Sources of Sound

Sound generation requires a mechanism to create time-varying disturbances in a medium. Understanding the fundamental types of acoustic sources—their physical principles, radiation characteristics, and efficiency—is essential for analyzing voice production, where multiple source mechanisms operate simultaneously. This section explores the physics of sound generation from basic principles through specific applications in phonation and speech articulation.

Fundamental Principles of Sound Generation

All sound sources share the common requirement of creating time-varying pressure disturbances that propagate as acoustic waves.

Requirements for Sound Generation

For a mechanism to generate sound, it must satisfy specific conditions:

Time Variation: Static conditions produce no sound. The disturbance must change with time—oscillating, pulsating, or varying in some way.

Medium Coupling: The source must couple effectively to the medium (air for speech). Energy must be transferred from the source mechanism to air particle motion.

Sufficient Frequency: Frequency must lie within audible range (approximately 20-20,000 Hz for humans) or at least within relevant range for analysis (for voice, typically 80-8000 Hz).

Adequate Amplitude: Disturbances must be large enough to propagate as recognizable waves rather than being immediately dissipated by viscous losses.

Energy Conversion

Sound generation involves converting other forms of energy into acoustic energy:

Mechanical → Acoustic: Vibrating solids (vocal folds, tuning fork, loudspeaker cone) transfer motion to air.

Aerodynamic → Acoustic: Flowing air creates disturbances through turbulence, vortex shedding, or flow modulation.

Thermal → Acoustic: Rapid temperature changes cause local pressure variations (rare in voice production).

For voice, both mechanical (vocal fold oscillation) and aerodynamic (turbulent airflow in constrictions) mechanisms contribute.

Efficiency Considerations

Not all energy supplied to a source converts to radiated sound:

Radiation Efficiency: Fraction of input energy that becomes propagating sound waves.

Losses: Energy dissipated as heat through viscous friction, thermal conduction, and internal material damping.

Impedance Matching: Efficient radiation requires good acoustic impedance match between source and medium.

Small, low-frequency sources typically have poor radiation efficiency, requiring large displacements or pressures to generate significant sound.

Source Types and Classification

Acoustic sources are classified by their spatial distribution and radiation patterns.

Monopole Sources

A monopole source is a localized point that alternately injects and withdraws fluid from the medium, creating spherically symmetric pressure fields.

Physical Model: Pulsating sphere whose radius oscillates sinusoidally.

Volume Velocity: Rate of volume change with time:

U = dV/dt

where V is the volume of the sphere.

Characteristics:

• Spherically symmetric radiation (equal in all directions)
• Pressure decreases as 1/r with distance r from source
• Intensity decreases as 1/r² (inverse square law)
• Most efficient simple source type
• Radiated power proportional to U²

Radiation Pattern: Omnidirectional—same amplitude regardless of angle.

Examples:

• Small loudspeaker (piston in baffle) at low frequencies
• Oscillating balloon
• Open end of pipe (approximate monopole at low frequencies)

Voice Application: The open lips approximate a monopole source at low frequencies where wavelength greatly exceeds mouth opening dimensions. Glottal volume velocity source can be modeled as monopole for some analyses.

Figure 5.7: Schematic representations of monopole, dipole, and quadrupole sound sources showing characteristic radiation patterns.

Dipole Sources

A dipole source consists of two monopoles of equal strength but opposite phase located close together (separation << wavelength).

Physical Model: Oscillating rigid sphere in fluid (one side pushes fluid, opposite side pulls) or two pulsating spheres 180° out of phase.

Force Application: Dipole results from alternating force applied to medium.

Characteristics:

• Figure-eight radiation pattern (maximum along dipole axis, null perpendicular)
• Pressure varies with cos(θ) where θ is angle from dipole axis
• Less efficient than monopole (cancellation effects)
• Radiated power proportional to force² and frequency⁴
• Strongly frequency-dependent: much weaker at low frequencies

Radiation Pattern: Maximum radiation along axis connecting the two monopoles (dipole axis), zero radiation perpendicular to this axis.

Examples:

• Vibrating tuning fork
• Oscillating sphere (no net volume change)
• Unbaffled loudspeaker (front and back out of phase)

Voice Application: Vocal fold oscillation has dipole-like characteristics when considering force oscillations during collision. However, the volume velocity source perspective (monopole) is usually more convenient for voice analysis.

