February 23, 2017 at 7:12 pm #4006
Animation Pagoda StaffModerator
- Topic Count 63
- Replies 0
Modeling can be done a few different ways. The easier additive method combines separate pieces together. The more advanced method models one object out of a single mesh.
A single mesh is preferred by professionals for anything that will be rigged or textured, namely soft body character rigging. Combining parts works for most everything else as long as the geometries are tangent and the joining seams can be concealed. There are ways to combine separate pieces into a single mesh, which is usually best to do once a model has been finalized and is ready for texture mapping.
The polygon count of a model factors in to the render time, so don’t go overboard with the complexity of a model and try to use only as many geometry as needed. The more polygon faces a mesh has, the more memory it takes up and the harder it is to edit. Undesirable sharp edges can be smoothed out to appear more organic using the smooth mesh preview mode.
Creating a realistic human character that can be rigged and animated is one of the more difficult and time consuming things to do in 3D animation. Create model sheets and practice modeling rough human forms before attempting to jump into creating a character from scratch. Don’t feel bad if your first character ends up with a face that looks more like a potato than a person. It happens. Just keep practicing.
Most animators recommend modeling objects with all quads that are roughly evenly-spaced and sized. This will make things much easier when texturing, rigging, and subdividing meshes. If you don’t plan on animating a mesh, topology doesn’t really matter that much, but employers sometimes look at the model topology on demo reels to see if the artist knows what they are doing.
It is good practice to avoid having lots of triangles or polygon faces, but in certain cases they are unavoidable. Don’t worry too much if complex curved surfaces need to have an extra vertex or two to work properly.
3D assets used in game design must be converted entirely into triangles in order for the game engine to be able to process the information. The general practice is to create a high-poly model and then retopologize down to medium and low-poly states.
The subdivision tool splits the mesh into more faces, which adds geometry and allows the model to appear a little smoother. Subdivided models can exponentially increase the amount of file memory, which may cause your computer to run slowly.
In general, models with high poly count are easier to sculpt with programs like zBrush, while low-poly models are easier to tweak vertices and make quick adjustments, so it is generally recommended to start with low poly modeling and gradually add geometry when needed. Only subdivide into a high poly model when you are ready to create a final model or advance to fine detail sculpting in another program.
It is time consuming to have to go back and make adjustments to a wireframe with lots of geometry. While there are tools designed to reduce the number subdivisions, such reductions might potentially result in undesired loss of detail in certain areas. Always save a backup copy of your low poly model before subdividing.
There are various techniques for creating clothing depending on the project requirements. Realistic clothing is usually easier to create in sculpting programs like zBrush as opposed to Maya, since fabric wrinkles usually require a high density mesh to deform smoothly. However, experienced modelers can sculpt polygons into the general shape of cloth using basic extruding techniques or the quad draw tool.
Crafting clothing that moves according to proper physics simulation, such as a cape or long skirt, may require special fabric simulation software such as Marvelous Designer or nCloth. It is also possible to use shaders, bump maps, and baked textures to simulate cloth, though in general this works better for scenes that will not be animated. Textures will not always cast shadows or flow correctly when in motion, so texture mapping is usually only done in certain types of game animation, not feature animation.
Usually productions will need to develop a large number of characters in various shapes and sizes, requiring interchangeable clothing meshes that can be applied quickly and efficiently. A common technique for creating clothing in a studio pipeline involves modeling clothing over a humanoid base mesh. The general practice involves creating at least one standard male and female base mesh along with several other body type variations. These base meshes can be used to copy and extrude the face topology to create clothing layers that will seamlessly hug the body mesh when constrained. The clothing types can easily be revised, adjusted, or switched out without having to create a character model from scratch for every single model.
It is possible to create a single clothing mesh for a character without first modeling a detailed body mesh if using a model sheet. This method might save a little modeling time in a crunch, but it is generally only recommended for custom characters, It is probably recommended to learn how to model human anatomy and topology before delving into costume design. If you spend the time to create one solid base mesh, you can reuse the model in other projects. The other option is to practice getting really good at churning out new models quickly for each new project.
Once you have created an amazing model, you’ll want to get a good render of it from multiple angles. Looped turntables allow a camera to rotate around a fixed axis. This is good for showing off details on models, characters, and scenes. It only works properly if the model is centered around the origin of the ground plane (0,0,0).
You must be logged in to reply to this topic.