But why? What is the point in applying an extra processing step that can be achieved simply by having a high-res model in the first place. As far as i can see, the only advantage of this method is that it is degradable by having a convenient 'on/off' switch.
Besides the real beauty here is not...