Hi,
I'm a computer engineering student with a soft spot for building loudspeakers. I am currently in the process of planning a solution for a loudspeaker simulator program, and despite what many seems to think, this isn't easy at all.
There are two obvious mysteries for me. I have been searching the web (including this forum) and asking my math teachers about both. This forum is my last hope
(Ok, maybe a bit dramatic there)
Mystery number one: Transfer functions. I assume this to be the functions from which I can draw the actually simulated graphs. For example, the transfer function for SPL would typically be a function of a specific frequency at a specific input level (1W/1m), using a specific driver in a specific box, etc. Does anyone know where I can find -detailed- information about which equations to use where? I have read through The Loudspeaker Cookbook and Testing Loudspeakers, but the only relevant info I can find is vague references to various fourier transforms.
Mystery number two: Banal as it may seem, how do I draw a semi-log graph paper, as seen in frequency response sheets? I have a fair idea about how to go forth to do this, but as always, a good reference beats trial and error, atleast in this matter
Information, or hints about where to obtain information, will be greatly appreciated. As mentioned, I currently own the cookbook 6th edition and Testing Loudspeakers 1st edition, in addition to The master handbook of acoustics 4th edition (in case anyone is going to reference pages in any of those books).
Kind regards,
Norway
I'm a computer engineering student with a soft spot for building loudspeakers. I am currently in the process of planning a solution for a loudspeaker simulator program, and despite what many seems to think, this isn't easy at all.
There are two obvious mysteries for me. I have been searching the web (including this forum) and asking my math teachers about both. This forum is my last hope
(Ok, maybe a bit dramatic there)Mystery number one: Transfer functions. I assume this to be the functions from which I can draw the actually simulated graphs. For example, the transfer function for SPL would typically be a function of a specific frequency at a specific input level (1W/1m), using a specific driver in a specific box, etc. Does anyone know where I can find -detailed- information about which equations to use where? I have read through The Loudspeaker Cookbook and Testing Loudspeakers, but the only relevant info I can find is vague references to various fourier transforms.
Mystery number two: Banal as it may seem, how do I draw a semi-log graph paper, as seen in frequency response sheets? I have a fair idea about how to go forth to do this, but as always, a good reference beats trial and error, atleast in this matter
Information, or hints about where to obtain information, will be greatly appreciated. As mentioned, I currently own the cookbook 6th edition and Testing Loudspeakers 1st edition, in addition to The master handbook of acoustics 4th edition (in case anyone is going to reference pages in any of those books).
Kind regards,
Norway
The best place to get the transfer functions for closed and ported box loudspeakers is probably Thiele and Small's original Audio Engineering Society (AES) papers. These papers are what most of the simulations programs available on the web use as a basis. But be prepared to do some homework to be able to understand the concepts being applied. I would recommend Acoustics by Beranak as a primer for the lumped parameter circuit models used to represent the acoustical and electrical equivalent circuits in the Thiele and Small papers. In fact, the first two volumes of the AES anthology are indespensible when studying speaker theory and include the Thiele and Small papers.
I have been where you are and these are the references I followed,
I have been where you are and these are the references I followed,
There is a good reson why you rarely see the actual equation for eg a bass-reflex box. Even the simplified version of it is rather nasty:
Down to the right you can see the sound pressure p=... which includes a H(s) which is described on the line below. H(s) contains k1, k2 and k3, which in turn is defined on the following 3 lines. These contain 5 greek letters which are defined at the top of the column, from the properties of the box, vent and driver.
And this is a simplified model that does not include voice coil inductance, baffle step etc.
The solution to this is not to describe the transfer function with a single equation, but to build a circuit emulator, and enter the mechanical analogue circuit into it. If you do so, it is much easier to add extra features, eg a lossy voice coil inductance, passive crossover components etc.
An externally hosted image should be here but it was not working when we last tested it.
Down to the right you can see the sound pressure p=... which includes a H(s) which is described on the line below. H(s) contains k1, k2 and k3, which in turn is defined on the following 3 lines. These contain 5 greek letters which are defined at the top of the column, from the properties of the box, vent and driver.
