Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
IEC 61043: Two-Microphone Sound Intensity Measurement
From Finite-Difference Pressure Gradient to In-Situ Noise Source Localization — A Complete Engineering Guide
Conventional sound pressure level (SPL) measurement with a single microphone tells you how loud a machine is, but says nothing about where the acoustic energy originates or where it flows. For decades, determining sound power accurately required anechoic or reverberation chambers — expensive, immobile, and wholly impractical for in-service industrial equipment. IEC 61043 changed this landscape by standardizing the two-microphone (p-p) sound intensity technique: a method that directly measures the acoustic energy flux vector using a pair of precision-matched pressure microphones. This vector quantity — sound intensity — opens the door to in-situ sound power determination under ISO 9614, noise source mapping, and advanced acoustic diagnostics that single-channel SPL meters simply cannot deliver.
The fundamental insight: Sound pressure is a scalar (magnitude only). Sound intensity is a vector (magnitude + direction). A sound intensity probe can therefore answer the question “which direction is the acoustic energy flowing?” — a capability that transforms how engineers approach noise control problems on the factory floor.
1. How the Two-Microphone Intensity Probe Works
1.1 The Physics: From Pressure to Particle Velocity
Sound intensity I is defined as the time-averaged product of instantaneous sound pressure p(t) and particle velocity u(t):
I = ⟨ p(t) · u(t) ⟩
Measuring p is trivial — any calibrated microphone can do it. The challenge lies entirely in determining particle velocity u without a dedicated velocity sensor. Under free-field plane-wave conditions, a simple relationship u = p/(ρc) holds, but real-world acoustic environments — with reflections, diffraction, near-field curvature, and flow noise — violate this assumption routinely.
The p-p method adopted by IEC 61043 circumvents this by invoking the linearized Euler equation of fluid dynamics:
ρ₀ · ∂u/∂t = −∇p
Integrating with respect to time and restricting to one spatial dimension (the probe axis r):
u = −(1/ρ₀) ∫ (∂p/∂r) dt
This is where the two-microphone configuration becomes essential: the spatial pressure gradient ∂p/∂r is approximated using the finite difference between two closely-spaced pressure microphones.
1.2 Finite-Difference Approximation: The Core Mathematics
Let two microphones be separated by a distance Δr, producing pressure signals p₁ and p₂. The pressure and its spatial gradient at the acoustic center (midpoint) are approximated as:
p ≈ (p₁ + p₂) / 2
∂p/∂r ≈ (p₂ − p₁) / Δr
∂p/∂r ≈ (p₂ − p₁) / Δr
Substituting into the Euler equation and transforming to the frequency domain via FFT yields the cross-spectral formulation — the computational heart of every modern sound intensity analyzer:
I(ω) = − (1 / ρ₀Δr) · Im[G₁₂(ω)] / ω
Here G₁₂(ω) is the cross-power spectrum between the two microphone signals, and Im[·] denotes its imaginary part. This equation reveals a profound fact: sound intensity is entirely determined by the imaginary part of the cross-spectrum — a quantity that captures the phase relationship between the two measurement channels.
Standard Probe Configurations per IEC 61043
| Configuration | Arrangement | Typical Use Case | Frequency Range |
|---|---|---|---|
| Face-to-Face | Two microphones facing each other, separated by a solid spacer | General-purpose noise surveys, sound power determination | 50 Hz ~ 10 kHz |
| Side-by-Side | Microphones mounted parallel, side by side | Near-field scanning, confined spaces | 100 Hz ~ 7 kHz |
Why face-to-face dominates: The solid spacer in the face-to-face configuration provides a precisely defined and mechanically stable separation distance Δr. This not only improves measurement repeatability but also reduces mutual scattering between the two microphone diaphragms. Standard IEC 61043 spacers are 12 mm (for 125 Hz to 10 kHz) and 50 mm (for 50 Hz to 1.25 kHz).
1.3 Frequency Bounds: Spacing Defines the Usable Bandwidth
The finite-difference approach imposes both low-frequency and high-frequency limits, dictated by the microphone spacing Δr:
| Spacer Δr | Low-Frequency Limit (Phase-Mismatch Dominated) |
High-Frequency Limit (Finite-Difference Error) |
Recommended Operating Range |
|---|---|---|---|
| 12 mm | ~125 Hz | ~10 kHz | 125 Hz ~ 10 kHz |
| 50 mm | ~50 Hz | ~1.25 kHz | 50 Hz ~ 1.25 kHz |
The high-frequency constraint follows from the condition k·Δr < 1, where k = 2π/λ is the wavenumber. In practice, keeping Δr < λ/6 limits the finite-difference bias error to within ±1 dB. At the low-frequency end, the limitation is not mathematical but instrumental — and it is the most dangerous error source in the entire technique.
The low-frequency trap: At long wavelengths, the true pressure difference between the two microphones is tiny. Even a 0.1° inter-channel phase mismatch can produce an apparent intensity reading comparable to — or larger than — the true value. This is why IEC 61043 devotes considerable attention to phase calibration requirements and the microphone-switching correction technique.
