With mass of 3D poly can be CPU resources hog. 2D image of normal/displacement map uses less 3D poly and use less CPU resources in 3D matrix calculation. 3D poly should be only used as guided poly in 3D space for textures/layers. They always can convert mass poly into fewer poly as long as if the models look ok. With LDNM, when the renderer get to a pixel; it picks up the poly face and then it work on all layers within the poly area. Displacement/Normal map is very detail and faster than a model with mass 3D polys. Displacement/Normal map is only peak pointers with intensity and the light sources hits it and convert it into 2D image process as texture map. A model with mass 3D polys is wasteful resources and drain on CPU because of matrix renderer scripts have to deal with all the complex numbers in 3D space to determine which color it should make for a single pixel. LDNM is basically a scissors script like post-production video editor but still with some 3D polys. LDNM Renderer calculates differently on 3D poly and used it as for light sources for the displacement/normal map in 2d image process. It may cut down poly when not needed because the displacement map can cover it.