Is it just me or is Kerbin's terrain shorter in ksp2?

[Mikhail Kerbachev]

It seems like mountains are much shorter than in KSP1, like the Kerbin's terrain was scaled down like 5~10x. Does anyone know if this is intentional, intended to be permanent, or the reasons why?

[JoeSchmuckatelli]

34 minutes ago, Mikhail Kerbachev said:

It seems like mountains are much shorter than in KSP1, like the Kerbin's terrain was scaled down like 5~10x. Does anyone know if this is intentional, intended to be permanent, or the reasons why?

Bunch of speculation - they shortened the mountains west for a better glide path to the runways... Also those may have been kept artifacts like the mohole in KSP1 that did not make it back into 2... And there is some indication from the terrain system rework post that they could more easily introduce interesting terrain as part of the rework 

Edited April 22, 2023 by JoeSchmuckatelli

[pandaman]

As I understand it all the celestial bodies got fully reworked and replaced.  But they reused the basic 'map' of Kerbin for familiarity reasons, and some 'old friends' from other bodies, like the Mohole and Dres canyon, were also remodelled.

[Lyneira]

Mountain ridges look great from high altitude and space, but I missed seeing some tall peaks and cliffs and valleys to fly through when I was just flying a plane near some mountains. I hope they'll get raised up again. And add some snow-topped peaks to them where appropriate, pretty please?

[Mikhail Kerbachev]

2 hours ago, Lyneira said:

Mountain ridges look great from high altitude and space, but I missed seeing some tall peaks and cliffs and valleys to fly through when I was just flying a plane near some mountains. I hope they'll get raised up again. And add some snow-topped peaks to them where appropriate, pretty please?

I totally agree. Looks great from space, but just not interesting airplane cruises.

[Jason_25]

Try going to the "Kimalayas" kind of southwest of the north pole.  You can see it below you and to the left when flying to the north pole from the space center.  It is a pretty high frozen plateau with many snowy mountains.  After I landed my F-32 near a peak, a cloud came by and "fog" rolled in.

[The Aziz]

Clouds are at 1km if I'm not mistaken, so 1km high mountain is.. quite small.

[Lyneira]

Did a quick flight to the north pole and while I found huge swathes of mountains, a few with snowy peaks, I was comfortably flying above them at around 1800 m altitude. The cloud layer seems to start at around this height too.

[VlonaldKerman]

The old mountains were disproportionately huge compared to the size of Kerbin. 6000+ m mountains on kerbin are like 60000+ m mountains on earth. The to-scale version of Mt. Everest in KSP would be ~800m tall.

That said, Kerbins atmosphere is also rediculously tall, and I’m not a stickler for perfect geological realism. I enjoyed the old mountains- I used to climb them on foot (because I’m weird and I’m a climber myself).

[Lyneira]

The surface gravity of Kerbin is the same as Earth's, so it seems to me like geological processes would be capable of forming mountains just as tall in absolute height. Maybe someone who knows a bit more about (exo)geology could chip in with some examples?

Approaching it from another angle: Kerbin's atmosphere height, being as tall as it is, is clearly aimed at being analogous to earth's atmosphere height. The KSC buildings are also similarly large in scale.  From this I gather that the terrain and atmosphere are intended to be similar to Earth's in scale from a ground point of view. Which would suggest mountains with a similar height range.

[Jason_25]

The cloud layer is closer to 2k I think.  That means these mountains were all higher than any mountain in the Appalachian range for example.  Because it was so far north it was snowy also.

There also looks to be a fairly solitary large mountain that you fly over on the next continent east from Kafrica.

[regex]

Looks pretty majestic to me...

Spoiler

But I'm probably easily pleased based on how harsh people are on this game. vOv

[Lyneira]

I was thinking more of steeper geometry like this that we're currently lacking:

Spoiler

 

Spoiler

 

[Jason_25]

This is a problem of "terrain resolution" and it is suprisingly common in flight simulators and other games to have bad terrain.  Microsoft FSX was bad about this making Mount Everest look like a lump of mashed potatoes.  Only maybe the latest versions of X-plane and Flight Simulator are getting better but all mountains in games also look too "smooth" like they are made of glass.  We know what in reality there are tiny cracks and jagged sections that people can climb on that are all but nonexistent in games.  But to have a whole planet like the pictures above would use a huge amount of memory.

[K^2]

4 hours ago, Jason_25 said:

But to have a whole planet like the pictures above would use a huge amount of memory.

