually about 20m or so and very obvious, it would be better if the detection occurred within a cm as it were, i.e. much closer to the boundary and I imagine that might be more feasible with non deformable structures. A lot of work to make a detection volume if you have not set the engine up to do that but maybe not if you can derive a detection volume from the model structure in relation to the center of mass.
Just thinking out loud.
I wonder if it is possible to impose a limit on the kind of output that can be generated in that way. Either by limiting output or input or both e.g. prevent divisors below a certain size, i.e. round them up.
This may be something to do with the leverage again. It does not easily move the CoM but it does cause rotation. So if the CoM is close enough to the point of contact, e.g. when the ship model is transiting the boundary of another collision volume aka glitching, the distance between the point of contact and the CoM can get very small and thereby generate high rates of rotation when plugged into the standard collision sim.
You may be onto something with differentiating inside and outside the model. It does seem as if it fails to do that. I remember this kind of thing did not happen in Jumpgate, your ship just used to be destroyed by collisions! But even if you had a well shielded ship which could survive collision it would always bounce off and not cross the boundary into the internal space of the colliding object. Maybe distance to CoM could help, or an internal no go volume just inside or congruent with the collision volume.
They have had problems like this in Empyrion where the deformable voxel based terrain mesh does not always form cohesively and avatars can glitch through and fall to the center of the planet, which they solved by adding another boundary inside the mesh and if the player falls inside a planet they are transported back to the surface magically, which is preferrable to avatar death and an irretrievable backpack, so noone complains! But it is not ideal as the fall is usually about 20m or so and very obvious, it would be better if the detection occurred within a cm as it were, i.e. much closer to the boundary and I imagine that might be more feasible with non deformable structures. A lot of work to make a detection volume if you have not set the engine up to do that but maybe not if you can derive a detection volume from the model structure in relation to the center of mass.
Just thinking out loud.
bloop
-
- Moderator (English)
- Posts: 4933
- Joined: Fri, 21. Dec 18, 18:23
Re: Keep ending up inside K's
Although tricky it is usually possible to escape a ship once inside.
- You can instruct your copilot to fly out. Let them take control of the ship and order them to fly away. After much bumping around they will usually break free. To avoid being destroyed I would recommend taking control of the ship as quickly as possible once free.
- Through some jank movement it can be possible to escape yourself. Disable collision avoidance, auto upright and auto roll. The fly to a mostly flat surface of the object you are trying to escape. Line up your ship so that it is facing parallel to the surface. Then roll and strafe into the surface, adjusting speed as required. Might take a few attempts but should eventually work. It is likely lower frame rates might make this easier to pull off.
- If both the above fail, you can teleport to another asset of yours and then immediately teleport back. By cycling high attention state of the area your ship should be displaced nearby. Taking control of the ship quickly might be a good idea to avoid being destroyed.
-
- Posts: 562
- Joined: Thu, 11. Nov 04, 22:25
Re: Keep ending up inside K's
Inside a Katana. Teleported to a ship which was coming out of storage and lifting off, it moved while the teleport was happening and the avatar ended up backstage!
https://steamcommunity.com/sharedfiles/ ... 2883464591
https://steamcommunity.com/sharedfiles/ ... 2883464591
bloop
-
- Posts: 6680
- Joined: Mon, 5. May 08, 00:05
Re: Keep ending up inside K's
The easiest way to escape from within an object:Imperial Good wrote: ↑Mon, 31. Oct 22, 16:10 Although tricky it is usually possible to escape a ship once inside.
- You can instruct your copilot to fly out. Let them take control of the ship and order them to fly away. After much bumping around they will usually break free. To avoid being destroyed I would recommend taking control of the ship as quickly as possible once free.
- Through some jank movement it can be possible to escape yourself. Disable collision avoidance, auto upright and auto roll. The fly to a mostly flat surface of the object you are trying to escape. Line up your ship so that it is facing parallel to the surface. Then roll and strafe into the surface, adjusting speed as required. Might take a few attempts but should eventually work. It is likely lower frame rates might make this easier to pull off.
- If both the above fail, you can teleport to another asset of yours and then immediately teleport back. By cycling high attention state of the area your ship should be displaced nearby. Taking control of the ship quickly might be a good idea to avoid being destroyed.
Changing PoV helps with this problem, as well

- Open map
- Select another object far away (~ more then 30-40km distance)
- Activate external view on that far away object (=your ship is moved to "low attention")
- Hit boost for a few seconds
- Set camera back to your own ship - freedom of movement restored

Spoiler
Show
BurnIt: Boron and leaks don't go well together...
Königinnenreich von Boron: Sprich mit deinem Flossenführer
Nila Ti: Folgt mir, ihr Kavalkade von neugierigen Kreaturen!
Tammancktall: Es ist eine Ehre für sie mich kennenzulernen...
CBJ: Thanks for the savegame. We will add it to our "crazy saves" collection [..]
Feature request: paint jobs on custom starts
Königinnenreich von Boron: Sprich mit deinem Flossenführer
Nila Ti: Folgt mir, ihr Kavalkade von neugierigen Kreaturen!
Tammancktall: Es ist eine Ehre für sie mich kennenzulernen...
CBJ: Thanks for the savegame. We will add it to our "crazy saves" collection [..]

-
- Moderator (English)
- Posts: 31785
- Joined: Fri, 16. Apr 04, 19:21
Re: Keep ending up inside K's
I think there are three aspects being discussed now that are of quite different issue contexts:
1. That collisions with a boundary can suck in all or part of the colliding ship rather than cause it to rebound outside the boundary.
2. That collisions can cause subsequent severe loss of control through:
1. That collisions with a boundary can suck in all or part of the colliding ship rather than cause it to rebound outside the boundary.
2. That collisions can cause subsequent severe loss of control through:
- a. Little/slow response to controls where assets are intermingled.
b. Uncontrolled rapid spinning where they are not intermingled.
A dog has a master; a cat has domestic staff.
-
- Posts: 116
- Joined: Sat, 6. Aug 22, 23:41
Re: Keep ending up inside K's
I don’t think so either, but it does… I’ve been stuck:
- inside my own ship when suit-boosting around the airlock
- inside all kinds of Paranid stations (especially)
- inside asteroids
- inside capital ships
My video card isn’t the greatest, so it may be an FPS calculation issue where things are getting out of sync… but in my case it happens almost exclusively at LOW speeds rather than HIGH speeds as others have suggested. Collision physics work (for me) pretty much 100% of the time when moving quickly at an object, and 85% of the time when very slow moving in contact or near-contact with an object.
-
- Posts: 230
- Joined: Tue, 11. Jan 22, 20:05
Re: Keep ending up inside K's
I have quite the opposite problem. Many S XEN ship get stuck inside my L/XL. I traveled half of Matrix 9 in my Asgard while having 5-6 xen inside and being unable to travel drive/kill them/use SETA
"Never listen to what people say. Just watch what they do."
Chrisjen Avasarala
Chrisjen Avasarala
-
- Moderator (English)
- Posts: 31785
- Joined: Fri, 16. Apr 04, 19:21
Re: Keep ending up inside K's
@ Pejot: Just teleport OOS away from the Asgard, wait a few seconds, and then teleport back. Or alternatively use this sort of attention-changing trick as above. Now the S ships would be outside your Asgard and probably under attack from its anti-fighter turrets.
A dog has a master; a cat has domestic staff.