I think that if there are likely to be planets in I:B with atmospheres, even some simple trickery clouds to pretty things up would make a big difference to immersion. Something to act as a placeholder until the real thing can be coaxed into running smoothly…
Trees can come later.
How about matching players by settings then or at least give them te possibillity to? 
That’s something, that’s missing in every game so far. An in many games it really does make a difference if your settings are high or low.
I really like the idea of fog layers.
Here’s what has been done with trees and clouds in Outerra:
http://www.outerra.com/forum/index.php?topic=2148.0
The trees seem to be just 2D-sprites and the volumetric clouds are not yet that refined (blobby, blurry). But it still adds a ton of value.
They have got some nice clouds (still WIP though).
@INovaeKeith Aren’t those clouds on the big planet? are they just hideous close-up?
Just curious.
one of the main infinity engine feature is seamless planet landing so you should expect that players will want to use this feature. clouds and various atmospheric conditions is very important to make planets not just rocky moons or gas giants. trees are not so important. on the next step you should think not just about clouds but also about lightings, winds, harm environments like heigh pressure and temperature, acid rains, atmospheric icing an other “fun” moments.
players should have many ways to kill themselves close to surface. not just hitting it.
Another good reason to try to implement clouds. Gas giants kinda need them to look good. Otherwise flying into the upper atmosphere of a gas giant is just flying into a layer of fog… instead of this.
Mmm, yes, would definitely be nice flying through clouds like that. After all, Jupiter has some of the most epic-scale storms in the solar system (barring the Sun), so decent clouds for gas giants is the way forward.
I disagree, clouds and basic vegetations is very important to any planet, just like terrain or lakes or asteroids or space stations. Obviously not every planet should have trees, in fact most of them should have no life at all, but volumetric clouds is a must. Especially for gas giants as you can simply make a tiny rock nucleus and simply have a HUGE atmosphere full of clouds which deals damage to you the lower you go (probably way more complicated than that.
So yeah IMHO clouds should be on the same priority as terrian fidelity and asteroid diversity. Trees can obviously go much lower since they only appear on life planets.
[quote=“Jinnay, post:10, topic:23, full:true”]Just remembering of how often i got stuck in a tree and died while trying to avoid getting shot or just have fun i’d say yes, it would add to the quality. But still not nearly top priority.
[/quote]Indestructable trees is just stupid. Obviously trees would get destroyed/broken in a very simplistic way (toppling with particle leaves is enough) and your ships takes some hull damage.
Here is what Cameni had to say about the clouds in Outerra:
The algorithms aren’t yet completely stabilized, as I tried to devise a new technique best suitable for OT and GPU processing. There were many problems along the way and at times it didn’t look as the right way at all, but most of them got resolved and I believe the remaining ones will be dealt with as well.
Once it’s all final and working as desired I shall write a blog about the technique.
Hope there will be something to learn from Outerra if you are still struggling with the clouds.
I wouldn’t say “struggling” is the right word, more like we haven’t bothered to try because we’ve got 1000 other things to do 
[quote=“INovaeKeith, post:31, topic:23, full:true”]I wouldn’t say “struggling” is the right word, more like we haven’t bothered to try because we’ve got 1000 other things to do[/quote]Are Outerra dx9? Is there anything in dx11 or dx11.1 that allows for an easier time to do clouds? Also about the art style, is battlescape aiming for photorealism or is there any stylisation (plastific, cartoony, caricature) to it, because those clouds look cartoony but would fit perfectly within a more stylised setting (outerra is photorealistic).
I know you guys are busy with other things but i got a few questions in reguards to clouds and trees:
Any ideas about using nvidia tech like physX, GI Works, OptiX or FlameWorks to generate the clouds (or equivalent from ATI or other companies)? I imagine FlameWorks being amazing at it. And to support non-nvidia hardware would it be possible to compute the clouds of planets on servers and send that to our clients via internet? Because if everyone uses the same star system generation seed code you know which bodies can use the cloud generator thus eliminating the need for that in clients. Perhaps this would be a “cheaper” system to deploy until you can develop a fully procedural cloud system later.
If you eventually start developing clouds do you think they should be fully interactive? Things like making holes in the clouds when a spaceship flies through them, making turbulence whirls as a ship passes through.
Do you plan to give each ship an atmospheric drag coefficent number based on ship design? A bulky squary spacecraft would be given a high number thus affected by air a lot more and possibly taking more damage and being destroyed on high speeds. Would love to get a small sleek ship and being able to get away from pirates in atmospheres where they can’t follow.
Any word on gas giants? Can we fly into the clouds like this?
What’s the status of the trees currently? Is it “tried to test things but didn’t really do anything yet” or “almost done actually, just need to figure out X and Y and a few bugs”
If trees happen will be be the typical pine tree x 10000000000 or will there be different trees, different sizes, different (alien) shapes?
