You're not really doing anything wrong--it's just that each process requires working with the raw data. You could try the following: - Perform the georeferencing - Perfom the Beta Nought calibration - Carry out an Image to Image registration (Map->Registration->Select GCPs:Image to Image) using both images Use the georeferenced image as the base and the Beta Nought image as the warp, then manually specify a series of tie points in each image. The goal is to give the registration process a good pool of points that represent the same locations in each image while keeping the RMS error low (less than 5.0, ideally). The end result will be a georeferenced Beta Nought image.