MUSIC 276: Psychoacoustics and the Science of Musical Sound

These notes draw on Thomas D. Rossing, Richard F. Moore, and Paul A. Wheeler’s The Science of Sound (3rd ed., 2002), Brian C. J. Moore’s An Introduction to the Psychology of Hearing (6th ed., 2012), Arthur H. Benade’s Fundamentals of Musical Acoustics (2nd ed., 1990), William M. Hartmann’s Principles of Musical Acoustics (2013), and supplementary material from Stanford University MUSIC 150 (Musical Acoustics) and MUSIC 151 (Psychoacoustics) course materials and MIT OpenCourseWare 21M.380.

---

Chapter 1: Vibrations, Waves, and Sound

The study of musical acoustics begins with the most elemental physical phenomenon in all of music: vibration. Every sound — the warmth of a cello, the attack of a snare drum, the shimmer of a cymbal — originates in the back-and-forth mechanical oscillation of some physical object. To understand the richness and variety of musical sound, we must first develop a precise and quantitative language for describing vibration, and then understand how vibrational energy propagates through air to reach a listener’s ear. This chapter lays that foundation, moving from the simplest idealization of oscillatory motion through the wave equation and the decibel scale, to the perceptual landmark of the equal-loudness contour. Throughout, we will keep one foot in the physics and the other in the experience of music, because the two are inseparable.

1.1 Simple Harmonic Motion

The simplest and most fundamental type of vibration is simple harmonic motion (SHM). A mass on a spring, a pendulum swinging through a small angle, the cone of a loudspeaker driven at a single frequency — all are well described by SHM. The essential ingredient is a restoring force: whenever the object is displaced from its equilibrium position, a force acts to push or pull it back, proportional to how far it has moved.

\[ \omega = \sqrt{\frac{k}{m}}. \]\[ f = \frac{1}{T} = \frac{\omega}{2\pi} = \frac{1}{2\pi}\sqrt{\frac{k}{m}}. \]

This formula encodes an important physical principle: stiffer systems (larger \(k\)) vibrate faster, while more massive systems (larger \(m\)) vibrate more slowly. Instrument makers have exploited this for millennia. Tightening a guitar string raises \(k\) (by increasing tension, which acts as an effective spring constant for transverse displacement) and raises the pitch. Winding bass strings with extra wire raises \(m\) per unit length, lowering the pitch without requiring an impractically long string.

\[ \dot{x}(t) = -A\omega\sin(\omega t + \phi), \qquad \ddot{x}(t) = -A\omega^2\cos(\omega t + \phi). \]

Notice that velocity is 90° out of phase with displacement (the particle moves fastest when it passes through equilibrium and is momentarily at rest at the extremes), and acceleration is exactly 180° out of phase with displacement (the restoring force is always directed toward equilibrium).

1.2 Damping and Resonance

\[ m\ddot{x} + b\dot{x} + kx = 0, \]\[ x(t) = A e^{-\zeta\omega_0 t}\cos(\omega_d t + \phi), \quad \omega_d = \omega_0\sqrt{1 - \zeta^2}. \]\[ Q = \frac{\omega_0}{2\zeta\omega_0} = \frac{m\omega_0}{b} = \frac{f_0}{\Delta f}, \]

where \(\Delta f\) is the half-power bandwidth of the resonance peak (the frequency range over which the response is within 3 dB of its maximum). Thus \(Q\) quantifies both temporal selectivity (how long the system rings) and spectral selectivity (how narrowly it responds in frequency): these are not independent properties but two faces of the same coin, related by the time-bandwidth product \(\Delta t \cdot \Delta f \approx Q/(\pi f_0) \cdot f_0/Q = 1/\pi\).

A high-\(Q\) resonator rings for a long time and has a very sharp frequency response; a low-\(Q\) resonator decays quickly and responds broadly. A guitar string may have \(Q \sim 1000\); a drum membrane \(Q \sim 10\); the ear canal resonance \(Q \sim 10\). The body of a violin, which must respond broadly enough to reinforce strings tuned to different pitches, has body resonances with \(Q \sim 20\)–50. In contrast, a tuning fork, designed to maintain a single precise frequency, has \(Q \sim 10{,}000\)–\(30{,}000\) depending on alloy and geometry.

\[ x_{\text{ss}}(t) = \frac{F_0/m}{\sqrt{(\omega_0^2 - \omega_{\text{drive}}^2)^2 + (b\omega_{\text{drive}}/m)^2}}\cos(\omega_{\text{drive}} t - \delta), \]

where \(\delta = \arctan\!\bigl(b\omega_{\text{drive}}/m(\omega_0^2 - \omega_{\text{drive}}^2)\bigr)\) is the phase lag. The amplitude is maximized when \(\omega_{\text{drive}} \approx \omega_0\) — this is resonance. The sharpness of the resonance peak is governed by \(Q\): the frequency bandwidth at which the amplitude falls to \(1/\sqrt{2}\) of its peak value is \(\Delta\omega = \omega_0/Q\). Every musical instrument exploits resonance: the body of a guitar resonates at and near the string frequencies, efficiently transferring vibrational energy to the surrounding air.

1.3 The Wave Equation

Sound is not a local phenomenon — it is the propagation of mechanical disturbance through a medium. When a vibrating surface displaces the air immediately in contact with it, a chain reaction of compression and rarefaction propagates outward. The mathematical description of this propagation is the wave equation.

For a one-dimensional medium (a string, or a narrow tube of air), the transverse displacement \(y(x, t)\) satisfies

\[ \frac{\partial^2 y}{\partial t^2} = c^2 \frac{\partial^2 y}{\partial x^2}, \]\[ y(x, t) = f(x - ct) + g(x + ct), \]\[ y(x, t) = A\cos(kx - \omega t + \phi), \]

where \(k = \omega/c = 2\pi/\lambda\) is the wavenumber and \(\lambda = c/f\) is the wavelength. At 440 Hz in air, \(\lambda = 343/440 \approx 0.78\) m — roughly the length of a standard ruler. At 20 Hz (the lower limit of hearing), \(\lambda \approx 17\) m; at 20 kHz (the upper limit), \(\lambda \approx 1.7\) cm.

\[ \frac{\partial^2 p}{\partial t^2} = c^2 \nabla^2 p. \]

1.4 Sound Intensity and the Decibel Scale

\[ I = \frac{p_0^2}{2\rho_0 c}. \]

The human auditory system responds to an extraordinary range of intensities — from the threshold of hearing near \(I_0 = 10^{-12}\) W/m\(^2\) to the threshold of pain near \(1\)–\(10\) W/m\(^2\), a ratio exceeding \(10^{13}\). A linear scale is entirely impractical for this range; instead, we use a logarithmic one.

\[ L_{\text{total}} = 10\log_{10}\!\left(10^{L_1/10} + 10^{L_2/10}\right). \]

Two identical sources each at 70 dB combine to give 73 dB, not 140 dB — a consequence that surprises many students encountering acoustic addition for the first time.

Typical SPL values in musical contexts: a single violin playing pianissimo at 2 m is approximately 40–45 dB; a full symphony orchestra at fortissimo in the front seats is 100–105 dB; a rock concert near the speakers can reach 110–120 dB. The OSHA standard for safe continuous exposure is 90 dB over 8 hours; each 5 dB increase halves the permissible exposure time.

1.5 Equal-Loudness Contours

The decibel level tells us the physical intensity of a sound, but loudness is a perceptual quantity — it describes how intense a sound seems to a listener. The crucial insight of Harvey Fletcher and Wilden Munson (1933) is that the ear is not equally sensitive at all frequencies. The auditory system is most sensitive in the frequency range 1–5 kHz, where the outer ear canal resonance (around 3–4 kHz) provides a boost of up to 10–15 dB.

The equal-loudness contours (standardized as ISO 226:2023) map the combinations of frequency and SPL that are perceived as equally loud. Each contour is labeled in phons, where \(N\) phons at any frequency means the sound is as loud as a 1 kHz tone at \(N\) dB SPL.

\[ S = 2^{(L_p - 40)/10}, \]

where \(L_p\) is the loudness level in phons and \(S\) is loudness in sones. Thus 50 phons \(\approx 2\) sones; 60 phons \(\approx 4\) sones; 70 phons \(\approx 8\) sones.

---

Chapter 2: The Harmonic Series and Timbre

If Chapter 1 gives us the physics of a single sinusoidal tone, Chapter 2 confronts a central fact of musical reality: almost no musical sound is a pure sinusoid. The rich, distinctive quality of a violin versus an oboe, of a bright trumpet versus a mellow French horn, cannot be encoded in a single frequency. It lives in the structure of the sound’s component frequencies — in the harmonic series and the distribution of energy among its members. Fourier’s theorem tells us that this description is not just convenient but mathematically complete.

2.1 Fourier’s Theorem

\[ f_n = n f_1, \quad n = 1, 2, 3, \ldots \]

The set \(\{f_1, 2f_1, 3f_1, 4f_1, \ldots\}\) is the harmonic series. The first harmonic (\(n = 1\)) is the fundamental; the second (\(n = 2\)) is the octave; the third (\(n = 3\)) is the perfect twelfth (octave plus fifth); the fourth (\(n = 4\)) is the double octave; the fifth (\(n = 5\)) is a major third above the double octave; and so on. Musicians have recognized the special consonance of these ratios intuitively for millennia — the harmonic series is the acoustic origin of just intonation.

2.2 Timbre and the Spectral Envelope

The sequence \(\{A_n\}\) of harmonic amplitudes is the amplitude spectrum of the sound. The timbre of a steady-state periodic tone is determined primarily by the shape of this spectrum — specifically by the spectral envelope, the smooth curve interpolating the peaks \(A_n\) as a function of frequency \(nf_1\). Two tones with the same fundamental frequency but different spectral envelopes will have the same pitch but very different timbres.

\[ c(n) = \mathcal{F}^{-1}\!\left\{\log\left|\hat{S}(f)\right|\right\}, \]

and the MFCCs are the first dozen or so coefficients of the cepstrum computed on a mel-frequency scale (a perceptually motivated frequency scale that is approximately linear below 1 kHz and logarithmic above). MFCCs are the dominant feature representation in automatic speech recognition and music classification systems, and they capture timbre more compactly and robustly than the raw amplitude spectrum.