Will trees be indestructable (i hate you, Planetside 2) or will they be interactive (broken/crushed/burned…)?
Using physics to generate the clouds is not a good idea, the grid resolution would not be sufficient at all to work at a planetary level, and those clouds would be generated using iterative simulation steps and not procedurally, which we need. I haven’t investigated those libs in too much details but my bulk feeling is that no, we can’t use them, we’ll have to find a solution that works procedurally.
If clouds are procedural it’ll be kindda hard to have dynamic turbulence and whirls based on ships movements. If clouds aren’t procedural, then the resolution isn’t good to cover an entire planet. So it’s not too likely to happen, sorry.
Yep, we’re planning for something similar to this for atmospheres.
Yeah, that’s what I have in mind for volumetric clouds on gas giants. With some technical limitations, of course.
I’ve tested procedural placement of tree meshes on planets in the past, but that was just as a placeholder. We need a better solution that can provide a higher density of trees, and better performance. So that’s R&D, but no, we haven’t done any significant work in that area yet.
In the future we’d probably like to generate the trees procedurally, based on their planet environment. For Battlescape, if we add trees ( which isn’t guaranteed, especially if we raise the minimum amount for the Kickstarter ) then it’ll probably be some statically spawned meshes on the planet surface.
They’d probably be indestructable.
Hey,
its nice to see, that we are finally back to the community interactivity with the development.
I think the community really likes to see clouds. And I also see, that they are not just added like putting a wallpaper on a desktop.
So what exactly is the hardest part in adding clouds? Would it help, if the community crawls the web to find a procedural algorithm which is capable of generating nice clouds with decent performance? What are the pre-requisites for such algorithm?
No. Flavien & Keith are quite capable of finding & studying the applicable papers & whatnot to derive the best solution. It’s a matter of time/priority. The implementation is non-trivial, and it is not on the schedule to be implemented before the crowd funding campaign, that is that basically…
I’m sure if another coder could be found who would work for free, and could handle the INE codebase, as well as have as good or better handle on the volumetric could implementation than either Flavien or Keith (highly unlikely to find that), only then could it be a realistic possibility to have clouds anytime soon.
So not before the campaign itself. But could it fit in the planning after that, but before I:B official launch date? Clouds wouldn’t be that important for the Kickstarter demo (if it is still planned), but they would add much to I:B itself.
[quote=“INovaeFlavien, post:34, topic:23”]If clouds are procedural it’ll be kindda hard to have dynamic turbulence and whirls based on ships movements. If clouds aren’t procedural, then the resolution isn’t good to cover an entire planet. So it’s not too likely to happen, sorry.[/quote]What if it is done locally on the client? The cloud itself doesn’t need to be interactive but the client will generate those effects on top of the static clouds and the effect would be completely independent to each client and unrelated to server/map. I’m sure it is way more complicated than this but [Player In Cloud->Generate Turbulence Effect]?
What about various levels of detail meshes? From orbit or high altitude it would be a flat texture, getting closer would look like this and at a few hundred meters or close it would actually start generating individual trees. Or would that method incur too much work/performance because you effectively need to make three separate generators that must work together? I’m not a programmer but i’m still curious on how you are going to solve that challenge.
[quote=“INovaeFlavien, post:34, topic:23”]They’d probably be indestructable.[/quote]And by that you mean they would not physically interact with ships and simply be “transparent”, right? Not even a simple 3-frame topple animation on contact?
Keep up the awesome work guys! <3<3
That depends on the success of the crowd funding most likely.
I was busy writing a long reply to this topic, and I still am, but I figure I might show one of the first screenshots (it’s pretty old); which I uploaded to Facebook. Don’t bother with the absurdly high resolution, or the fact that the branches are placeholder. I was testing the invariance of the max-res for the LOD system. It’s a wireframe so I can see the geometry
https://scontent-a-dfw.xx.fbcdn.net/hphotos-ash3/941406_388157824626462_1713662591_n.jpg
Just wondering here, why insist on volumetric clouds?
I mean, I agree that volumetric clouds would be awesome, probably look great if you could get it to run. You can’t. Well, at least it seems that what you’re saying is that you’re very unlikely to get them to work on a significant percentage of your potential costumer base’s computers and thus would rather avoid spending time on them, which makes sense.
So why not put in a simpler version? Something like what Guns of Icarus Online does* - not very pretty, perhaps, but pretty enough to provide suspension of disbelief, a significant boost to gameplay, and can run on a potato.
*) I don’t actually know what they do, exactly, but from the way clouds always seem to turn around to face me I’m guessing it’s something like sprites and regions of increased atmospheric blur. You’ve already got atmospheric scattering, how hard can adding sprites be?
