Developer Insights #9 – Orbit Tessellation

[JB182]

wow! so awesome!

[swjr-swis]

21 hours ago, Intercept Games said:

If you enjoyed this technical deep dive and want to see more stuff like it please let us know!

Letting you know...

 

[Yellowburn10]

Wow, I love stuff like this! Pointy orbits have always been something I've noticed but never really took thought into how or why it happens, so this explanation was really enlightening!

[Dr. Kerbal]

On 4/15/2021 at 3:49 PM, Anth12 said:

@Johannes Incredible Post. I would love to see more of these types of dev posts.

Me too. The way @Johannes explains this is very interesting.  A non coder like me understood this!!! Keep up the good coding work!

Edited April 22, 2021 by Dr. Kerbal

[shdwlrd]

Very informative. I liked how clear and concise your explanation was. I never realized how much work was involved in just drawing smooth curves in a game.

[Grimmas]

It goes without saying but I'd love to see more technical posts about KSP2 like this.

[pandaman]

Thanks @Johannes

This was great, very interesting, both from 'What can we expect to see?'  and 'How do they do it? ' perspectives.'

It's also good to be able to see and appreciate how much thought and effort goes into things we see and just take for granted, unless they don't look perfect. 

[Iamsodarncool]

On 4/15/2021 at 4:00 AM, Intercept Games said:

If you enjoyed this technical deep dive and want to see more stuff like it please let us know!

I enjoyed this technical deep dive and I want to see more stuff like it!

 

[Jack Mcslay]

I would love for this to be backported to KSP, it might help the ridiculous performance drop when displaying asteroids

[Redneck]

On 4/15/2021 at 3:00 AM, Intercept Games said:

Hello all, I’m Johannes Peter – A programmer on Kerbal Space Program 2 – and I love solving interesting problems!

The problems we face in game development rarely have a single “correct” answer. The more specialized your game is, the more specialized your problems are, and the more creative your solutions need to be. As a rocketry simulator on an interplanetary scale, Kerbal Space Program has already tackled a wealth of unique programming challenges.

Today I want to share a solution I’ve worked on for a problem that is fairly unique to KSP: How to draw accurate orbits that look stellar regardless of where they’re viewed from?

I’ll briefly cover a standard approach for drawing orbits, touch on some of the issues with that approach, and then look at the solution that KSP2 is using now: screen space orbit tessellation.

This dev diary will get a bit into the technical side of KSP2’s development, with diagrams to help illustrate some of the core concepts. We will mainly focus on this test scene of Kerbin orbiting Kerbol:

 

just as long as this works with multiplayer positioning then whatevs. You guys did consider multiplayer positioning when thinking about this correct?

[MechBFP]

17 hours ago, Redneck said:

just as long as this works with multiplayer positioning then whatevs. You guys did consider multiplayer positioning when thinking about this correct?

It has to do with what the camera displays when looking at the map view. It doesn't have anything to do with multiplayer, at all.

[Johannes]

On 4/18/2021 at 8:00 PM, Redneck said:

You guys did consider multiplayer positioning when thinking about this correct?

Wait, this game has multiplayer?

(yes that was a joke)

[MKI]

something as simple as drawing a curved line in 3D space = complex math + code

 

6 minutes ago, Johannes said:

Wait, this game has multiplayer?

(yes that was a joke)

You got me XD

[Johannes]

Joking aside though: As far as the orbit tessellation system is concerned, multiplayer orbit trajectories are no different from local ones  

Edited April 21, 2021 by Johannes

[JB182]

5 hours ago, Johannes said:

Wait, this game has multiplayer?

(yes that was a joke)

im surprised like "WHAT?! THE DEVELOPER DONT KNOW IF THE GAME WILL HAVE MULTIPLAYER!?" until i see a smol letter below XD

[OutInSpace]

So this is why the orbits in KSP1 look funky and inaccurate. I feel like I've just had some great cosmic truth revealed to me...

[Redneck]

On 4/20/2021 at 4:26 PM, Johannes said:

Joking aside though: As far as the orbit tessellation system is concerned, multiplayer orbit trajectories are no different from local ones  

just curious because the positioning system in KSP1 caused ALOT of issues for the multiplayer mod. Jitter,orbits and trajectories off, etc. Thanks for replying 

[MechBFP]

12 hours ago, Redneck said:

just curious because the positioning system in KSP1 caused ALOT of issues for the multiplayer mod. Jitter,orbits and trajectories off, etc. Thanks for replying 

Keyphrase here is “multiplayer mod”. Of course it’s going to be glitchy as high heaven. That has nothing to do with how KSP was built and everything to do with what the mod makers had to work with instead. 

[ItsJustLuci]

Wow... This already seems a lot more optimized than ksp was. I'm willing to wait years for optimizations alone XD 

[Redneck]

6 hours ago, MechBFP said:

Keyphrase here is “multiplayer mod”. Of course it’s going to be glitchy as high heaven. That has nothing to do with how KSP was built and everything to do with what the mod makers had to work with instead. 

oh it has EVERYTHING to do with how KSP1 was built. The code is/was a nightmare!

[MechBFP]

3 hours ago, Redneck said:

oh it has EVERYTHING to do with how KSP1 was built. The code is/was a nightmare!

No, it doesn’t. Not having access to the source code means you need to hook into things not intended to be used for multiplayer purpose. KSP code has nothing to do with that. 

[Stormpilot]

Very cool! (Waiting for this game is going to be so hard ).

[Incarnation of Chaos]

Is this done on the CPU or GPU? If the latter, what's the implications for AMD GPU which generally struggle with tessellation in games? 

Also, does this also mean that it might be possible to use a "tiled" approach to rendering KSP2 in the far future? 

Between this and the way your team is handling extrasolar systems with their own floating origins, there's a lot of potential for very interesting things to increase performance.  

Amazing writeup though, this is the stuff that I want to see!

[MechBFP]

1 hour ago, Incarnation of Chaos said:

Is this done on the CPU or GPU? If the latter, what's the implications for AMD GPU which generally struggle with tessellation in games?

This kind of tessellation an integrated GPU could eat through without breaking a sweat. Its not a graphically intense thing to tessellate a single circle.

[Incarnation of Chaos]

1 hour ago, MechBFP said:

This kind of tessellation an integrated GPU could eat through without breaking a sweat. Its not a graphically intense thing to tessellate a single circle.

I'm aware, but it's still something I'd like to know. With multiple systems, and the conics tessellated in multiple places it would potentially add up fairly quickly.