2. Error Sources and Calibration Strategy
2.1 Phase Mismatch: The Number-One Enemy of Intensity Measurement
The severity of phase mismatch can be understood mathematically. In the frequency domain:
I(ω) ∝ Im[G₁₂(ω)] = |G₁₂| · sin(φ₁₂)
At low frequencies, the true inter-microphone phase difference φtrue is extremely small (often fractions of a degree). The measured phase equals φtrue + φerr, where φerr represents the aggregate phase error of both measurement channels (microphones, preamplifiers, cables, ADC). Since sin(φ) ≈ φ for small angles:
Imeas / Itrue ≈ 1 + φerr / φtrue
When φerr becomes comparable to φtrue, the relative error can exceed 100%. IEC 61043 requires that inter-channel phase error remain below ±0.3° at 1 kHz for a 12 mm spacer — a demanding specification that ordinary lab-grade microphones cannot meet without special pairing.
2.2 Complete Error Taxonomy
| Error Type | Mechanism | Dominant Band | Mitigation |
|---|---|---|---|
| Phase Mismatch | Group-delay differences between the two transducer + signal-conditioning chains | Low frequencies (most severe) | Matched microphone pairs; microphone-switching calibration; phase correction transfer function |
| Finite-Difference Bias | Truncation error from approximating the derivative with a first-order difference | High frequencies | Appropriate spacer selection; high-frequency correction algorithms |
| Near-Field Curvature | Assumption of linear pressure variation between microphones breaks down near compact sources | All bands, near source | Maintain measurement distance > 2·Δr; higher-order difference schemes |
| Reactive Field Error | In standing-wave fields, the simple gradient-velocity relationship fails; intensity is underestimated | Low-frequency standing waves | Increase source-to-measurement-surface distance; identify and avoid strong reactive regions |
| Scattering / Diffraction | Probe body disturbs the very sound field it is intended to measure | High frequencies | Smaller probes; numerical diffraction correction factors |
2.3 The Microphone-Switching Technique
IEC 61043’s recommended calibration method is elegantly simple:
- Reference measurement: Measure cross-spectrum G₁₂(1) with Mic A on channel 1, Mic B on channel 2.
- Switched measurement: Physically swap the microphones (A to channel 2, B to channel 1) and re-measure G₁₂(2) at the same position.
- Geometric averaging: The corrected cross-spectrum is the geometric mean of the two measurements, effectively cancelling inter-channel amplitude and phase differences.
Real-world impact: This simple procedure reduces effective inter-channel phase error from the typical ±1~2° of unpaired microphones to better than ±0.1° — making the difference between unusable data and precision-grade intensity measurement. Modern digital analyzers automate the switching procedure and store phase-correction functions for real-time application.
3. Engineering Applications: From Sound Power to Source Mapping
3.1 In-Situ Sound Power Determination — Bypassing the Anechoic Chamber
The single most valuable industrial application of IEC 61043 probes is sound power determination under the ISO 9614 series. Three accuracy grades are defined:
| Standard | Method | Grade | Enveloping Surface | Best For |
|---|---|---|---|---|
| ISO 9614-1 | Discrete points | Precision (Grade 1) | Measurement surface subdivided into segments; at least one point per segment | Laboratory reference measurements, sound power arbitration |
| ISO 9614-2 | Continuous scanning | Engineering (Grade 2) | Probe swept at constant speed across the entire surface | Industrial shop-floor surveys, production-line QC testing |
| ISO 9614-3 | Discrete points (engineering) | Engineering (Grade 2) | Discrete grid with relaxed field-indicator criteria | Broadband high-frequency sources, geometrically complex surfaces |
The sound power is computed by integrating the normal intensity component over the enclosing measurement surface:
W = ∫ₕₓ₀ Iₛ dS ≈ Σ Iₛₓ · Sₓ
Why scanning wins in the factory: ISO 9614-2 scanning is the method of choice for production-line use. An operator sweeps the intensity probe at 0.1~0.5 m/s across the measurement surface, covering the entire envelope in 2~5 minutes per surface. This is an order of magnitude faster than discrete-point measurement and provides inherent spatial averaging. Handheld analyzers (e.g., B&K 2270, Siemens Simcenter Soundbrush) provide real-time scanning feedback — a green LED illuminates when scanning speed is within the acceptable range.
3.2 Noise Source Localization — The Vector Advantage
Because intensity is a vector quantity, a measurement grid yields far richer diagnostic information than an SPL map alone:
- Iso-intensity contour maps: Normal intensity component plotted on a 2D measurement grid. Positive contours indicate outward energy flow (radiation from the source); negative contours reveal inward flow (energy entering the source plane).
- Intensity vector maps: Full 3D vector field showing both magnitude and direction at each grid point. Dominant noise sources appear as high-magnitude regions with outward-pointing vectors.