Quadrupole Sources

A quadrupole source consists of two dipoles of equal strength but opposite polarity located close together.

Physical Model: Four monopoles arranged in specific phase relationships, or lateral oscillation of fluid (shear flow).

Characteristics:

• Complex, directional radiation pattern
• Even less efficient than dipole
• Radiated power proportional to frequency⁶
• Extremely weak at low frequencies
• Important for jet noise and turbulence

Radiation Pattern: Multiple lobes and nulls creating complex directionality.

Examples:

• Turbulent jets (jet engine noise)
• Free shear layers
• Vortex interactions

Voice Application: Turbulent airflow in constrictions (fricatives, aspiration) generates quadrupole-type sound. This is why fricatives are relatively weak acoustically compared to vowels—quadrupole sources are inefficient, especially at low frequencies.

Volume Velocity as a Source

The concept of volume velocity provides a particularly useful description for voice acoustics.

Definition

Volume velocity (U): The volume of fluid displaced per unit time.

U = A × v

where:

• A = area through which fluid flows
• v = velocity of fluid through area

Units: m³/s or cm³/s

Physical Meaning: How rapidly volume is being added to (positive U) or removed from (negative U) a region of space.

Glottal Volume Velocity

During phonation, air flows through the time-varying glottal opening:

Flow Waveform: Typically characterized by:

• Rapid increase during glottal opening
• Maximum flow at peak glottal opening
• Rapid decrease as glottis closes
• Zero flow during closed phase (if closure is complete)

Typical Magnitudes:

• Peak glottal flow: 200-800 cm³/s (0.2-0.8 L/s)
• Average (DC) flow: 100-200 cm³/s during speech
• AC (oscillating) component creates sound

Source Strength: The time-varying (AC) component of glottal volume velocity acts as the primary acoustic source for voiced sounds.

Volume Velocity and Acoustic Pressure

For a monopole source with volume velocity U:

p = (ρω²U)/(4πr) × sin(ωt - kr)

where:

• ρ = air density
• ω = angular frequency (2πf)
• U = amplitude of volume velocity oscillation
• r = distance from source
• k = wave number (2π/λ)

Key Relationships:

• Pressure proportional to U (source strength)
• Pressure proportional to ω² (frequency squared)
• Pressure decreases as 1/r (inverse distance)

Implication: Higher frequency components of volume velocity waveform radiate more efficiently than low-frequency components (ω² factor).

Aeroacoustic Sources

Airflow without solid boundaries can generate sound through several mechanisms.

Turbulence Noise

Turbulent flow generates broadband noise through chaotic velocity fluctuations:

Mechanism: Random pressure fluctuations in turbulent eddies create quadrupole-type acoustic sources.

Characteristics:

• Broadband (continuous) spectrum
• Radiated power proportional to flow velocity to high power (V⁶ to V⁸)
• Relatively inefficient (quadrupole radiation)
• Frequency content increases with velocity

Voice Application:

• Aspiration noise from turbulent glottal flow
• Fricative consonants (/s/, /f/, /θ/, etc.)
• Breathy voice quality
• Stop burst release

Spectral Properties: Typically show gradual spectral roll-off without discrete harmonics. Fricative spectra reflect both turbulence generation characteristics and cavity resonances (filtering).

Flow-Obstacle Interaction

When airflow encounters an obstacle or edge, vortex shedding and flow separation create sound:

Vortex Shedding: Alternating vortices shed from obstacle edges create periodic pressure fluctuations.

Frequency: Vortex shedding frequency depends on flow velocity (V) and obstacle dimension (D):

f ≈ SV/D

where S is Strouhal number (typically 0.2-0.3).

Edge Tones: Flow impinging on sharp edge creates self-sustained oscillation through feedback between acoustic waves and flow instability.

Voice Application:

• Potential contribution to voiceless sounds
• Interaction of supraglottal flow with vocal tract structures
• Possible role in whistling and some voice qualities

Jet Noise

High-velocity air jets generate noise through turbulent mixing:

Mechanism: Shear layer between jet and surrounding fluid becomes turbulent, creating quadrupole sources.

Characteristics:

• Very inefficient radiation (∝ V⁸ for subsonic jets)
• Directional (maximum downstream)
• Broadband spectrum

Voice Application: Relevant for understanding high-velocity airflow sounds (forceful fricatives, glottal attack transients).

