Calibrating microphone with air hose hiss.

[Circlotron]

The other day I downloaded a trial version of Spectrogram 7 http://www.visualizationsoftware.com/gram/gramdl.html very useful even in the demo version BTW. It is an audio spectrum analyser with a wiggly line, 1/3 octave, and coloured mess display. For best results you need to use a calibrated microphone of course, and the software has an input equaliser to flatten the respone of the mic you are using. Ater thinking long and hard, I hit upon the idea of using the air hose from my compressor and blowing the air past (not into!!) the mic and recording the results. IMHO it is almost a perfect white noise source that rises at 12dB per octave as it should. Just a matter of fiddling the mic equalisation till you have a upward ramping straight line, then away you go.

[dhaen]

Circlotron,

How did you determine that it was a true white noise source, and didn't have some other similar law?
Doesn't the sound change with the nozzle?


Cheers,

[Circlotron]

I did it with the nozzle fully open.
First, this is proper white noise generated with CoolEdit, saved to a CD and played into the sound card from the CD player.

[Circlotron]

This is the unequalised microphone response.

[Circlotron]

This is the equalised microphone response. I don't really know that it is true white noise but it certainly sounds like it. The equalised microphone measurements now make a lot more sense than what they did before.

On this last pic the vertical scale is 30dB per division, not 20dB like the other two pics.

[dhaen]

Circlotron,

Yep, I'd go with that too.
Few things in nature have a "droopy near the top end" law.
(Except my anatomy)

Cheers,

[Circlotron]

One thing I should mention is that I used the bare microphone insert. It used to be inside this plastic housing complete with clip, and there was a 2mm hole about 2mm deep through which the sound travelled on it's way to the mic. This made the response just rotten. Shown below is the original response to white noise.

[halojoy]

Originally posted by Circlotron
This is the equalised microphone response. I don't really know that it is true white noise but it certainly sounds like it. The equalised microphone measurements now make a lot more sense than what they did before.

On this last pic the vertical scale is 30dB per division, not 20dB like the other two pics.

Hey, Circlotron

So what is your conclusion?

Can we say that a hair drier can be used
as a poor mans - white noise generator?

Can you describe to all of us,
what use we can have of white noise.
Is this of no interest if we have no Oscilloscope
and other "fancy" test instrument to go along with it.
---------------------------------------------------------

Are you discussing with the tiny expert-panel?
or with the vaste majority, who have solder irons & a multimeter
but not much more.

-------------------------

Please, Explain to us many readers:
What is it all about? This WHITE NOISE thing.

halo - limping his way - in half-truths

[Circlotron]

Everyone here has a pc, most of us can afford a $2 electret microphone, all of us like audio. You can plot frequency response graphs of loudspeakers if you feed them pink noise and use the software mentioned in the first post. Trouble is, the results are only as good as the microphone you are using. You need a known source of wideband noise to apply to your microphone first and see how it behaves so you can set your spectrum analyser software so the microphone frequency response appears to be flat with pink noise. (actually rising in the case of white noise) High pressure hissing air (100+ psi) seems to be a reasonable and reproducible example of true white noise .


-------------------------


/Circlotron - being most patient with halojoy.

[mrfeedback]

Graham, Behringer or AKG or somebody sell a calibrated electret capsule for $AUS30 or so.
I'll find out today.

Eric.