Barkhausen noise power versus size of a minor hysteresis loop

It has been observed that Barkhausen noise is quite low in the Rayleigh region and that it appears rather abruptly when the hysteresis loop departs from the Rayleigh region. These facts are explained in terms of the interaction between domain walls and crystalline defects. Only defects which have saturated (pressure independent of wall position) contribute to Barkhausen noise. The noise power is proportional to the cube of the defect strength. On a small minor hysteresis loop only weak defects saturate. Thus the noise is low. On large minor loops substantially all defects saturate, and the noise is high. The transition between the two cases is relatively abrupt. The noise is always small near the initial end of a branch of a loop of any size and reaches a maximum at the final end.