"Peter Webb" wrote in
u:
"Skywise" wrote in message
...
wrote in
:
Which leaves building something like a solid state video camera for
audio.
One could build a 32X32 pixel audio imager with 1,024 cheap
microphones like these for $0.27 in quanties of 1,000:
http://www.mouser.com/catalog/specsheets/KT-400332.pdf
Follow those with a little amplification/buffering, a switch matrix,
and dump the output into a PC sound card with appropriate software.
And voila, you have an audio imager.
With this design you'd need an audio input for every microphone.
A typical sound card only has 2 inputs (stereo) so you'd need
512 audio cards. And then processing all those signals? Yikes!
Rather, have a microcontroller scan your 1024 inputs and send
the values to the PC via USB, where software then assembles the
input into an image.
OH, very important point. Each pixel would be generating a value
based on sound volume. Frequency would have to be discarded, so
this would be a 'monochrome camera'. You could expand to three
inputs for each pixel with each of a different bandwidth, then
that could give you 'color' when represented as RGB.
Disagree. You could measure the sound pressure wave, which will give you
frequency information.
Indeed, I suspect that you would have to anyway, for a couple of
reasons:
1. If you just try and measure average intensity from each mike over
some period, you are effectively sticking in a bandpass filter which
removes high frequencies, which is where most of the directional
information comes from.
2. The cheapest/easiest/most effective of building such a device is by
attaching them to a flat surface (eg a wall) and use it is as a phased
array (eliminates mirrors and lenses and the attenuation and distortion
they create). This will require the baseband signals to be correlated.
I think you could actually build a very accurate imaging device in this
manner. If you place the microphones (say) 10 metres apart, you should
get a resolution at 10 kHz of about 1/100 of a radian. Same principle as
used in radio astronomy when linking radio-telecopes seperate by a large
distance to improve resolution.
I do not disagree with what you describe, but you described a completely
different method than what was originally proposed, which was to use
acoustic lenses and mirrors to form an image. Phased array has no need
for these structures.
Brian
--
http://www.skywise711.com - Lasers, Seismology, Astronomy, Skepticism
Seismic FAQ:
http://www.skywise711.com/SeismicFAQ/SeismicFAQ.html
Quake "predictions":
http://www.skywise711.com/quakes/EQDB/index.html
Sed quis custodiet ipsos Custodes?