Browse the wiki

Model LODs

This doc will give a run-down of how we name and handle model LODs at CAVE.

Introduction

LODs (Level of Detail) are a great way of managing the complexity of an asset. If you are new to LODs, check out the following Wikipedia page: Level of detail, which has plenty of good info on what LODs are, how they are used and how they came about.

When creating assets, a discussion should take place with the supervisor to decide which LODs should be created as not all assets will require multiple levels of detail. Primarily,  we’ll be using LODs for two purposes:

  1. We will switch the model LOD depending on the assets proximity to the camera (for example, an asset closer to the camera will require a higher level of subdivision and detail).
  2. Allow artists to work with LODs at a lower resolution for better performance (for example, an animator will want to work with a LOD that allows for real-time playback).

In many studios, you’ll find that they’ll have a ‘number’ or ‘letter’ based system for LODs. For example, from low-res to high-res meshes, they may go: 0, 1, 2, 3, 4… Or they may go 100, 200, 300, 400…, Or maybe A, B C, D.

At CAVE, we’ll be going with a number based system (0, 1, 2, 3…) where 0 is the highest LOD for custom hero topology and all other resolutions will either work down from that or be created from an existing base mesh.

Please note that all LODs are based on a 4K render (3840 x 2160).

LODs Breakdown

Although the notes below refer to modelling, we will also use a similar LOD system for rigging and textures where required. More details on this to come soon.

LODDESCRIPTIONSHOT TYPE
lod6Proxy geo or calimari geo for rigging purposes.n/a
lod5For crowd assets. Extreme Wide Shot (EWS)
lod4For low-resolution assets.Wide Shot (LS)
lod3For mid-resolution assets Full Shot (FS)
lod2For high-resolution assets.Medium Full Shot (MFS)
lod1For ultra-high resolution assets. Medium Close-up (MCU) and Close-up (CU)
lod0For assets that need custom topology (wrinkles, skin folds, etc.)Extreme Close-up (ECU)

Model LODs in Detail

Here we will break down each LOD to give you a clearer idea on how to approach your assets:

lod6

This LOD will primarily be used as proxy geometry allowing the artist to have a lighter 3d scene. Very rarely will this model be rendered. This LOD will also be used by the rigging department to create calamari geo.

lod5

Models at this LOD will primarily be used for crowd agents. This LOD should support Extreme Wide Shots. Bevelled edges (and creases) will not be required at this level. If need be, you can delete polygons where they are not visible in render.

Additional detail could be added to this LOD via normal maps but may not be required.

When creating crowd assets, make sure to check in with the FX team who will be tackling the crowd simulation. You’ll want to ensure your assets are going to be suitable for simming. For example, ask them what the optimal poly count will be for them to handle the data. They may ask for 8000 polys for combined body and costume. And maybe they want the costumes modelled seperately to the body mesh, so they can mix and match.

lod4

This LOD should be suitable for Long Shots. Bevelled edges (and creases) may not be required. The geometry should be usable for projection work. This LOD does not need to have details such as wrinkle lines modelled into the topology.

Additional detail could be added to this LOD via normal and displacement maps but may not be required.

lod3

This LOD should be suitable for Full Shots. Bevelled edges (and creases) may be required. The geometry should be usable for projection work. This LOD does not need to have details such as wrinkle lines modelled into the topology.

Additional detail could be added to this LOD via normal and displacement maps but may not be required.

As with every asset, it is important that you test your models in context. So as you create a version, make sure to pass it to the next stage of the pipeline, so they can get it into shot as soon as possible and check if the model holds up to the shot requirements.

lod2

This LOD should be suitable for Medium Shots and support the full asset being viewed on screen. If creating a hard surface asset for this LOD, bevelled edges (or creases) will be required. This LOD does not need to have details such as wrinkle lines modelled into the topology.

Additional detail can be added to this LOD via normal and displacement maps.

 

lod1

This LOD should be suitable for Medium Close-up and Close-up shots, and support head and shoulders, and neck and head shots. If creating a hard surface asset for this LOD, bevelled edges (or creases) will be required. This LOD does not need to have details such as wrinkle lines modelled into the topology.

Additional detail can be added to this LOD via normal and displacement maps.

At this LOD, the model should be suitable for hald body renders, however, you should aim to use the CAVE base mesh for characters and maintain the topology.

This LOD is also for assets that need to be suitable for destruction. For such assets, you will also need to model the interior aspects (for example, when modelling a wall for FX, you will need to model the plaster and the brick, and anything else in-between).

lod0

This is our hero LOD and is for extreme close ups and where you will need to create a custom topology to account for skin folds and wrinkles. When used in combination with displacement/vector displacement and normal maps, you should be able to pull off any shot.

For characters, you should start with our base mesh but be prepared to modify the topology to hit the requirements of the model. For props, vehicles, etc. you should still aim to create a mid-resolution model and then work into it. That way, you’ll have somethign that is great for viewport viewing and the fancier hero LOD for fancy renders.

Further Reading

Useful Links

The following links are mainly for your reference and to give you an idea of how other companies approach and tackle LODs:

Support CAVE Academy

Here at CAVE Academy the beauty of giving and sharing is very close to our hearts. With that spirit, we gladly provide Masterclasses, Dailies, the Wiki, and many high-quality assets free of charge. To enable the team to create and release more free content, you can support us here: Support CAVE Academy

Join the discussion

Comments 2
  1. SOH Lagrange on 22.03.2023 at 06:53 BST

    This is very usefull. Now I will no longer use the highest resolution for my renders only for close shots. Really appreciate your work.
    Thanks

    • Jahirul Amin on 22.03.2023 at 09:23 BST

      You are welcolm SOH Lagrange. When you set up your renders, do test your meshes. You’ll always want the best result, so if a _fg_ asset actually looks better for an _mg_ shot, then use that. The core idea is that you can flip between meshes which gives you flexibility.

Leave a Comment