My scene is organized with a simple scene graph where the child multiplies its transformation with the one of the parent.
For normal objects, likes meshes, this works as expected, but I also want to insert the camera into the scene graph. To get the view-matrix I need to take the inverse of the global transformation matrix, which works except that this messes up the near/far plane dimensions. Because if a node above the camera has a scaling of 0.1 (across every dimension) the inverse will scale the whole world by a factor of 10.
Is there a simple way to get rid of the scaling or do I need to update my near and far plane based on the final view scaling?