Types of modeling. Basics of Sculpting, Retopology, and Sweep
In this article, we’ll talk about sculpting, retopology, and scanning. But first you need to decide on the goal. What will we model, and in what way?
Suppose we decide to create a character for the game, but what if it is an environment, architecture, or something else? First you need to find out what modeling methods are and understand which one is more suitable for us.
1. Polygon modeling
This is probably the most popular way to develop a 3D model. The bottom line is to create and edit a grid of polygons, which consist of vertices and edges. By clicking on the left mouse button we create a new vertex, which is connected by an edge.
Such a modeling process can be represented as a shape, for example, of a face covered with rectangles with varying degrees of perspective distortion.
In simple terms: “We create a grid that consists of primitive figures (primitives).” The result is a polyhedron. To better associate, we can recall how in films and animations many small robots turned into one big one. So it is with polygonal modeling. Many primitives make up one large model.
By the way, most often the polygon has four faces, but there are also three. Primitives with three faces are used only in certain cases, but there can be no more than four faces. There is no reference to real units of measurement, so the model is inaccurate. Accordingly, this method is not suitable for modeling any details or architectural projects where every millimeter is important.
You simply move the vertex, edge or the entire polygon, focusing on the appearance. Polygon modeling works well if you are modeling an art thing and the exact dimensions are not important to you. It can be a character, a game level location or an animal.
2. NURBS modeling
The main difference between this method and polygonal modeling in smoothness. The NURBS model does not consist of polygons, but of curves (splines), however, during visualization it is still converted to polygons, although it remains in the curves inside the modeling system. Used to create flowing organic forms and patterns.
Suppose you want to imagine a complex three-dimensional surface of natural origin. It can be described by vertices and divided into primitives, but it will take a lot of time, and changing the curvature of the surface in polygons will simply not be convenient. In such cases, just apply the NURBS modeling method.
3. Exact modeling in Saprah
In this method, the model is defined by mathematical formulas, so the surface of the model will be absolutely smooth at any approximation, and you can adjust it with an accuracy of millimeter.
Used when accuracy is important, not artistic expression. Purely theoretically, in this way you can create a character, but it will take a huge amount of time and effort, unlike polygonal modeling and sculpting.
Go back to the polygon. Why are there more of them in some models, but fewer in some?
The fact is that the models are divided by the number of polygons:
- Hi-poly - a large number of polygons (approximately 1 - 3 million).
- Mid-poly - the average number of polygons.
- Low-poly - a small number of polygons (approximately 5 - 10 thousand).
Accordingly, the more polygons, the more detailed the model, but requires more resources. So hi-poly is the most detailed.
In games on PC and console, they most often use mid-poly, sometimes hi-poly (in AAA projects), and in mobile games low poly.
We proceed to the part of modeling called “Sculpting”. This is the first 3D stage in creating a character. In it, the character is molded like clay, hence the name. At this stage, you can plunge into creativity and sculpt without thinking about the training grounds. And there will be many of them. Do not worry, we will simplify the whole thing later. By the way, for teaching sculpting, it would be nice to buy a graphic tablet.
And so, smoothly move on to the main topic of the article. But I didn’t say anything about the programs. I suggest taking Zbrush for sculpting and 3D Coat for retopology and sweep. The first thing you need to do is run Zbrush. We get to the main screen of the program. First we need to create a sphere.
In the Tool tab, which is located on the right side of the screen, select "Sphere3D". Pull the sphere in the workspace by clicking LMB.
To move we use hot keys:
- LMB - rotation.
- LMB + alt - move.
- ctrl + pkm - approximation.
Click "Load Next User Interface Layout".
A panel with brushes appears. To edit the sphere, click "Edit".
In order not to receive such a message, you need to click “Make PolyMesh 3D” in the Tool panel.
Now you can sculpt. It remains only to turn on the symmetry on the "X" key. When you press alt, the brush starts to work in the opposite direction. If the brush squeezed, then vice versa it will press. For ears, eyes, nose, it is best to create separate spheres and sculpt them separately. Over time, the grid will change and you will need to count the polygons. To do this, in the Geometry tab, click DynaMesh.
If you work with only one sphere, then all the details will come to naught when recounting polygons. So it is better to do this as separate objects, and at the end of the work to combine.
If there are no ideas for sculpting, you can go to artstation.com and find your favorite work. You can find concept art, add Zbrush, and use it as a reference for training.
In the process of working on a model, you may need a mask tool. You can apply it by clicking on ctrl. A mask is an area that the brush does not respond to. Thus, you can sculpt a lot of interesting things.
At the end of the work on the model, approximately 1-3 million polygons will be obtained. It will be difficult to open such a model in another program, so you need to reduce the number of polygons. To do this, in the Zplugin tab, select the item "Decimation Master" and click on Pre-process All. Zbrush will start the process and reduce the number of polygons.
The finished model can be displayed in obj format. Almost all editors eat it. You can save by clicking on “Export” in the Tool bar.
After sculpting, we will have a file with the model in the .obj extension. Open 3D Coat and drag the file there.
In our model there are still quite a lot of polygons. To simplify, you need to cover the model with polygons manually, preserving the shape. This is the process of retopology. To get started, go to the Retopology tab. Here we need to actually draw polygons manually. Here is an example of how they should be arranged.
Around the eyes and mouth, we build polygons in a circle. At the place of the bends we add more polygons, and on the fixed parts the polygons can be large and in small quantities, for example, on the back of the head and forehead.
Scanning or UV mapping is a very important process in model development. At this point, we are already preparing the model for texturing. What is the point? Imagine a cardboard box that is laid out on a plane. The box was dismantled and now it is in the form of a single sheet of cardboard. So with our model, we laid it out in 2D space.
Click on the uv mapping tab. 3D Coat shows the current scan of the model. Artifacts are marked in blue and red. In order for the texture to fit on the model without problems, only the gray color should be in the UV Preview tab. How to do it? On shift + LMB, you need to remove unnecessary seams so that the model is divided into parts. UV scanning is needed for the convenience of texturing and saving resources.
After there are no artifacts left on the model, you need to click pack UV, and then apply the UV layout. Congratulations, the model is finally ready for animation and texturing.