Isotropic noise generation

[ptengineer]

Anyone know how or have experience with generating isotropic noise? I've used normally distributed noise but I can't seem to turn up any material on isotropic. Thanks.

 

[GregLocock]

http://www.lmgtfy.com/?q=isotropic+noise+generation

Cheers

Greg Locock

1.79 x 10^12 furlongs per fortnight- that's as fast as I do anything

 

[VEBill]

Amusingly, the very first hit is now this exact thread.

(...Looks skyward, waves at passing Google robot...)

 

[VEBill]

Put the search term in quotes to exclude misses and the ONLY hit is this exact thread.

As we all probably already know, isotropic means equal in all directions. Perhaps the OP could provide more details of the application. Is it related to acoustical measurements?

 

[GregLocock]

I'm guessing he means EM radiation, so I'd start with an EM interference testing suite. Might help if he defined what he really meant by isotropic, it has several meanings.

Cheers

Greg Locock

I rarely exceed 1.79 x 10^12 furlongs per fortnight

 

[ptengineer]

Sorry about not providing some more background. Yes this is in regards to acoustics and I'm looking at the behaviour of some system composed of hydrophones when stimulated with isotropic versus normally distributed noise.

 

[VEBill]

Isotropic in this case must mean that the noise is arriving from all directions. Tricky.

 

[GregLocock]

I think you had better give a proper explanation of what you want, your use of the phrase "normally distributed" worries me.

Assuming VE1BLL is right then isotropic noise is easily generated on land, in a reverberant room.




Cheers

Greg Locock

I rarely exceed 1.79 x 10^12 furlongs per fortnight

 

[VEBill]

He's testing hydrophones. So must be in a big tank of water.

Relying on reflections to fill the test space with noise may affect the spectrum of the noise.

I was also wondering about the use of that phrase.

 

[ptengineer]

Currently I simulate the noise applied at the elements (hydrophones) of the array as normally distributed noise with a mean of zero and standard deviation of 1. For example in matlab if I have 10 elements (hydrophones) with a sampling rate of 1K the data coming from the array would be

time_data = randn(10, 1024); % 1 second of data assuming Fs is 1K

This looks like it is going in the right direction:

http://books.google.com/books?id=YI...noise generation&pg=PA24#v=onepage&q=&f=false

I'm trying to do this for noise that is now an isotropic field instead of a 1-D field.

Thanks again.

 

[GregLocock]

you need to model each sound source and then work out the contribution of each source at each hydrophone, allowing for attenuation with distance, and cancellation or reinforcement from the other sources.

Might have been a bit quicker getting to here from there if you had explained you were only simulating the whole experiment.





Cheers

Greg Locock

I rarely exceed 1.79 x 10^12 furlongs per fortnight