How to create 3D models

While there are a number of ways to create 3D models, most of them boil down to two basic methods: building a model in 3D modeling software, or taking an object from the real world and turning it into a digital model using a 3D scanner. Let’s look at the advantages and disadvantages of both methods, and learn how to determine which method is best suited for your task.

Method 1: 3D modeling

The first way to create 3D models is to start from scratch using specialized 3D modeling software. This method is widely used by professionals across various industries: engineers, industrial designers, architects, CGI artists, and many others.

What’s great about this method is that it allows you to design something that doesn’t yet exist – something completely unique like a new component for a car or a fantasy creature in a video game – or something that already exists but is not available for scanning.

For example, if you need a model of a world-famous building located remotely or far from where you are, it can be much easier and more cost-effective to create a 3D model from scratch, using reference materials such as photos and videos, rather than traveling to that place and arranging a 3D scan (which could be a challenge in itself when it comes to historical buildings!). Or, if you need to design something completely new that hasn’t yet been created, 3D modeling is the perfect way to let your imagination run wild and bring your ideas to life.

There are different 3D modeling techniques and different modeling software where you can design a model from scratch. It all depends on the particular object you want to create, and its intended application.

Parametric modeling

Parametric 3D modeling or CAD (Computer-Aided Design) is the No. 1 method used by engineers and designers to build realistic computer models of future parts and assemblies. Nearly every modern-day product we interact with on a daily basis was created using 3D CAD modeling.

With this method, a designer creates a 3D model of an object that can have all of the same parameters as an actual physical object: material, weight, size, optical parameters, physical parameters, etc. These models can be then 3D printed or CNC machined, as well as used to run complex simulations. For example, you can create assemblies of parts to see how they fit together, test how they will react to forces applied to them, monitor how fluids will flow through them, evaluate how they will be manufactured using simulations, and more.

Given the labor intensity of this method, it is used only for scientific purposes, in production, and in some professional collections of 3D models. For instance, this method was used to create and update the 3DModels.org library.

Polygonal modeling

This modeling technique is at the heart of almost every video game or sci-fi movie you have ever seen or played. A polygonal model is built of polygons: flat, two-dimensional shapes, triangles, or quads which the artist is modifying to build a 3D mesh. Unlike CAD modeling, this technique is more concept-driven rather than measurement-driven. Animation and video game studios use polygonal modeling to design everything from movie and game characters to various 3D assets such as weapons, armor, vehicles and entire virtual worlds.

Digital sculpting

This method, also used by 3D artists in games and animation movies, works best for creating hyper realistic objects with organic and smooth shapes. It’s also used for creating prototype models for design, sketches and 3D printing. The process is very similar to sculpting with real materials like clay or stone. By using brush-like sculpting tools you can manipulate the polygonal mesh of an object, pushing, pulling, and twisting various parts of its geometry or adding extra geometry to mimic an organic structure. Digital sculpting requires even more artistic skills – and is more meticulous and time-consuming – than polygonal modeling. That’s why in many cases these methods are used together: first the object is modeled, and then sent to a 3D sculptor for detailing to then take its final shape.

Whatever method you’d like to go with, it’s important to choose the right modeling software that will have all the necessary tools to help you create your 3D model.

Method 2: 3D scanning

3D scanning is a process of analyzing an object from the real world, to collect all the data in order to recreate its shape and appearance, digitally. Thanks to this process, the object can become a 3D model, which could help you as a base for the 3D project you are about to develop, but it can also be useful to reconstruct, analyze, or simulate ideas.

Different machines and methods exist to 3D scan objects. You might not know it, but there are different ways possible to create a digital version of a real object. We are going to see more in detail of how each of them works. There are a lot of different 3D scanning methods, but today, we will focus on three of them, that we can consider as the main ones: Laser 3D scanning, photogrammetry, and structured light scanning. The choice of the 3D scanning technique will be made regarding your project or its context.

Laser scanning

As the name suggests, laser scanning uses targeted laser beams to capture the shape of products. By bouncing beams off surfaces, handheld scanners can build detailed images. And they can usually do so in real-time, allowing users to create 3D visualizations whenever they need to.

This is a major advantage in some situations, and laser scanning can work well when small batches of products need to be processed. However, as with photogrammetry, some limitations can apply.

The major problem associated with laser scanning involves shiny and extremely flat surfaces. In these cases, the reflective beam can become less precise, leading to significant distortions in the 3D image.

If you are dealing with small batches, these flaws can be remedied via modeling software. But the imprecision of laser scanning can rule it out for bulk eCommerce projects.

Photogrammetry

Photogrammetry involves using multiple cameras to capture an array of images of the same product. By combining many different angles, it’s possible to build a model which is virtually 3D.

Camera rigs for photogrammetry tend to be large and may require as many as 200 lenses. On top of that, the process also requires powerful software to amalgamate and process every image.

This means that photogrammetry comes with significant start-up costs and requires a high degree of technical knowledge to function effectively.

Some items may also be poorly suited to capture via photogrammetry. For example, glass products may reflect light in ways that compromise the capture system. Anything involving crystal or metals could be excluded, which in practice limits the technology’s eCommerce potential.
However, there are some very important advantages. Most importantly, photogrammetry is an extremely fast way to capture 3D images. Assuming you have the processing power available, images can be rendered in very short time periods.

The quality of images captured via photogrammetry can also be superior to other techniques. With many cameras involved, it’s possible to capture complex textures. And the accuracy level tends to be high as well. So for high-level visualizations, it’s definitely an attractive option.

Moreover, photogrammetry can be great for capturing very large objects such as buildings, where resolution isn’t important. For example, a single camera drone can deliver enough images to capture a golf course. But that’s not ideal for sellers of items like furniture or decorations.

Structured light scanning

Structured light scanning uses targeted light beams to build digital profiles of physical objects, making it similar to laser technology.

Generally, these systems employ multiple beams which are directed at the target object. These beams create a specific pattern on the object’s surface, which will change depending on how flat the surface is.

Cameras inside a scanning unit capture the moment these light beams hit the object and relay the image to structure light scanning software. This software decodes the meaning of each pattern and combines every reference point to build a digital representation of the target.

This can result in accurate images, but it creates a problem. At any one time, structured light systems can only capture one perspective. To truly represent 3D objects, the object will need to be rotated and recaptured.
If the software is good enough, it can use algorithms to calculate the relation of each perspective and create a composite model of the whole object. And if flaws emerge, it can clean these up automatically.

Structured light scanning is very fast and with handheld or automated scanners available, it’s also very flexible and accurate. However, conditions need to be right to capture images properly, and the cost of building high-end scanning setups can be beyond smaller eCommerce companies.

If you choose structured light scanning, it’s a great option for irregularly shaped, complex objects like figurines or tools, especially smaller objects that can be placed on indoor worktops.

As you can see, there are various ways to create 3D models. In cases where you need to model something that doesn’t exist, or something completely new, you’ll probably want to design it in 3D modeling software. However, if you need to model something that already exists, consider 3D scanning. Not only you can save time and effort, but you’ll get precise results which you might never achieve with conventional modeling tools.