Currently working on a project where we have a single model that is being showcased on a turntable with multiple virtual cameras (from 3 different angles, namely Top-down with UI overlay, left, and right) placed to look at 6 different points of interest on the model, and we just turn off and on the correct virtual camera based on an enumerator ID and let the CM Brain handle blend given a pre-defined CinemachineBlenderSettings. The issue we're running into is that the blends will often clip through the model, despite the model having collision volumes (correctly layered and everything). Given some research, it seems like blending between cameras doesn't calculate collision, so I am at somewhat of a loss. I considered a ClearShot camera to find the most 'matching' child virtual camera (among the preset POV virtual cameras) and just blend to that one, but I am unsure. Any help is greatly appreciated.