Dissect This Screenshot For Me To Identify Meshes

Basically, I’m having a really hard time identifying how to build up a city from the ground up. I just keep looking at other games and I don’t really get it. Tell me if I’m on the mark here:


5a8065fe9411b16e536714917e77e9df819e1677.jpeg

GTAV is an extremely complex game, because of all the data that has to be managed there are many things for them to consider in how it’s constructed.

Thanks, Darth. While GTA is complex, I think my question isn’t. I’m just wondering, where I’ve colored things, would you assume those are different meshes, or do you think that it’s even more granular. Do you assume these textures on the sidewalk were created JUST for this area, or they are probably tiled elsewhere?

The streets/sidewalks textures are definitely used a lot, when you zoom out more you can see tiliing, though they probably do some type of decal thing and blending to break it up a bit. The maintenance things on the sidewalk might be decals for instance.

Thank you, this was very helpful. I think the answer I was looking for was decals. Is it common to use a ton of decals, like I’m doing here?

Thanks again!

I’m not sure of the performance cost of decals, but remember that you can also use vertex painting to blend textures and break up tiling.

The performance cost of decals in UE4 is extremely low, so you can use quite a lot of them without fear of tanking performance like in UE3.

Thanks for the tip! Is it common to do it the way I’m doing in my screenshot below, or would it make more sense to put those cracks in the material network somehow?

Most likely ground is single object and things you colored differently are a single entity. Textures most likely are reused throughout the city.

Ground and buildings might be handled in different fashion, though, which depends on their engine.
For example, they could have “City Block” object that acts as single entity, contains ALL unmovable objects (ground and buildings) merged into one entity AND determines what to feed into GPU depending on your position and precomputed occlusion data. Or they could use separate “static mesh” per every background object.

GTA 5 is not using Unreal engine, keep that in mind.