Using the stud cavity as a return is perfectly legal, and exceptionally common in Canada... Ditto for the cardboard 'pan' material - it's used to close off joist or stud spaces for use as air ducts.
As for the firewall; double layer 1/2" is typical in residential use; 5/8" type X is used in commercial/institutional applications. One side of the double wall should be insulated, for sound transmission...
If I had to guess, the insulation is the part missing... If the air returns line up; the sound would transfer easily...
Putting filter material (like a cut-to-fit air filter) behind the return grate would help; but don't know if it would be enough - likely not, I'm afraid to say...