Different instrument families have characteristic spectral shapes. The clarinet in its chalumeau register suppresses even harmonics (for reasons we will explain in Chapter 4), giving a spectrum dominated by the odd harmonics \(f_1, 3f_1, 5f_1, \ldots\) — a hollow, reedy quality. The trumpet at fortissimo has a brilliantly rich spectrum with significant energy in harmonics 10–20, produced by nonlinear wave steepening in the bore at high amplitudes. The flute at soft dynamic levels has a spectrum dominated by the fundamental with comparatively little energy in the upper harmonics — a cool, pure quality. The oboe has a highly complex spectrum with strong harmonics 3–8 relative to the fundamental, contributing to its penetrating nasal character.

2.3 Transient Timbre and the ADSR Envelope

Steady-state timbre — the spectrum of the sustained portion of a note — is only part of the story. The attack transient, the first few tens of milliseconds of a sound, carries disproportionate perceptual weight. In classic experiments (Grey, 1977), listeners who heard instrument tones with the attack portions removed or spliced onto mismatched sustained tones identified instruments far less accurately than with intact tones. When the attack of a piano was prepended to the sustained body of a harpsichord tone, listeners most often identified the sound as a piano — the attack dominated perception.

The amplitude envelope of a musical tone is often modeled in synthesis and electroacoustic contexts by the four-stage ADSR model.

A piano note has an extremely short attack (a few milliseconds), a rapid decay, effectively zero sustain (the hammer rebounds immediately from the string), and a long natural decay determined by string damping. A bowed string can maintain indefinite sustain and shape the release deliberately with a diminuendo. A pipe organ can maintain perfectly flat sustain (as long as the key is held and the bellows supply sufficient air pressure) with an essentially instantaneous release. These envelope differences are perceptually decisive in shaping the expressive character of each instrument.

2.4 Spectrograms and Time-Frequency Analysis

The spectrogram is the standard tool for visualizing how the spectrum of a sound evolves in time. It displays time on the horizontal axis, frequency on the vertical axis, and intensity (in dB) as a color or gray value at each point. Formally, the spectrogram is derived from the short-time Fourier transform (STFT):

\[ \mathrm{STFT}(t, f) = \int_{-\infty}^{\infty} s(\tau)\, w(\tau - t)\, e^{-2\pi i f \tau}\, d\tau, \]\[ \mathcal{S}(t, f) = \left|\mathrm{STFT}(t, f)\right|^2. \]\[ \Delta t \cdot \Delta f \ge \frac{1}{4\pi}, \]

with equality for a Gaussian window. In practice, a window of 20 ms gives a frequency resolution of roughly 50 Hz — adequate for separating harmonics of a low note but merging the harmonics of a high note. Spectrograms of music reveal the onset and decay of individual notes, the presence of vibrato (a periodic frequency modulation that appears as a wavy harmonic track), and the formant structure of the voice.

---

Chapter 3: Acoustics of String and Keyboard Instruments

The vibrating string is among the most ancient and universal of musical sound sources, and it is also among the most tractable analytically. The physics of the ideal string leads directly to the harmonic series, explaining why plucked and bowed string instruments blend so naturally with one another and with the voice. The departure from the ideal — through string stiffness, bridge coupling, and soundboard resonances — is where the distinctive character of individual keyboard instruments emerges.

3.1 Modes of the Ideal Vibrating String

\[ c_s = \sqrt{\frac{T}{\mu}}. \]\[ y_n(x, t) = A_n\sin\!\left(\frac{n\pi x}{L}\right)\cos(2\pi f_n t + \phi_n), \quad n = 1, 2, 3, \ldots \]

The mode shapes \(\sin(n\pi x/L)\) are the normal modes of the string: the first mode has no internal nodes (maximum displacement at the center); the second mode has one node at the center; the \(n\)th mode has \(n-1\) internal nodes. The general motion is a superposition of all modes, with coefficients determined by the initial shape and velocity of the string (i.e., where and how the string was plucked, struck, or bowed).

The formula \(f_n = \frac{n}{2L}\sqrt{T/\mu}\) encodes the three classical means of controlling pitch: length (stopping), tension (tuning pegs, capo), and linear density (wound strings for bass). A string twice as long, at equal tension and density, vibrates one octave lower. A string under twice the tension vibrates \(\sqrt{2}\) times faster — roughly a tritone (six semitones) higher in equal temperament.

3.2 Inharmonicity and Piano Tuning

Real strings are not perfectly flexible — they have bending stiffness, which adds a restoring force proportional to the curvature of the string rather than its tension. This stiffness raises each partial above its ideal position by an amount that grows with harmonic number.

For typical piano treble strings, \(B \sim 10^{-4}\); for bass strings wound with copper or steel, \(B\) can reach \(10^{-3}\) or even \(10^{-2}\). For the lowest bass notes on a concert grand, the inharmonicity is large enough that the 16th partial may be 60–80 cents sharper than the ideal 16th harmonic — more than half a semitone.

This inharmonicity creates the phenomenon of octave stretch in piano tuning. Because the second partial of note \(A_3\) is somewhat sharper than exactly twice the fundamental of \(A_3\), tuning \(A_4\) so that its fundamental coincides with that raised second partial gives an \(A_4\) that is slightly higher than exactly twice the frequency of \(A_3\). This effect accumulates across the full range, so a well-tuned piano has its highest notes about 30–40 cents sharper, and its lowest notes about 30 cents flatter, than strict equal temperament would specify. Octave stretch is not an error — it is what makes a piano sound in tune with itself.

3.3 Wolf Tones and Body Resonances in Bowed Instruments

The interaction between string and instrument body is not always benign. In bowed string instruments, when a string’s modal frequency coincides closely with a strong body resonance, the result can be a wolf tone: a note that refuses to sustain steadily, instead wavering and “howling” between two nearby pitches. The physical mechanism is a coupling instability between the string (driven by the bow) and the body resonator: energy transfers back and forth between the two resonators at a rate comparable to the bowing frequency, causing the amplitude to fluctuate at low frequency.

The wolf tone most commonly appears on cellos around E3–F3 (approximately 165–175 Hz), where a strong body resonance (the “main body” or “main air” mode) coincides with a string fundamental. Cellists manage the wolf tone through a combination of subtle bowing technique (adjusted bow speed and pressure), the use of a wolf suppressor (a small metal cylinder with a piece of felt, clamped to the after-length of the G string to detune the string’s secondary resonance and shift the coupling away from the wolf frequency), and instrument selection (some instruments are worse than others depending on the exact coincidence of body modes and string pitches).

3.4 Coupling, Soundboards, and Instrument Families

The string alone radiates very little sound: its cross-sectional diameter is far too small to push air efficiently, and the acoustic radiation resistance of a string-diameter object at audio frequencies is negligible. The bridge transfers the string’s vibrational energy to a much larger radiating surface — the soundboard (in guitars and violins) or the rim and back plates (in pianos).

The bridge is not a passive transmission element but an active filter: its mechanical impedance, which varies strongly with frequency, shapes how efficiently the string loses energy at each partial. A string coupled to a bridge loses energy most rapidly at frequencies near the bridge resonances, which shortens the sustain at those frequencies. The ideal bridge has an impedance high enough that the string vibrates freely (low energy loss per cycle, long sustain) but low enough that sufficient power flows to the soundboard to produce adequate volume — a compromise that Stradivari and the great violin makers resolved empirically through centuries of experiment.

---

Chapter 4: Acoustics of Wind and Brass Instruments

Wind instruments produce their tones by establishing standing-wave resonances in a column of air. Unlike strings, where the resonator and radiation mechanism are distinct, in wind instruments the air column is both the resonator and — through its open end or bell — the primary radiator. The bore geometry determines which resonant frequencies are excited and which notes and timbres are available to the player.

4.1 Standing Waves in Cylindrical Pipes

Consider a rigid cylindrical pipe of length \(L\) containing air at speed of sound \(c\). The acoustic pressure \(p(x, t)\) satisfies the one-dimensional wave equation. The boundary conditions at each end depend on whether it is open (pressure node: \(p = 0\)) or closed (velocity node, equivalently pressure antinode: \(\partial p/\partial x = 0\)).

In real pipes, the open end is not a perfect pressure node — the effective acoustic length of the pipe is slightly longer than the physical length, because the acoustic field extends a short distance beyond the open end before it can radiate freely. This end correction is approximately \(\Delta L \approx 0.6\, a\), where \(a\) is the radius of the pipe. For a clarinet with bore radius \(a \approx 7\) mm, the end correction adds approximately 4 mm to the effective length — a small but acoustically significant amount that must be accounted for in the design of fingering systems.

4.2 The Flute and the Clarinet

The flute is an open-open cylindrical pipe: the embouchure hole acts as one open end, and the foot joint (or the first open tone hole) acts as the other. It therefore supports all harmonics and overblows at the octave (moving from first to second register by increasing air speed). The flute’s spectrum, particularly in the low and middle register, is dominated by the fundamental and the second harmonic, giving its characteristic pure and cool tone.

The clarinet behaves acoustically as a closed-open pipe in its fundamental (chalumeau) register: the reed end is a nearly rigid pressure antinode, and the first open tone hole (or the bell) acts as an open end. It therefore supports only odd harmonics (\(f_1, 3f_1, 5f_1, \ldots\)) and overblows at the twelfth — the interval of an octave plus a fifth — corresponding to the jump from the fundamental to the third harmonic \(3f_1\). This is why the clarinet has three distinct registers (chalumeau, clarion, altissimo) rather than the two of most other woodwinds, and why the gap between the high chalumeau and low clarion registers (the so-called “break”) requires a different fingering strategy.

4.3 Brass Instruments: Impedance and the Bell

Brass instruments are neither simple open nor simple closed pipes — they are complex hybrids of cylindrical and conical tubing, widening to a dramatically flared bell. The acoustic behavior is most naturally described through the input impedance \(Z_{\mathrm{in}}(f)\), the ratio of acoustic pressure to volume velocity at the mouthpiece. High impedance peaks correspond to frequencies at which the mouthpiece pressure builds up most efficiently — these are the resonant frequencies that the player can “lock on to” with the lip reed.

