What is the reason of hysteresis loss? Consider a ferromagnetic material is subjected to a gradually increasing external magnetic field and magnetized to saturation. Then a reverse field is applied to demagnetize the material and then the magnetization is allowed to saturate in the opposite direction. Then the field is again increased in the positive direction until it saturates and a hysteresis loop is completed. 
Why is it that at the end of the process the sample heats up? What happens inside the material so that heat is generated?
 A: Hysteresis is a kind of non-time-reversible transition, so it is
associated with entropy (which must, in theory, generate heat).   
Because magnetic materials have crystal structure, the internal
magnetic polarization is not uniform, but aligns in small patches
to the crystal, and these patches (magnetic domains) shift in
size and shape during a magnetic change.   That changes the
internal forces between the domains, which causes flexing of the
material, in nearly random distribution, i.e. generates acoustic
white noise.
The losses in moving a magnetic domain wall are increased by 
flaws and inclusions which 'pin' the boundary of the domain,
and which determine whether the material is a 'hard' (highly
hysteretic) or 'soft' magnet.
Even soft magnetic materials have losses with magnetism change, 
and part of that is induced electrical current (eddy current)
which depends on the electrical resistivity, and on dB/dt,
so is greatest in conductive materials, and at high frequencies.
Iron is an effective transformer-core material at low frequency
(50 to 400 Hz) but nonconductive ferrites are preferred for
low loss at higher frequencies.
There is also a small dimensional signature to magnetic changes in
ferromagnets (called magnetostriction), and this makes microscopic
acoustic events every time there is a magnetic change.
Any internal acoustic energy will decay to thermal equilibrium by Umklapp scattering, becoming heat.
