Acoustic Camera Buyer’s Guide: How to Read Specifications
A BMW engineer spent 3 days hunting a 400 Hz hum with accelerometers. An acoustic camera found it in 12 minutes. But only one that could see below 800 Hz.
The difference? Most portable acoustic cameras work well for single sources above 700-800 Hz—suitable for medium to high-frequency noise. But they miss critical low-frequency sources including HVAC systems (100-500 Hz dominant range), which require the same low-frequency capability as automotive NVH work. The critical frequencies for automotive NVH live below 800 Hz: the 100-500 Hz range where engine harmonics, road noise, and structural vibrations concentrate. Even mid-size stationary arrays start at 500 Hz, missing the frequencies that matter most. This is where most systems go blind.
The Challenge of Automotive NVH Testing
Noise, vibration, and harshness issues cost automotive manufacturers millions. A single NVH problem discovered late in development can delay production, trigger recalls, or damage brand reputation. Premium vehicle buyers expect near-silence at highway speeds. Electric vehicles have raised the bar even higher—without engine noise masking, every whir, hum, and rattle becomes audible.
Traditional NVH testing relies on accelerometers placed at dozens of points across a vehicle’s structure. These sensors excel at measuring vibration amplitude and frequency. But they can’t tell you where the noise originates. That requires hours of detective work: move sensors, retest, analyze, repeat.
Microphone arrays offered some improvement, overlaying acoustic data onto optical images. But traditional fixed arrays face their own limitations: they’re heavy (15+ kg), expensive (€20,000-€100,000+), and require specialist training. Most critically, they’re stationary—inconvenient to move between your wind tunnel, anechoic chamber, and test track.
The Low-Frequency Challenge in Automotive NVH
Why 100-500 Hz Matters for Automotive Applications
The most problematic automotive noise sources cluster in the low frequency range:
- Engine harmonics: 100-500 Hz (combustion orders, firing frequencies)
- Road noise: 200-800 Hz (tire-road interaction peaks at 300-600 Hz)
- Wind buffeting: 15-50 Hz (open window/sunroof resonance, peak ~20 Hz)
- Cabin wind noise: 50-400 Hz (wind-induced structural vibration)
- Electric vehicle motor noise: 40-2000 Hz (fundamental whine, speed dependent), with harmonics extending to 10+ kHz
- Suspension noise: 100-400 Hz (structure-borne sound from road inputs)
Customer complaints overwhelmingly target low frequencies, with structure-borne road noise (dominant below 600 Hz) being the primary source of dissatisfaction in premium vehicles. Yet most portable acoustic cameras can’t visualize anything below 800 Hz.
Why? The physics is unforgiving.
The Portability Paradox
Sound wavelength equals the speed of sound (343 m/s) divided by frequency. At 100 Hz, the wavelength is 3.4 meters. At 400 Hz, it’s 0.85 meters.
To reliably localize a sound source, your microphone array diameter must be at least as large as the wavelength you’re trying to capture. This creates an impossible trade-off:
Portable handheld cameras (12-35 cm diameter):
- ✅ Easy to carry between test sites
- ✅ Quick setup (minutes)
- ✅ Simple operation
- ❌ Lower frequency limit: 800-3000 Hz (misses critical automotive frequencies)
Mid-size stationary arrays (50-100 cm diameter):
- ✅ Better low-frequency reach (500-1500 Hz)
- ✅ Moderate microphone count (40-80 channels)
- ❌ Still misses most automotive NVH (below 500 Hz)
- ❌ Heavy (10-20 kg) and less portable
- ❌ Expensive (€20,000-€50,000)
Large stationary arrays (1.5-3 meter diameter):
- ✅ Can capture down to 100-250 Hz
- ✅ High microphone count (80-150 channels) for good resolution
- ❌ Very heavy (20-40 kg with mounting hardware)
- ❌ Very expensive (€40,000-€100,000+)
- ❌ Requires fixed installation or dedicated cart
- ❌ Can’t move between wind tunnel, test track, and anechoic chamber
Traditional “hybrid” systems try to split the difference by using modular arrays that add microphone elements when you need larger aperture. But this approach has its own problems: as diameter increases, so does microphone count (and weight, complexity, and cost). The expanded configuration becomes just as immobile as a large fixed array.