The bell of a brass instrument transforms the impedance curve in two crucial ways. First, it raises the cutoff frequency \(f_c\) above which the impedance peaks flatten out and sound radiates efficiently outward; at frequencies below \(f_c\), energy is largely reflected at the bell and remains in the resonator. Second — and more importantly for musical usability — the curved taper of the bell corrects the otherwise non-harmonic spacing of the impedance peaks of a simple truncated cone, pushing them toward an even harmonic series. The player can thus play the entire harmonic series with predictable, musically useful pitches.

4.4 The Lip Reed and Embouchure

The air column resonances do not self-oscillate — they must be driven by the player’s lip reed. In brass instruments, the player’s lips are pressed against the mouthpiece cup and form a valve that opens and closes in synchrony with the pressure oscillations in the mouthpiece. This is a regenerative feedback system: the pressure peak in the mouthpiece pushes the lips apart, admitting a pulse of high-velocity air; the subsequent pressure trough tends to close the lips; and the cycle repeats at the resonant frequency of the air column.

The lips behave as an outward-striking reed (opening when mouthpiece pressure is high), which is the appropriate valve type to maintain oscillation at a pressure impedance peak. By adjusting lip tension (via embouchure muscles), the player shifts the natural resonant frequency of the lip system. When the lip resonance is near one of the air-column impedance peaks, the two resonators couple and the system oscillates stably at that frequency. Moving from one note to another in a harmonic series requires the player to shift lip tension to the neighborhood of the next impedance peak, a skill developed through years of practice.

The mouthpiece plays a specific acoustic role beyond simply holding the lips in place. The mouthpiece cup volume acts as a Helmholtz resonator in conjunction with the backbore, creating a strong resonance (the “popping frequency”) in the range 500–900 Hz. This resonance, positioned between the second and third harmonics of the instrument’s playable range, helps the player lock onto notes in the upper register and contributes to the characteristic “bite” and projection of the brass tone. Different mouthpiece depths and cup volumes shift this resonance, and players adjust mouthpiece choice to suit their embouchure and preferred register range.

4.5 Woodwind Tone Hole Acoustics

In woodwind instruments, the pitch is controlled primarily by opening and closing tone holes — openings in the bore that effectively terminate the resonating air column at the position of the first open hole. When a tone hole is open, the acoustic pressure is approximately zero at that location (open boundary condition), shortening the effective resonating length. When all tone holes are closed, the full tube length resonates at the lowest note.

\[ \Delta L \approx \frac{a^2}{a_h^2}\left(t + 0.7 a_h\right), \]

where \(a\) is the bore radius at the hole location. Larger holes give shorter length corrections (more effective cutoff), while smaller holes are less effective at terminating the bore. In the saxophone, the large tone holes (approaching the bore diameter) are extremely effective and give excellent octave behavior; in the baroque flute, the smaller holes produce significant length corrections and require careful adjustment of finger position and embouchure to achieve correct intonation.

---

Chapter 5: Percussion and the Human Voice

The previous chapters dealt with one-dimensional resonators — strings and tubes — whose modes are harmonic or nearly so. Percussion instruments introduce richer geometry: the two-dimensional membrane and the three-dimensional plate or bell. The modes of these systems are generally not harmonic. The human voice is in a category of its own: it uses a biological source-filter system to produce the extraordinarily nuanced and flexible timbres of speech and singing.

5.1 Vibrating Circular Membranes

\[ \frac{\partial^2 z}{\partial t^2} = c_m^2 \left(\frac{\partial^2 z}{\partial r^2} + \frac{1}{r}\frac{\partial z}{\partial r} + \frac{1}{r^2}\frac{\partial^2 z}{\partial \theta^2}\right), \quad c_m = \sqrt{\frac{\sigma}{\sigma_s}}. \]\[ z_{mn}(r, \theta, t) = J_m\!\left(\frac{z_{mn}^{(m)}}{a}\, r\right)\cos(m\theta + \psi_{mn})\cos(2\pi f_{mn} t + \varphi_{mn}), \]

where \(J_m\) is the Bessel function of the first kind of order \(m\), and \(z_{mn}^{(m)}\) is the \(n\)th positive zero of \(J_m\).

This non-harmonicity is why a simple drum does not produce a clearly defined musical pitch. The ear, hearing inharmonic partials that do not fit a common fundamental, does not assign a definite pitch sensation (or assigns one only vaguely based on the lowest mode).

5.2 Tuned Percussion

Despite the inherent inharmonicity of membranes and plates, many percussion instruments produce clear pitches. The solution in each case is to use the physical system to bring certain modes into harmonic (or approximately harmonic) ratios while suppressing or damping the remaining inharmonic ones.

The timpani (kettledrum) is the classic example. The air cavity enclosed by the kettle below the drumhead couples to the membrane, shifting several of the lower \((1,n)\) and \((2,n)\) modes upward so that their frequencies approximate the ratios 1 : 1.5 : 2 : 2.5 — a harmonic-like series that gives the timpani its characteristic definite pitch. The player also deliberately strikes the head at roughly one-quarter of the diameter from the rim, which preferentially excites the \((1,1)\) mode (which participates in the near-harmonic series) while suppressing the \((0,1)\) mode (which does not).

The marimba bar achieves tuning through geometry: the wooden bar is undercut in a curved arch along its underside. By adjusting the depth and shape of this arch, the maker can precisely control the frequency of the second transverse mode, pulling it to exactly four times the fundamental (two octaves above). This gives the marimba its characteristic warm, pure tone — the ear hears two harmonically related partials and fuses them into a single rich, pitched percept. The vibraphone uses a similar undercut to tune the second partial to three times the fundamental (a twelfth), and additionally employs resonating tubes and a motor-driven vane system to produce vibrato.

5.3 Bells and Inharmonic Spectra

Orchestral bells (tubular bells, orchestral chimes) and cup bells (church bells, handbells) produce tones with approximately harmonic partial structures achieved through their three-dimensional geometry rather than through deliberate undercuts. A church bell’s partials — the hum, prime, tierce, quint, nominal, and superquint — are tuned through centuries of foundry craft to approximate specific intervals, giving the bell a complex but recognizable pitch identity.

The hum tone of a bell is approximately one octave below the prime, which defines the perceived pitch. The tierce is a minor third above the prime; the quint is a fifth above; the nominal is an octave above the prime. Modern bell founders verify and adjust these partials by selectively removing metal from specific regions of the bell’s inner surface, a process called tuning that requires both acoustic measurement and experienced craftsmanship.

5.4 The Human Voice: Source-Filter Model

The voice is perhaps the most sophisticated acoustic instrument in existence. Gunnar Fant’s source-filter model (1960) provides the theoretical framework for understanding vocal sound production.

\[ F_n \approx \frac{(2n-1)c}{4 L_{vt}}, \quad n = 1, 2, 3, \ldots \]

giving \(F_1 \approx 500\) Hz, \(F_2 \approx 1500\) Hz, \(F_3 \approx 2500\) Hz. Real vocal tract shapes, modified by tongue, jaw, lip, and velum movements, shift these formant frequencies to cover a wide range.

Vowel identity is determined primarily by the positions of the first two formants \((F_1, F_2)\) in the acoustic space. The vowel /a/ (as in “father”) has high \(F_1 \approx 700\)–800 Hz and relatively low \(F_2 \approx 1100\)–1200 Hz. The vowel /i/ (as in “see”) has low \(F_1 \approx 270\)–300 Hz and very high \(F_2 \approx 2200\)–2300 Hz. The vowel /u/ (as in “moon”) has low \(F_1\) and low \(F_2\). These \((F_1, F_2)\) coordinates define the vowel space, and the entire vowel inventory of a language can be visualized as a set of points in this two-dimensional plane.

5.5 Formants in the Singing Voice: Soprano and Vowel Modification

An important additional complication arises for soprano singers, who sing at fundamental frequencies (\(F_0\)) that may exceed \(F_1\) for certain vowels. The vowel /a/ has \(F_1 \approx 800\) Hz; a soprano singing A5 (880 Hz) has \(F_0 > F_1\), meaning that the first harmonic (the fundamental) lies above the first formant. In this situation, the harmonic nearest to \(F_1\) is the fundamental itself, and the vocal tract amplification of the fundamental is determined by how close \(F_0\) is to \(F_1\).

Sopranos routinely modify their vowels — adjusting tongue height, jaw opening, and lip configuration — to keep a vocal tract formant aligned with the fundamental or with one of the lower harmonics, maximizing vocal power at high pitches even at the cost of vowel intelligibility. The phenomenon, known as formant tuning, results in the characteristic indistinctness of soprano vowels above about E5 (659 Hz): the perceptual identity of the vowel is sacrificed for loudness and ease of production at high pitch. This is an acoustic constraint that every composer writing high soprano lines must understand.

5.6 The Singer’s Formant and Vocal Acoustics

Classical singers — particularly tenors, baritones, and basses — develop a characteristic singer’s formant: a broad spectral peak centered near 2500–3500 Hz, arising from the clustering of the third, fourth, and sometimes fifth formants. By adjusting the larynx height (lowering it, which lengthens the pharynx) and the geometry of the laryngeal tube, skilled singers can bring \(F_3\), \(F_4\), and \(F_5\) close together in frequency, creating a single broad high-amplitude peak in the vocal transfer function \(H(f)\).

This boost of 10–15 dB in the 2.5–3 kHz range allows the operatic voice to project over a full orchestral accompaniment in a large opera house. The orchestra’s long-term average spectrum has a centroid well below 2 kHz; the 2.5–3 kHz region is relatively unoccupied. The singer’s formant exploits this acoustic window to make the voice audible and intelligible without electronic amplification — a remarkable physiological and acoustic adaptation developed through intensive training.

---

Chapter 6: Room Acoustics and Concert Hall Design

The experience of music in a room is shaped not only by the direct sound from the instrument but also by the many reflections that arrive at the listener’s ears afterward, and by the slow decay of sound energy as it is gradually absorbed by the room’s surfaces. Room acoustics is the discipline that studies this interaction, with direct consequences for the design of concert halls, recording studios, and rehearsal spaces.