Sound Generation in Phonation

The laryngeal source involves both primary and secondary acoustic generation mechanisms.

Primary Glottal Source

Mechanism: Modulated airflow through oscillating vocal folds.

Source Type: Primarily volume velocity (monopole-like) source at glottis.

Waveform: Quasi-periodic glottal flow pulses with characteristic shape:

• Relatively slow opening
• Rapid closure
• Closed or nearly closed phase

Spectrum: Harmonic structure with fundamental frequency F₀ and overtones at integer multiples. Spectral slope typically -12 dB/octave (amplitude decreasing as 1/frequency²).

Efficiency: Relatively efficient due to monopole-like radiation and confinement by vocal tract (acoustic coupling).

Secondary Laryngeal Sources

Aspiration Noise: Turbulent flow through partially open glottis.

• Broadband, continuous spectrum
• Increases with glottal gap area and flow velocity
• Contributes to breathy voice quality

Vocal Fold Collision: Impact forces during medial contact.

• May contribute high-frequency components
• Relevance debated—effects may be minor
• Potentially important for some voice qualities (pressed, creaky)

Supraglottal Flow Separation: Flow separation from ventricular folds or epiglottis.

• May generate additional noise
• Typically minor compared to glottal source

Sound Generation in Speech Articulation

Different consonant types employ distinct acoustic generation mechanisms.

Fricatives

Mechanism: Turbulent airflow in narrow constriction.

Source Type: Quadrupole (aeroacoustic turbulence).

Location: Constriction site determines fricative type:

• Labiodental /f, v/: lips-teeth
• Dental /θ, ð/: tongue-teeth
• Alveolar /s, z/: tongue-alveolar ridge
• Palatal /ʃ, ʒ/: tongue-palate
• Glottal /h/: glottis

Characteristics:

• Broadband, noise-like spectrum
• Relatively weak intensity (inefficient quadrupole source)
• Spectrum shaped by cavity resonances anterior to constriction
• Voiceless fricatives: turbulence source only
• Voiced fricatives: turbulence plus glottal source

Spectral Properties: Center frequency and spectral shape depend on constriction location and geometry. Front fricatives (/s/) have higher spectral peaks than back fricatives (/ʃ/).

Stops (Plosives)

Mechanism: Sudden pressure release following complete closure.

Source Type: Transient monopole (burst) plus possible aspiration.

Phases:

1. Closure: Complete constriction, pressure buildup
2. Burst: Rapid release creating transient
3. Aspiration: Brief turbulent flow before voicing onset (in voiceless stops)

Characteristics:

• Brief, transient excitation
• Broadband burst spectrum
• Weaker than fricatives (shorter duration)
• Spectrum shaped by cavity anterior to constriction

Place of Articulation: Burst spectrum and transitions in adjacent vowels cue stop identity (/p/ vs. /t/ vs. /k/).

Nasals

Mechanism: Glottal source with nasal cavity coupling.

Source Type: Volume velocity (glottal) radiating through nasal passages.

Characteristics:

• Glottal harmonic source (like vowels)
• Nasal cavity resonances (nasal formants)
• Anti-resonances (zeros) from coupling of nasal and oral cavities
• Generally weaker than vowels (smaller nasal radiation opening, losses in nasal cavity)

Spectrum: Harmonics shaped by nasal formant pattern, with characteristic anti-formants (spectral valleys) due to side-branch resonator effect of closed oral cavity.

Approximants and Glides

Mechanism: Glottal source with relatively unconstricted vocal tract.

Source Type: Volume velocity (glottal).

Characteristics:

• Similar to vowels but with more constriction
• Formant transitions convey phonetic information
• Acoustic energy intermediate between vowels and fricatives

Examples: /w/, /j/, /r/, /l/

Source-Filter Theory

Understanding sound sources is essential to source-filter theory of speech production.

Source-Filter Independence

Source-Filter Theory (Fant, 1960) posits that:

Source: Generates acoustic excitation (glottal volume velocity for vowels).

Filter: Vocal tract modifies source spectrum through resonances.

Independence: Source and filter operate independently (approximately).

Output: Acoustic output = source spectrum × filter transfer function.

