the surface normals don't change when the mask rotates from the front to the back so the lighting appears to "change direction". I think looking at the edges gi

the surface normals don't change when the mask rotates from the front to the back so the lighting appears to "change direction". I think looking at the edges gives you more cues to see that the mask is rotating vs only having the reversing lighting cues in the center