The linear drag coefficient is a function only of object size and medium viscosity. Multiply sources say viscosity of air is independent of pressure and density. How can the linear drag be independent of density of the medium? It seems like less dense medium should mean fewer collision and in turn less drag force. What am I missing here?