How do many sounds get received by an ear?

[Graeme M]

I was just struck by a physical effect that I don't know anything about - how does an ear detect many sounds simultaneously? I did a quick Google but nothing came quickly to light.

 

I know generally how the ear works, but I guess something I don't understand is how different sounds are differentiated simultaneously. I suppose it's the same question as how a microphone detects sounds. When various sounds strike the ear membrane, it vibrates and transmits the vibrations via the ossicles to the cochlear and thence a variety of fine mechanisms to eventually result in signals to the brain. That much is easy enough as a general description.

 

What I don't understand is how a membrane vibrating can pass that much information. When I consider this, all I can see is successive additional sounds simply creating more intense vibrations. I'm not sure what the words might be to describe that, but I'm seeing a simple linear action - a membrane moving in and out in response to air pressure fluctuations. More sounds increase the frequency of vibrations and vice versa. I can't see why I hear more than a simple pitch change of a single sound.

 

Yet my ear (and a microphone) can detect the various sounds of an orchestra, or the sounds of a bushland setting.

 

What exactly is happening here?

[Strange]

You can add up the (relatively) simple vibrations of each sound to give the more complex vibrations of the combined sound.

 

Computationally, you can do the reverse - a Fourier transform - where you take that complex vibration and turn it into a list of individual frequencies and their amplitudes.

 

The ear has thousands of tiny hairs in the cochlea. These respond to different frequencies and so, effectively, do a Fourier analysis of the sounds entering the ear. The brain then uses the information from these hair cells to turn the vibrations into meaningful sounds.

 

How the brain is able to pick out one voice in a crowd, or the flute in an orchestra is, I think, still a bit of a mystery. I assume it is some sort of complex pattern matching - it knows what combination of frequencies make up that particular voice or instrument and filters them out from the background. Despite certain TV cop shows, I don't think that can be done by computers yet.

Edited August 12, 2015 by Strange

[studiot]

 

How the brain is able to pick out one voice in a crowd, or the flute in an orchestra is, I think, still a bit of a mystery. I assume it is some sort of complex pattern matching - it knows what combination of frequencies make up that particular voice or instrument and filters them out from the background. Despite certain TV cop shows, I don't think that can be done by computers yet.

 

 

Fourier transforms are not enough.

They provide an instantaneous spectrum of frequencies, ie a spectrum at every instant that (in general) varies from instant to instant.

 

Real sounds have a time envelope (attack, duration, decay) which also characterises them.

Varying these is used in electronic synthesisers to turn one basic oscillation into copies of many different sources.

[Graeme M]

My apologies if I seem particularly dense.

 

You can add up the (relatively) simple vibrations of each sound to give the more complex vibrations of the combined sound.

 

<snip>

 

The ear has thousands of tiny hairs in the cochlea. These respond to different frequencies and so, effectively, do a Fourier analysis of the sounds entering the ear.

 

I can follow what you say about the simple vibrations adding up to complex vibrations - in the case of sounds travelling through the air that seems obvious. I'm not sure it's so obvious to me about how that translates into the vibrations of a membrane as transmitted through three small bones. The hairs of the cochlear aren't responding directly to the sounds entering the ear, they are responding to the vibrations in a bone. Obviously it does happen, I'm not questioning that. I just can't see how these complex vibrations are still present in the output of the stapes. Clearly I don't understand something about vibrations!

 

When we have the vibrations or air pressure pulses of many sounds, we have a certain amount of information encoded there. Once it vibrates a membrane, and that membrane sympathetically vibrates a bone, it seems to me that the information is flattened out into a far simpler form (ie how many ways can a bone vibrate?). Put another way, generating the sound of a trumpet or a guitar requires a lot of information to be encoded - we change the frequency of the sound by a variety of mechanical means. Yet the same sound is received and reproduced by a completely inert bone.

Edited August 12, 2015 by Graeme M

[StringJunky]

Think of it as vibrational patterns in the fluid of the semi-circular canal and the hairs are placed strategically for frequency. The time difference between the two ears gives the positional information and partly whether they are from the same source or not.

 

This Yamaha Tutorial gives a nice overall view of ear anatomy, sound capture and processing. There's 3 sections.

Edited August 12, 2015 by StringJunky

[Strange]

 

Fourier transforms are not enough.

They provide an instantaneous spectrum of frequencies, ie a spectrum at every instant that (in general) varies from instant to instant.

 

You can only do a Fourier transform over some sample period (ideally it would be infinite but that isn't usually practical).

 

 

Real sounds have a time envelope (attack, duration, decay) which also characterises them.

Varying these is used in electronic synthesisers to turn one basic oscillation into copies of many different sources.

 

The envelope, and any other characteristics, are encapsulated in the range of frequencies generated by the Fourier transform.

