Why Duna is like Mars and Eve isn't like Venus ?

[Whirligig Girl]

I'm okay with a purple eve, just one with a more pale dull color and MUCH thicker atmosphere.

[Deadweasel]

Of course you can generate low detailed planets, like current KSP planets, with a simple algorithm and get whole galaxy with billions planets with different sizes, colors, surface roughnesses, and gas pressures. But I think that it would not be very interesting if planets lacks of fine details. Producing fine details, like weather, small scale topography, all kind of stuff on surface etc. is very hard. You must precalculate whole planet, which needs insane amounts of RAM and storage memory, or create details in real time, which means huge computing costs.

I prefer that Squad should develop current planets instead of create a lot of new ones which are practically similar except size, color etc. On every real planet there are different regions but KSP's planets have maybe couple of anomalies and couple of interesting topographic places. I have driven hundred kilometers on Eve's surface and there was not any interesting details. I would like to see rivers, lakes, stones, small scale hills, rains, sunny times etc. and take into account their effects to kerbals, vehicles (for example cooling, wet soil ..) instead of ten new detailless planets of moons. KSP have already small, medium, large and impossible celestial bodies. Easy, hard, and impossible places to visit etc. Plain planets and mountainous moons. Places with different atmospheres. In my opinion there should be more variables before increasing number of celestial bodies would be reasonable.

It is true that FTL can be limited to interstellar transitions. But game's current simplified physics model can not handle really different solar systems, for example multiple stars, twin planets etc. which would be interesting challenges to orbit planning. Other solar systems would be very much like Kerbol's system.

I see what you're getting at with the planet details, and you're right: it would take a ton of power that most people don't have to simulate active conditions on planets, but to your later point regarding simplified physics, Unity isn't going to be able to do that sort of detail. Harvester has been working on a neat process to get more details out of a surface, as currently all of the surfaces on the planets are simply hand-painted height maps, which results in fairly low details (seriously, would you like to paint an entire planet's-worth of interesting topography? It'll take months for one person to produce the kinds of details you're talking about.)

That said, again, planetary bodies don't really have physics applied to them on a system-wide scale. None of the orbiting bodies have any influence over one another. The flip when transitioning from one sphere of influence to another is abrupt and instantaneous, which is why there are no lagrange points between Kerbin and the Mün, for example.

Because of this, and because none of the terrain details are loaded when at a distance, it should be possible to have just about any layout you can imagine, though I would think overlapping spheres of influence might just result in the birth of another kraken.

Here's a direct rip of what's loaded and running in the game, with one Kerbal standing on the runway, and nothing more:

That's a decent chunk of the planet's surface that's in view of the camera at the time. The player location is in that tiny square with the black artifacts in the lower right frame. The KSP itself occupies a teeny tiny corner of that square.

The rest of the mesh is composed of two layers. There is the general planet's spehere, and directly on top of that is another layer containing the topography. In the upper left frame, you can see the KSP's deep red quad, and the rest of what's in the camera's view is produced outward from there. Notice how the mesh gets larger and larger as it gets further away from the origin point. The game is showing progressively less detail the further you get from the camera's point.

Here's the important part:

At the time of this capture, the Mün was at a point in its orbit where, if it were being rendered at all physically, it should have been visible in one of the views of this capture, yet there is no other mesh visible, demonstrating that the game does not load and maintain objects beyond a certain point. After that point, it seems they're little more than a 2D texture in the sky.

Anyway, all this is to show that hugely improved terrain detail and more and varied worlds are absolutely possible, it just takes a lot of time to actually create it. "Rome wasn't built in a day", and all that.

[the_bT]

[...]

Here's the important part:

At the time of this capture, the Mün was at a point in its orbit where, if it were being rendered at all physically, it should have been visible in one of the views of this capture, yet there is no other mesh visible, demonstrating that the game does not load and maintain objects beyond a certain point. After that point, it seems they're little more than a 2D texture in the sky. [...]