6.1 Reverberation and Sabine’s Formula

When a sound source in a room suddenly stops, the sound energy decays gradually as it bounces repeatedly off walls, ceiling, and floor, losing energy at each reflection. The reverberation time \(T_{60}\) is defined as the time required for the sound energy density to fall by 60 dB (a factor of \(10^6\) in intensity) after the source ceases.

\[ T_{60} = \frac{0.161\, V}{-S\ln(1 - \bar{\alpha})}, \]

where \(\bar{\alpha} = A/S\) is the mean absorption coefficient averaged over the total surface area \(S\).

6.2 The Structure of Room Impulse Responses

\[ p(t) = (s * h)(t) = \int_{-\infty}^{\infty} s(\tau)\, h(t - \tau)\, d\tau, \]

the convolution of the source signal with the impulse response.

The impulse response has a characteristic structure in three temporal zones. The direct sound arrives at time \(t_0 = d/c\) (where \(d\) is the source-to-receiver distance), as a single sharp peak. The early reflections arrive in the interval roughly 0–80 ms after the direct sound, as a series of discrete echoes from nearby surfaces. The reverberant tail consists of exponentially decaying diffuse energy, extending for a time roughly equal to \(T_{60}\).

6.3 The Precedence Effect

The precedence effect has major implications for the acoustic design of concert halls. Early reflections from the side walls, ceiling, and balcony fronts, arriving within the 1–30 ms window, enhance the perceived loudness (adding 3–10 dB of loudness without disturbing localization), contribute to envelopment (the sensation of being surrounded by sound), and increase intimacy (the perceptual closeness to the source). In the great shoebox halls, the narrow width ensures that first-order side-wall reflections arrive within 20–25 ms of the direct sound, maximizing this beneficial early energy.

6.4 Acoustic Measurement Methods

\[ h(t) = (r * x^{\text{rev}})(t), \]

where \(r(t)\) is the recorded response and \(x^{\text{rev}}(t)\) is the time-reversed excitation signal. Swept-sine measurement has become dominant because of its immunity to nonlinear distortion artifacts that affect MLS measurements, and because the sweep rate can be adjusted to allocate more measurement energy at low frequencies where room noise is highest.

From the measured impulse response, acoustic parameters are computed: \(T_{60}\) (reverberation time) from the energy decay curve; C80 (clarity for music, the ratio in decibels of energy arriving in the first 80 ms to energy arriving after 80 ms); D50 (definition, the fraction of energy in the first 50 ms — used for speech); IACC (interaural cross-correlation, measuring spatial impression); and G (strength, measuring the total energy relative to the direct sound at a reference distance). These parameters form the objective acoustic profile of a hall and can be compared systematically across different venues.

6.5 Optimal Reverberation and Hall Geometry

Different musical uses require very different reverberation times, as determined by both the intelligibility requirements of the content and the temporal density of the musical texture.

For speech and theater, clarity (the ability to distinguish successive phonemes and words) is paramount. Reverberation smears successive phonemes together; the optimal \(T_{60}\) for a lecture hall or theater is 0.6–0.9 s. For chamber music, where harmonic progressions unfold rapidly and individual voices must remain distinct, \(T_{60} \approx 1.3\)–1.6 s is ideal. For Romantic orchestral music, with its large textural contrasts, powerful climaxes, and sustained harmonic blocks, the optimal \(T_{60}\) is 1.8–2.2 s. For pipe organ music and choral polyphony in large liturgical spaces, the continuous legato of vocal lines blends beautifully under \(T_{60} = 3\)–5 s, though intelligibility of text suffers severely.

Three hall geometries have dominated concert hall architecture. The shoebox (rectangular) hall — exemplified by the Musikverein Wien (1870) and Boston Symphony Hall (1900) — is consistently rated among the finest acoustically. Its narrow width (typically 20–24 m) ensures strong lateral reflections, and the high ceiling (typically 15–18 m) sends reflected energy back to the audience on a slower path, extending early energy without creating flutter echoes. The ornate plasterwork — caryatids, coffered ceiling panels, deep niches — acts as geometric diffusers, scattering sound energy in many directions and preventing harsh specular echoes.

The vineyard design, pioneered by Hans Scharoun in the Berlin Philharmonie (1963) and later adopted by Gehry’s Walt Disney Concert Hall (2003) and Rem Koolhaas’s Guangzhou Opera House, surrounds the orchestra with terraced audience platforms. This creates an immediate visual and psychological intimacy — no audience member is more than about 30 m from the platform — but the acoustics are more challenging to control. Lateral early reflections from side walls are less consistent than in the shoebox, and the complex surface geometry requires careful modeling to avoid acoustic dead zones and flutter echoes.

The fan-shaped hall (wide, with audience fanning out from the stage) was fashionable in the mid-20th century for its large seating capacity but has generally poor acoustics: lateral early reflections are weaker because the side walls are distant from most of the audience, and the wide, shallow room produces less spatial impression.

---

Chapter 7: The Auditory System and Pitch Perception

Having studied sound production and propagation in considerable detail, we now turn to the other end of the acoustic chain: the biological machinery by which the auditory system converts pressure variations into the rich perceptual world of pitch, timbre, loudness, and space. The ear is, in many respects, a more sophisticated signal processor than any device human engineering has yet produced — dynamically compressive, exquisitely frequency-selective, and capable of operating over more than 13 orders of magnitude in intensity.

7.1 Anatomy of the Auditory Periphery

The auditory periphery is conventionally divided into the outer ear, middle ear, and inner ear.

The outer ear consists of the pinna (the visible cartilaginous flap) and the external auditory meatus (ear canal), approximately 2.5 cm long and terminating at the tympanic membrane (eardrum). The pinna’s irregular, folded shape introduces direction-dependent and frequency-dependent filtering — the head-related transfer function (HRTF) — that the brain uses for vertical localization (elevation perception) and front-back discrimination. The ear canal acts as a quarter-wave resonator with resonant frequency near \(f = c/(4 \times 0.025) \approx 3{,}430\) Hz, boosting sound pressure at the eardrum by up to 10–15 dB in the 2–5 kHz range and contributing to the shape of the equal-loudness contours.

The middle ear contains three tiny ossicles — the malleus (hammer), incus (anvil), and stapes (stirrup) — which form a mechanical lever system. This system achieves an impedance match between the low-impedance air (acoustic impedance \(\rho c \approx 420\) Pa·s/m) and the high-impedance fluid of the cochlea (\(\rho c \approx 1.5 \times 10^6\) Pa·s/m). Without impedance matching, approximately 99.9% of incident acoustic energy would be reflected at the air-fluid interface — a 30 dB transmission loss. The middle ear recovers approximately 25–28 dB of this through two mechanisms: the area ratio between the tympanic membrane (\(\approx 55\) mm\(^2\)) and the stapes footplate (\(\approx 3.2\) mm\(^2\)) provides a pressure amplification by a factor of about 17, and the lever action of the malleus-incus combination provides an additional factor of about 1.3.

The inner ear (cochlea) is a fluid-filled, snail-shaped structure coiled approximately 2.75 turns in the temporal bone of the skull. The cochlea is divided along its length by the basilar membrane (BM) and Reissner’s membrane into three fluid-filled chambers: the scala vestibuli, scala media, and scala tympani. The basilar membrane spans the length of the cochlea (\(\approx 35\) mm) and is graded in mechanical properties: stiff and narrow near the base (oval window end) and flexible and wide near the apex.

7.2 The Basilar Membrane and Tonotopic Organization

When the stapes footplate moves in response to a sound, it creates a pressure wave in the cochlear fluid that sets the basilar membrane into traveling-wave motion. Georg von Békésy (Nobel Prize, 1961) directly observed these traveling waves in cadaver cochleae and found that each frequency produces maximum basilar membrane displacement at a characteristic location: high frequencies at the base, low frequencies at the apex.

The tonotopic organization of the basilar membrane is preserved throughout the auditory pathway: the auditory nerve fibers, cochlear nucleus, inferior colliculus, medial geniculate nucleus, and primary auditory cortex (A1) are all organized tonotopically, with neurons responding to different frequencies arranged in a systematic spatial order.

7.3 Hair Cells and Neural Transduction

Atop the basilar membrane, within the organ of Corti, lie two populations of sensory cells: approximately 3,500 inner hair cells (IHC) arranged in a single row, and approximately 12,000 outer hair cells (OHC) arranged in three rows. Each hair cell has a bundle of stereocilia — tiny hair-like projections — on its apical surface.

When the basilar membrane deflects, the stereocilia are bent due to the shearing motion between the membrane and the overlying tectorial membrane. Bending in the excitatory direction (toward the tallest stereocilia) stretches fine tip links connecting adjacent stereocilia, mechanically opening ion channels. Potassium ions (K\(^+\)) rush into the hair cell from the endolymph (which has an unusually high K\(^+\) concentration maintained by the stria vascularis), depolarizing the cell and triggering calcium-dependent vesicle fusion at the ribbon synapses below the cell. The inner hair cells release glutamate, which excites the fibers of the auditory nerve, generating action potentials that travel to the brainstem.

The outer hair cells, far more numerous, serve primarily as electromotile amplifiers. Their cell bodies contain prestin, a voltage-sensitive motor protein that contracts and expands the cell body in synchrony with the membrane potential, creating active mechanical forces that amplify the traveling wave on the basilar membrane — particularly for quiet sounds. This cochlear amplifier adds approximately 40–60 dB of gain for soft sounds, providing the ear’s enormous sensitivity and its remarkable frequency selectivity (sharp tuning curves). Hair cell damage, whether from noise exposure, ototoxic drugs, or aging, destroys this amplification irreversibly, resulting in sensorineural hearing loss.

7.4 Auditory Scene Analysis and Stream Segregation

In real musical situations, the auditory system faces the inverse problem: given a mixture of sounds from multiple sources arriving simultaneously at the two ears, reconstruct the individual source streams. Albert Bregman (1990) termed this capacity auditory scene analysis (ASA) and identified a set of organizational principles by which the auditory system groups acoustic features into coherent perceptual streams.