 

For example, a single frequency is only really a single frequency if it is infinite in duration. If it is of shorter duration then there are more frequencies present. If it has a sharper attack and decay, then there are relatively more high frequency components.

 

See also: the uncertainty principle.

Edited August 12, 2015 by Strange

[Graeme M]

Thanks StringJunky, that tutorial is great. Such an amazingly complex system built on basic principles. However I can't make the necessary mental leap and as I noted it's clear that I am foundering on the matter of vibrations. Once the stapes vibrates the oval window membrane of the cochlear the form of the vibrations in the fluid causes the basilar membrane to vibrate at tuned frequencies (ie different locations) and stimulate the hairs. So far so good.

 

But I come back to the stapes. How can so much information be encoded in the simple vibration of a bone? Intuitively all I see is the bone vibrating faster or slower or with greater or lesser intensity. It's a 2-dimensional response that then creates a 3-dimensional response in the fluid cavity. The stapes activates the membrane at a particular rate with a particular intensity - I don't see the difference between that and a tuning fork. How does that simple motion encode for all the sounds of a complete orchestra?

 

I suppose the question is now more one of physics. I failed physics at school...

[StringJunky]

Thanks StringJunky, that tutorial is great. Such an amazingly complex system built on basic principles. However I can't make the necessary mental leap and as I noted it's clear that I am foundering on the matter of vibrations. Once the stapes vibrates the oval window membrane of the cochlear the form of the vibrations in the fluid causes the basilar membrane to vibrate at tuned frequencies (ie different locations) and stimulate the hairs. So far so good.

 

But I come back to the stapes. How can so much information be encoded in the simple vibration of a bone? Intuitively all I see is the bone vibrating faster or slower or with greater or lesser intensity. It's a 2-dimensional response that then creates a 3-dimensional response in the fluid cavity. The stapes activates the membrane at a particular rate with a particular intensity - I don't see the difference between that and a tuning fork. How does that simple motion encode for all the sounds of a complete orchestra?

 

I suppose the question is now more one of physics. I failed physics at school...

I like the article too... perfectly condensed. On the question of how such complexity transmits through the stapes... It does! Think of all the information embedded in the undulating line of a vinyl record; a needle is following a groove and vibrates left/right and up/down, varying a current.

 

Part of your problem, I think visualising this, is you are experiencing what appears to be a completely seamless symphony of sounds, when to another more highly-tuned organism it's patchy and rough as heck! We don't know what we don't know ...if that makes sense... it's only amazing because that's all we know. Our brain will 'digitise' any gaps to make our sonic experience smoother as well.

Edited August 12, 2015 by StringJunky

[Strange]

The stapes activates the membrane at a particular rate with a particular intensity - I don't see the difference between that and a tuning fork. How does that simple motion encode for all the sounds of a complete orchestra?

 

The difference is that a tuning fork resonates at a single frequency. The stapes simply transfer the complex vibrations going on in the air to the inner ear.

 

So a tuning fork will vibrate like a sine wave, but the stapes will vibrate like:

[Graeme M]

Well, that's got to be one of the more amazing things I've come across. Of course I always knew that sound is transferred by vibrations and roughly how an ear and a microphone (and a record stylus) work, but to actually sit down and think about it shows me an underlying complexity that to be honest I can't get my head around. It still seems weird to me that an entire orchestra (for example), doing so many complex movements to generate a whole host of separate sounds, can be replicated by a simple vibrating membrane.

 

I tried to read up on Fourier transforms and how they can unpick the components from a complex waveform and it escapes me completely.

 

On a slightly different tack, the energy used to vibrate the air or fluid is doing work, does that imply that the ear should heat up in loud/complex sound fields? Does air heat up from sound vibrations?

[studiot]

 

 

Graeme M

On a slightly different tack, the energy used to vibrate the air or fluid is doing work, does that imply that the ear should heat up in loud/complex sound fields? Does air heat up from sound vibrations?

 

 

Not significantly.

The energy in sound is small compared to the heat capacity of materials like the solids and liquids forming the ear.

The standard atmospheric pressure on the eardrum is far, far greater than the pressure variations that are the sound waves.

If the eardrum had to support anything approaching that magnitude of pressure difference it would burst.

That is what the Eustachian tube is there for.

To equalise the standing pressure on either side of the eardrum.

 

 

Strange

You can only do a Fourier transform over some sample period (ideally it would be infinite but that isn't usually practical).

 

Yes exactly, and then it is not exact, and is called a finite fourier transform, except the sample period is zero for an analytical Fourier Transform. You mean the interval of integration , which runs from minus infinity to plus infinity, not the sampling period.

 

In fact the ear is a measurement device that measures an amplitude time plot or graph.

