In RAID5, the "parity" data is actually an XOR function of the four sets of data in the stripes. In RAID6, this is only the first operation. The second is much more complex involving both an XOR function and a bit shift function. BOTH generate "parity" data that are stored in the stripes of the array. I don't have anywhere near the math knowledge to understand it, but maybe you can grasp more of it from the link and other references.
I agree it's confusing. I don't understand it, and that is why I never tried to do it. the point is that the RAID controller is supposed to do all the work for you. You are only supposed to read the RAID manual for your system and follow the instructions for how to monitor for problems and how to use the software utilities provided with it to restore a degraded array.