Digital 3D models are all around us, with uses in everything from art to engineering. Sometimes referred to as digital twins, they also form the basis of the immersive experiences – such as augmented reality (AR) – that are transforming sectors including retail, ecommerce, manufacturing, and more.
As the building block of the technologies that are reshaping customer journeys and business operations, it’s crucial that companies get it right when creating 3D models of their products. This one-stop guide, written in plain English, outlines how brands can create the ideal 3D model for their needs.
What is a digital twin?
In simple terms, a digital twin is a digital 3D model of a physical asset or object that accurately mimics its real-world counterpart. 3D models have a wide array of uses in modern business, including prototyping, design, marketing, and much more.
By creating 3D models of their products, businesses can open the door to a host of innovative approaches that benefit the bottom line. When combined with AR, for example, 3D models enable brands to elevate their customer journeys by offering product visualizations and live configuration functionality, which have been shown to boost conversion rates, improve buyer confidence, and reduce returns.
Read more: The benefits of digital twins for retailers
Establishing a 3D strategy
A number of considerations go into creating the right 3D model. With the wide range of applications for the technology, it’s essential that businesses identify their own specific use-cases in advance and develop their plans accordingly.
A key component of this is the establishment of a strategy for a 3D pipeline – one that makes certain the models created are right for the needs of the business. The best approach is to ensure models are designed in multiple levels of detail, in the most versatile file formats: .FBX, .OBJ, .GLB, .DAE. This ensures that the models can be adapted for various uses once ready. Furthermore, we advise creating models in the highest resolution and level of detail possible as this allows them to be repurposed for different channels later on.
How to create a high-quality 3D model
This section of the guide takes a deeper dive into the specifics of what goes into the creation of a superior 3D model. We’ve identified six key areas to focus on:
1. Mesh topology
3D models can be thought of as a collection of small 2D shapes joined together to form a 3D shape; a cube, for example, can be constructed from six 2D squares. The small 2D shapes are known as polygons, and when joined together in large numbers form what’s called a “mesh of polygons” or even just a “mesh”. Mesh topology, therefore, simply refers to the geometric arrangement, placement, and distribution of those polygons, to ensure the creation of a model that accurately represents the real-world object.
The ideal topology is where the mesh of a 3D model consists mostly of quads (polygons with four sides), and the direction of those quads is consistent – as shown in the example below:
This type of mesh topology enables 3D artists to easily make changes to the 3D model. This is not the case when the model is made up of tris (triangular polygons) or n-gons (any polygon with more than four sides), making the mesh inconsistent. See below for an example:
2. Optimal polygon count
Determining the optimal number of polygons to create a digital model is a balancing act. First and foremost, the polygon count must be high enough to ensure that the product is accurately represented. Having too few polygons can cause a number of issues, notably with curved surfaces and shading, while a higher number of polygons results in a more detailed model. However, having too many polygons can mean the file size of the model being too large. Take a look at the three examples below:
A huge file size can mean the 3D model is unsuitable for various applications, resulting in slow loading times and a clunky or glitchy user experience. As such, finding the sweet spot that ensures the model is realistic and accurate without sacrificing customer experience is key. For augmented reality uses, a polygon count of under 200,000 is ideal.
When creating a 3D model, the ideal shading design should ensure curved surfaces appear smooth and edges look sharp, as shown in the examples below:
Bad shading can result in a number of serious flaws in the model, including artifacts (unwanted bumps, dents, or other imperfections) and unintended reflections. Bad shading is often caused by poor mesh topology.
4. Texture quality
The first priority when creating any 3D model is ensuring it accurately reflects the real-world product, and that is particularly the case when it comes to texture. However, representing texture accurately doesn’t simply mean going for the highest resolution possible. Take a look at the two examples below:
Rendering lifelike texture depends on the level of detail the model achieves. To do so accurately, it must be created using a number of special images called “texture maps”– images that 3D artists use to represent different characteristics of various materials. These include normal maps, which recreate the distinct surface texture of materials like wood or leather; roughness maps, which define how shiny the surface is; ambient occlusion maps, which define areas of shadow; and Albedo – or base color – maps, which define the color of the surface.
Combining all these maps ensures the model will appear realistic, right down to the finer details, and that when rotated and viewed from different angles by the user, the model’s surface reacts to the light falling on it.
5. File size
We’ve already touched on the importance of file size, but mesh topology and resolution are not the only factors that impact this. If reducing the mesh resolution doesn’t result in a manageable overall size, it may be that image textures are the issue. In .GLB format, where images and mesh are combined into one .GLB file, image textures can represent a substantial proportion of that file.
One solution is to use texture tiling for textures with a repetitive pattern, while reducing the size of the image texture can also result in space savings. Another potential solution is to change the image format from .PNG to .JPG, which is more compact. For AR applications, a file size between 2MB and 5MB is ideal.
6. Baked ambient occlusion
As mentioned in point four, ambient occlusion (AO) maps define areas of shadow in an object, for example small gaps between sofa cushions, or the interior of a deep cabinet where light cannot fully penetrate. As some software that renders 3D models can calculate shadows using a technique called light bouncing, AO maps are sometimes ignored by designers. However, in real-time environments such as games and AR, this can be the difference between a model appearing realistic rather than cartoonish:
The obvious difference in the level of realism and quality achieved illustrates the importance of AO maps.
In a world increasingly defined by digital innovation, creating high-quality 3D models stands as a pivotal need for businesses navigating the realms of technology and consumer engagement. These models are transformative assets, evident in their ability to elevate brand interactions and streamline business functions.
This guide provides the basis for a strategic approach to creating 3D models, with best-practice tips for creating high-quality 3D models that align with specific business objectives. With an emphasis on precision without compromising efficiency, the approach highlights the importance of striking a balance between realism and user experience, which paves the way for seamless and effective integration across a host of applications.