Grouping operates on two levels. Simultaneous grouping (within a single moment of time) uses cues such as common onset and offset (components starting and ending together are more likely to belong to the same source), harmonicity (partials fitting a common fundamental tend to fuse into a single percept), and common spatial location. Sequential grouping (across time) uses cues such as proximity in pitch (notes close in pitch tend to form the same melodic stream), similarity in timbre, and continuity of trajectory (a tone gliding smoothly upward groups with itself).

7.5 Place Theory and Temporal Theory of Pitch

Two complementary theories have historically described how the auditory system encodes pitch.

The two theories are not mutually exclusive; modern understanding holds that both mechanisms operate and cooperate across different frequency ranges. Phase locking in the auditory nerve is robust up to about 4–5 kHz in humans; above that frequency, the nerve can no longer follow the rapid oscillations, and place information dominates. Below about 1 kHz, temporal information is so precise that just noticeable differences in frequency can be as small as 0.1–0.2% (roughly 2–3 cents). Between 1 and 5 kHz, both mechanisms contribute. Above 5 kHz, pitch perception becomes increasingly vague and musical melodies lose their clear tonal quality.

7.5 The Just Noticeable Difference in Frequency

The just noticeable difference (JND) in frequency is the smallest change in frequency that a listener can reliably detect. For sinusoidal tones, the JND depends on frequency, duration, and intensity. At 1 kHz and 60 dB SPL with tones of duration 200 ms or longer, the JND is approximately 2–3 Hz (0.2–0.3%). At low frequencies (100–200 Hz), the JND in absolute hertz is smaller (about 1 Hz) but represents a larger percentage. The JND expressed in cents is roughly constant across the musical frequency range at about 3–6 cents for pure tones, and 1–3 cents for complex tones (which provide additional harmonic information for pitch encoding).

These values have direct implications for the perception of mistuning. A piano tuned 5 cents sharp in one octave would be detectable by a trained listener in a controlled test, though unnoticed in ensemble playing. The equal temperament major third, 14 cents sharp of just, is well above the JND and is distinctly audible as beating in a sustained pianissimo chord.

7.6 Virtual Pitch and the Missing Fundamental

One of the most compelling demonstrations in auditory science is the phenomenon of virtual pitch, or the missing fundamental.

The missing fundamental explains why a telephone (which transmits frequencies above roughly 300 Hz) reproduces male speech with recognizable voice pitch near 100–150 Hz: the harmonics above 300 Hz are sufficient for the auditory system to reconstruct the pitch. It explains why a small speaker with poor bass response still sounds musical. It also underlies several important musical illusions, including the bass pedal tones of pipe organs: many large organ pipes speak at frequencies below the lowest resonance of the room or the acoustic range of the listener’s speaker system, and the virtual pitch supplies the perceived bass.

7.7 Pitch Circularity: The Shepard Tone

The Shepard tone (R. N. Shepard, 1964) is an auditory illusion producing a pitch that appears to ascend (or descend) endlessly through successive semitones without ever actually becoming higher (or lower) in absolute frequency.

A Shepard tone is constructed as a superposition of sinusoids at frequencies that are exact octave multiples of one another (e.g., \(f, 2f, 4f, 8f, \ldots\)), with amplitudes shaped by a fixed bell-shaped spectral envelope that emphasizes middle frequencies and fades out the very high and very low components. As the pitch class ascends by one semitone (a factor of \(2^{1/12}\)), the entire pattern shifts: the highest component fades out and a new component fades in one octave below, so the overall spectral distribution remains identical. After 12 semitone steps the pattern has returned to exactly its original frequency distribution — yet the perceived pitch has “ascended” continuously.

The illusion exploits the ambiguity of octave equivalence in pitch perception. Because all components are octave-related, the auditory system cannot assign an absolute octave register — only a pitch class (the note name, e.g., “A” regardless of octave). The gradual brightening of the spectral envelope as the tone rises a semitone biases the pitch judgment toward “higher,” even though the total acoustic pattern is cyclic.

---

Chapter 8: Consonance, Dissonance, and Tuning Systems

The final chapter addresses a question that has occupied music theorists, mathematicians, and acousticians for over two millennia: why do some combinations of tones sound pleasant and stable (consonant) while others sound tense and unstable (dissonant)? And what tuning systems best serve the diverse requirements of different musical genres and instruments? The answers draw on everything we have developed about the harmonic series, critical bands, beats, and pitch perception.

8.1 Beats Between Pure Tones

Before analyzing the complex case of two musical tones with full harmonic series, we consider the simplest case: two pure sinusoids at nearby frequencies \(f_1\) and \(f_2\). Their superposition is

\[ \cos(2\pi f_1 t) + \cos(2\pi f_2 t) = 2\cos\!\left(\pi (f_1 - f_2) t\right)\cos\!\left(\pi (f_1 + f_2) t\right), \]

which is a rapid oscillation at the average frequency \(\bar{f} = (f_1 + f_2)/2\), amplitude-modulated at the beat frequency \(f_{\text{beat}} = |f_1 - f_2|\). When \(f_{\text{beat}}\) is less than about 10–15 Hz, the ear resolves the individual amplitude pulsations as slow, regular beats — the slight rhythmic waxing and waning that a piano tuner uses to check interval purity. As \(f_{\text{beat}}\) increases toward 15–40 Hz, the beats are too rapid to resolve individually but create a percept of roughness or dissonance — an unpleasant, harsh quality. Above about 50–60 Hz, the two tones begin to be heard as separate pitches, and roughness decreases again.

This pattern — maximum roughness at beat frequencies of 15–30 Hz — is the psychoacoustic basis of Helmholtz’s theory of dissonance.

8.2 Helmholtz’s Theory of Dissonance

Hermann von Helmholtz (1863) proposed that the consonance or dissonance of a musical interval is determined by the roughness arising from beating between nearly coincident partials of the two tones. A musical interval is consonant when the low-order partials of the two tones either coincide exactly (producing no beats) or are separated by more than the roughness range (producing no roughness). It is dissonant when multiple pairs of partials fall within the roughness range and produce simultaneous fast beats.

Helmholtz’s theory predicts the traditional hierarchy of consonances in Western music — octave, fifth, fourth, major third, minor third — in essentially the correct order, based on the frequency of partial coincidences and the magnitudes of beating between non-coincident partials. However, the theory is incomplete: it predicts that the octave (2:1) should be maximally consonant (all partials of the upper tone coincide with partials of the lower), which is correct, but it cannot fully explain why certain simultaneous intervals sound beautiful in a harmonic context even when some roughness is present, or why the same interval sounds different in different registers.

8.3 Tonal Consonance versus Cultural Consonance

Helmholtz’s roughness-based theory of consonance is a sensory theory: it predicts which intervals produce physical roughness from coinciding partials, independent of any musical context or enculturation. But musical consonance and dissonance are not purely sensory — they are also contextual and syntactic. A diminished seventh chord, which has relatively high sensory dissonance, sounds dramatically tense and unstable in tonal context; the same four pitches presented in isolation without context might still sound rough, but the urgency of their resolution comes from learned tonal expectations.

Huron (2006) distinguishes five components of what we loosely call “consonance”: roughness (Helmholtz’s sensory factor), sensory interference from spectral overlap, pitch commonality (how many partials are shared between simultaneous tones), expectation (how well the interval fits learned tonal patterns), and pleasantness. These five components partially but not perfectly correlate with one another and with listeners’ judgments of consonance. The cross-cultural evidence is mixed: some studies find that individuals from cultures without Western tonal music rate interval consonance in broadly similar rank orders to Western listeners (suggesting a sensory basis), while others find significant cross-cultural differences for certain intervals (suggesting cultural learning).

The resolution seems to be that roughness provides a sensory foundation for the consonance hierarchy — particularly for the extremes (unisons and octaves are universally preferred; minor seconds and tritones universally produce more discomfort) — while the intermediate intervals (thirds, sixths) show significant cross-cultural variability and are more strongly modulated by enculturation.

8.4 Plomp and Levelt: The Consonance Curve

Plomp and Levelt (1965) conducted systematic psychoacoustic experiments with pairs of pure tones and derived the empirical consonance curve for intervals between sinusoids.

\[ D(f_1, f_2) = \sum_{m=1}^{N}\sum_{n=1}^{N} d(m f_1,\, n f_2,\, A_{1m},\, A_{2n}), \]

where \(d(f_a, f_b, a, b)\) is the Plomp-Levelt dissonance for two pure tones at frequencies \(f_a\) and \(f_b\) with amplitudes \(a\) and \(b\). Plotting this total roughness as a function of the frequency ratio \(f_2/f_1\) for a sawtooth-like spectrum reveals minima (consonance maxima) at exactly the just intervals: 2:1 (octave), 3:2 (fifth), 4:3 (fourth), 5:4 (major third), 6:5 (minor third), 5:3 (major sixth), and so on — in close agreement with the traditional hierarchy of Western consonances. This was a major triumph of psychoacoustic theory: the consonance hierarchy emerges from first-principles auditory mechanics, without invoking cultural convention.

8.4 Pythagorean Tuning and Just Intonation

The history of Western tuning is a history of compromise between competing ideals. We trace the major systems chronologically.

\[ f_n = f_0 \cdot \left(\frac{3}{2}\right)^n, \quad n = 0, 1, 2, \ldots, 11, \]\[ \Pi = \left(\frac{3}{2}\right)^{12} \Big/ 2^7 = \frac{3^{12}}{2^{19}} = \frac{531441}{524288} \approx 1.01364 \approx 23.5 \text{ cents}. \]

The Pythagorean major third has frequency ratio \((3/2)^4 / 4 = 81/64 \approx 1.2656\), which is 21.5 cents sharp of the pure major third \(5/4 = 1.250\) (the syntonic comma \(\approx 21.5\) cents). Pythagorean tuning gives beautiful fifths and fourths but harshly impure major thirds — suitable for medieval organum and early polyphony that avoided thirds, but problematic for Renaissance music that emphasized the major third as the defining interval of the triad.

\[ 1 : \frac{9}{8} : \frac{5}{4} : \frac{4}{3} : \frac{3}{2} : \frac{5}{3} : \frac{15}{8} : 2. \]

The two whole tones are the major tone (9:8 \(\approx 203.9\) cents) and the minor tone (10:9 \(\approx 182.4\) cents), differing by the syntonic comma (21.5 cents). Just intonation gives maximally consonant triads in a single key but creates serious problems for modulation: the same pitch may need to be tuned differently depending on whether it functions as the major third of one chord or the fifth of another. Just intonation is most naturally suited to a cappella choral singing, where singers can continuously adjust pitch to minimize roughness.

