mkoijn
Structural
- Jan 7, 2003
- 28
Could anyone tell me if I am calculating sound intensity properly? The main difficulty I am having is when it comes to calculating the PI index. In theory the Intensity level in dB should equal the sound pressure auto spectrum average of the two mikes in dB when in perfect free field conditions. To test my method and as I don't have access to free field conditions I simply synthesised data with a known phase relationship and from that phase difference, calculated the distance between the microphones that would give exactly that phase shift.
The only way I can get the Intensity dB and the pressure dB data to match is by using the following method:
1) Run the data through the analyser and save the imaginary part of the cross spectrum and the auto power spectrum of both channels as text files ( this data is all in Pascals squared).
2) Calculate the intensity by -cross Imag/D*P*w
D= distance between microphones
P= air density in Kg/m3
w= frenquency in Hz*2*3.141592
( note that the cross imaginary data is still squared when calculating the intensity level)
3) convert intensity to dB using 10*log10(intensity/1pW)
4) Average the two auto power spectra and take the square root.
5) convert the auto power spectra to dB using 20*log10(aps/20uPa)
Using this method the intensity level is almost the same as the pressure level. At frequencies aproaching a phase difference of 60 degrees for that microphone distance, the intensity is slightly lower, but I understand this should be so because of the finite difference approximation.
If I use any other method i.e taking the square root of the cross imaginary data first before calculating the intensity level, or using 10*log10(aps/20uPa), or leaving the aps's squared, the resulting intensity dB level is far higher than the pressure dB level which isn't correct.
The other point I'm not clear on is why in the intensity level calculation air density is taken into account but is not for the pressure?
I've read a lot of articles and papers about sound intensity but I've not found anything that states explicitly the exact method for calculation. As I am working alone on this, completely self taught without any real physics background and on a very tight budget any help in this would be gratefully received.
Finnigan
The only way I can get the Intensity dB and the pressure dB data to match is by using the following method:
1) Run the data through the analyser and save the imaginary part of the cross spectrum and the auto power spectrum of both channels as text files ( this data is all in Pascals squared).
2) Calculate the intensity by -cross Imag/D*P*w
D= distance between microphones
P= air density in Kg/m3
w= frenquency in Hz*2*3.141592
( note that the cross imaginary data is still squared when calculating the intensity level)
3) convert intensity to dB using 10*log10(intensity/1pW)
4) Average the two auto power spectra and take the square root.
5) convert the auto power spectra to dB using 20*log10(aps/20uPa)
Using this method the intensity level is almost the same as the pressure level. At frequencies aproaching a phase difference of 60 degrees for that microphone distance, the intensity is slightly lower, but I understand this should be so because of the finite difference approximation.
If I use any other method i.e taking the square root of the cross imaginary data first before calculating the intensity level, or using 10*log10(aps/20uPa), or leaving the aps's squared, the resulting intensity dB level is far higher than the pressure dB level which isn't correct.
The other point I'm not clear on is why in the intensity level calculation air density is taken into account but is not for the pressure?
I've read a lot of articles and papers about sound intensity but I've not found anything that states explicitly the exact method for calculation. As I am working alone on this, completely self taught without any real physics background and on a very tight budget any help in this would be gratefully received.
Finnigan
