I wondered why George was complaining about the speed of the zernike smoothing operation. That is because it runs at least 4 times slower than the 7.3.2 one does. I suspect when the zernike process was refactored and slightly modified that introduced the slow down.