8.5 Meantone Temperament and the Wolf Fifth

\[ 4 \times \left(\frac{3}{2} \cdot 2^{-1/(4 \times \ln(81/80)/\ln 2)}\right) \approx \frac{5}{4} \times 4 \quad \text{(over two octaves)}. \]

Meantone gives beautifully pure major thirds (the characteristic warmth of Renaissance and Baroque keyboard music) in the most common keys — C, G, D, F, B-flat — but produces a severely out-of-tune fifth in one position of the cycle of fifths: the “wolf fifth” between G-sharp and E-flat (or enharmonically equivalent notes), which may be 35–40 cents flat of pure. Playing in keys that require the wolf fifth is distinctly unpleasant, restricting music to a subset of keys.

8.6 Equal Temperament

12-tone equal temperament (12-TET) distributes the Pythagorean comma evenly among all twelve fifths, tempering each by one-twelfth of the comma (approximately 2 cents):

The equal-tempered fifth is \(2^{7/12} \approx 1.49831\), which is 1.955 cents narrow of the pure fifth 3:2 \(\approx 1.50000\) — a discrepancy audible only in a direct comparison with a pure reference but essentially imperceptible in normal musical contexts. The equal-tempered major third is \(2^{4/12} = 2^{1/3} \approx 1.25992\), which is 13.7 cents sharp of the pure major third 5:4 = 1.25000. This is a much more substantial impurity, clearly audible as approximately 6–7 beats per second in a sustained piano major third in the middle register, and more rapid (more conspicuous) in the treble.

The equal-tempered minor third (300 cents) is 15.6 cents flat of the pure minor third 6:5 (315.6 cents). The major sixth (900 cents) is 15.6 cents sharp of the pure major sixth 5:3 (884.4 cents). These systematic deviations give 12-TET its characteristic compromise sound: slightly tense thirds and sixths, with nearly perfect fifths and octaves.

8.7 Comparison of Tuning Systems

The practical choice of tuning depends on the musical genre, instrument, and compositional style. A cappella choral singing and string quartets naturally gravitate toward just intonation and can dynamically adjust pitch in real time. Keyboard instruments with fixed pitch must choose a static tuning; equal temperament has dominated keyboard practice since the mid-19th century. Wind orchestras rely on the players’ ability to adjust intonation continuously to minimize roughness in sustained chords. Barbershop quartet singing has developed a strong preference for just intonation: the “lock and ring” of a perfectly tuned seventh chord (with a 7:4 harmonic seventh) is the genre’s most celebrated sonic ideal.

8.8 Well Temperament and the Affective Theory of Keys

Between the extremes of pure just intonation and 12-TET lie a family of well temperaments popular in the 17th and 18th centuries, including Werckmeister III (1691), Kirnberger III (1779), and the many temperaments proposed by Neidhardt, Vallotti, and others. These systems make all 24 keys usable (unlike meantone) but preserve different degrees of purity in different keys: the most common keys (C, G, D, F) are close to meantone and have nearly pure major thirds, while the remote keys (C-sharp, F-sharp, B) are closer to Pythagorean and have slightly sharper, more tense thirds.

This variation in key color was widely perceived in the Baroque and early Classical periods as giving each key an affect or emotional character: C major pure and simple; D major brilliant and triumphant; B minor dark and melancholy; E-flat major noble and solemn. Johann Mattheson (1713) and other theorists catalogued these key affects in detail. Whether Bach’s Well-Tempered Clavier was composed for a specific well temperament (in which these key colors are acoustically grounded) or for equal temperament (in which they vanish) remains contested among music historians and tuning theorists, and is one of the most actively debated questions in the history of musical acoustics.

8.9 Inharmonic Timbres and Alternative Consonance

The consonance theory developed in this chapter assumes instruments with harmonic spectra. When instruments with inharmonic spectra — bells, metallophones, certain electronic sounds — are combined, the standard harmonic intervals no longer minimize roughness, and the traditional hierarchy of consonances breaks down.

The unity of the science of musical sound lies precisely in this web of relationships. The harmonic series, born from the physics of vibrating strings and air columns, shapes the consonance hierarchy through the psychoacoustics of critical bands and roughness. The architecture of concert halls is designed to deliver this sound, shaped by centuries of aesthetic refinement, to ears whose anatomy encodes the logarithm of frequency along the basilar membrane — matching, perhaps not coincidentally, the logarithmic structure of the musical scale. The cultural and the physical, the biological and the mathematical, are woven together in every moment of musical experience.

---

Supplementary Material: Selected Topics

S.0 The Bowed String: A Limit-Cycle Oscillator

The interaction between a bow and a string is one of the most complex oscillator dynamics in all of musical acoustics. Unlike the plucked or struck string — which receives an initial energy input and then decays freely — the bowed string is a self-sustained oscillator, maintained in periodic motion by the continuous transfer of energy from the moving bow.

The mechanism, described by Lord Rayleigh (1877) and modeled quantitatively by McIntyre, Schumacher, and Woodhouse (1983), is the stick-slip cycle. The bow hair, coated with rosin (a pine resin that has a high static friction coefficient and a lower kinetic friction coefficient), exerts a force on the string. In the sticking phase, static friction causes the string to travel with the bow, building up potential energy in the string’s displacement. When the tension in the string exceeds the maximum static friction force, the string slips, shooting back rapidly toward its equilibrium, and the kinetic friction during the slip phase is much lower, allowing this rapid motion. When the string returns to near equilibrium and its velocity matches the bow velocity again, sticking resumes — and the cycle repeats.

The remarkable result, predicted by the reflection function model and confirmed experimentally, is that the stick-slip cycle locks onto the fundamental period of the string, so that the string executes a nearly perfect sawtooth-wave displacement: linear rise during the stick phase, rapid fall during the slip phase. The fraction of the period spent in the slip phase (approximately equal to the bow-to-bridge distance divided by the string length, normalized by the period) is the beta ratio \(\beta = d_{b}/L\), where \(d_b\) is the bow-to-bridge distance. The resulting Helmholtz motion produces a spectrum with all harmonics present with amplitudes proportional to \(1/n\) (a perfect sawtooth), weighted by the amplitude spectrum of the corner at the stick-slip transition.

S.1 Nonlinear Acoustics and Combination Tones

The acoustic theory developed in the preceding chapters is largely linear: we assumed that pressure fluctuations are small enough that the wave equation can be derived from a linear approximation to the equation of state of the gas. In practice, all acoustic systems exhibit some degree of nonlinearity — and in certain musical contexts, this nonlinearity becomes perceptually important.

\[ f_{mn} = |m f_1 \pm n f_2|, \quad m, n = 0, 1, 2, \ldots \]

The most perceptually prominent is the cubic difference tone at \(2f_1 - f_2\) (for \(f_2 > f_1\)), and the quadratic difference tone at \(f_2 - f_1\). Tartini (1754) described the audibility of the difference tone \(f_2 - f_1\) in quiet sustained intervals played on the violin, and he used these “Tartini tones” as a check on the purity of just intervals: a pure fifth (3:2 ratio) produces a difference tone exactly one octave below the lower note (\(3f - 2f = f\)), while a mistuned fifth produces a difference tone slightly off from that position, revealing the detuning.

The cochlea itself is a nonlinear system (due to the active mechanics of the outer hair cells), and it generates combination tones internally. The distortion product otoacoustic emission (DPOAE) at \(2f_1 - f_2\) is measurable with a sensitive microphone in the ear canal even without any external stimulus at that frequency — it is generated within the cochlea and radiates back through the middle ear. DPOAEs are widely used clinically as a non-invasive measure of cochlear health, particularly for screening neonatal hearing.

For musical instruments, nonlinear wave propagation in the bore of a brass instrument at high playing levels causes a progressive steepening of the acoustic pressure waveform — a phenomenon analogous to the formation of shock waves in supersonic aerodynamics, though much milder. The steepening enriches the high-frequency content of the spectrum, contributing to the dramatic brightening of brass tone at fortissimo that is one of the most characteristic sounds of the symphony orchestra.

S.2 Loudness Summation and Masking

The decibel level of a complex sound is not a sufficient predictor of its perceived loudness, because the auditory system integrates loudness differently across frequency. The loudness of a complex sound is approximately the sum of the loudnesses of its components within each critical band, with critical bands acting as semi-independent loudness channels.

Simultaneous masking occurs when a loud sound (the masker) reduces the audibility of a softer sound (the target) presented at the same time. The masker spreads its masking effect broadly upward in frequency (upward spread of masking) and narrowly downward. The masking pattern in the frequency domain is called the excitation pattern, and its shape is governed by the frequency selectivity of the basilar membrane. A narrowband noise masker centered at 1 kHz masks tones much more effectively above 1 kHz than below, because the traveling wave on the basilar membrane has a gradual basal slope and a steep apical slope.

Temporal masking extends the masking effect into time: forward masking (post-stimulatory masking) occurs when the masker raises the threshold for a target presented after the masker has ended, for up to 100–200 ms. Backward masking (pre-stimulatory masking) can occur when the masker follows the target by up to about 20 ms — a counterintuitive result explained by the time required for neural processing to reach its decision threshold. Temporal masking plays an important role in the perception of the complex onsets and releases of musical notes, and in the psychoacoustic model underlying perceptual audio coding algorithms such as MP3 (MPEG-1 Audio Layer III) and AAC.

