In the usual derivation of Planck's radiation law, the energies or frequencies $\omega$ of the oscillators depend on the measurements $L$ of the black body. The model is such that the only characteristic energy is given by these oscillator excitations and in terms of the temperature $T$. Also, the oscillation amplitude vanishes on the walls. Specifically, the structure of the atom walls doesn't enter the computation. After all, that's pretty much the definition of a black body.
Are there computations (maybe QED + statistical physics?) for more realistic systems, which might model the interaction of photons with the wall atoms and which give results such that one can see the limiting case to an ideal black body?
The question came up when I wondered about the theory behind emission and absorptions of photons on the walls of a black body.