Why can we not obtain a unique solution for orientation from Essential Matrix?


In computer vision, why cannot we obtain a unique solution for the projection matrix given an Essential Matrix. Why is it said that it can only be obtained upto ‘scale’ and what does this mean?