singing

References

Summary

Sound is longitudinal pressure wave. It contains areas of high density (compression) and low density (rarefaction)

When the glottis opens, air particles are pushed through, by subglottal pressure. The actual molecules themselves don't make it very far, and travel relatively slowly. What creates the sound is the molecules colliding with each other: the collision energy travels along

Note: The mean free path of air at atmospheric pressure is only about 68 nm. So air molecules are colliding all the time

The actual pressure waves produced by the voice are only about 0.1 Pa. But the standard atmospheric pressure is 101,325 Pa.

The smallest sound a human can hear at 1000 Hz is about 20  × 10-6 Pa.

Air

Subglottal Pressure

The diaphragm and intercostal muscles regulate the air flow. Support is essentially the condition of maintaining constant subglottal pressure as volume of air in the lungs decreases.

Subglottal pressure determines the loudness of the voice, by altering the amplitude of the

As the pressure increases,

Maximum Flow Deceleration Rate

The slope of the glottal flow waveform as it falls back to zero.

Acoustic power is propertional to the square of MFDR; if the folds snap shut quickly, the glottal flow stops quickly

Acoustic Radiation Pressure

The air pressure coming out of the mouth

Par ∝ Ps2.

Perceived Loudness

This is the sound pressure level

Vocal Tract

Lungs

Trachea

A rigid tube made of cartilage rings. These are extremely sturdy and prevent your airway from collapsing inward when air gets pulled into the lungs, and the pressure in your trachea decreases.

Subglottic Space

This sits right below the larynx. The air from the lungs pools here, while the vocal folds are closed.

Eventually the air pressure forces the folds open.

Larynx

This contains all of the main pieces of your voice:

  1. Vocal folds
  2. Glottis
  3. Epilaryngeal space

The larynx is a protective cage made up of cartilage, that sits on top of the trachea.

Vocal Folds

The vocal folds actually aren't very thin; they're more like a block.

Glottis

The space between the vocal folds. This is initially shut, but air pressure from the subglottic space pushes it open from the bottom.

Since the folds have some depth to them, they open from the bottom to the top, which takes a bit of time.

Now, as the air passes through the glottis, the elasticity of the folds wants to push the glottis closed. But the column of air being pushed through them wants to keep it open.

Bernoulli's Principle

The energy associated to a fluid in motion (like anything) is made up of kinetic and potential energy. The potential energy of the fluid corresponds to the pressure. So

$E = \dfrac{1}{2} \rho v^{2} + P$

If the energy is constant, Bernoulli's law says that increasing the velocity decreases the pressure. Inside the glottis things are a bit different.

These more nuanced features are actually a good thing — they allows the pressure inside the glottis to decrease, which leads to its closing.

As the air molecules near the edges bump into the vocal folds, some of that energy is lost as heat — what that means at the molecular level, is that instead of the kinetic energy going straight into the wall and coming back out, it gets distributed in a number of different random directions. This is inelastic scattering

As a consequence, the energy that comes back into the air molecules when the vocal fold pushes back is actually less than it started with. In this way, the air molecules on the boundary lose more and more energy, decreasing the internal pressure in the glottis, allowing the vocal folds to start closing

This has nothing to do with Bernoulli's principle.

Meanwhile, the further you get into the bulk of the fluid, the faster the air travels, because it's not doing inelastic scattering with the folds. It ends up ripping away from the boundary air, and becoming a free jet that isn't attached to the walls anymore.

The free jet then creates vortices and turbulence, introducing all kinds of complicated physics.

  1. What is Bernoulli's Principle?

    Bernoulli's principle says that in a fluid with steady flow the higher the velocity of the fluid, the lower the pressure.

    So for example, if air or water is accelerating (i.e. if a force is acting on it), it will drop the pressure. But this only works once the fluid is already flowing.

    When the glottis opens, the air that was blocked by it initially accelerates, but this is irrelevant, because the air wasn't flowing just before then (because it was blocked by the closed glottis). So as the air passes through the glottis, it doesn't actually drop the pressure.

Supraglottal Pressure

Now, as the air passes through the glottis, it eventually meets the air above the glottis; the new air crashes into the air that's sitting there, creating a pressure wave that passes through the rest of the vocal tract. These pressure waves are what we interpret as sound.

Because of atmospheric pressure, there's always some air in our vocal tract already

Rarefaction

As the glottis begins to close, and less air is coming through, the air above the glottis continues moving forward, creating a rarefaction zone (lower pressure) above the glottis.

Now, air molecules aren't all traveling the same direction; there are always some of them are traveling backwards, towards the glottis. They'll fall back into the glottis, further empasizing the rarefaction zone.

This rarefaction zone sitting right at the top of the glottis allows the myoelastic forces to start to close the glottis from the top

This allows the air to get cut off even faster.

