Creating a fairly large map to fly through with jets etc. and running into "problems"

So i am trying to create a map which basically looks like this:

http://i.imgur.com/kypjOj0h.jpg?1

Not the same size (i am not crazy) and obviously not much detail (especially on the ground),
it will consist of maybe 50 distinct buildings (like skyscrapers and other distinct architecture) and a whole lot of repeated instances of smaller buildings (with very low tris)

if anybody here has ever played a flight simulator, ace combat, hawx or any other similar flight games with real looking locations will know what kind of size and detail i will be going for, minus the art style (i am not going for realism)

basically it has to be big enough for fly through with jets and similarly fast aerial vehicles without giving the feeling of a tiny world

i have started to model some of the buildings but before i continue doing so, i want to make sure i am on the right path and maybe even get some much needed advice from people who are far more experienced with UE4 than a newbie like me

**i understand that i will have to:

  • keep tris as low as possible for the distinct architecture and even lower for the filler buildings & architecture

  • create LODs

  • use instances for repeated buildings and architecture

  • use level streaming / world composition ??? **
    (question because i am wondering if that would stream fast enough for fast aerial vehicles without visible texture and mesh pop in etc.)

the problem i am running into is when i create my buildings with real world sizes (skyscrapers 300+ meters etc.) and it turns out that i would need really big worlds (who would’ve guessed that…), so i was wondering wouldn’t it be a better idea just to model everything smaller ?

60 meters instead of 300 for my skyscraper and so on… smaller mountains, trees, humans etc. - i am just inexperienced to the point where i am worrying that doing so could bite me in the youknowwhat later on, i will gladly take any advice you guys might have for me :slight_smile:

UE4 has been amazing and i am sure there is a better way to do this, than with what i have come up

If I were making a flight sim in UE4 (let’s pretend it’s not a crazy idea for an indie game) these are the things I would do:

First thing you should do is nail down the scale - don’t worry if you are doing it right/wrong, you will have to redo everything anyways.
I don’t care about real world units. You are really looking for feel. Does the world feel big or small? Then adjust.
You can scale terrain actors. I would do this instead of tiled worlds at least for now. Start with one. You get that working perfect, then go for tiled worlds.
You don’t need crazy ground collision- this is a good thing.
Bad thing- you don’t really have any way of engineering your level for visibility except for distance culling (maybe later you can do something with a cloud layer, etc.)
Get your pawn feeling like your game. Attach a rough cockpit to the camera. Play with speeds. Fly around your level, tweak, tweak, tweak.

Stop worrying and just do. You will never ship this game, trust me, it’s about learning at this point. It’s like riding a bike. You won’t learn by trying to figure out all the things that will go wrong. Once you know the concept (peddle with your feet) you just have to do it.

that is indeed a crazy idea, one i am not too keen on trying out, i just dropped those video game names to give an idea for what i am going for scale & detail wise for the map, i am not attempting realism, so your suggestion for getting the right feel is more important than real world sizes, still if anybody who has tried such a thing could give me any insights, i would be very grateful

yeah i have been thinking about just creating a relatively quick test version with the landscape tools and brush actors and trying to nail all the other parts before i commit (quick being relative because that would probably take a week at the very least), there is quite a bit of work that would be done twice though, hence me hoping for someone who has tried such a thing to save me from some common pitfalls, i understand if there are mistakes i have to make to learn from them :slight_smile:

just there is never enough time…

true, i really don’t plan on shipping anything though, just want to create a prototype and see how it plays out - no more or less