They're not, your capture is just incomplete.

As far as I remember, before the main game scene is rendered, the game renders a separate scene with all the planets. After that the depth buffer is cleared and everything else is drawn on top. This capture apparently only shows the second part of that process. Needless to say, the scene that is rendered first with all the planets in it shows only low detail versions. It's basically a 3d Skybox with moving objects in it.

[Deadweasel]

They're not, your capture is just incomplete.

As far as I remember, before the main game scene is rendered, the game renders a separate scene with all the planets. After that the depth buffer is cleared and everything else is drawn on top. This capture apparently only shows the second part of that process. Needless to say, the scene that is rendered first with all the planets in it shows only low detail versions. It's basically a 3d Skybox with moving objects in it.

Ah, well that makes perfect sense too. Either way, I'm still not seeing how there'd be a problem with adding a bunch more celestial bodies to even the Kerbin system (as long as there wasn't an expectation for active events like real-time weather and waving trees and other related nonsense)

[Erlendftw]

Of course you can generate low detailed planets, like current KSP planets, with a simple algorithm and get whole galaxy with billions planets with different sizes, colors, surface roughnesses, and gas pressures. But I think that it would not be very interesting if planets lacks of fine details. Producing fine details, like weather, small scale topography, all kind of stuff on surface etc. is very hard. You must precalculate whole planet, which needs insane amounts of RAM and storage memory, or create details in real time, which means huge computing costs.

I prefer that Squad should develop current planets instead of create a lot of new ones which are practically similar except size, color etc. On every real planet there are different regions but KSP's planets have maybe couple of anomalies and couple of interesting topographic places. I have driven hundred kilometers on Eve's surface and there was not any interesting details. I would like to see rivers, lakes, stones, small scale hills, rains, sunny times etc. and take into account their effects to kerbals, vehicles (for example cooling, wet soil ..) instead of ten new detailless planets of moons. KSP have already small, medium, large and impossible celestial bodies. Easy, hard, and impossible places to visit etc. Plain planets and mountainous moons. Places with different atmospheres. In my opinion there should be more variables before increasing number of celestial bodies would be reasonable.

It is true that FTL can be limited to interstellar transitions. But game's current simplified physics model can not handle really different solar systems, for example multiple stars, twin planets etc. which would be interesting challenges to orbit planning. Other solar systems would be very much like Kerbol's system.

I really, really agree with what you are writing. When I read your text, I tried to imagine landing on planets with rivers and wet soil, finding the perfect landing-area, analyzing the planets, sending rovers, building bridges / bases in areas. Imagine if the developers added a new planet similar to Kerbin, and adding forests, rain forests, mountains, lakes etc, it would be fantastic!

[Hannu]

I see what you're getting at with the planet details, and you're right: it would take a ton of power that most people don't have to simulate active conditions on planets, but to your later point regarding simplified physics, Unity isn't going to be able to do that sort of detail.

OK. I do not exactly know what is that Unity and why it has some limitations compared to common programming languages, for example C.

It'll take months for one person to produce the kinds of details you're talking about.)

I know. I have sometimes tried to program some kind of algorithms to produce game worlds, but it is quite impossible to get fine results as a hobby. It needs huge amounts of time.

That said, again, planetary bodies don't really have physics applied to them on a system-wide scale.

Yes, it is clear restriction that dynamic physical models are impossible. But I think that by using some kinds on noise algorithms it would be possible to generate nice "pseudophysical" details automatially. For example small scale topography, weather etc. It is possible to compute values of variables at certain place and time without computing previous values by using suitable noise functions.

But I do not know is that possible in Squad's game system. Maybe not, if it has some limitations. But it would be possible by using any universal programming language. However, maybe it is not possible under gamemaker's economic resources. Of course KSP is just a game (in my opinion one of best ones) and some hopes are unrealistic.

[Deadweasel]

OK. I do not exactly know what is that Unity and why it has some limitations compared to common programming languages, for example C.

