Weather and environmental visual effects

[mcwaffles2003]

10 minutes ago, Pthigrivi said:

Doesn't this get a little weird because your flights are modeled continuously to the surface? I think it would be a problem if you're looking down from orbit and see a hurricane swirling but when you descend into it disappears to a randomized clear day (or worse, the opposite). Seems like if you're gonna do it you gotta go all the way.

Why would you see a hurricane from orbit though if there isn't one there? It's not like clouds from orbit will be static features and just because a storm animation starts with a clear day doesn't mean the animation has to begin from start if you intercept a storm in progress. I mean Im going out on a limb here as I'm not a game dev, but couldn't a storm animation like I've stated earlier be broken up into segments and have it begin from the appropriate one to the environment as it appears from orbit?

 

MSFS2020 has real time weather but that doesn't necessarily mean that the game itself is running weather simulations. I suspect it has external queues being fed to it and its just procedurally generating storm animations from where its being told to put them.

[Vl3d]

8 minutes ago, mcwaffles2003 said:

Why would you see a hurricane from orbit though if there isn't one there? It's not like clouds from orbit will be static features and just because a storm animation starts with a clear day doesn't mean the animation has to begin from start if you intercept a storm in progress. I mean Im going out on a limb here as I'm not a game dev, but couldn't a storm animation like I've stated earlier be broken up into segments and have it begin from the appropriate one to the environment as it appears from orbit?

 

MSFS2020 has real time weather but that doesn't necessarily mean that the game itself is running weather simulations. I suspect it has external queues being fed to it and its just procedurally generating storm animations from where its being told to put them.

Read about the implementation above. It only needs the visuals.

Edited June 30, 2022 by Vl3d

[Pthigrivi]

12 minutes ago, mcwaffles2003 said:

Why would you see a hurricane from orbit though if there isn't one there? It's not like clouds from orbit will be static features and just because a storm animation starts with a clear day doesn't mean the animation has to begin from start if you intercept a storm in progress. I mean Im going out on a limb here as I'm not a game dev, but couldn't a storm animation like I've stated earlier be broken up into segments and have it begin from the appropriate one to the environment as it appears from orbit?

 

MSFS2020 has real time weather but that doesn't necessarily mean that the game itself is running weather simulations. I suspect it has external queues being fed to it and its just procedurally generating storm animations from where its being told to put them.

Ah, missed the subtle distinction. I agree you do need to match patterns on the surface to the simplified, animated texture you see from a distance, but you don't need to do complex atmospheric modeling or anything, just a set of animated regions that trigger various visual effects. Obviously complex modeling is not gonna happen. Even the former seems like a whole lot of work for a visual effect that isn't important to gameplay. 

[jost]

1 hour ago, Vl3d said:

Read about the implementation above. It only needs the visuals.

Well the implementation is actually not a full blown simulation but the use of precalculated data because you would need high performance (like used in clusters in big data centers for scientific purposes) systems for a real simulation. Which is fine for it's intended use case: Teaching how to do atmosphere science with MPAS (the software to do the simulations): https://www.kerbalwxproject.space/presentations  
The developer also has an image of the server to do the calculations: It took around six hours:
https://www.kerbalwxproject.space/background

Please note, that this is just for Kerbin: For every other body with weather it must be done too!

So this kind of simulation is obviouvsly not possible in realtime. But let's say we stick with precalculated data just doing visuals:
With the detail you envision (by comparing to other games with different goals/gameplay) they will need appropriate hardware. Which will propably be not what every potential customer have. 

Not to be misunderstood: This is NOT intended as a bashing against KWP. I think it's a great project and  can't image how much work it needed to get it running plus the writing and publishing the papers based on it. The developer has been invited to scientific conferences because of it and quite rightly and highly deservedly so. So: Big kudos and respect to cmet24!
 
But: It's just not a good base  to talk about the idea of weather in KSP2

[mcwaffles2003]

12 hours ago, Pthigrivi said:

Ah, missed the subtle distinction. I agree you do need to match patterns on the surface to the simplified, animated texture you see from a distance, but you don't need to do complex atmospheric modeling or anything, just a set of animated regions that trigger various visual effects. Obviously complex modeling is not gonna happen. Even the former seems like a whole lot of work for a visual effect that isn't important to gameplay. 