Note: If the folds are open, the air after the glottis still crashes into the next bit of air, but there's no area of rarefaction, so instead of a wave with peaks and troughs, you just get a steady block of pressure, which doesn't correspond to sound (or really it's a kind shush or breath-like sound; white noise basically).

Fourier Series

Fourier theory tells us that the sharper the edges on a periodic function, the more sine waves (i.e. the more overtones) you need to form it.

If you can clamp the glottis shut quickly at the top and bottom, your pressure wave has a sharp drop — it's basically a sawtooth wave! Which means that the sound that will be produced when you have a whole bunch of these sawtooth pieces will have a ton of higher overtones

But why doesn't everyone sound operatic all the time then?

The key is the way the air column above the glottis moves, leaving the rarefaction zone behind.

Opera singers basically make it so that the air above the glottis rockets forward

Ventricular Folds

Also called the false vocal folds.

Epilaryngeal Space

Sits right on top of the larynx. You need it to act as a smooth transition between the vocal folds and the pharynx (which is just open air inside your throat)

When sound travels between materials with different impedances, a lot of energy reflects at the boundary between the media. This is because the two different media have different densities and elasticities.

If you start with loose, light molecules (low density, high elasticity) and go to tight heavy ones

If you start with a dense, inelastic material and go to a light, elastic one

Physics Derivation

Why does this happen? Well if we look at conservation of momentum:

Say the heavy molecule has speed v0M and mass M and the light one is initially at rest and has mass m and gets accelerated to velocity $vfm$nilnil. Then by conservation of momentum:

Mv0M = vfMM + vfmm

Now use conservation of energy:

m(vfm)2 + M(vfM)2 = M(v0M)2

We rewrite both equations to get M(v0M−vfM) = vfmm m(vfm)2 = M(v0M−vfM)(v0M+vfM)

Yielding vfm = v0M + vfM; plugging into the conservation of momentum and rearranging gives

$v_{f}^{m} = \dfrac{2M v_{0}^{M}}{M + m}$

which shows that the final velocity of the smaller particle can never exceed 2v0M, and in fact decreases as m decreases.

  1. Relativity

    By the way, close to light speed, this stops holding (this isn't relevant for sound waves, but it's just a cool fun fact). It turns out that if you use the relativistic versions of the equations, heavy molecules can transfer essentially all of their energy to lighter ones in a collision!

    This is actually something that produces cosmic ray air showers in our upper atmosphere.

Acoustic Impedance Matching

So we need the epilaryngeal space to help us match the densities of the vocal folds and the pharynx; otherwise sound will be lost through reflections.

One way around this it to just try to increase the amplitude / energy of the sound — this is what untrained singers do

Acoustic Impedance

Acoustic impedance (Z) is acoustic pressure (Δ P) divided by the acoustic particle velocity (u) and the cross-sectional area of the tube.

$$ Z = \dfrac{\Delta P}{u A} $$

Pressure: The acoustic pressure is the difference between the baseline pressure of the medium, and the local pressure making up that region of the sound wave; in other words, the difference of the bunching up of the molecules as the sound wave passes through them, from when they're left alone.

Why does the pressure change as you pass through different media? The pressure wave travels along as a macroscopic phenomenon, due to the molecule motion carried forward by kinetic energy. Depending on the medium, it might be harder or easier to move molecules.

When the medium is dense, the molecules are tightly packed. They can't move very much at all, so as the energy passes through them, they stick together locally, giving a high pressure

In a low density medium, the molecules have a lot of room to move, and so as the energy passes through them, you only have a few molecules moving together at a time, until they crash into the next group, and those few ones move together, and so on. So the pressure is lower.

Velocity: But there are 3 different velocities of the sound:

  1. Bulk flow velocity: the speed of the individual molecules traveling along the vocal tract
  2. Wave velocity: the speed of the pressure wave (the speed of sound in the medium)
  3. Acoustic particle velocity: the speed of the molecules oscillating back and forth as the pressure wave passes through them.

Vocal folds: High impedance

Pharynx: Low impedance: the cross sectional area is huge, and the pressure is much lower

Epilaryngeal: If we make the tube very narrow, we reduce the cross-sectional area, increasing the impedance.

Aryepiglottic Sphincter

This is a ring of muscle that sits on top of the vocal folds. They connect to the arytenoid cartilage at the back of your vocal folds, all the way up to the sides of the epiglottis (which is what closes when you swallow)

When you activate these muscles, you narrow the vocal tract at that point

Sensation: The feeling of closing the glottis is like when you hold your breath. But tightening the aryepiglottic sphincter feels like the sound is brightening; more pointed; you aren't changing the folds, you're just forcing the sound through a smaller hole

Impedance Matching

We use the sphincter to narrow the epilaryngeal space, increasing the impedance to match the vocal folds. This prevents us from losing energy in the transition from the folds to the epilarynx.

But what about going into the pharynx? There should still be a reflection problem since the impedance drops.

