Programs used by REW

[Antoine B]

As a recent REW user, I am disappointed by the absence of a formalisation of the procedure.
In the help, it is described pedagogically but without using mathematical terms.
A paper is cited, I red it and didn't find more than a block chain explaining the main steps of the process.

I could write a piece of code on this basis, but I would prefer to rely on a published program.

In fact, I don't want to use REW as a black box but as a scientifical tool to make experiments !

So I would like to know where I can find this formalisation of the measure treatment : generation of sweep, acquisition, signal processing and graph plot, and maybe if there is a python equivalent.

Thank you !

 

[John Mulcahy]

Reading the papers and then writing code worked for me, so you could try that. If you like python there appear to be plenty of python acoustics libraries.

 

[Antoine B]

Can you please link me a more detailed paper than the one cited in the help ?

 

[John Mulcahy]

There are several papers listed in the Reference material section of the REW site.

 

[Antoine B]

These ones :

• Guy-Bart Stan, Jean-Jacques Embrechts, Dominique Archambeau, "Comparison of different impulse response measurement techniques", JAES Volume 50, Issue 4, pp. 249-262, April 2002.
• Swen Müller, Paulo Massarani, "Transfer-Function Measurement with Sweeps", JAES Volume 49 Issue 6 pp. 443-471, June 2001.
• Angelo Farina, "Simultaneous measurement of impulse response and distortion with a swept-sine technique", 108th AES Convention, February 2000.
• Nikolaos M. Papadakis and Georgios E. Stavroulakis, "Low Cost Omnidirectional Sound Source Utilizing a Common Directional Loudspeaker for Impulse Response Measurements", Appl. Sci. 2018, 8(9), 1703;

?

Probably the third one contains more formalisation

 

[John Mulcahy]

Yes. In essence the task is to determine the system transfer function. That can be expressed in the frequency domain as FFT(response)/FFT(stimulus), the inverse FFT of that ratio gives the transfer function as an impulse response. A helpful property of an exponential sine sweep as the stimulus is that harmonics are collected at negative time, making it easy to window them and determine the harmonic distortions.

An alternative approach is Fast Subband Adaptive Filtering, which is available in the current beta build. There is a thread on that at https://www.avnirvana.com/threads/fsaf-fast-subband-adaptive-filtering-measurement.13810/, the MATLAB source has extensive documentation PDFs.

 

[Antoine B]

I will look at the sources, however if I already see the big picture, I would like to know the detail, which mathematical function is used to generate the sweep and how the signal is combined with source to get frequency response. I suppose it is a simple division in frequential domain but the time reference is also important.

I hope I will see them in sources, or else I'll ask here ;')

 

[John Mulcahy]

If you want me to spoon-feed you the information to produce your own implementation I don't think that's a good use of my time. For information on sweep generation look at Synchronized swept sine.

 

[Antoine B]

Yeah I think I wasn't clear enough due to english issues, but I said I will first look at the sources and come back if I still don't have something