in the vertex shader I have the vertex position vpos which is in object space and I want to get the vector going from the camera to vpos. Can I just do vpos - cameraPos or do I need to do some matrix multiplication with the model-view inverse because vPos and cameraPos are not in the same space?