WTF!? Wavelet TransForm for audio measurements - What-is? and How-to?

[mige0]

The port length is the hottest candidate as far as I can see (pipe "resonance") - you do not need the anechoic room to do a zero distance measurements at the port.
Most probably IF its from there, you will also see it at close mic'ing at the bass unit - probably attenuated by the dampened box volume


Michael

 

[Elias]

Hi Kees,

Thanks for finding some good use for the tool


About the measurements of yours, was the chamber truly anechoic, since it seems to have a reflection at 5ms? Would you try to analyse the impulse response with constant Q (without multiresolution) and having a value of 3 for Q. It should point the possible reflections/diffractions very well.


Then for the room analysis it is certainly possible to extend time/freq limits, but with computational load. I chose the limits so that a normal few years old laptop can handle the calculations in a reasonable time.

At the moment you could try Bark wavelet, it has time range up to 200ms.


Or better yet modify the limiting values to your needs:

for example constant Q wavelet function wt_cq.m (in your folder path)

values for lower freq limit are here, 500Hz is the default, range is between 200 to 2000Hz: You could change 200 to 20 to be able to start from 20Hz.

return_input_params{2, 1} = "range";
return_input_params{2, 2} = "Start frequency [Hz] ?";
return_input_params{2, 3} = 500;
return_input_params{2, 4}{1} = 200;
return_input_params{2, 4}{2} = 2000;


stop time parameters are here, default value 10ms and range 5 to 100ms: You could try to change 100 to a higher value.

return_input_params{4, 1} = "range";
return_input_params{4, 2} = "Stop time of the 2D plotting [ms] ?";
return_input_params{4, 3} = 10;
return_input_params{4, 4}{1} = 5;
return_input_params{4, 4}{2} = 100;



Maybe better not to try 3D for extended limits, or you need a powerfull PC. Or maybe my PC is too poor


- Elias

Hi all,

first of all I would like to thank you for making this very nice tool available!

I am playing with it, feeding it some of my IR's and have some questions:

1 - please see the attached picture, the measurement of one of my Genelec 8050 studio monitors. http://www.genelec.com/pdf/DS8050a.pdf . This measurement was performed in a fully anechoic room (at least above 40Hz) at 1m distance.
I would like your opinion on the wavelet, what do you think it says about the speaker?
I know that the phase response can be seen in the first arrival, but what do all the resonance peaks up to 5ms tell me? This cabinet is know for it's low diffraction design and stiff cabinet, so is it all still driver/cabinet resonances or a processing artefact of some kind?

2 - Would it be possible to extent the analysis down to 20Hz? I understand that this takes bigger FFT filters and thus calculation time but... Or is it a limitation in the time domain?

3 - I am a professional acoustical consultant and studio designer and I feel that this could also be a great tool for analysing room IR's, because it will show you the amount of energy per frequency which is absorbed with a reflection AND the phase of the reflection (which is changed if it is caused by a panel absorber such as all walls) this is vital for making a good LF wavefront throuhgout the room.
I have tried some measurements with the wavelet but the frequency limits and time limits are now a constraint to do this, do you have any plans to make them bigger?

again, many thanks for making this available!

Kees