And this is a simplified model that does not include voice coil inductance, baffle step etc.
The solution to this is not to describe the transfer function with a single equation, but to build a circuit emulator, and enter the mechanical analogue circuit into it. If you do so, it is much easier to add extra features, eg a lossy voice coil inductance, passive crossover components etc.
I don't know exactly what kind of program you are going to write but I guess its going to be based on T/S parameters so you could look up Thiele & Smalls original paper for a starter. If there is an engineering program at your school check out the literature they're using for the electro-acoustical course.
The other issue is more trivial... As a CS student I guess you're familiar with MATLAB. You'd probably want to check your math routines in MATLAB first before writing any code anyway.
Look under 'semilogy' and 'semilogx' for drawing semilog plots.
If writing in C++ or other language you probably have to implement these functions yourself. I don't know as I have not done any "high level" programming for the last couple of years. If you know your logarithms it isn't difficult.
Det var en gymnasist ifrån Kumla
Som med log-lagar brukade fumla
Log a + log b
Log ab det é
Han numera alltid hörs mumla!
/Magnus
The other issue is more trivial...
As a CS student I guess you're familiar with MATLAB. You'd probably want to check your math routines in MATLAB first before writing any code anyway.Look under 'semilogy' and 'semilogx' for drawing semilog plots.
If writing in C++ or other language you probably have to implement these functions yourself. I don't know as I have not done any "high level" programming for the last couple of years. If you know your logarithms it isn't difficult.
Det var en gymnasist ifrån Kumla
Som med log-lagar brukade fumla
Log a + log b
Log ab det é
Han numera alltid hörs mumla!
/Magnus
I know where Small's papers for Direct Radiator and Closed Box is posted. Email for the address-I am not sure if the individual wants the URL posted.
I can send you the GIF copies of Thiele's papers for the ported box, including the transfer function. I send them in GIFs about 200KB each. Just print them out-it's like getting a Xerox copy sent long distance. Email me if you want them sent.
Make sure you have a lot of space in your email box!
I can send you the GIF copies of Thiele's papers for the ported box, including the transfer function. I send them in GIFs about 200KB each. Just print them out-it's like getting a Xerox copy sent long distance. Email me if you want them sent.
Make sure you have a lot of space in your email box!
kelticwizard,
I wrote in my email to you that I'm happy for you to point people at the webpage with the articles on, but thanks for being considerate about it I actually posted it on the thread where we originally talked about the articles, but here it is again
If you send me the vented papers then I will host them as well.
I wrote in my email to you that I'm happy for you to point people at the webpage with the articles on, but thanks for being considerate about it
I actually posted it on the thread where we originally talked about the articles, but here it is again If you send me the vented papers then I will host them as well.
Thiele's article is in two parts.
I had already scanned the first part some months ago.. I scanned the second. Then I tried to attach the pages to an email I was sending to richie00boy.
The Yahoo mail showed it was attaching the whole 25 pages or so to the email. I left for awhile.
I came back about an hour later, and the screen still showed it was attaching the pages to the Email. I checked the Status of my dialup connection, and it showed that nothing was being uploaded or downloaded.
Just love Yahoo mail.
I will try again later. Both parts are scanned, it's just a question of getting them over to richie00boy.
I had already scanned the first part some months ago.. I scanned the second. Then I tried to attach the pages to an email I was sending to richie00boy.
The Yahoo mail showed it was attaching the whole 25 pages or so to the email. I left for awhile.
I came back about an hour later, and the screen still showed it was attaching the pages to the Email. I checked the Status of my dialup connection, and it showed that nothing was being uploaded or downloaded.
Just love Yahoo mail.
I will try again later. Both parts are scanned, it's just a question of getting them over to richie00boy.
A bibliographic selection of AES papers about Transfer Function, loudspeaker measurements, etc.. is here:
http://purebits.com/bibliog.html
http://purebits.com/bibliog.html
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.