Convert from floating point to signed integer by rounding to the nearest integer value then clipping to max and min of a 16 bit integer