You have to go procedural for fine features. One thing I've seen that shows a lot of promise is AI-generated erosion, which can take a low-res height map, and generate the fine features that can be old and smooth, new and jagged, and cut with water channels as appropriate. In combination with virtual textures, you can get a very realistic looking terrain with no memory overhead (since you want virtual texturing for visuals anyways) and limited impact on performance.

There are challenges associated with it. Running neural networks on GPU is still a significant workload, even if you split the work between frames using virtual textures. And this does alter the heightmap, so you need to integrate it into your physics somehow. That means having a compute pass for the terrain, or worse, moving all of your physics to the GPU, which would increase GPU workload even further.

Caveats above in mind, the proof-of-concept demos I've seen look great. With ray tracers being a hot topic in GPU hardware, we're getting more and more dedicated silicone for working with AI and BVH, which can both be leveraged for terrain upsampling and physics respectively. So I expect that we'll be seeing tech like this integrated into large open world games in the near future.

[BowlerHatGuy3]

On 4/24/2023 at 6:03 AM, Lyneira said:

I was thinking more of steeper geometry like this that we're currently lacking:

  Hide contents

 

  Hide contents

 

Oh yes! Cliffs in both games are so overly round. It’d be nice to see some drop offs like this.

[Wetzelrad]

Is Eve also shorter? I flew around for a bit and found the cloud cover to be at around 3000m. The highest hills I saw did not penetrate the cloud cover and only rise to around 1600m above sea level.  By comparison, the wiki for Eve in KSP1 says this under updates:

Quote

• Terrain tweaks — the tallest points are now about 6 km in altitude, compared to 11 km before.

[18Watt]

In my brief visits to Eve, I found the terrain spectacular.  Possibly not as extreme and sudden elevation changes as in KSP1, but spectacular.

Just did some rovering around Minmus, and the terrain there is much more extreme than it was in KSP1, in my opinion anyway.

So far, in general, I'm happier with the terrain in KSP2 than I was in KSP1.  I do think it is interesting comparing the terrain in the two games.

Edit:  I've only seen small portions of Eve, Minmus, and Duna in KSP2.  Haven't really even explored Kerbin in KSP2 yet.  So I do not mean to dispute anybody's observations.  Just adding my $0.02 worth..

 

[JoeSchmuckatelli]

On 4/29/2023 at 11:47 PM, 18Watt said:

In my brief visits to Eve, I found the terrain spectacular.  Possibly not as extreme and sudden elevation changes as in KSP1, but spectacular.

Just did some rovering around Minmus, and the terrain there is much more extreme than it was in KSP1, in my opinion anyway.

So far, in general, I'm happier with the terrain in KSP2 than I was in KSP1.  I do think it is interesting comparing the terrain in the two games.

Edit:  I've only seen small portions of Eve, Minmus, and Duna in KSP2.  Haven't really even explored Kerbin in KSP2 yet.  So I do not mean to dispute anybody's observations.  Just adding my $0.02 worth..

 

I agree - and would almost think it's a general consensus.  I think the thing some would like to see come back is a few places with somewhat improbable terrain transitions to make atmospheric flight fun (the old 'towers' west of KSC were always a great place to test spaceplanes). 

[Jason_25]

Please see the results of my recent mission which is pertinent to this thread:

 

[Lyneira]

Less than a third the height of the real Mount Everest of 8848.86 m. Even the much more modest Alps have many peaks above 4000 m. Although what also plays a role here is the "roundedness" of the terrain that seems to be affecting current mountains. You can easily stand on a peak and barely be able to look out over nearby terrain below you because the peak is actually a giant, slightly rounded, plateau.

[dr.phees]

Currently many people build planes, well for ... reasons. And the devs apparently like to fly them, too, based on their videos.

It should be obvious, that planes on a computer scream for interesting landscapes, cool mountains, peaks etc. It should be hysterically gorgeous.

Why the terrain has been flattened instead is beyond me.

[RocketRockington]

It's because of the limitations of the technology.  Making cliffs that are based on a height map look good is difficult for a variety of reasons to do with both texturing them and how the geometry of the cliffs is generated.  KSP1 mostly went with the 'its ok to be ugly, just make the mountains challenging'.  KSP2 is much more focused on visuals, especially visuals at a distance/from above for pretty vista screenshots, and as such they decided to reduce the mountain heights to make those look better.