S.3 Spatial Hearing and the HRTF

The perception of the spatial direction and distance of a sound source is achieved through three primary cues: interaural time difference (ITD), interaural level difference (ILD), and spectral cues from the pinna (HRTF).

\[ \mathrm{ITD} \approx \frac{r}{c}\left(\theta + \sin\theta\right), \]

which reaches a maximum of about 650–700 \(\mu\)s for a source directly to one side (\(\theta = 90°\)). The auditory system is exquisitely sensitive to ITD: JNDs for ITD can be as small as 10–15 \(\mu\)s (corresponding to about 1° of azimuthal resolution near the median plane). Phase locking in the auditory nerve makes this precision possible for frequencies below about 1500 Hz; above that frequency, phase locking is insufficient for ITD extraction, and the system relies instead on ILD.

The ILD is the difference in sound pressure level between the two ears, caused by the acoustic shadow cast by the head. For frequencies below about 500 Hz, the head is much smaller than the wavelength and casts negligible shadow (\(\mathrm{ILD} \approx 0\)). For frequencies above 2–3 kHz, the head shadow can produce ILDs of 10–20 dB for lateral sources. The combination of ITD and ILD gives unambiguous lateral localization but is insufficient to distinguish front from back or to determine elevation — both of which require spectral cues from the HRTF.

In concert hall acoustics, the binaural reproduction of music via headphones (or two-channel stereo loudspeakers) fails to reproduce the natural HRTF of the listening room and the performer positions. The development of binaural room impulse response (BRIR) measurement and convolution-based auralization allows realistic simulations of acoustic spaces, enabling architects to “hear” a hall before it is built.

S.4 Musical Intervals and the Perception of Harmony

Beyond the pairwise consonance of dyads, harmony in tonal music involves the simultaneous sounding of three or more notes (chords) and the succession of chords over time (harmonic progressions). The psychoacoustics of chord perception extends the dyadic consonance theory in important ways.

For a major triad in root position (frequencies \(f\), \(5f/4\), \(3f/2\) in just intonation), the combined harmonic series of the three notes has a strong common fundamental at \(f\) and extensive partial coincidences among the upper harmonics. The chord is perceived as a unified sonority — the three tones fuse into a single perceptual object rather than being heard as three separate pitches — when the partials of the components reinforce a common virtual pitch. The degree of fusion depends on the complexity of the frequency ratios: the simpler the ratios (3:4:5 for a just major triad), the stronger the fusion.

The harmonic tension of a chord in tonal music — its sense of needing to resolve to a more stable chord — is partially explained by the roughness of its partial interactions and partially by learned syntactic expectations built up through musical enculturation. The dominant seventh chord (e.g., G–B–D–F in C major) has a higher roughness score than the tonic major triad (C–E–G) because of the tritone between B and F, whose partials interact in the peak roughness range. This increased roughness corresponds to the strong tendency of the dominant seventh to resolve to the tonic, a relationship that forms the cornerstone of tonal harmony.

S.5 Digital Audio and the Sampling Theorem

Modern music production, distribution, and analysis is almost entirely digital. The conversion of a continuous acoustic signal into a digital representation, and the faithful reconstruction of the original from its digital form, rests on one of the most important results in applied mathematics.

\[ \mathrm{SQNR} \approx 6.02B + 1.76 \text{ dB}, \]

giving \(6.02 \times 16 + 1.76 \approx 98\) dB for 16-bit audio — well within the dynamic range of human hearing for undithered signals.

Dithering — the addition of low-level noise before quantization — converts the deterministic quantization distortion (which creates audible harmonic artifacts in low-level signals) into random noise with a flat spectrum, trading a small increase in noise floor for a dramatic improvement in signal quality at low amplitudes. Properly dithered 16-bit audio is widely considered audibly transparent (indistinguishable from higher bit-depth formats) for most musical signals at typical listening levels.

S.6 Acoustic Impedance and Transmission Lines

The concept of acoustic impedance is the fundamental tool for analyzing the behavior of sound in bounded media — tubes, rooms, horns — and at interfaces between different media. It is the acoustic analogue of electrical impedance in circuit theory.

\[ Z_{\mathrm{in}} = Z_0 \frac{Z_L + i Z_0 \tan(kL)}{Z_0 + i Z_L \tan(kL)}, \]

where \(k = \omega/c\) is the wavenumber. This formula predicts the resonant frequencies as the poles of \(Z_{\mathrm{in}}\) (for an open end, \(Z_L \to 0\), giving \(Z_0 \tan(kL) = 0\), i.e., \(kL = n\pi\), or \(f_n = nc/2L\) — the open pipe result).

The impedance formalism is particularly powerful for analyzing the behavior of brass instrument bores, which consist of multiple sections of different bore profile joined end to end. The bore can be discretized into short sections, each modeled as a short transmission line element, and the total input impedance computed by cascading the element impedance matrices — a method implemented in modern brass acoustics simulation software (e.g., the BIAS and BORE programs used in research and instrument design).

S.6b Active Room Acoustics and Electroacoustic Enhancement

For certain architectural or functional constraints, passive acoustic design alone cannot achieve the desired acoustic character. A multipurpose hall that must serve both unamplified orchestral concerts (needing \(T_{60} \approx 2\) s) and amplified theatre or opera (needing \(T_{60} \approx 0.8\) s) cannot optimally satisfy both requirements with fixed passive surfaces. Similarly, a historic building may have architectural constraints that preclude adding the absorbing or diffusing surfaces needed for good speech intelligibility.

Electroacoustic enhancement (EAE) systems address these challenges by using arrays of microphones, digital signal processing, and loudspeakers to modify the perceived acoustic character of a space. The two main approaches are reverberation enhancement (adding synthetic reverberation to make a room “sound larger”) and direct-sound reinforcement (boosting the direct sound level without adding reverberation, as in conventional sound reinforcement).

The LARES (Lexicon Acoustic Reverberation Enhancement System) and SIAP (System for Improved Acoustic Performance) are early examples of reverberation enhancement systems. More sophisticated modern systems, such as VRAS (Variable Room Acoustics System) and CARMEN, use feedback between microphone arrays and loudspeaker arrays to create controllable gain in the late-reverberant field. The gain must remain below the Larsen threshold — the feedback gain at which the system becomes unstable and produces a howl — which limits the maximum achievable increase in \(T_{60}\). For practical installations, increases of 50–100% in \(T_{60}\) are achievable without audible artifacts.

A subtler application is acoustic glare reduction using active diffusion: loudspeakers positioned at specular reflection points play back delayed, decorrelated versions of the direct sound, creating the effect of a diffuse surface and reducing harsh specular reflections from flat parallel walls.

S.7 The Acoustics of Rooms at Low Frequencies

Sabine’s diffuse-field theory (Chapter 6) becomes inaccurate at low frequencies, where the room dimensions are comparable to or smaller than the acoustic wavelength. In this regime — typically below a few hundred hertz, depending on room size — the acoustic field is dominated by a small number of room modes, the standing-wave resonances of the enclosure.

\[ f_{mnp} = \frac{c}{2}\sqrt{\left(\frac{m}{L_x}\right)^2 + \left(\frac{n}{L_y}\right)^2 + \left(\frac{p}{L_z}\right)^2}, \quad m, n, p = 0, 1, 2, \ldots \]\[ f_S \approx 2000\sqrt{\frac{T_{60}}{V}} \text{ Hz}, \]

modes overlap so densely that the field becomes diffuse and Sabine’s formula applies. For a room with \(V = 200\) m\(^3\) and \(T_{60} = 0.5\) s, \(f_S \approx 2000\sqrt{0.5/200} \approx 100\) Hz.

Below \(f_S\), low-frequency acoustics in a room is a mode problem, not a diffuse-field problem. The audible consequence is bass coloration: certain bass frequencies are strongly reinforced (where the source is at an antinode of a room mode and the listener is also at an antinode) and others are strongly attenuated (where source or listener is at a node). This is why a home listening room sounds radically different at different positions — the bass response can vary by 20–30 dB across the room. Low-frequency room treatment (bass traps — thick, dense absorbing panels placed at pressure antinodes, typically room corners) is the most effective mitigation.

S.8 Psychoacoustics of Vibrato and Tremolo

Vibrato and tremolo are two of the most common expressive devices in music, often confused but acoustically distinct.

\[ f(t) = f_0 + \Delta f \sin(2\pi f_v t), \]\[ \phi(t) = 2\pi \int_0^t f(\tau)\, d\tau = 2\pi f_0 t + \frac{\Delta f}{f_v}\cos(2\pi f_v t), \]

so the modulation index is \(\beta = \Delta f / f_v\). For a violin vibrato with depth \(\Delta f = 20\) Hz (about \(\pm 13\) cents at A4) and rate \(f_v = 6\) Hz, the modulation index is \(\beta = 20/6 \approx 3.3\). Fourier analysis of an FM tone shows sidebands at \(f_0 \pm n f_v\) for all integers \(n\), with amplitudes given by Bessel functions \(J_n(\beta)\) — so a deep vibrato has significant energy spread over many sidebands, enriching the timbre markedly.

Perceptually, vibrato at rates of 5–8 Hz with depths of 25–50 cents is characteristic of classical string and vocal performance. It enriches the timbre (FM sidebands spread energy around each harmonic), helps the voice or instrument project over accompaniment (frequency smearing reduces cancellation by room modes), and contributes to the perceived warmth and expressivity of the tone. Interestingly, vibrato also serves as a perceptual glue in ensemble intonation: the constant frequency fluctuation makes small pitch deviations between instruments less salient, partially masking the beats that would otherwise reveal mistuning.

\[ s(t) = \bigl[1 + m\sin(2\pi f_t t)\bigr]\cos(2\pi f_0 t), \]\[ s(t) = \cos(2\pi f_0 t) + \frac{m}{2}\cos\bigl(2\pi(f_0 + f_t)t\bigr) + \frac{m}{2}\cos\bigl(2\pi(f_0 - f_t)t\bigr), \]

showing that amplitude modulation creates exactly two sidebands at \(f_0 \pm f_t\). Tremolo at rates of 4–8 Hz produces a pulsating, heartbeat-like effect used in organs (the tremulant mechanism), electric guitars, and certain orchestral passages. The perceptual distinction between vibrato and tremolo is important: vibrato in classical performance tends to be centered on the notated pitch (the time-averaged pitch is close to the target), while tremolo does not change the perceived pitch at all for slow modulation rates.