Unity is the "pre-packaged" engine on which KSP runs. It provides the basis for the sound, graphics and physics-handling capabilities of the game. As such, it offers a shortcut to getting a game idea off the ground much more quickly and easily than if they were to attempt to code their own core engine from-scratch (which requires the talents of a lot of people, not to mention money).

I know. I have sometimes tried to program some kind of algorithms to produce game worlds, but it is quite impossible to get fine results as a hobby. It needs huge amounts of time.

Well, again, remember that the worlds we're currently playing with have terrain that was NOT generated by a program, but in fact hand-painted. Harvester has shown off some work he's been doing with generating more detailed craters using a programmed algorithm, which we may actually get to see on the Mün with the next update.

Yes, it is clear restriction that dynamic physical models are impossible. But I think that by using some kinds on noise algorithms it would be possible to generate nice "pseudophysical" details automatially. For example small scale topography, weather etc. It is possible to compute values of variables at certain place and time without computing previous values by using suitable noise functions.

You might be misinterpreting what's going on when people are talking about "dynamic terrain generation". See, terrain creation is done beforehand, during development. All that's being displayed is a surface mesh on a sphere that was created and exported during design time. What we see in the game is a terrain mesh that increases in detail as we get nearer to it (so the game doesn't have to bother with rendering polys that would most likely not be visible to the player from a distance anyway). Otherwise, that terrain detail is static and pre-defined. What folks are getting antsy about is seeing more details in the terrain, which would be made possible by creating an algorithm to produce more detail and taking the time demand of hand-painting heightmaps off the developers' shoulders, which can then be exported and applied to the planets/moons.

But I do not know is that possible in Squad's game system. Maybe not, if it has some limitations. But it would be possible by using any universal programming language. However, maybe it is not possible under gamemaker's economic resources. Of course KSP is just a game (in my opinion one of best ones) and some hopes are unrealistic.

Many things are *possible*, yes, but there has to be a balance struck between "realism" and play-ability. Too much detail can bring the game to a crashing halt just as certainly as a thousand-part ship. Every polygon requires computational time to render, and the engine's desire to work out current physics interactions with each and every part of a ship at any given moment only add to the load.

My original point was that, because the planetary terrain meshes are static and pre-defined, and because those details aren't even loaded and displayed until the player comes near enough to one of them anyway, it's not as system-damning to add even dozens more celestial bodies as some people insist on claiming. At a distance, they're all just simple spheres with textures on them, if that.

All that said, it's going to be interesting to watch the outcry when more detailed terrain does become a reality, and all those people already going on at-length about bad framerates on their 7 year-old business class laptops will suddenly see the game become almost unplayable because of the additional poly count when in orbit or at ground level. They're already having "issues" with frame lag in some cases just by looking at Kerbin.

Edited July 13, 2013 by Deadweasel

[spikeyhat09]

doublepost (why the hell is there no delete function? or am i just incompetent..)

Edited July 13, 2013 by spikeyhat09

[spikeyhat09]

i think its very simple. the kerbin-mun system resembles the earth-moon system because its what we home in the kerbal universe. also, they were to first and second bodies added to the game (other than Kerbol), so it makes sense that it would look alot like earth-moon. mars is a very favored planet, especially since the curiosity landing was close to the 1.7 update (little over a month afterwards, so the hype was still close to its peak).

[ZelvaCZ]

Why there are Minmus then? Why that green jupiter has Mun with oceans and atmosphere. I don't meant to disrespect you, but this is only game that is set in fictional universe.

Eh.. Gas giants can be green - ice giants in our solar system (Uranus & Neptune) are blue-green. If it's composition was slightly different, they can be green.

And Laythe? Its atmosphere can contain greenhouse gases, e.g. carbon dioxide.