The gap is obvious: automotive NVH engineers need 100-500 Hz capability in a portable form factor. For years, that was physically impossible.
How Coherence Scanning Holography Solves This
Seven Bel’s Sound Scanner uses a fundamentally different approach: Coherence Scanning Holography (CSH). Instead of adding more fixed microphones to increase aperture, CSH uses five microphones mounted on a rotating arm.
How it works:
- Five MEMS microphones rotate on a 254 cm diameter arm
- The rotation creates 400+ virtual microphone positions along concentric circles
- Coherence-based holographic processing (not traditional beamforming) analyzes the data
- Cloud processing handles the computational load
The result:
- 125 Hz lower frequency limit (P254 module) in a portable system
- 250 Hz lower frequency limit (P132 module) for most automotive applications
- <5 kg total weight even at maximum 254 cm scanning diameter
- No additional microphones needed when scaling up measurement area
- One device works in your wind tunnel, test track, AND anechoic chamber
This solves what the acoustic camera industry calls “the impossible triangle”—low-frequency capability, portability, and affordability. Traditional systems force you to choose two of three. CSH delivers all three.
Good to know: The Sound Scanner requires audio data spatially acquired across a quarter revolution (pizza slice) to compute an acoustic image. At a typical rotational speed of 2 revolutions per second, this translates into a minimum transient duration of the sound event of 125ms, i.e. 1 divided by 8 quarters covered per second. So, the sources are not required to be quasi stationary during the measurement. On top, advanced algorithms are available for achieving sub millisecond time resolution for reproducible, highly transient sound events like door slams or clutch engagement clicks.
How Acoustic Imaging Changes NVH Analysis
Acoustic imaging makes sound visible. Using an array of microphones and an optical camera, the system overlays a color-coded acoustic “heat map” onto a real image of your test object. Hot colors (red, orange) indicate louder sound sources. Cool colors (green, blue) show quieter areas.
The advantage over point-by-point measurement is speed and completeness. Instead of guessing where to place sensors, you see the entire acoustic field at once. What took hours with accelerometers takes minutes with acoustic imaging.
Traditional method:
- Mount 50+ accelerometers on vehicle structure (2-3 hours)
- Run test protocol
- Analyze time/frequency data to identify problem frequencies
- Reposition sensors to triangulate source (another 1-2 hours)
- Repeat until source is localized
Acoustic imaging method:
- Position acoustic camera viewing the area of interest (5 minutes)
- Run test protocol
- See exactly where sound radiates from in real-time
- Done
This speed advantage becomes exponential in development cycles. Earlier problem detection means fewer prototype iterations and lower overall development costs.
Key Applications in Automotive NVH
Wind Noise Testing (200-1000 Hz)
Wind noise is a primary quality indicator for premium vehicles. It typically manifests in multiple frequency bands:
- Wind buffeting/booming: 15-50 Hz (open window/sunroof pressure oscillations, peak ~20 Hz)
- Low-frequency cabin wind noise: 50-400 Hz (wind-induced structural vibration)
- Wind whistle: 800-3000 Hz (seal leaks, mirror turbulence)
Most handheld acoustic cameras capture the whistle but miss the low-frequency cabin wind noise—exactly the frequency range customers complain about most. Even mid-size stationary systems starting at 500 Hz miss the critical 200-400 Hz range where wind-induced structural vibration causes interior noise.
Case example: An automotive OEM identified A-pillar turbulence generating 396-978 Hz noise in a premium sedan. Using a Sound Scanner P132 (250 Hz – 20 kHz), engineers pinpointed the source in 12 minutes of wind tunnel time. Traditional accelerometer-based testing would have required 2+ hours of tunnel time at €1,500/hour—a €2,850 time savings on a single test.
