Dwarf City

[Falconius]

So a continuation of my Dwarf City from a dusty ancient challenge. It occurs to me that this is covers a bit larger area than a city map since it includes some close by localities. What would such a map be called? In any case it's primary purpose is to represent the city.
### Latest WIP ###
DwarfCityProgress1.jpg
Some things I discovered are that even 20 buildings isn't really enough. If you do simple buildings it's hard to build variety, and each time you place them it takes more work (since you may have to add other parts to them to help differentiate them. And that if you make more unique types of buildings they are too unique to use too many of them. I'm trying to work out a way to make more modular pieces. The first step of which I guess is having standard sizes...

I figured out the length of that wall (roughly) and it's about 4.2 kms long. And it so far has 139 towers. Which will be adjusted for gates and water outlets etc so it may end up being less. I'm not sure if I should have a smaller or older wall inside that one.

There should be trees appearing on all the lower green levels. I'm not sure why they haven't. I added the wall after doing the trees on the level they are visible, so obviously that will have to be adjusted. That level has 25,000 trees, apparently, Which is a scary prospect as it still looks somewhat sparse. This means that if I did more trees, say 50,000 for the lower levels and 25,000 for the others I'd end up with about 375,000 trees (since I think the treeline is at about level 10 or 11). Most of the trees average about 500 faces, but they go up to over a thousand for more complicated ones. So it could end up that my computer has to deal with well over 300 million faces (including all the other stuff) for a full shot. Not sure if it can. I need to figure out someway to mitigate the impact. Maybe use 2d trees? If I could make them track the camera somehow...

[Francissimo]

What you plan to do with trees is really ambitious! I've tried that too with sketchup, but my computer fainted before i can make a real forest. You can either try low poly trees, 2D trees or simply avoid tree in 3D and paint your forest in photoshop after you render your topography and building in 3D (good way to avoid your cpu melting).
Good luck with your project, if you computer can stand a forest with so much trees i'll be glad to see the result

[Falconius]

Found my trees. They were so small they were invisible. Which explains why the view port in Blender was lagging. The computer seems to be doing fine rendering with this amount so it shouldn't be a problem to scale up the intensity a bit. They are set up so that they are only visible in the render and there it doesn't seem to be a big drawback. I'll admit that part of the motivation for going 2d was this artwork here. But I'd like to keep everything as simple and in the same sort of simple tone and style as it already is.
### Latest WIP ###
TreesFound1J.jpg

[Sobacodaw]

I am curios - how long does it take to render such highpoly scene ? Incredible work, by the way

[Falconius]

This was rendered in Blender's Cycle's engine. Provided my computer doesn't freeze (which it has been doing under some certain 3D graphical tasks for some reason) the timer I think said it took 15 minutes. IU can't render on the CPU without the computer freezing but it does go faster. My current set up is Ryzen 7 1700, GTX 1080 with 8 gb of video RAM, and I've got 16 gb of system RAM installed. But I've only got it turned up to 64 samples, and the denoising is on. The ones on the previous thread averaged about 40 minutes I think, but that was with an older computer (that never froze). I tried this out on Blender 2.8 with the new EEVEE renderer and it was right quick, just a couple of minutes, but there were some other problems with 2.8 (like for some reason none of the mirroring modifier were working so my houses were only showing one corner. Blender 2.8 also has a very cool clay render option. There is also a noticeable quality difference between Cycles and EEVEE but that could have more to do with the settings I was using.

[randigpanzrall]

Wow, this is a great project which needs lots of time, doesn´t it? I totally respect this.

[Mouse]

Oh man! I've been totally asleep for the past week!

Wonderful to see you continuing this map

And... Oh NO! I haven't even got the hang of Cycles yet and their changing it all again?!

I think this is the point at which Blender takes a very back seat for me. I don't have time to keep up with it.

But you make it all look totally simple

[Falconius]

Wow, this is a great project which needs lots of time, doesn´t it? I totally respect this.

Thank you Randigpanzrall. It does seem to take a lot more time than I first envisioned. I can never tell if it's because I'm an idiot and have terrible work flow, or if I'm just really bad at judging the scope of a project (because I'm an idiot).

Oh man! I've been totally asleep for the past week!

Wonderful to see you continuing this map

And... Oh NO! I haven't even got the hang of Cycles yet and their changing it all again?!

I think this is the point at which Blender takes a very back seat for me. I don't have time to keep up with it.

But you make it all look totally simple

Thanks Mouse.

There is nothing particularly new to get used to between Cycles and EEVEE. The share the same nodes and materials. Cycles is a lightpath tracing renderer, and it can give extremely high quality renders, it's slow and noisy (grainy). EEVEE is more like a game renderer and can give very good quick renders that approximate what cycles will look like. The intention is for it to be able to be used in the view port and to replace Blender Internal. All this stuff isn't happening till Blender 2.8 which is intended to be released later this year.

[randigpanzrall]

LOL - I think, it has nothing to do with idiocy (never know, that this word excists in english language) but with idealism Don´t ever give up !!

[MidcenturyMr]

For practical purposes, I would imagine they would remove any trees close to the walls to avoid breaching/scaling opportunities. Something maybe to consider depending on the application purposes of the map. Either way, awesome work! Ahead of my abilities, for sure. I seriously hope to match that quality some day.