High-Tech Microphone Picks Voices From a Crowd 221
JerryQ writes with news of an impressive audio detection system from a company called Squarehead that was demonstrated during a professional basketball game. According to Wired, "325 microphones sit in a carbon-fiber disk above the stadium, and a wide-angle camera looks down on the scene from the center of this disk. All the operator has to do is pinpoint a spot on the court or field using the screen, and the Audioscope works out how far that spot is from each of the mics, corrects for delay and then synchronizes the audio from all 315 of them. The result is a microphone that can pick out the pop of a bubblegum bubble in the middle of a basketball game..."
Re:Would work on stored sound too (Score:4, Informative)
Re:Would work on stored sound too (Score:5, Informative)
It occurs to me that if you store all 325 audio streams with accurate time-codes and the relative positions of the microphones you would be able to do this at any time later on the stored sound as well. You could probably get away with much fewer than 325 microphones at some cost in quality.
Yes. And that's already part of the system.
Sounds like beamforming (Score:5, Informative)
This sounds like beamforming [wikipedia.org]. Submarines do this. Works great.
Re:come on people... (Score:4, Informative)
and if you actually read *all* of TFA the scientist says 300 as well. so now we have three numbers: 325, 315, 300. If they keep this up they'll get down to 1 and their product will be a lot cheaper ;)
Of course we can assume he rounded there for ease of explaining.
Prior Art (Score:4, Informative)
My father, would tell me stories when I was growing up about helping design a surveillance tool for ease-dropping on restaurant conversions that used the same principle. They had a map of the table layouts and you would place a pointer over the table you wanted to listen to. Mics hidden around the edge of the restaurant would capture the sound. This was back during the early 60's so they used a mechanical delay mechanism. Said it worked as well as if you had planted the mic at the table, plus you didn't have to worry about where they sat. Like many things, this is more powerful and versatile but hardly anything new.
Re:Data harvesting? (Score:3, Informative)
This is the classic phase array antenna approach from radar tech applied to sound. Cool application though.
In fact it is easier for sound because the amount of data per element is much smaller than in let's say a radar.
big deal. (Score:1, Informative)
this was done by under grads from umass 6 yrs ago. I have no Idea if this was the first time it was done, but big deal
http://www.ecs.umass.edu/ece/sdp/sdp04/goeckel/ [umass.edu]
Re:come on people... (Score:3, Informative)
With only two microphones it's all about how they are located. If you can locate two mics so they both pick up the noise but only one of them picks up the signal then with modern "adaptive filtering"* DSP techniques you can achieve huge noise reduction. IIRC this technique is used for micing helicopter pilots among other things.
*I put adaptive filtering in quotes because you don't directly use the filter to remove noise. Instead you use the filter to eliminate magnitude/phase differences in the noise picked up by the two microphones. After adaptive filtering the "noise" can then be subtraced from the "signal+noise" to give just the signal (that's the idea anyway, in practice the adaptive filter is not perfect so some noise is left)