The thermoacoustic sound generation offers a promising wideband alternative to mechanically driven loudspeakers. Over the past decade, the development of nanomaterials with new physico-chemical properties promoted a wide interest in the thermophones technology. Indeed, several thermophone structures based on suspended nanowires, graphene sheets, highly porous foams or sponges have been investigated. At the same time, theoretical models have been developed to predict the frequency and power spectra of these devices. However, most of models have taken into consideration a solid homogeneous material for representing the thermophone generating layer, and its microstructure was therefore neglected. If this assumption holds for thin dense materials, it is not acceptable for thick and porous thermo-phone devices. Hence, a model able to describe the behavior of highly porous foam-or sponge-like generating layers is proposed. It is based on a two temperature scheme since the thermal equilibrium is not typically attained between the foam material and the embedded air. To do this, the fluid equations for the air are coupled with the heat equation for the solid foam through boundary conditions mimicking the energy exchange at the contact surface between them. The behavior of the main physical variables within the porous generating layer is explained and comparisons with recent experimental results are thoroughly discussed.