Not even the click of a pen or the rustle of a shirt goes undetected by Signal Interface Group’s acoustic imager. But how about a person’s, um, rude noises?
“We don’t record those,” the company’s president, Neil Fenichel, says with a smile.
The gizmo that Fenichel demonstrated this week at Signal Interface Group’s office in Bellevue, Wash., is designed for higher purposes: to find out why an elevator is whining, where an air-conditioning system is leaking, which fluorescent light is buzzing, why a car’s engine is making a funny sound, or even how a hummingbird does its buzz.
The imaging system, developed in cooperation with Bellevue-based OptiNav, combines several tricks of the acoustic trade. It starts with a square metal plate, about 2 feet wide, which is dotted with an array of 40 tiny digital microphones.
The impulses from each microphone are fed into OptiNav software that analyzes how much sound is coming from which direction. A graphical rendering of sound intensity is then superimposed on a video image from a camera that’s mounted on the plate.
“One of the goals is to make this something like a point-and-shoot,” Fenichel said.
The system costs about $9,000, including $6,000 for the hardware and $3,000 for the software. Comparable acoustic imagers from other providers can cost as much as $50,000, Fenichel said.
Fenichel, who was trained as a mathematician and engineer, has been working in the field of PC-based data acquisition for decades. He founded Microstar Laboratories in 1982, and six years ago, he founded Signal Interface Group as a spin-off that eventually focused on acoustic imaging.
“It was a very good choice, because this is a lot of fun,” he said.
One of the tests was conducted by OptiNav’s founder, Robert Dougherty, who used the microphone array to take a look under the hood of his car and identify the source of the engine’s weird whining noise. It turned out that the culprit was a serpentine belt. “You can see the belt and hear the whining,” Fenichel said.
Now Fenichel is heading to Germany, along with his son (and employee) Abraham, to spread the word about the acoustic imager at next week’s Berlin Beamforming Conference.
The Fenichels are already making plans to offer an API to developers that will let them apply their own algorithms to data from the microphone array. And they’re working on a version of the array that’s about five times as wide as the current model, which would be capable of detecting lower frequencies.
That should make it possible for architects to determine how sound waves propagate through a building’s walls and windows – or for party poopers to figure out which neighbor has the music turned up too loud.
So have any of the machines been sold as novelty items? “The answer to the question is, not yet,” Fenichel said.