The video links are：
We now have a very large factory model (over 10GB) that needs to be loaded in HoopsCommiunicator, but it is still very large after lightweighting, I see that the model has been simplified and some details removed in the video, I wonder how it is done?
The video links are：
Our current lightweighting method is to use HoosCommunicator’s converter to reduce the size of the data after lightweighting using the lowest precision, but I feel that there is a limit to do so, can not remove the excess inside the model does not need to render the part, is it necessary to manually delete the excess part, so the workload is a bit large
We also want to do the same as in the video where the entire scene is loaded first with a simplified model, and then load the full high-precision model when viewing a part of the model separately
Hey Rui, I think that might be @guido’s demo, so he may have some insights here.
the lods for the models in this video are created using a combination of a custom internal tool and the blender plugin from Simplygon. The demo is also using an internal library built on top of HOOPS Communicator for digital twin authoring we are planning to release this year. This library supports automatic LOD switching among many other things. You can see an early version of it in action in a live demo here:
To be more specific about the simplification: the workflow used in this demo is to convert the original CAD model to a tessellated format (gltf, fbx), load it into blender and then use the Simplygon plugin for blender for the simplification of the models. The results are then exported back to gltf from blender and converted to Stream Cache via converter. Obviously, this is a manual and time consuming process currently but as far as I know improved LOD generation is on the longer term roadmap for HOOPS Communicator.
Our Integration partner ITI also provides simplification technology specifically for the factory use-case which you might find interesting in this context: