I haven't used YPAO myself yet but I'm guessing there might be different microphones involved here (assuming YPAO provides it's own) and different positions. Most decent room correction suites are going to request measurements from multiple positions and the results will be tailored as an averaging of the measurements across those positions. If you measure the applied curve at a single position, it's not necessarily going to be perfectly neutral even if the algorithm is capable of completely dealing with a deviation at that frequency (room correction is much better at dealing with peaks than it is with dips).