Well, when you use the aryepiglottic sphincter, it doesn't give you a harsh sudden narrowing into the pharynx; it's more tapered. As you travel along, and the tube widens, the impedance does still change, and you get mini reflections along the way, but they end up adding to less than if you just did a hard transition from the folds right into the unnarrowed pharynx.

Resonance and Amplification

Because the epilarynx is closed at one end (the vocal folds) and open at the other (the top of the sphincter), it's a quarter wave resonator — this means it will amplify any sound whose wavelength is exactly 4 times the length of the tube.

If you work out the calculation for the 2-3 cm long epilaryngeal tube, you get that it's about 3000 Hz (!)

At the opening to the pharynx, the molecules hit by the sound wave get thrown into the pharynx due to the low impedance (this is called an acoustic pressure node). As we mentioned above, it's impossible for the light molecules in the pharynx to carry all of the energy of the pressure wave, so a bunch of it does reflect back (about 50% even in the best case scenario)

So if you aren't singing opera, and don't have those amplified 3000 Hz frequencies, you just end up with 50% of the sound you started with

But if you do have those 3000 Hz, look what happens:

You very quickly reach equilibrium though, which is why the sound doesn't just get louder forever

Pharynx

Goes from the top of your larynx to the back of your nose.

Highly muscular, allowing fine control

In opera we want to widen the pharynx (this is the famous open throat). We need a nice gradual taper along the pharynx, to make sure that the impedance mismatch doesn't destroy the sound.

Inferior/Middle Pharyngeal Constrictor Muscles

In speech, these are squeezed inward. In opera, we keep them widened.

This creates a large chamber that amplifies lower frequencies

But we need this to give a gentle taper. So in fact, the inferior muscles can't just go completely slack. You have to figure out the perfect way to slowly loosen the muscle as you move upwards (this is one of the many reasons why opera is so hard)

By nature, the pharyngeal constrictors all fire together; they're meant to push food into the esophagus

You need specific training to overcome this

  1. Incipient Yawn

    You practice the very start of a yawn — where the lower throat opens, but the top doesn't

  2. Semi-Occluded Vocal Tract

Infrahyoid Muscles

Your pharynx starts where the larynx stops. You can lengthen it by moving the larynx down. Lengthening a tube shifts its resonances lower, so this allows you to emphasize the low frequencies even more. But it also bunches the higher frequencies closer

Normally, your pharynx has resonances around 2500 Hz, 3000 Hz, and 3500 Hz. When you lengthen the pharynx, these regions coallesce into one big region

The sternohyoid and sternothyoid muscles pull the larynx downward.

Velopharyngeal Closure

The soft palate or velum sits between the pharynx and the nasal cavity. Opera singers will actually try to completely seal off the nasal cavity

The levator veli palantini muscle contracts, lifting the soft palate and slamming it against the back of the throat.

This is very surprising, because in singing, you're often taught to activate nasal resonance. Turns out, there's no such thing as nasal resonance — you're never actually meant to resonate the sound in your nasal cavity (if you do get the sound in there, the spongey tissue and hairs there will actually kill it; especially the high frequencies). When passing through the nasal cavity, the sound actually winds up muffled

The reason it feels like it's in the nose, is because if you manage to get the higher frequencies into your mouth, they resonate the hard palate, which then reverberates up into the nose.

Tongue Root

The back of the tongue actually sits inside the pharynx. In untrained singing, the tongue will fall backwards and close off the pharynx.

The genioglossus muscle moves the root of the tongue forward and down, away from the back wall of the throat

Oral Cavity

Jaw

Opera singers drop the jaw down and back, rotating it at the TMJ (temporomandibular joint); this helps with that gentle taper: they keep space in the back of the throat into the mouth, and then out

The wider the mouth, the higher the F1 formant

Tongue

Controls the F2 formant

Lips

The tip of the tongue and the lips are used for articulation

Acoustic Inertance

M = ρL/A M is the inertance, L is the length of the tube, and A is the cross-sectional area

Narrowing the epilaryngeal tube increases the inertance. When the air reaches the wide pharynx, the inertance drops

Muscles

Interarytenoid Muscles

Close the glottis

Lateral Cricoarytenoid Muscles

Close the glottis

Cricothyroid Muscle

Tilts the thyroid cartilage forward from the cricoid cartilage. This increases tension on the vocal folds and thins their mass.

Thyroarytenoid Muscles

Resonance

Resonance inside of a cavity is really a property that emerges due to multiple waves — the reason the epilaryngeal space is able to boost the 3000 Hz frequency, is because when the pressure waves reflect back off of the pharynx and then off of the vocal folds, they constructively interfere with the next group of waves coming through the folds.

More specifically, the 3000 Hz frequencies constructively interfere. It's very much not that the epilaryngeal space is generating 3000 Hz frequencies itself; it's just that it provides the right environment for any existing 3000 Hz frequencies to constructively interfere.