Are modular assets still needed for Nanite?

Personally, building a scene, especially a huge continuous wall or floor, with modular assets feels very counterintuitive. From my understanding, the reason for using modular assets is that they can be occluded when you are not looking at them. For example, if a building was a single asset, it would always be fully rendered even if I see just a corner of it, whereas modular assets allow my computer to render only the parts I am looking at.

I don’t know if I misunderstood the mechanism behind Nanite, but it seems it would break up any meshes into smaller pieces and occlude them anyway when they are out of sight even if they were only a single mesh. So my question is, are modular assets still needed for occlusion? Is it viable to model buildings or any other objects in big chunks now?

1 Like

There’s tons of reasons to use modular assets. They work better with distance fields and lumen. It’s easier to quickly build and modify scenes. You can use semi procedural methods of assembling and placing modular assets. Iteration is easier and quicker, troubleshooting is easier and quicker. Easier to add variety and variations.

1 Like

I understand there is more than one reason to use modular assets, but as with anything, there are pros and cons. Personally I find it hard to do because I keep forgetting to make precise measurements for modular assets. I am wondering with nanite, if it’s viable to make huge assets while keeping optimal performance.

100% would still recommend Modular assets. Modular assets will help a lot with loading and unloading meshes and materials, reduce draw calls when instanced correctly, and in a production scenario, really help with building the environments from a level design/level art stand point. and even better, you can create modular assets that utilize nanite, so its the best of both worlds :slight_smile:

Modular assets are not only used for optimization reasons (that might or might not hold up anymore with the advent of Nanite) but also for workflow reasons. Modular assets allow you to create complex scenes without having to spend an impossible amount of time on them and to iterate quickly on level design, for example. There are still plenty of reasons to use them, and with the increase of open world games and Epic targeting them with Unreal, there will be even more going forward.

Alright, thank you. Do you guys have any sort of guidelines on how small a modular asset should be? That’s the most confusing part about making modular assets for me. For example, should a huge long continuous wall be made out of many instances of a modular asset, or should it be just one mesh?

1 Like

Modular workflows work really well for any kind of architecture. for one long wall, it would be good to have it split up into a few smaller walls. as long as your textures/mesh tile corrects (adhering to 5.12 texel density) you wont notice a difference in the look. this will give you more flexibility if you wanted to change something. and as said earlier in the thread, for various optimization reasons, its good to have them split into multiple meshes. If you instance it correctly, you will have less draw calls too. :smile:

1 Like