Acoustic Measurements

[SL_in_Salado]

Question for Acoustics Experts.
Trying to calibrate the signal from a microphone using an acoustic calibrator that generates 94dB & 114 dB signal at 1000Hz.

Data from the microphone is collected in Volts, converted to Pascals using the microphone sensitivity (mV/Pa).
Options from this point are:
1) Convert the Pa data to dB using the zero reference of 20uPa. (SPL(in dP) = 20 X log10(p/pref).
2) Perform FFT with dB on the Y-axis. Look at peak level at 1000Hz.
3) Perform 1/3 octave analysis on FFT output and look at the dB in the 1000Hz band.

The problem is that only option 3 gives values anywhere near the 94 or 114dB generated by the calibrator. Where is the correct place to make the comparison? And once I find the correct place, how is a correction made?

One other question: The Voltage or Pascal data from the microphone are a low level (94dB = 1.002Pa = 0.053V). When the data is saved as a WAVE file, the volume is too low to hear. Is there a standard way to go from a voltage waveform from a microphone to a 16 bit WAV file that can be heard through something like Media player?

Hardware is SCXI 1000, 1600 & 1530 with single microphone.
(System Info: LabVIEW 7.1.1, DAQmx 7.4, Windows XP)

[chericks1]

Hi SL in T,

Are you using the Sound and Vibration Toolkit VIs to analyze the data?

Chad Erickson
Applications Engineer

[SL_in_Salado]

Have Sound and Vibration Toolkit Version 3.1. Once I replaced the standard FFT with the one from SVT I now get correct/matching dB levels for both FFT and Octive Analysis. >> Problem solved!

Have still not been able to figure out what I need to do to convert a waveform file containing microphone data (+-1 Pa) to a 16 bit WAV file with an amplitude large enough to be heard in Windows Media Player.

I tried multiplying the Pa waveform data by a multiplier and was able to get audio out. A 114db signal at 1000Hz sounded OK, but the 94 dB signal was a little distorted. (These signals produced by a Larsen Davis Precision Acoustic Calibrator Model CAL200)

Someone implied that SVT contains a VI for this conversion but I can't seem to find it.

Any suggestions?

[chericks1]

Hi SL in T,

There is an example in SVT called SVXMPL Wav Recorder that shows how to write to a WAV file. Please navigate to:

C:\Program Files\National Instruments\LabVIEW 7.1\examples\Sound and Vibration\Wav

Let me know if you have further questions or if this does not resolve your issue!

Chad Erickson
Applications Engineer

[SL_in_Salado]

Hi Chris,
I had acqually used the example you pointed out when writing my VI to write a WAV file. However, it does not seem to address the issue of how to get the data in the WAV file to be of sufficient size to generate an audible signal. (I also tried playing the file with "Resampling Wave File Player.vi" which was included in LabVIEW and got the same result.)

[chericks1]

Hi SL in T,

Sorry for the short responses. You can output up to an amplitude of 32768 to a WAV file. It is 16-bit meaning that it has a range of -2^15 to 2^15. -32768 to 32768. So if you write a tone with an amplitude of 32768, it is using all of the range of the WAV file. A possible consequence to amplification past this value is the signal being clipped in hardware and the sound being distorted.

Other than this, there is the volume control for the sound card which controls how loud the signal comes out. You probably know this, though 😄

Let me know if you have further questions on this issue!

Chad Erickson
Applications Engineer

[SL_in_Salado]

Yes, I continue to have questions.
How do I talk to a live preson at NI? I left a message on your voice mail...
Please give me a call.

Message Edited by Support on 04-29-2005 02:13 PM

[chericks1]

Hi SL in T,

I'm currently looking into this issue to see if there is a VI that does the conversion from microphone data to wav file. Thanks for your patience!

Chad AE
Applications Engineer

[Ian_Ren]

Try the vi attached.

1) Assume your signal is already in "Pa".
2) "Wave file full range" defines what the full range of the wave file represents, eg, 94dB means the full range of the wave file is 1Pa.

Hope it helps

Ian

[imperial-aero]

Hi,

I'm trying to convert the Voltage values from a microphone to Pascals. I'm wondering whether this is a straight forward conversion (using the mV/Pa specified on the microphone) or do I need some sort of acoustic calibration.

Any help would be very much appreciated.

Ruf.