0 / 60 seg.

Doing this kind of synthesis or generation of imagery over that entire surface, varying y over the surface, you make a kind of map -- a visual map of all the things the network knows how to recognize.