In this situation, we compared the insertion loss (IL in dB) between the source and receiver rooms. This insertion loss is calculated by averaging the sound level at 5 receivers inside each room for several configurations of the separation wall (with or without a door). This insertion loss can be compared to the theoretical value, obtained by applying the theoretical theory of reverberation of coupled rooms.

The results are given in the following table.

• Note that the case R=0 dB means that the surface is perfectly transparent (no wall separation).
• The IL obtained using the theory and the numerical simulations, always differ from the values R imposed on the door and on the separation wall, due the theory hypothesis and due to the estimation of the average sound level in a room. 

In theory, regarding the two first lines of the table (cases 1 and 2), if the numerical codes are perfectly coherent with the expected physical behavior, the case without a separation wall (#1) should be equivalent to the case (#2) of two rooms coupled through a transparent wall. In practice, this is verified for the SPPS code, but not for the MD code. Considering the diffusion model, the implementation considers two distinct volumes for calculating the diffusion coefficient (that is the main parameter of the MD code); the diffusion coefficient is necessary different to the one calculated with a single (and larger) room,which should justify the fact that values should be different for cases 1 and 2 for the DM_Comsol as well for the DM_Octave. By observing the case 8, where a surface present a value of 0 dB for R, one can see that the DM_Comsol give similar results than SPPS, which should confirm that the DM_Comsol gives the expected result. In this case, the DM_Octave seems not relevant.

Concerning the configurations with a transmission loss that is not equal to 0 (cases 2 to 7), we observe a very good agreement between the SPPS/MD codes and the theory of coupled rooms.