Ideally natural frequencies should be outside band of audio frequency beyond one octave. However in really case, I think the membrane has them within audio band so that there are rings or negative influences from those nearby natural frequencies which are generated by the mass and tension of the membrane. I think the membrane can be relaxed by the impressed vibration of audio pressure and recover the designed sound quality. Those kind of abnormally might be from high differentials over the thermal expansion or from chemical loading like humidity if the membrane is made out of hydrophilic materials then material can be more or less stressed on its equilibrium point. We might give it more time, which is not necessarily active powered burn-in but simply adjusting time for chemical stress releasing time.