[princess venus]

I think that purple liquid evaporates and rains down and leaves purple color to soil. I suggest that Squad should add thick clouds, heavy rains, rivers, and smaller lakes. Purple color should be lighter on areas where rainfall is low.

that makes sense thanks for that feed back and i agree on the rainfall thing

[Annikovka]

I think it is good as purple because it makes it all more interesting, jool for example is interesting to me because it is green, I think it adds differences to the game however it is fairly similar in that it is hot, and high gravity

[Kryssandra]

I don't understand why everybody wants to precisely model our solar system in KSP. Our solar system is boring.

Our solar system is incredibly fascinating if you actually take the time to learn about the wonders it holds. If you honestly believe our solar system is -boring-... that saddens me, and I feel deeply sorry for you.

[Rage097]

Duna does in fact seem like mars. HOWEVER. Nothing in the Kerbol system is exactly the same as our solar system.

For example, Duna has only one moon while Mars has two. Earth has a single moon while Kerbin has two. Dres is Dres. Jool doesn't have anywhere near the amount of moons that Jupiter does. Eeloo is close to pluto, but not quite. Eve has a moon while venus doesn't, not to mention it's purple.

The Kerbol system is not an exact replica of our solar system, therefore Eve =/= Venus.

[Sirius Celestia]

Eve actually is quite Venus-like.

1. high temperature atmosphere (check the 2hot thermometer)

2. high pressure atmosphere (check the barometer)

anyway, some of it is a bit different

1. bigger than kerbin (which it means it was bigger than earth. even actually kerbin is smaller than earth, but actually kerbin is earth-analog)

2. Purple (what kind of chemistry is that?)

3. strange liquid water, what is that thing? molten rock/metal? (if it's actually molten metal or rock, i will call that venus-like)

[GDJ]

Well, if Squad made the Kerbol system just like ours, it would be MASSIVELY HUGE compared to what it is right now. The Sun would be large enough to engulf Eve, Moho would be were Eeloo is right now, and from there you can extrapolate the scale.

It's a tiny system really, and eventho' the Kerbol star is nowhere massive enough to be a star in reality (mass-wise, it's big enough to be a white dwarf.....maybe), it's just fine the way it is.

I think there is enough planets to keep us occupied......altho another star with 2 little planets orbiting would be cool......say 1 LY away.......

Edited December 24, 2013 by GDJ

[lajoswinkler]

Eve actually is quite Venus-like.

1. high temperature atmosphere (check the 2hot thermometer)

2. high pressure atmosphere (check the barometer)

anyway, some of it is a bit different

1. bigger than kerbin (which it means it was bigger than earth. even actually kerbin is smaller than earth, but actually kerbin is earth-analog)

2. Purple (what kind of chemistry is that?)

3. strange liquid water, what is that thing? molten rock/metal? (if it's actually molten metal or rock, i will call that venus-like)

1. Temperature on Eve is -40°C to 150°C. Completely unrealistic and impossible. Venus has an average temperature of 460°C on the surface.

2. Pressure on Venus is around 92 atmospheres, on Eve it's 5. That's nothing compared to Venus. Eve really whould have extra pressure so that it crushes down probes and Kerbals, unless the equipment and Kerbal suits are pimped up (costly!) to withstand it. That would be a nice building and economy challenge.

3. Its diameter really doesn't matter here. The difference is negligible.

4. The color is unexplainable, but I'm ok with it. If it ever gets a thick, opaque shroud with low visibility on the order of a kilometre (I hope so, it would be very challenging because you'd have to use a radar upon landing), it could be somewhere between pink and violet.

5. Liquid bodies on Eve might be water, if the environment is below water's critical point of 374°C and the pressure sufficiently high. In fact, that would be a nice thing for Eve.

Venus doesn't have molten metals or rocks on the surface, except where potential volcanism occurs, for which we don't have direct evidence at this moment. Venus is probably a lot less volcanic than Earth.

