In this short tutorial I will show you three different Remeshing methods in Blender 5, specifically Voxel-based methods.
Video Transcript
Hello everyone! In this short tutorial I will show you three different Remeshing methods in Blender 5, specifically Voxel-based methods, which with the default settings produce the same result but are suited to different uses and situations. Along the way, we will also look at when to use one method rather than another.
To begin, let us clarify what is meant by Voxel Remeshing. It is a technique used to rebuild the topology of an object, generally producing quadrangular faces, in which the space containing the object is divided into cubes called Voxels. Voxel stands for Volume Pixel, which is the 3D equivalent of the pixels in an image. The smaller these cubes are, the higher the resolution of the final mesh will be.
While I was speaking, on screen I showed several results obtained by changing the size of the Voxels on the same starting object, which was created using a 3D scan and therefore had many vertices and poor topology, as I showed you in Edit Mode. The goal is therefore to obtain both a lower number of vertices than the original and a topology that is easier to manage for the next operations.
The first of the three methods we will examine in this tutorial is the one I already used on screen to explain the general concepts, so now it is time to present it properly. This is the Remesh tool available in Sculpt Mode.
In Sculpt Mode, at the top right there is the Remesh menu, where you can set the size of the Voxels used to subdivide the volume. The sampling tool allows you to set the Voxel size equal to the size of the face on which it is used, ensuring Voxels that are at least as small as that detail. In my case, the model is 24 centimeters tall and I am setting Voxels with a size of 1 millimeter.
The model went from about 360,000 faces to about 124,000, which is less than half, while preserving the overall shapes and creating a topology made only of quadrangular faces. With larger Voxel sizes, the mesh will be less detailed and can be used, for example, as a starting base for baking Normal Maps.
This method is therefore fast and practical, but it is destructive, meaning the original geometry is not preserved. In addition, to try different variations we must undo the operation each time and try again. It is therefore suitable only in certain cases, such as the previously mentioned Normal Maps baking.
If we need a non-destructive method, for example to use other modifiers while still keeping the ability to modify the original geometry or the extent of the Remesh, then we can use the Remesh modifier in Voxel mode.
By setting the Voxel Size field to the same value used with the Remesh tool in Sculpt Mode, which is 0.001, we will obtain a result practically identical to the one seen in Sculpt Mode, with about 123,000 quadrangular faces.
The difference in this case is that the method is non-destructive, meaning we can add other modifiers before or after Voxel Remesh and we can also modify the geometry in Edit Mode and then see it updated by the modifier. If you are interested in this possibility, then this is the tool to use.
The third method is non-destructive and allows you to modify the data even in the middle of the Remeshing process, in a certain sense. This method uses Geometry Nodes, and specifically the Mesh To SDF Grid node and the Grid To Mesh node, which I am adding to the node setup.
The first node breaks the volume into a 3D grid and includes a parameter called Voxel Size. The second node reconstructs the grid into a mesh, and here the parameter we are interested in is Threshold, to which we assign the same value as the Voxel Size parameter. By connecting these parameters to the same input of the Group Input node, we can set the size of the Remeshing Voxels. I set it, again in this case, to 0.001.
This solution results in about 128,000 quadrangular faces. This is a result similar to the other two solutions and, as with the Remesh modifier, this method is non-destructive, so we can work on the mesh in Edit Mode, in Sculpt Mode, or also use other modifiers. But that is not all.
Between the Mesh To SDF Grid node and the Grid To Mesh node we can insert nodes that operate on Grid-type data, such as nodes from the Volume group, and modify the 3D grid before reconstruction.
Many of the nodes in the Volume Operations group implement smoothing operations, which are useful if you want to remove noise or imperfections or, in the case of a model like this, create geometry for Normal Maps baking.
That is all for this video as well! As with my other videos, the text version of this tutorial and other resources are available on my website at the link you can find in the description. If this video was helpful to you, you can thank me with a Like and by subscribing to the channel. See you soon!