AFMG Network Forum

Bernard · Member

Hello to everyone,
I don’t know how to assign scattering coefficients for materials I am using for a concert hall. How does the scattering wizard work? The manual says that according to the shape and size I can calculate the scattering coefficient but doesn’t say much more than that. How do I use the options of Stripes, Squares and rectangles to specify the size and the shape? If the wall is rectangular but has the shape of the steps on the bottom, how should I introduce this shape into the options of the scattering wizard or calculator?
If my concert hall has many beams and trusses on the ceiling, how can I calculate the scattering coefficient of those beams? Would it be enough to just draw surfaces on the ceiling and the assigning a very high scattering coefficient? Wouldn’t this add a lot of uncertainty to the model?
Thanks for your help.
Bernard

Stefan · Member

Hello Bernard,

(unfortunately) you raise some good points. The scattering coefficient is indeed still related to a set of open questions that need to be addressed in the future. In particular a single-valued coefficient cannot take into account the scattering behavior of long vertical or horizontal stripes that is greatly different with respect to the angle. Several other limitations exist as well.

In general, the choice of the scattering coefficient is a matter of experience. It should take into account the following points:

1. There are no real-world surfaces with 0% or 100% scattering.

2. In many cases a surface does not scatter much (maybe 10%) up to a certain cut-off frequency above which it scatters a lot (maybe 80%). The primary goal of the wizard is to help determining this transition.

3. Scattering always consists of two major effects:
a) the acoustic properties of the wall material itself
b) the geometric properties of the wall within the room
While the measurement of the first is already difficult, the second contribution is hard to derive without experience and some trial and error.

4. In many cases it is not sensible to explicitly model relatively small and random surfaces, such as beams, steps of stairs, pews etc., but rather assign an approximate scattering coefficient to a representative surface.

5. Note that the actual exact value of the scattering coefficient is not really very important. In the simulation results the differences between a surface scattering 50% or 60% of the incident sound are often negligible.

For further reading please consult text books and DEGA, AES, ASA literature. There are many publications about the topic by various well-known researchers.

Stefan

Ron Sauro · Member

Bernard.... When this was written Stefan pointed you in the right direction according to what we knew.... in some ways that has changed ... in some ways signifcantly.... Our Lab is the only place doing Full scale scattering coefficients in the world at this time.... We have measured about 60 different items now and have a significant library of info... IF you are speaking of just things like RT the methods described are beter than nothing... but one thing now verified over and over is that scattering is only a measurement of surface Roughness.... not a way of seeing spectral reflections.... A flat plane that is parallel to a test floor in the ISO 17497-1 testing method wil show 0 scattering.... BUT take the flat plane and incline it 20 degrees or more and it will show scattering of close to one.... EVEN though it is still a flat plane.......

If the deviation in the depth of a sample exceeds 1/16 of the test sample size,,,,, then the scattering coefficient disappears and it become a spectral reflector and can then be simulated by geometry... so in this case ... you have to know the limitations of your tools.. IF you model the planes of the seats in your model and include surface detail to the size that is 1/2 the wavelength of the highest frequency of interest then you will get the correct information regarding 1st and 2nd order reflections.... Faces not reflecting potential 1st or 2nd order reflections from the sound source do not need to be modeled in detail.....As I said before ... if your only interest is RT then this may or maynot be necessary but if you are interested in reflections or echos then I feel it is....

As a general rule flat surfaces that do not have a scattering coefficient measured can be assumed to have between .1 to .2 scattering....if they are hard....

Ron Sauro
NWAA Labs