For flipping, the origin should probably right at the middle of the
image. The single argument to "-t flip" is the direction of
the reflection vector; things move along this vector. Thus,
"-t flip:0" means flip along the X axis, and
"-t flip:90" means flip along the Y axis.