- Negative-intensity zones: Areas where the net intensity points toward the test object. These are diagnostic gold — they may indicate panel absorption, acoustic short-circuit paths, or intrusive background noise from adjacent machinery.
Interpreting negative intensity: A negative normal-intensity reading on a measurement surface does not automatically invalidate the data. Three common causes: (1) the surface region is absorbing rather than radiating sound (e.g., an acoustic treatment panel); (2) extraneous background noise from nearby equipment is transmitting through the measurement surface; (3) near-field reactive energy oscillating back and forth across the surface. ISO 9614-2 introduces the F₂ field indicator to quantify this effect: the ratio of positive-only intensity to the sum of absolute intensity values must stay within specified limits for the target accuracy grade.
3.3 Field Indicators — Quantitative Data-Quality Criteria
IEC 61043 and ISO 9614 jointly define a set of field indicators that provide objective go/no-go criteria for measurement validity:
| Indicator | Definition | Criterion | What It Tells You |
|---|---|---|---|
| F₁ | Pressure-Intensity level difference: Lp − LI | 0 ~ 10 dB depending on grade | Indicates how far the measurement surface is from free-field conditions; smaller is better |
| F₂ | Positive-to-negative intensity ratio | Grade 1 ≤ 1.5, Grade 2 ≤ 3 | Quantifies the reactive character of the local sound field; high values warn of standing waves or external noise intrusion |
| F₃ | Directivity non-uniformity | Grade 1 ≤ 1.5, Grade 2 ≤ 6 | Measures the spatial uniformity of the intensity distribution across the measurement surface |
| F₄ | Temporal pressure-intensity stability | Application-dependent | Monitors whether the source’s acoustic output is stationary during the measurement period |
F₁ exceeds 10 dB? Before abandoning the measurement, investigate three possibilities: (1) is the source operating at a much lower level than expected? (2) is there strong background noise from an unaccounted source? (3) is the measurement surface too far from the source, capturing more reverberant energy than direct? Shrinking the measurement surface closer to the source often resolves elevated F₁. If the problem persists, consider hybrid pressure-intensity methods or switch to surface-velocity (accelerometer-based) techniques.
FAQ
Q1: Can I build a sound intensity probe by duct-taping two measurement microphones to a spacer?
A: In principle, yes — a p-p probe is just two microphones and a known separation. In practice, off-the-shelf microphones carry phase tolerances of ±2° or worse at 1 kHz, whereas IEC 61043 compliance demands inter-channel phase matching better than ±0.3° at low frequencies. Commercial intensity probes (e.g., GRAS 50AI, B&K 4197) use factory-matched microphone pairs with individually measured phase-correction functions stored onboard (TEDS). A home-built setup will almost certainly produce intensity readings dominated by phase-mismatch artifacts rather than genuine acoustic energy flux.
Q2: How does sound-intensity-based sound power compare in accuracy to the traditional anechoic/reverberation room method?
A: Under favorable conditions (low background noise, F₁ and F₂ within recommended limits), the intensity method achieves approximately ±1.5 dB expanded uncertainty at Grade 2 (engineering) and ±1 dB at Grade 1 (precision). Traditional precision anechoic-room methods achieve around ±0.5~1 dB. The intensity method trades a small accuracy margin for dramatic practical advantages: the equipment stays in place and operating, background noise from adjacent machinery is largely rejected (intensity vectors from extraneous sources tend to cancel over a closed surface), and no specialized acoustic facility is required. For the vast majority of industrial noise compliance tasks, Grade 2 accuracy is more than sufficient.
Q3: How sensitive is the intensity probe to airflow and wind?
A: Significantly more sensitive than a single SPL microphone. Turbulent airflow across the probe generates low-frequency pressure fluctuations (pseudo-sound) that the finite-difference processor interprets as genuine acoustic intensity, producing large spurious readings at low frequencies. At wind speeds above 2 m/s, a dedicated foam windscreen is mandatory, and the probe axis should be aligned parallel to the airflow direction whenever possible. IEC 61043 provides an informative annex with wind-induced error estimation methods. For high-wind environments (>10 m/s), consider nose-cone windscreens or alternative measurement techniques such as accelerometer-based surface intensity.
Q4: p-p (dual microphone) vs. p-u (Microflown) — which intensity probe technology should I choose?
A: Both have their place. The p-p method (IEC 61043) offers mature standardization, broad frequency coverage (50 Hz ~ 10 kHz with spacer swapping), and decades of proven field reliability with condenser microphone technology. The p-u method (Microflown sensor) provides a much smaller probe form factor ideal for near-field work and confined spaces, superior low-frequency performance (down to 20 Hz), and freedom from finite-difference bias error — but it is not yet covered by an IEC measurement standard and requires more careful handling (the MEMS-based velocity sensor is fragile). Practical recommendation: for ISO-compliant sound power determination, choose p-p. For near-field source imaging and very-low-frequency diagnostics, consider adding a p-u probe to your toolkit. Many advanced acoustic labs now maintain both.