But it has multiple sensors, each responsive to different frequencies. It does not perform a Fourier transform even a finite one since the sensitivity of these receptors is not constant. The finite transform it does generate is still subject to the vagaries if aliasing and windowing as with any finite time domain to frequency domain transform. It is also sensitive to 'crossover distortion' for sounds that whose frequency spectrum have significant components with the range of more than one sensor.

The receptors themselves comprise fine hairs growing along the basal membrane, which is just ove 30 mm long) within the cochlea.

The higher frequencies are detected at the end near the exciting membrane, called the oval window, and the lower frequencies at the remote end.

 

Edited August 13, 2015 by studiot

[jamieoverton727]

Sound waves travel into the ear canal until they reach the eardrum. The eardrum passes the vibrations through the middle ear bones or ossicles into the inner ear. The inner ear is shaped like a snail and is also called the cochlea. Inside the cochlea, there are thousands of tiny hair cells. Hair cells change the vibrations into electrical signals that are sent to the brain through the hearing nerve. The brain tells you that you are hearing a sound and what that sound is.

[studiot]

Sound waves travel into the ear canal until they reach the eardrum. The eardrum passes the vibrations through the middle ear bones or ossicles into the inner ear. The inner ear is shaped like a snail and is also called the cochlea. Inside the cochlea, there are thousands of tiny hair cells. Hair cells change the vibrations into electrical signals that are sent to the brain through the hearing nerve. The brain tells you that you are hearing a sound and what that sound is.

 

Yes this is all true, but misses important things out.

[CPG]

I'll try to add some anatomical perspective that I think could help tie things together with some of the physical and mathematical perspective on acoustics that has already been provided (good posts!).

 

The mechanical vibrations of the hair cells along the cochlea are able to open ion channels in order to transduce vibration into a neural (electrical) signal.

 

 

http://www.nature.com/nrn/journal/v15/n9/images/nrn3786-f3.jpg

Hudspeth, A. J. Integrating the active process of hair cells with cochlear function. Nat. Rev. Neurosci. 15, 600–614 (2014).

 

 

 

Let's zoom out from the cellular level to see the structure and layout of the cochlea. Different locations along the structure are differentially sensitive to the frequency content of these vibrations. This is mostly a result of the cochlear structure rather than having hair cells that are sensitive to different frequencies. I'm oversimplifying a bit here but you can approximate it like so:

 

Image Source:

https://sites.google.com/site/jayanthinyswebsite/_/rsrc/1380059170661/workshops/different%20aspects%20of%20organ%20of%20Corti.png

 

 

 

These neurons then project sensory information to the brain. If you're already familiar with how our sense of touch is somatotopically organized in the brain (different regions of the brain represent different parts of the body), this acoustic information is handled in a similar way. The auditory cortex in the brain has a tonotopic organization that allows the incoming signals to be processed in different places, and you end up with an organization in the brain something like this:

 

 

Image source, University of Calgary:

http://ucalgary.ca/pip369/mod6/sound/auditorybrain

 

 

 

 

A few side notes:

 

- Sound does not travel in air as it does the aqueous environment of the cochlea. The bones of the middle ear are needed to perform impedance matching, so that vibrations resonating at the eardrum will be mechanically transduced to vibrate appropriately in fluid of the inner ear.

 

- As long as you're thinking about the structure of the cochlea, notice that the high-frequency hearing range is towards the opening where all sounds must pass through. As people age, we tend to lose our hearing most dramatically from this end, but it is very loud and low frequency vibrations that do the most damage. (ie blasting the bass too loudly in your headphones when you listen to music).

 

- I tried to provide a simplified overview, but I would like to take this opportunity to apologize to nature for glossing over her staggering complexity and elegant solutions to so many real-world problems in acoustic perception.

 

 

CPG

 

 

Edit to add: The first image I linked is visible to anyone visiting the nature website, but unfortunately the full text of that review is not. Sorry

Edited August 17, 2015 by CPG

[studiot]

Cpg, your post is a welcome extension of mine, but one small point needs some clarification.

 

 

As long as you're thinking about the structure of the cochlea, notice that the high-frequency hearing range is towards the opening where all sounds must pass through

 

Along with your picture, this could be taken to imply that the end of the cochlea is open. It is not open but covered or closed by a membrane.

 

 

https://en.wikipedia.org/wiki/Oval_window

Edited August 17, 2015 by studiot

[CPG]

 

Along with your picture, this could be taken to imply that the end of the cochlea is open. It is not open but covered or closed by a membrane.

 

 

True, I did not intend to give the impression that the fluid of the inner ear is continuous with the outside.

 

For anyone wishing to visualize why this anatomical clarification matters, here's a brief animation of the physiology at work:

 

http://blausen.com/?Topic=8814

Edited August 17, 2015 by CPG