What software to model complex 3d baffle shape?

Jump to Latest
N

needtubes

Member
Joined 2003
#1
I am aware of modeling software to predict the response of flat baffles of various shapes, but was wondering if anyone knew of software that could account for a third (depth) dimension. The motivation is the experiment from years ago (was it Dick Olsher?) where the response of a speaker was measured on many baffle shapes including some, like a hemisphere and pyramid, that are 3d in nature. Since it would be expensive, time-consuming, and difficult to built various complex shapes to measure, I was hoping someone was aware of software which could be adapted to this purpose. Does any exist?
 
#2
Yes any 3d linear acoustic software can perform the task for you. There are two common approaches: the first performs a sweep through time in the manner of an experiment and the second assumes simple harmonic motion and solves a stationary problem for a set of frequencies. There is a lot software around but the open source and free stuff tends to be research-type software requiring the user to know not only the relevant maths but also how to generate grids, plot and manipulate input and output files. The commercial software is easier to use but tends to be expensive and still requires the user to learn a bit about the maths in order to drive it sensibly.

A commercial BEM code with a restricted free version that seems to be used a fair bit by DIY folk here is AKABAK. There are a number of fairly lengthy threads discussing it's use on the speaker forum and one or two here. This is possibly your best bet in terms of finding help to get started. For example here is a current thread with 2M views discussing the software.
 
#5
As we would see it these days, in all the shapes tested by Olson the acoustic power would be the same. An axial response plot is not a clear metric of the problem, nor a clear indicator of what's going on. While those tests were an eye opener at the time, diffraction response simulation is not the whole aim in itself. Even assuming it was, you want something that is capable of performing a statistical analysis of the diffraction in isolation.
 
You must log in or register to reply here.
Top Bottom