Beyond the FRAME environments you can choose in our environment picker, you can upload your own 3D models to be the environment for your FRAMES. We have some requirements and some tips and tricks to get you off the virtual ground with your own beautiful models.
For starters, your model must be a .glb model and be no more than 15MB in size. We will occasionally make exceptions to this if you want to upload a larger model (email email@example.com if you do), but we encourage you to try to get your model well below that size limit to ensure good performance and decent load times for your FRAME.
Also, the environment modelmust have a bounding radius of 500m or less.
FRAME runs from a web browser, and if you expect any of your audience to join from VR headsets, mobile devices, or older desktops, you should do your best to get your file size as low as possible and follow the guidelines below. We are confident that with the right techniques and optimizations, you can bring gorgeous spaces into Frame.
If you are struggling to get your model in a state where it can imported as a Frame environment, we may be able to help. Email firstname.lastname@example.org
As mentioned, we only support .glb models. .glb is the binary, single-file form of a gltf model. You can learn more about the gltf format here. It's a flexible, open-source format that is specifically designed for web environments like FRAME.. Most 3D design tools have native .glb export tools or plugins/extensions that let you export your models to .glb
We make most of our models in Blender, or they at least end up there before we do our final exports, because we have found that Blender has the best support for the gltf/glb format. That said, Microsoft makes an exporter for 3ds Max, and many other tools besides Blender have native glb exporting available. You can ensure that your model is a valid gltf/glb model here.
Overall, if you're new to GLTF, and especially if you also use Blender, this is a great article with some good tips.
The origin point (0, 0, 0) of your 3D scene will be the default spawn location for users that enter your FRAME, so make sure that you strategically position your origin point. If your model has an origin point that is far away from the model itself, it may not even work at all. Also, avoid having the origin point of your scene directly inside any enclosed geometry in your environment. Ideally, your origin point is at an open area with not too much going on around it, to help your users not feel overwhelmed when they first spawn in your space.
You should limit the amount of textures in your model, and definitely avoid really heavy textures. If you can avoid using textures where you can instead use vertex colors or color palettes, then do so. If you're having a hard time reducing the heaviness of your textures and you want help, let us know at email@example.com and we'll see how we can help.
Reduce your poly count as much as possible. For geometry that you can duplicate, set them up as linked duplicates in Blender.
Use back-face culling where appropriate.
If you want to use a more advanced technique to compress the geometry in your model, you can apply Draco compression to your model. You can build the Draco compression library yourself, or you can use tools like Cesium's gltf-pipeline library.
For models with very complex geometry, Draco compression will make a dramatic difference. Otherwise, you will only get some small gains.
There is no real-time lighting in Frame (yet), so we suggest using unlit materials and baked lighting/shadows. To create unlit materials in Blender, use the background material type.
3D creation tools sometimes have different scale systems, and getting this right can be tough. In FRAME, we try to respect "1 unit is 1 meter".
To make this easier to visualize, here's a model of one of our avatars so that you can scale your model accordingly. Keep in mind that the avatar doesn't represent the full physical height as it is just the torso and the head. We typically set the avatar eye height to be 1.6 meters.