So you decided to make 3d model, but don’t know where to start? In this article I will share my own 3d modeling workflow, which covers all stages, from idea to production ready model.
First of all, you need to know what kind of object do you want to make. Sometimes you already have a vision, sometimes you need to create a specified model for a client. But if you don’t know what to do, then try to brainstorm new ideas and select the best one.
You need to collect a bunch of photos with a similar object to get a clear vision of different parts for your model. It’s also a good idea to watch works from other artists on websites, like Artstation or Sketchfab. You can make a concept art or simple sketches of the future model if you’re making a model of non-real things.
In this stage, you need to make a plan on how you can make your model. Each model could be split into separate simple primitives. Your goal is to imagine how you can create a complex model from simple pieces.
Now you need to add some basic primitives into your scene. Your goal is to create base shape of your future model. Be sure that size and proportions of your model are correct. Split your model into different pieces and start working on each one separately.
Refine low poly model
If you decided to make a high poly model from low poly by adding bevel and smooth, then you probably need to add additional details into your low poly to make it look better. You should keep your model with lowest possible polygon count, but quite enough to make good bake from your high poly model.
You can also make a high poly model first, and add so many details as you want. You can use sculpting techniques or use some standard hard-surface modeling tools. When your model is ready, you need to build your low poly model from high poly by using retopology tools. It’s your decision what to do first: low poly or high poly. But keep in mind, that you can add additional details to your model in 3D painting software, like Substance Painter. So there is no need to add so many details in your high poly model in 3d modeling software, because you can add them later with alpha masks, quick and easy.
You need to unwrap your model by using UV editor so you can put 2D texture onto it. Sometimes you need to do this manually, sometimes automatic method is work well. Try to learn existing tools in your 3d modeling software, so you can speed up your work.
In this stage, you need to build a skeleton for your model and add control elements to simplify animation process. Key terminology to learn: joints, driven keys, blend shapes, inverse and forward kinematics, control curves, constraints, deformers.
After rigging, you need to bind your rig to your 3d model. Then you can move the joints or control elements and the mesh will follow them. You also need to paint weights to setup how much influence each joint should have over each vertex of your model.
When you’re done with rigging and skinning, you can start making animations for your model. You can make key-driven animation, or use motion capture data to make your model alive.
To achieve good performance in your games, you should keep polygon count for your model as low as possible. You can make different models for a different level of distance. And when your model is far from the player, you can show extremely low poly version(for example billboard for tree). When a player is near, you can show high poly version. You can make LOD variations by using different tools and modifiers in your 3d modeling software. Or you can use plugins inside the game engine(like symplygon or instaLOD), so they will produce LODs automatically.
Now you should bake different maps, like normal, ambient occlusion or thickness, to add details from high poly model to your low poly mesh. You can use different software for that, for example, XNormals(which is free). Also, Substance Painter has pretty good bakers, which works well even without a cage. If you’re planning to texture your model in SP, then it would be better to bake your maps in it.
After baking you should create textures for your model. You can use different software, like Substance Painter or Desinger, Quixel Suite, Mari or Photoshop, for that.
You can import your model into 3D renderer or game engine, even if you don’t have textures and animations. It depends on what you need. It will be better to import your model as soon as possible and then update it later when you add new details.
You need to set up materials for your model in 3D renderer or game engine. So it’s time to import your textures, add additional effects to your material and make some test renderers. You can also skip texturing stage and add existing materials to your model if you have them(for example brick material to your wall model).
You can bake lightmaps if you don’t plan to use dynamic lighting with your model. You can also add additional sources of light and bind them to your model.
If it’s a game model, then you probably need to write some code to make your model alive: add AI controller, animation controller, etc.
That’s all main stages of 3d modeling workflow. Of course not all of this steps are necessary for each model. I hope this list of steps will be helpful for newcomers, so they can dive deeper into each one later.