S.9 The Piano Action: Mechanics and Acoustics

The piano action — the mechanical system between the key and the string — is an engineering marvel of the 19th century and represents the most mechanically complex part of any common musical instrument. A modern grand piano action contains approximately 12,500 parts, and each of the 88 notes has an independent action assembly with roughly the same number of moving components as a fine mechanical watch.

The key function of the piano action is the escapement: a mechanism that allows the hammer to be thrown toward the string with the velocity imparted by the player’s finger, then to escape from the key mechanism and fall back freely after striking, so that the string can vibrate undamped while the key is still depressed. Without the escapement, the hammer would damp the string immediately after contact, killing the tone. The repetition mechanism (or double escapement, invented by Sébastien Érard in 1821) allows the hammer to return to an intermediate position after the first blow, enabling rapid re-striking of the same note without fully releasing the key — essential for fast repeated-note passages and trills.

The acoustic consequence of the hammer’s mechanical properties is significant. A new, unworn piano hammer has a felt covering of graded hardness — softer on the outside, harder on the inside (near the molton core). At low velocities (soft playing), the outer soft felt is compressed during contact, and the contact duration is long (roughly 4–8 ms). Long contact acts as a low-pass filter on the spectral content transmitted to the string: harmonics with half-wavelengths shorter than the contact duration are suppressed. At high velocities (loud playing), the harder inner felt contacts the string with a short duration (1–2 ms), and the high-frequency harmonics are transmitted strongly — producing the bright, sharp tone of a fortissimo piano attack. This velocity-dependent spectral balance is the fundamental mechanism of the piano’s dynamic expressivity, and it is precisely what distinguishes a well-regulated piano with properly voiced hammers from a poorly maintained one.

The total sound of a piano note is also shaped by the duplex scale — short lengths of string beyond the bridge that are not the primary speaking length but are connected sympathetically to it. These segments, tuned to harmonics of the note by adjusting their length, add a faint but characteristic halo of overtone resonance to the piano’s sound, particularly at soft dynamics. The duplex scale is present in most modern concert grand pianos and is considered by many builders to contribute to the instrument’s characteristic richness and sustain.

S.9b Perceptual Audio Coding and Musical Quality

The development of perceptual audio codecs — systems that compress digital audio by discarding information that the ear cannot perceive — is one of the most successful applications of psychoacoustic theory to engineering. The MP3 (MPEG-1 Audio Layer III) codec, introduced in 1993, achieves compression ratios of 10:1 or higher over uncompressed PCM audio by exploiting three psychoacoustic phenomena: simultaneous masking, temporal masking, and the critical band structure of auditory frequency analysis.

The encoding process divides the audio spectrum into subbands corresponding approximately to critical bands (32 subbands in Layer III, further refined by a modified discrete cosine transform). Within each subband, it computes the masking threshold — the minimum level of a signal that would be audible given the levels of other signals in adjacent bands and in adjacent time frames. Any signal component below its masking threshold is inaudible and can be quantized with fewer bits (or discarded entirely) without perceptible degradation. Bits are then dynamically allocated across subbands, concentrating quantization precision where it is most needed.

At bitrates of 128 kbps (kilobits per second), MP3 is often indistinguishable from uncompressed CD audio (\(\approx 1411\) kbps) in casual listening. In critical listening by trained subjects using high-quality playback systems, artifacts become audible — particularly a pre-echo on transient sounds (a brief burst of quantization noise that precedes a sharp attack, arising from the MDCT’s finite time-frequency uncertainty) and tonal noise under complex musical passages where the masking model is violated. At 320 kbps, MP3 is considered transparent (indistinguishable from CD) by the overwhelming majority of listeners.

The AAC (Advanced Audio Coding) codec, introduced in 1997 and used by iTunes and streaming services, improves on MP3 through more efficient entropy coding, improved joint stereo coding, and a more accurate perceptual model, achieving CD-transparent quality at bitrates around 128–192 kbps for most musical content.

S.10 The Physics of the Singing Bowls and Struck Idiophones

Tibetan singing bowls, crystal bowls, and the Chinese bianzhong (set of tuned bells) represent a distinct category of struck idiophones whose acoustic behavior illuminates several of the principles discussed throughout this course. The singing bowl, made of an alloy of five to seven metals, is struck with a padded mallet and sustained by rubbing the rim with a wooden or leather-covered stick. The friction between stick and rim excites the bowl into oscillation through a mechanism closely analogous to the bowed string: a stick-slip interaction that preferentially excites the fundamental mode, with the frequency determined by the bowl’s geometry and material.

\[ f_{20} \approx \frac{h}{2\pi a^2}\sqrt{\frac{E}{3\rho(1-\nu^2)}}, \]

where \(\nu\) is Poisson’s ratio of the bowl material. The Chladni pattern of a singing bowl struck and dusted with fine powder reveals four antinodal regions alternating with four nodal regions around the circumference, corresponding to the four-lobed deformation of the \((2, 0)\) mode.

The acoustic radiation from a singing bowl is directional: because the bowl radiates as a quadrupole (four antinodes alternately expanding and contracting), the sound field has four lobes of maximum intensity at 45° to the nodal lines and four nodes in the directions of the Chladni nodal lines. A listener slowly circling the bowl will hear four maxima and four minima of loudness per revolution. This directional pattern also means that the bowl is a very inefficient radiator: the quadrupole radiation falls off more rapidly with distance than a dipole or monopole, which is why singing bowls are quiet instruments best experienced in close proximity.

S.11 Bioacoustics, Hearing Damage, and Conservation of Musical Hearing

No account of the science of musical sound would be complete without attention to the fragility of the auditory system and the occupational hazards faced by musicians. The cochlear hair cells that transduce acoustic energy into neural signals are among the most metabolically active cells in the body and are exquisitely sensitive to damage from excessive noise exposure and ototoxic chemicals.

\[ \mathcal{D} = \int_0^T I(t)\, dt, \]

and the risk of damage is approximately constant for a given total energy dose, regardless of whether it is delivered rapidly (high intensity, short time) or slowly (lower intensity, long time).

The perceptual consequences of hearing loss extend beyond reduced sensitivity. Cochlear damage destroys not only the passive mechanics of the basilar membrane but also the active cochlear amplifier provided by the outer hair cells. The result is a broadening of auditory filters (reduced frequency selectivity), which impairs the ability to separate concurrent sounds in a complex auditory scene — the phenomenon known as the cocktail party problem. For musicians, the consequences of reduced frequency selectivity include difficulty in hearing individual lines within a polyphonic texture, impaired tuning sensitivity, and reduced ability to detect small intonation errors — the very skills most central to musical excellence. Preserving musical hearing is not merely a quality-of-life issue but a professional necessity, and the acoustical science of hearing protection deserves the same careful attention as the science of concert hall design or instrument acoustics.

S.12 Consonance in Spectral Music and Microtonal Composition

The 20th and 21st centuries have seen composers systematically interrogate the assumptions of the Western tonal tradition, including the harmonic series, equal temperament, and the consonance hierarchy. Spectral music, originating in Paris in the 1970s with composers such as Gérard Grisey and Tristan Murail, takes the harmonic series as its primary compositional material: chords are derived directly from the partial series of a chosen fundamental, and the microtonally precise frequency ratios of the partials are approximated using quarter-tones, eighth-tones, or other microtonal notation.

Grisey’s Partiels (1975), for large ensemble, begins with the sound of a low E on the trombone and then unfolds an orchestrated “slow-motion” version of its harmonic series — each partial assigned to a different instrument, with microtonal adjustments to bring all partials to their just positions rather than their equal-tempered approximations. The result is a chord of extraordinary resonance and fusion, because all partials of the chord are exact multiples of the fundamental: the entire ensemble sounds as if it is a single giant instrument producing its natural overtones.

Microtonal equal temperaments other than 12-TET provide alternative tuning environments with different patterns of consonance and dissonance. 19-TET (dividing the octave into 19 equal steps) approximates the pure major third (386.3 cents) very closely: the 19-TET major third is \(6 \times 1200/19 \approx 378.9\) cents, only 7.4 cents flat of just — much closer than 12-TET’s 13.7 cents sharp. 19-TET also has a near-pure minor third and reasonable approximations to the seventh harmonic (7:4), making it attractive for just-intonation-like harmonies in a closed, equal-tempered system. 31-TET approximates quarter-comma meantone almost exactly (each fifth is \(18 \times 1200/31 \approx 696.8\) cents, compared to 696.6 cents in quarter-comma meantone) and provides near-pure approximations to nearly all just intervals up to the 11th harmonic, at the cost of a more complex notation system.

The development of electronic and computer music has liberated composers from the physical constraints of fixed-pitch mechanical instruments. Arbitrary frequency ratios, including irrational ones, can be specified and realized with perfect accuracy. Composers such as Harry Partch (who built custom instruments in 43-tone just intonation), La Monte Young (who explored sustained just-intonation drones), and James Tenney (who used rational frequency ratios derived from the harmonic series in a rigorous compositional framework) have demonstrated that the tonal resources of just intonation far exceed those of any equal temperament, at the cost of requiring purpose-built instruments or electronic realization. The science of tuning systems is therefore not a closed chapter from the 18th century but a living field at the intersection of acoustics, mathematics, and compositional aesthetics.

S.13 Summary of Key Physical and Perceptual Parameters

The following table collects the key quantitative parameters that have appeared throughout these notes, for quick reference.

These numbers are not mere cataloguing; each one is the crystallized result of a physical derivation, an experimental measurement, or both, and each connects directly to the practical decisions made by instrument makers, concert hall architects, recording engineers, and performing musicians every day. The science of musical sound is, ultimately, a science of these connections — the web of relationships linking the restoring force in a vibrating string, the nonlinear mechanics of a bowed bow hair, the logarithmic mapping of the basilar membrane, the roughness of a mistuned third, the reflections from the plasterwork of a 19th-century Viennese concert hall, and the subjective experience of beauty in organized sound.

---