In Sweden, the minimum impact sound insulation requirements in dwellings is LnT,w + CI,50-2500 = 56 dB. Something worth noting about the spectrum adaptation term is that it is just a logarithmic addition of the 1/3 octave bands 50-2500 Hz. This is somewhat misleading because the way it is written it looks as if one should first calculate LnT,w, then the spectrum adaptation term and finally add them. It might have been more pedagogic to call it LnT,w,50 or something similar (which they are called now in the Swedish standard, by the way).
When doing logarithmic addition, they largest terms will determine the sum. This is a challenge in timber constructions because they always tend to have high impact sound levels in the lowest frequencies due to the resonant nature of double constructions. Heavy constructions do not display this characteristic. They have their highest third octave bands higher up in the frequency spectrum. We have seen first-hand in a very large measurement series that LnT,w + CI,50-2500 was entirely determined by frequencies below 160 Hz in 66 out of 70 dwellings. When even lower frequencies from 20 Hz were included, 62 out of 66 dwellings were entirely determined by the frequency range 20-40 Hz.
We now know that the lowest 1/3 octave bands completely determine LnT,w + CI,50-2500. This brings us to the second challenge: Measurement uncertainty. Depending on the room size, we will have huge variations in the sound field in a room in this very same frequency range because of room resonances. For normal dwellings, this will certainly fall in the 50-100 Hz range in most rooms, and the smaller the room, the worse it gets. We are talking about more than 30 dB spatial difference in a single third octave band all in the reach of one arm’s length. Thus, if you do a sloppy measurement, you can potentially end up with a very bad or very good result, depending on which measurement positions you choose. In smaller rooms, it is common that the consultant measures the sound pressure level in the middle zone of the receiving room, but closer to the side opposite the loudspeaker in the sending room. This can clearly introduce systematic errors in the measurements if one is not careful.
So now we know that the 50-100 Hz range is crucial for the calculation, and it is also the range with the highest measurement uncertainty. But it does not stop here. Very low frequencies also correlate strongly with perceived annoyance. That means that the “sound that you hear” when your neighbor is walking above you, is primarily in the 50-100 Hz range – and perhaps even more so when frequencies from 20 Hz are included. It is no wonder that many people have claimed that impact sound insulation in wooden constructions is challenging!
Nowadays, we know all this and there are of course technical solutions that will fulfill all the requirements and the next generation of wooden constructions are on the way with even better performance. What you need to remember though, is that you cannot build a concrete house in wood! You must build a wooden house in wood. It is a completely different product. This goes for all disciplines, and for us acousticians it is very important to remember the above conclusions especially regarding measurement uncertainty. Small mistakes by the acoustician can produce very large consequences. We cannot change the laws of physics, but we can change and optimize the design of our buildings accordingly.
The first step is to measure correctly.