|Originally posted by dave7
Since a suspended ceiling allows a lot of sound penetration, then, if there is thick batting behind it, there is an opportunity for more sound absorption, including low frequencies due to the depth of the batting (8" or more). Could the tiles themselves could act as sort of membranes for low frequency absorption? I am not sure about that because they would not have an air tight seal.
I've measured batting + acoustical tile ceilings which had extremely good wide-band absorption.
Membrane absorbers are typically relatively air-tight. This allows the mass of air behind the membrane to act as a cushion, and create resonance in combination with the mass of the vibrating membrane. There is then an easy formula to compute the resonant frequency of the absorber, based solely on the mass and air space depth.
Adding a porous absorber to this air space does a couple of things. It absorbs resonant sound energy, reducing the "Q" and therefore broadening the absorption curve. It also effectively adds greater volume to the airspace, further lowering the resonant frequency. This has to do with adiabatic gas compression (and so forth). Speaker builders use the phenomenon all the time in the design of enclosures.
On the other hand, without an air-tight membrane, porous absorbers act in a different fashion. They absorb different sound frequencies depending on their thickness, depth, specific acoustical resistance (and so forth), as well as their distance to other non-porous surfaces.
Building some combination of the above absorbers is certainly possible. Knowing its properties is another matter. There are two ways:
1. Build it, and test it under laboratory conditions.
2. Simulate it, using one or more computer methods (transfer matrix method, finite element method, and so forth).
1 has the advantage that you don't have to know anything about the properties of the materials. The disadvantages are that you need access to a testing lab, and also the results with your specific materials may not be generalizable to other materials and geometries.
2 has the advantage that you can play with parameters for the simple cost of a computer run. On the other hand, you need to develop and program the model, and the results depend upon the verisimilitude of this model and a knowledge of the relevant material physical properties.
This is of course just a complicated was of saying that we don't know the general answer to your question. ;)