I'm trying to compare two audio samples of speech from two different Android smartphones to determine if the people are co-located. I'm doing this using the frequency bins of an FFT between 300 Hz - 3400 Hz. When the two devices are sitting on a table, I have good results.

However, when the devices are placed within pockets, the audio is recorded differently than when they are placed on a table.

Does anyone know exactly what happens to the audio frequencies/signal when the microphones are covered by a pocket or a bag? Are certain frequencies filtered?

  • $\begingroup$ well, the bag isn't the problem – it's the giant meaty sound absorber the bag is worn on. Basically, whereas two microphones on the same plane which has the "open space" on one side are pretty much guaranteed to have the same "dominant" components, being the "directest" paths from source of sound to microphone, at leg height, that won't be the case, because not only is one side of the microphone the leg, isolating much better than a table, but also is it to be expected to not have mostly direct, but mostly indirect or heavily multipathed sounds. So, much much harder problem. $\endgroup$ – Marcus Müller Mar 30 '17 at 12:56
  • $\begingroup$ In addition to what @MarcusMüller said, if the person is moving, the frequency response (corresponding to the pocket and the leg) can be rapidly and constantly changing. $\endgroup$ – user304584 Mar 30 '17 at 15:49

The distance emitter-receptor and the presence of obstacles in the trajectory(clothes for example) will afect the recorded signal for sure. So some frequencies can be attenuated or just disappear.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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