Whether or not weather (is that a pun?) is important to gameplay is subjective and I would tend to disagree. As I mentioned earlier in the thread I believe atmosphere beautification is at a similar priority as terrain beautification. Both give a planet it's own identity and appeal. There's a reason no one visits Dres and that's because it both is not in the Kerbin system, and is plain dull. The gameplay of KSP certainly centers around building craft and having them fly about, but I would argue giving a desirable destination to reach enhances the core gameplay by incentivizing the player to actually engage with the gameplay. 

Personally, I don't think that KSP 1 should set the bar/standard for gameplay visuals, at all. As great a game as it is, the visuals have always been severely lacking. Personally, I don't think that just because clouds were never in the stock game that getting just clouds brings the game up to modern standards. Also, who here was put off seeing the updated terrain visuals and saying to themselves that it's an unnecessary effort that will hurt performance? I remember a general attitude akin to "wow, that's amazing, I can't wait to land on that." Heck the witcher 3, a game from 2015 had pretty weather effects, why not this one? Now if implementing it actually caused a significant performance hit I understand not introducing these kinds of systems, but that said, I don't believe the performance hit would be as significant as some here may make it out to be. Not to mention, just as all things graphics... if this were implemented, wouldn't it be expected there would be a toggle? Every other game with weather effects has one, why not this?

Now would something like this take a lot of effort to create? Undoubtedly. 

But my opinion is it would be worth it. I'd love to fly down to the surface of a planet and see a hurricane in action, or even possibly some seasonality to planets on tilted axes. 

10 hours ago, jost said:

Well the implementation is actually not a full blown simulation but the use of precalculated data because you would need high performance (like used in clusters in big data centers for scientific purposes) systems for a real simulation. Which is fine for it's intended use case: Teaching how to do atmosphere science with MPAS (the software to do the simulations): https://www.kerbalwxproject.space/presentations  
The developer also has an image of the server to do the calculations: It took around six hours:
https://www.kerbalwxproject.space/background

Please note, that this is just for Kerbin: For every other body with weather it must be done too!

So this kind of simulation is obviouvsly not possible in realtime. But let's say we stick with precalculated data just doing visuals:
With the detail you envision (by comparing to other games with different goals/gameplay) they will need appropriate hardware. Which will propably be not what every potential customer have. 

Not to be misunderstood: This is NOT intended as a bashing against KWP. I think it's a great project and  can't image how much work it needed to get it running plus the writing and publishing the papers based on it. The developer has been invited to scientific conferences because of it and quite rightly and highly deservedly so. So: Big kudos and respect to cmet24!
 
But: It's just not a good base  to talk about the idea of weather in KSP2

With my vastly limited experience running 3D simulations and CFD, the the speed at which a simulation can run is entirely dependent upon the resolution it's operating under. So I honestly wonder if cranking down the resolution on a system like shown above could have it run with reasonable performance, giving a very basic weather system from which possible centers and magnitudes of storm systems could be extrapolated as seed for a lighter running system of procedural generation of storms.

Probably the most efficient way to run a weather system though would probably just be a repeating system that animates over a set time period kind of like a giant global 3D gif.

[darthgently]

23 hours ago, mcwaffles2003 said:

I don't think full on weather sims with accurate storm systems and consistency are required

Yeah, I was just wondering. Visual only would be a doable in a mod I'd think. It doesn't have to be integral to the game

[Vl3d]

17 minutes ago, darthgently said:

Visual only would be a doable in a mod I'd think. It doesn't have to be integral to the game

Are you going to create that mod?

[Bej Kerman]

24 minutes ago, Vl3d said:

43 minutes ago, darthgently said:

Visual only would be a doable in a mod I'd think. It doesn't have to be integral to the game

Are you going to create that mod?

There's no point in asking these kinds of questions up to a year before release.

Edited July 1, 2022 by Bej Kerman
Better phrasing

[gussi111]

It would be nice if they could at least implement those footprints, that would be awesome, I would love to take a screenshot of my first footprints on Mun in KSP 2

Edited July 1, 2022 by gussi111

[darthgently]

On 7/1/2022 at 10:13 AM, Vl3d said:

Are you going to create that mod?

I thought you were given your hype level for the idea

[Vl3d]

If I'll ever make a mod I'll just do it for my personal enjoyment and utility. I won't say "that should be left by the devs to the modders to work on for free".

Edited July 5, 2022 by Vl3d

[darthgently]

On 7/5/2022 at 3:01 PM, Vl3d said:

If I'll ever make a mod I'll just do it for my personal enjoyment and utility. I won't say "that should be left by the devs to the modders to work on for free".

