1
$\begingroup$

Possible a duplicate but I could not find a direct answer. I used a sound card (Focusrite, at 96 kHz sampling frequency) connected to a calibrated hydrophone to record an ultrasonic signal, with a known frequency at known power. I have all specification of the hydrophone including the frequency sensitivity. The gain on the sound card was kept at 0. I use scipy.io.wavfile.read to read in the wave in python. So what the further processing steps to get the voltage? The recording is in 24 bit. I read that I need to normalize with the number of bits, which is 2**(bit-1), but how exactly? I also have the exact specification of the sound card. I am sorry if the question is trivial or not well formulated, I am a biologist trying to learn the basics of signal processing.

$\endgroup$

1 Answer 1

5
$\begingroup$

I would be sceptical about prosumer hardware having reliable absolute level reproduction. Is it not safer to record a reference signal (in air acoustics you would have a reference sound source that you wrap around your microphone, record the signal, and that becomes your reference level. Multiply or divide your signal by any constant to make it into «1» or some other preferred representation, and when you later record your unknown acoustic source, multiply/divide by that same value.

$\endgroup$
1
  • $\begingroup$ Totally agree with this. The best way is to record a calibration signal from the calibrator using the hydrophone - this becomes your reference for scaling later. $\endgroup$
    – jojeck
    Aug 3, 2023 at 16:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.