This separation enables independent analysis of source properties (vocal fold function) and filter properties (articulation).

Implications for Analysis

Source Analysis: Inverse filtering removes vocal tract effects to isolate glottal source waveform, revealing:

• Glottal flow waveform shape
• Open quotient
• Closing velocity
• Source spectral characteristics

Filter Analysis: Linear predictive coding (LPC) or other methods extract vocal tract transfer function, revealing:

• Formant frequencies
• Formant bandwidths
• Vocal tract shape information

Separating source and filter is essential for both theoretical understanding and clinical assessment.

Limitations

Source-filter independence is approximate:

Source-Tract Interaction: Vocal tract acoustics can influence vocal fold oscillation through:

• Intraglottal pressure affected by supraglottal impedance
• Acoustic coupling modifying oscillation threshold and frequency
• Nonlinear effects at high amplitudes

Nonlinear Phenomena: Strong source-filter coupling in special cases (throat singing, Tibetan chant, babies crying at high intensity).

Despite limitations, source-filter theory provides excellent approximation for most speech and singing conditions.

Summary

Sound sources generate acoustic waves through time-varying disturbances in a medium, converting mechanical or aerodynamic energy into acoustic energy. Sources are classified by radiation characteristics: monopole sources (pulsating volume) radiate omnidirectionally with high efficiency, dipole sources (oscillating force) radiate in figure-eight patterns with moderate efficiency, and quadrupole sources (turbulent flow) radiate complex patterns with low efficiency. Radiated power scales as frequency² for monopoles, frequency⁴ for dipoles, and frequency⁶ for quadrupoles, making high-frequency sources generally more efficient.

In voice production, the glottal source acts primarily as a volume velocity (monopole-like) source, creating periodic flow pulses that generate harmonic spectra with characteristic spectral slope. Secondary sources include aspiration noise from turbulent glottal flow and potential contributions from vocal fold collision. Speech articulation employs multiple source types: fricatives use aeroacoustic turbulence (quadrupole sources) creating broadband noise spectra, stops use transient burst sources from pressure release, nasals use glottal sources with nasal cavity coupling, and approximants use glottal sources with moderate vocal tract constriction.

Source-filter theory separates acoustic generation (source) from spectral shaping (filter), enabling independent analysis of phonatory function and articulatory configuration. While approximate (source-tract interactions exist), this framework successfully explains most speech and voice phenomena. Understanding source mechanisms is fundamental to analyzing normal and disordered voice production, interpreting acoustic measurements, and developing voice synthesis systems.

Key Takeaways

• ✅ Sound sources create time-varying pressure disturbances through mechanical or aerodynamic energy conversion
• ✅ Monopole sources (volume velocity) radiate omnidirectionally with highest efficiency, while dipoles and quadrupoles are progressively less efficient and more directional
• ✅ Radiated power scales as frequency² (monopole), frequency⁴ (dipole), and frequency⁶ (quadrupole), favoring high frequencies
• ✅ The glottal source acts as volume velocity (monopole-like) source generating periodic flow pulses with harmonic spectra
• ✅ Fricatives use aeroacoustic turbulence (quadrupole) creating weak, broadband noise; stops use transient bursts; nasals use glottal source with nasal coupling
• ✅ Source-filter theory separates acoustic generation from spectral shaping, enabling independent analysis of phonation and articulation
• ✅ Understanding source types explains relative intensity differences between vowels (efficient monopole) and fricatives (inefficient quadrupole)

Related Topics

• The Glottal Source Function
• Secondary Sound Sources in the Larynx
• Time-Varying Glottal Pressure and Flow
• Frequency Spectra & Spectral Slope
• Vocal Tract Resonance

Further Reading

1. Titze, I. R. (2000). Principles of Voice Production (2nd ed.). Iowa City: National Center for Voice and Speech.
2. Fant, G. (1960). Acoustic Theory of Speech Production. The Hague: Mouton.
3. Stevens, K. N. (1998). Acoustic Phonetics. Cambridge, MA: MIT Press.
4. Lighthill, M. J. (1952). On sound generated aerodynamically. I. General theory. Proceedings of the Royal Society of London. Series A, 211(1107), 564-587.
5. Flanagan, J. L. (1972). Speech Analysis, Synthesis and Perception (2nd ed.). New York: Springer-Verlag.