Find it on the workshop Christmas Morning.
https://www.youtube.com/watch?v=VLx18FNGifk
Oh my god
I can't wait to play on it
Looks awesome, but I problem I have with it, why are you using tree card textures rather than trees themselves, looks really fake with those there, I get that it's got to do with optimisation but that forest, or those mountain roads look really fake with those there. But either way, I'm excited to play on it.
Your assumptions would be right. It is too expensive to line all of the roads with a high density of real trees. I wanted to ensure that most of the roads in the map were in higj fps areas cause we all know how horrible it is driving in multiplayer. When I make RP maps, i aim more for function and what's going to be fun than beauty anyway.
Yep, I understand, but I do have a suggestion, you can follow the same technique gm_fork follows, essentially, rather than using the tree models, it uses a custom model with 2 polygons and an image of the tree itself as the texture, basically a billboard alternative to the original prop, it's cost effective, and looks relatively better than a card. Just some food for thought.
Definitely something to try. Since the tree walls are quite thin, i don't think that exact method would work. There's skybox walls right behind a lot of them. I wonder if it's possible to combine both methods. Have the 2d tree wall where it is with some card trees sticking out of it so it doesn't look so 2D. I don't think detailprops will be generated on vertical faces but i haven't messed around with them enough to know for sure.
You have sparked my interest at the very least. Thanks!
I don't see why not, could even LOD the models so they either fade or pop-out at a distance.
I'd make probably a 1024 segment of staggered trees, each with 4 planes in a 8 pointed star formation so 8 polys. Probably could fit 16 trees into it which we're still only looking at 128 polys per segment. I could make a lod model of 2 planes which brings it down to 64 if i neefed. Still. One car in most servers is 200,000+ polys. That's over 1500 segments without the lod. Would a lod even be needed at that point? Im not up to date with how hard polys are on players of 2018.
Hey @Statua.
A community I'm a member of is having a vote at the moment of what map that we will change to next. Would it be possible if you could post some more in depth pictures of the map? We're also quite curious about the size of the map. It would be awesome if we could see a level overview of the map.
I love the work and all the details you have put into this. Truly looks like a really good map!
Regards, Jompe.
Check out some of the photos I posted to my twitter. There's a decent level overview in there too.
https://twitter.com/Statua
Thanks! Appreciate it!
Polys are generally pretty cheap, even on garbo older GPUs. What can cause problems is drawcalls, and fillrate. Fillrate shouldn't be an issue if you're just using a few transparent sheets, but if particles or stuff get involved in can get bad. Drawcall is a big problem since it's CPU-side and a hard engine limitation.
Blueberry_pie did some benchmarks to figure out what worked best some time ago:
Model batching performance survey
I'd probably break out the tree-walls into a model editor, then create big batch sheets from that reference.
I've been considering converting them into models using propper so they cast shadows using a hitbox. It wont be a perfect cast unless I use textureshadows, but it'll at least look better than no shadow. I could probably also just slap on a bunch of nocull planes sticking out with flat brushes so it looks like some trees poking out. Would save some frames by batching as mentioned in the thread.
Drawcalls are more difficult to handle. Lots of assets can make things go out of hand quite quickly, especially in diverse roleplay maps that have a lot of different, unique areas. I'm sort of aware of what it is but its pretty hard to keep things slim sometimes and I don't have a method for keeping them down, yet.
I forgot about those links. As for linking my old account, it finds my old account ok but when I hit merge, it says Invalid Merge. Weird.
The flat planes are probably better off left as displacements or brushes, since you'll run into the issue of super-low lighting samples for static-lit props, especially on something that big I'd think.
I'd just go with adding planes "sticking out" of the forest wall, and adding more batched trees to cover it up. You can export the tree-lines using something like Craft, export to OBJ, and then import as reference geometry to model the planes in Blender or something.
Definitely something I'll have to spend some time on messing with. Maybe in an update youll see it. Or my next map, whatever it may be. Thanks a ton for the info!
I'll be looking out for this!
I'm also curious about what people are going to do at the railroad crossing with props and vehicles.
Its unbreakable. I've tried everything. The train tries to push it and if it gets stuck, it pauses for a second, then no collides itself with whatever it's stuck on and proceeds like nothing happened.
Wow, it's great seeing a new RP map with a tangible setting that actually looks nice. Looking forward to playing around on this!
*waves Canadian flag happily*
It's really clever how the the mountain region is actually a second level placed directly above the rest of the map.
Thanks! I'm quite proud of it. It's something I've been testing and experimenting with for awhile. Mostly getting it to feel seamless and still being able to see the 3D skybox.
Suggestions:
Add a Power Plant
Add a secret Combine base under the Police station
Add a Radio Station with a radio tower
I think improving on what's already in place is a better policy.
If anything gets added or changed majorly, it'll be for a v2 update. I don't want to be a dick like i did with rockford and start making huge forced changes. But your suggestions have been noted thank you.
Sorry, you need to Log In to post a reply to this thread.