[boomerdog2000]

We have a completely brand new solar system to design and play with. It seems perfectly reasonable, if not expected, to make things different from our current one. I have zero problems with there not being analogues for everything in our solar system or for planets being completely different from our system.

[Wallace]

That there picture doesn't tally with the real pictures of the surface of Venus, which looks more like a badly made car park, with flat slabs of rock. about 80% of the surface is flat volcanic plains, which is what you'd expect with huge flood basalts and no rain on the surface to erode it into valleys.

Also, the volcanism is likely much less than Earth. Venus does not have plate techtonics, instead it is believed that heat and pressure build up until there is a huge volcanic event that entirely destroys the crust, then it subsides. The crust is all a uniform age of about 300–600 million years old. Venus does have far more volcanoes than Earth, and much much larger, but that is because on Earth, extinct volcanoes are continually destroyed by the plate movement and erosion, whereas extinct Venusian volcanoes are not.

lastly, volcanoes on Venus do not look like most volcanoes of Earth, well, certainly not the pointy composite cone ones like in that picture. Venusian volcanoes are smooth, either shield volcanoes or pancake domes, and they are stupendously large, but mostly really flat compared to earth volcanoes.

http://en.wikipedia.org/wiki/Volcanism_on_Venus

[iDan122]

In Space Pioneer, they actually told about plans to make Venus habitats. At first I though the guys were mad, but then they said that there would be no surface base, it would be hanging in the atmosphere on oxygen-filled balloons. Wish we could do that on eve...

[panfish5]

If Eve was more like Venus I think it would be a bit boring. Some volcanoes would actually be really cool and a cloudy atmosphere but anything more than that it would be boring in my opinion. My favorite part of the game is probably the sur-real planets give it a good twist and I hope ksp sticks to this.

[tntristan12]

I think the reason Eve doesn't look that imposing is because of the current lack of atmospheric effects in the game. Re-entry on Eve tends to be extremely hot and violent, but you don't really notice because deadly re-entry has yet to become a thing. Furthermore, you don't get a sense of the thickness of the atmosphere because there's no visual distortion due to the atmosphere (this is the same reason why night on Kerbin looks so unnaturally clear). It'd be cool if the thicker, hotter, and heavier the atmosphere was, the blurrier vision became. Then Eve would be really forboding!

[zxczxczbfg]

It was previously thought that Venus was covered in planetwide oceans, and Squad may be acknowledging that. Also, it is similar to Venus in that the atmosphere is thick as Elmer's glue and the temperature is blisteringly high.

[lajoswinkler]

Here's a phase diagram of water with current Eve's properties on zero elevation.

The vertical stripe represents the temperature fluctuation, and the horizontal line represents the constant pressure. Dark line where they meet is how the water will behave with the change of temperature.

So you can see that water will melt at basically 0°C, but it will remain liquid until few degrees above 150°C, something which doesn't happen on Eve. Conditions on Eve are worse than in a pressure cooker (~1 atm higher than normal pressure), but really nothing compared to Venus.

I'd like Eve to have a constant (or with 10 degree oscillations) temperature on zero elevation of ~250°C. As for the pressure, at least 50 atm. It really doesn't have to be just like Venus, that would be unimaginative, but current values are simply too weak.

Atmospheric distortions are hard to implement because they're different at different heights. That isn't really required, but heavy fog would be nice. 500m visibility would look rather scary, and would force you to use a radar. To spice things up and make it more realistic, the fog might be densest at a certain height, with clear view at zero elevation, so that the mountains and hills would gradually turn to nothing as you look up their slopes.

Because Venus in true color looks pretty much like a featureless bright ball

and that's really boring, it would be cool if Eve's opaque atmosphere would look like Venus in ultraviolet, with false color.

To sum up, Eve needs rougher surface conditions and atmospheric stratification.

[TheGeoninja]

I think that eventually most of the planets in KSP will become much deadlier with radiation and such and temperature is a bit wonky at the moment so I would say eventually Eve will become somewhat more like Venus but with a different color.