h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

did you realise that hearing orks in a similar manner.

all those little hairs in your ear act as a phased array radar.

it is these that fail when you start to lose sensitivity in part of your hearing (ie top end only).

hi xyroth

I'm afraid I tend to disagree... The signals received in a phased array are superimposed on each other to derive a directional pattern of the desired shape: let's assume that the beam is set up to point at boresight with 1° beamwidth. Then anything received from, say, 10° to the left enters the antenna elements at times t(i) and is subjected to an additional time delay (or phase shift) delta_t(i). When combined, these individual contributions cancel each other out because any contribution with a phase p has a counterpart with phase -p and (ideally) nothing is left at the end. A signal from boresight enters all elements at the roughly same time t(i) = t(0) and the contributions add because the effect of delta_t(i) is to make them appear at the summation point with the same phase p0. Thus, a phased array makes up a device for *spacial* selection.

Within an ear, the hairs are arranged in a line and all of them are exposed to the *same* signal, at different times t(i) because of the time delay along the structure. There are no phase shifters or somesuch. What makes those hairs different from each other is their frequency response: the ones closest to the entry point are tuned to the low frequencies and those further down the chain react to higher frequency. Thus, the ear is built like a spectrum analyser, or call it a filter bank receiver. One reason for deafness or imparted hearing is the lack or destruction of some of these hairs, so that the ability to react to a part of the spectrum gets lost. Thus, the ear basically works in the *frequency* domain.

Human ears - taken individually - don't react to phase shifts. A striking example is music being replayed from a tape recorder. A magnetic tape does in no way preserve the phase relations within a piece of music, across the frequency range. If you compare the signal before and after recording music to a tape then you'd find it rather impossible to tell that it was the same thing. It's only the amplitude vs. frequency relation that is preserved. However, experience tells us that tape recordings are 'good'.

On the other side, the *pair* of ears could be called a phased array because it's the time difference between left and right that makes us realise that a source is located in some horizontal direction or other. But, consisting of only two receiver elements, the aural system should be better addressed as a monopulse system, built from two omnidirectional elements.

Apparently, this is true for the horizontal plane only. Regarding the vertical plane, it's the shape of the outer parts of the ear that matters: those wrinkles of an ear serve to apply some distinct distortions to the spectrum of an incoming signal, and it's only the experience gathered with time that enables us to tell whether some sound originates from somewhere above or below. - This latter point is a bit of guesswork from my side (I found something to this effect when doing the research for the 'bat's sensors' piece).



Bossel

true, phased arrays are usually used for spatial descrimination, but it was proved many years ago by astronomers that you could take the same basic system and optimise it for frequency descrimination rather than spatial descrimination.

this is what happens in the ear. because the hairs can only send a basic on-off signal, the brain processes this multitude of signals to massively improve frequency descrimination as a seperate but complementary system to the eardrum.

the other problem is that the hairs are only long enough to spot certain frequencies, but by spoting harmonics and subharmonics during this processing it can access a greater range of frequencies than it might otherwise be able to do.

Hmm, this sounds very much like the description of a filter bank receiver and not of a phased array.

What I forgot to point out earlier is that *all* elements of a phased array are identical with respect to their frequency response and directional pattern. It's only the time delay (caused by phase shifters and geometry) and the weighing factors (which make up the amplitude taper function) that defines the directional pattern of the arrangement which, multiplied by the element pattern, yields the overall directional pattern. Phased arrays only work in a narrow frequency band, say, +/- 5% off the centre frequency.

The ear's receiver elements (ie: hairs down the cochlea) differ in their frequency response, which makes them completely different from the elements of a phased array. The whole ear covers some 9 octaves which is far beyond the capabilities of a phased array.

What I can see as a common principle is that both systems derive their results by combining contributions from a multitude of sensors. Or am I missing the point?

Perhaps your are both right, in part. Have you considered the ear as not only having both a filter bank (the cilia) but one which, through its construction, is able to correlate the multipath phasing effects that come from a received signal arriving from different directions relative to the outer ear's canal? The multipath effect that occurs is primarily composed of phase changes (phased array principles) but the positioning of the individual cilia reacting to the sound wavefront in its final form inside the cochlea acts like a correlative system (filter bank, but spatially correlated as well) to indicate rough positioning of the incident sound wave.

Food for thought.

Cheers,
DT

Oh thanks DNAT!

If I get that right then there is a direction finding capability built into each individual ear, right?



Hey, if you click the 'edit' button on your space and write anything (*anything*, just so it isn't the blank page any more) then other people will get to see a 'discuss' button there and are able to contact you!

by george I think he's got it.

well done dnat.