The computation cost of applying the 3-D AMO or DMO operators is an order magnitude higher than that for 2-D ones. However, the computational cost is also directly dependent on the size of the operator. Therefore, for a simple Kirchhoff-type implementation Deregowski (1995), the cost of the 3-D AMO operator is proportional to the product of the inline and crossline grid points needed to represent the operator. Luckly, AMO operators are generally small in size, and as a result, the number of grid points in which the operator covers are relatively few. In the application of 3-D operators, the reduction in the size of the operator is far more important than in 2-D media.