We consider depth movies involving a human performer in this paper. We use 16-bit integer depth values for the depth maps. This gives a compact and exact representation than using floating point without compromising on the range of values in practice.