Acoustic Position Measuring Device
sears at cs.berkeley.edu
Tue Aug 19 19:37:19 CEST 2008
Cédric Berger wrote:
> On Tue, Aug 19, 2008 at 16:49, Charles Pax <charles.pax at gmail.com> wrote:
>> I have a few questions
>> 1) Are the any technical obstacles to this idea?
Time synchronization, sensitivity / volume of audio devices.
>> 2) What is the frequency range of the speaker?
>> 3) What is the frequency range of the microphone?
From google (speed of sound / 22000hz), here's the accuracy we're
speed of sound at sea level / (8000 hertz) = 4.253625 centimeters
speed of sound at sea level / (22 000 hertz) = 1.54677273 centimeters
>> 4) Any other comments of ideas?
(Assuming that we're using the FR speaker as the source, and a headphone
mic as the recording device):
Have the speaker send a long-ish sound whose frequency varies, from
(say) 4khz - 22khz. After recording, convolve the mic signal with the
original signal, and look for the spikes. The highest point on the
first big spike is the one you're looking for.
If the FR can simultaneously record from the FR mic and the headphone
mic, then you can get two spikes from one noise. One is for calibration
(it's at ~ 0 meters), the other is the measurement. This way you don't
need to know how much latency the soundcard has introduced. Otherwise,
you'd need to have the user do the process twice, once with the headset
mic near the FR, and a second time for the measurement.
I think the signal processing is simple enough to allow for continuous
measurement of distance. (Though it'd make a very irritating noise
while measuring. ;)
> I would think main problem is microphone sensitivity/speaker strength.
> Hard to get over noise.
> But I am curious to know to what point this could work !
> Openmoko community mailing list
> community at lists.openmoko.org
More information about the community