Well what I wrote was that it could be done in a mod. Presumably for the modder's personal enjoyment and utility. Your enthusiasm seemed to indicate that you would enjoy it and find utility in it.  Not sure where this is going. But my secondary implication is that something like this could probably be more easily incorporated into an already existing EVE for example for which anyone could do a pull request for their personal enjoyment and utility

Edited July 9, 2022 by darthgently

[Vl3d]

Oh my...

Hat off. This combined with Parallax 2 sets the bar in orbit.

[t_v]

Many people have said this before: there is no bar. At least, not one set by these mods. KSP 2 can look worse than this and still be a good game, just like KSP 2 can look worse than MSFS and still be a good game. The quality of these mods does not put any pressure on KSP 2 because reasonably, KSP 2 should be judged on its own merits rather than expected to keep up with products specifically tailored towards a specific aspect. 

[Vl3d]

If individual modders can make something this incredibly beautiful for KSP 1, then of course I'm going to expect better from the actual dev team for KSP 2. It's actually in the name: KSP 2=1+1. There's a +1 that automatically invites comparisons.

[The Aziz]

Does KSP2 look better than KSP1? Yes.

Case closed.

[BowlerHatGuy3]

10 hours ago, Vl3d said:

Oh my...

Hat off. This combined with Parallax 2 sets the bar in orbit.

This is amazing! Hope we see something similar KSP2.

[tilliepops]

10 hours ago, BowlerHatGuy3 said:

This is amazing! Hope we see something similar KSP2.

we will take a look at this:

 

Edited November 5, 2022 by tilliepops

[Vl3d]

can you imagine? you don't have to..

Edited December 1, 2022 by Vl3d

[LHACK4142]

16 hours ago, Vl3d said:

can you imagine? you don't have to..

Plot twist: I don't have a great computer. I do have to imagine. Please do stop setting arbitrary expectations for KSP2. You're comparing the base game of KSP2, which has to run on cheaper PCs while still looking good, to a heavily modded KSP1 that can only run very smoothly with high part-counts on a really quite nice PC. In other words, you're pointing out that KSP1 on super-ultra-insane graphics looks better than KSP2 with medium graphics, which is rather pointless. This is all not to mention the fact that KSP2 will be better optimised than KSP in other aspects such as dealing with big ships. KSP2 with all the graphical  mods that will come out eventually will look as much better than KSP1 as you're expecting. 

Quote

if individual modders can make something this incredibly beautiful for KSP 1, then of course I'm going to expect better from the actual dev team for KSP 2. It's actually in the name: KSP 2=1+1. There's a +1 that automatically invites comparisons.

You shouldn't really expect better from KSP2's dev team. KSP2's graphics are supposed to be a lot better than KSP1's, which they are, and not KSP1++++.

Edited December 1, 2022 by LHACK4142
added thing

[Strawberry]

As LHACK mentioned, Id rather devs focus on things that 100% of players can enjoy instead of something that only 5% of the players can enjoy. 

Optimization concerns aside, as pretty as weather is, I dont see what it adds from a gameplay perspective. Since colony management is intended to be something thats passive, I think weather that has effect on those wouldn't be fun as having to stop each of your timewarps 3 seperate times when some of your 50 colonies has a weather event that effects them. For normal rocketry, see this thread for the issues:

Having to timewarp a day just to land your spaceplane because it happens to be raining and thus you cant land on your runway is cool in paper but in practice itd get tedious quickly. 

[magnemoe]

Weather has in game effects, landing a plane with an low cloud cover is a problem for one. Wind also affect planes, its also hard to use parachutes to help land tall rockets if its wind, its an reason its only planned for rarely. 

[Cytauri]

On 7/1/2022 at 7:17 AM, Bej Kerman said:

There's no point in asking these kinds of questions up to a year before release.

Hi, I'm from the future, and I am going to inform you that on July 1st, the date in which you have written that, it will be exactly, 8 months, 23 days, 13 hours, 34 minutes, 12 seconds, 9 milliseconds, 92 microseconds, 17 nanoseconds, and 51 yoctoseconds, until Kerbal Space Program 2 comes out. Also, I hope that KSP2 will have burn marks on space craft and debris. 

(I reread this post exactly on December 20th 2023, this post came out on December 20th 2022! The game went full circle for me and because the For Science update came out, I feel just as excited as I was exactly a year ago. Great Job Devs!

Edited December 21, 2023 by Cytauri