Either a bad speaker or more likely simply reflects the change in room. Nearfield may not mean much in a room when the wavelengths approach and exceed room dimensions. Seriously doubt it is speaker break-in.
Especially with P/A speakers one shouldn't dismiss the idea of a substantial break-in so easily (by just unfounded speculation). There is data in my drawer:
A bass/mid driver sized 8", coil 2", fs nom. 78Hz, actually before break-in fs is 120Hz. After 2 hours of full excursion, happily eating up some 100Watts, though, it drops close to nominal values. After a few weeks it restored to the initial 120hz. Eventually it settled at very close to OEM's data.
In general some 30% or so doesn't harm, since optimum enclosure design follows the whole parameter shift, getting to the same enclosure, virtually same response. But 50%+ really suck. In HiFi such would be inacceptable, in P/A a driver gets beaten anyway.
So, what is seen so often here on audioscience are discussions without any supporting data. It puzzles me why that is.