It's funny to see a technique being heralded as 'novel' which has been use for well, I personally have known it for at least 20 years. In earlier days it simply consisted of connecting two dynamic mikes anti-polar (connect the "minus" terminals together, and take the output from both "plus" terminals). The idea is that noise from a distance will arrive at the same time at two mikes next to eachother, thus cancelling out signal in that configuration. Noise from proximity (i.e. your mouth) would induce a difference in pressure and would thus create a signal.

You can still see it used when technicians have to quickly rig mics without the benefit of a sound check or equaliser - just check for mike pairs.

The difference is now that such balancing has to be done by electronics (opamps) because electret microphones cannot be set up in that configuration - until someone gets smart and sticks it in one house..

New? Not by a long shot..

The only place the word 'new' appears in the article is in the sentence "This multi-mic idea is not new" - so you are correct in your comment when you state "New? Not by a long shot". Perhaps a second reading of the article is in order before you post comment upon't?

We know the concept isn't new. The article explicitly says that it isn't new. And I'm not seeing the word "novel" used either.

This is about a development: putting the concept to work in a Bluetooth headset.

"This multi-mic idea is not new, but it requires memory and a fair chunk of compute power to process the audio streams. Dynamic Hearing said it chose CSR because CSR's BlueCore5-Multimedia chip has both memory and a DSP (digital signal processor) built-in."

This is the quote in the article. Granted the technique isn't new, however, the fact that you're doing it in a bluetooth headset is a novel idea.

The common way to use two microphones is just to wire them together. So why are they using a signal processor? Are they using a delay loop so they can try mounting the two microphones further apart?

The more interesting thing that you can do with two microphones and a signal processor is, you can cancel noise at the earphone, to make it easier to hear the conversation. People know about that, but no one ever actually does it.

The most interesting thing you can do is predictive noise cancelation of cyclic signals - like car engine and road noise. If they are trying that, I wish someone would say so.

Check out the Jawbone bluetooth headset, it has 3 mics and noise cancellation and has been on sale since last year. I bought one and it works as advertised.

It's not new tech (as noted) but its not even novel in a bluetooth headset. I own a Jawbone Bluetooth (http://www.jawbone.com/) and it does the same thing...

Even if there's nothing new, chances are, they'll probably manage to get a patent for it in the USA where the patent office almost seems to specialise in granting patents for what the rest of the world very rightly accepts as unpatentable.