The advent of remote work as well as the many travel restrictions across the globe have left companies looking for more efficient and future-proof ways to train staff. One of the latest technologies that has been seeing a lot of demand is Microsoft’s newest AR glasses, the HoloLens 2.
Many companies have already incorporated AR into their workflow. Imagine a technician installing the electrical wiring in an aircraft, typically this would require extensive training and the help of two-dimensional drawings to help guide them through the steps. It’s a complex process that leaves a lot of room for error. However, equipped with HoloLens AR glasses, technicians can be guided by real-time instructions overlaid on machinery and parts in their filed of view, without any prior training.
As more companies look to incorporate the HoloLens into their training, we have been seeing an increasing need for re-purposing CAD design models for use with the HoloLens. While it all sounds really cool, working with the HoloLens presents some challenges that many companies may not be familiar with. In this post, we’ll go over the best ways to prepare and optimize CAD files for HoloLens 2.
How to repurpose CAD for a HoloLens visualization
All CAD files have to first be converted to polygonal mesh to be used for any real-time visualization, including the HoloLens. CAD files are very detailed, with hundreds of parts. When a CAD file is converted to mesh, the resulting 3D model will have thousands of faces (or triangles). The HoloLens 2 has limits in terms of how many objects and faces it can render. This means that the mesh file must be optimized to meet the performance requirements of HoloLens.
In summary, preparing your CAD file for HoloLens 2 has two main steps:
Step 1: Converting to GLB format: CAD files must be converted to GLB file format to be compatible with HoloLens
Step 2: Optimizing for HoloLens: there are specific performance targets that need to be met to ensure models are compatible with HoloLens, including:
- Limiting the scene to 2-3 objects: Keeping a low number of objects helps speed up render times.
- 1-2 materials per object: Materials control the color of the object and how reflective or dull they appear, but they also increase complexity.
- Less than 10,000 triangles: Too many triangles, or faces will mean a higher resolution, but slower performance.
As you can see, there are strict limits that must be met for a 3D model to be successfully rendered in the HoloLens. And the challenge is that a 3D modeler will have to manually optimize the 3D file by deleting parts and creating lower resolution objects, or LOD’s. This is where an automated 3D file optimization software like Meshmatic can help save time. Meshmatic automates the conversion and optimization steps, making sure your files are ready for HoloLens, in a matter of minutes. Let’s go over how Meshmatic can expedite your workflow from days or weeks to minutes.
First step: Converting CAD to GLB
As we mentioned, HoloLens 2 is compatible with GLB mesh file format and all CAD files must first be converted to this specific file format. This step is easily and automatically done with Meshmatic. Meshmatic can read CAD files directly – including STEP, SolidWorks, CATIA, Creo, and SketchUp – and accurately convert them to GLB mesh format. Meshmatic can also convert other mesh formats, such as OBJ and FBX, to GLB format for HoloLens.
Second step: Using Meshmatic to optimize CAD for HoloLens
As we explained earlier, after conversion from CAD to GLB, the resulting 3D model has to be optimized to meet HoloLens’ performance requirements. Optimizing a GLB file with millions of faces is no easy task and requires deep knowledge of 3D modelling. This is where a software like Meshmatic comes in handy because it automates all optimization steps necessary to make sure 3D CAD content is suitable for HoloLens 2.
Strict face count:
Meshmatic sets a hard limit on face counts when optimizing for HoloLens. This means that:
- Based on a threshold value, small parts are automatically deleted.
- Internal faces that are not visible in the visualization are automatically deleted.
Automatically removing unnecessary meshes reduces the overall face count to make the file more suitable for MR.
Unlike Unity or Unreal Engine, HoloLens renderer doesn’t automatically switch between LOD levels in real-time; it only renders a single mesh. Therefore, it’s necessary to generate the correct LOD for each object in your scene and import them as one compact file into HoloLens.
Based on the size of a part, Meshmatic will reduce the quality of the mesh while trying to preserve the overall shape. Smaller objects are heavily LOD, while larger objects more important to the scene are lightly LOD.
Merging into a single mesh
Finally, Meshmatic combines all the parts with the correct LOD’s into a single mesh. This reduces overall file size and makes the import process into HoloLens more streamlined.