After modification to the A-pillar seal geometry, the peak noise level dropped 4 dB(A)—audible improvement to the human ear.
Powertrain Noise Localization (100-2000 Hz)
Internal combustion engines radiate noise across a wide spectrum, but the most problematic frequencies concentrate in the low range:
- Combustion harmonics: 100-500 Hz (firing frequency and orders)
- Timing chain/belt noise: 500-1500 Hz
- Accessory drives: 400-2000 Hz
For electric vehicles, the challenge shifts but doesn’t disappear:
- Electric motor fundamental: 40-2000 Hz (varies with RPM), with harmonics extending to 10+ kHz
- Inverter switching noise: 10-20 kHz (IGBT switching frequency and harmonics; SiC/GaN designs operate at 100+ kHz)
- Gearbox whine: 500-4000 Hz (gear meshing frequencies, with resonances often occurring in 600-800 Hz and 1300-1700 Hz bands)
The EV transition actually increases the need for low-frequency acoustic imaging. Without engine noise masking, every motor harmonic and gearbox resonance becomes audible. Premium EV buyers expect near-total silence—making the full 40-2000 Hz motor noise spectrum a critical design challenge.
Module recommendation: P254 (from 125 Hz) for complete powertrain coverage including diesel engine fundamentals below 200 Hz. P132 (from 250 Hz) sufficient for most gasoline and electric powertrains.
Road Noise & Structure-Borne Sound (100-500 Hz)
Tire-road interaction generates broadband noise with peak energy at 200-800 Hz. This noise radiates both through the air (airborne path) and through the suspension structure (structure-borne path).
Critical frequency bands:
- Tire pattern noise: 400-800 Hz (tread geometry excitation)
- Cavity resonance: 200-250 Hz (air chamber inside tire)
- Road texture interaction: 300-600 Hz (varies with pavement type)
Regulatory pass-by noise testing focuses heavily on the 200-1000 Hz range. EU regulations measure A-weighted sound levels with emphasis on frequencies where human hearing is most sensitive (500-4000 Hz).
Key insight: Customer road noise complaints overwhelmingly target frequencies below 600 Hz, where structure-borne transmission dominates. If your acoustic camera can’t see below 800 Hz, you’re testing the wrong frequency range.
Squeak & Rattle (BSR) Testing
BSR testing requires understanding both the excitation source and the resulting audible noise:
Excitation Frequencies (What Triggers BSR):
- Vibration input: 5-100 Hz (industry standard BSR testing range)
- Dashboard/IP excitation: 10-80 Hz (test input to trigger buzzes)
- Seat mechanism excitation: 50-80 Hz (typical trigger frequencies)
Audible BSR Noise: While the vibration that triggers BSR occurs at 5-100 Hz, the resulting audible noise (buzzing, squeaking, rattling) can manifest across a broader frequency range as components resonate and vibrate against each other.
Acoustic cameras can help identify which panels or components are radiating BSR noise once triggered. However, the fundamental BSR phenomenon is driven by low-frequency vibration excitation (5-100 Hz), which means addressing BSR requires understanding both the excitation source and the acoustic radiation pattern.
Acoustic Cameras vs. Traditional NVH Methods
Each method has its place:
Accelerometers remain the gold standard for:
- Quantifying vibration levels (exact amplitude measurements)
- Analyzing structure-borne transmission paths
- Modal analysis and resonance identification
- Compliance testing requiring vibration data
Acoustic cameras excel at:
- Fast source localization (WHERE is the noise?)
- Testing airborne noise radiation
- Identifying hidden sources (behind panels, in cavities)
- Visual communication with management/clients
The optimal workflow: Use acoustic imaging to quickly localize the source, then deploy accelerometers to quantify vibration paths and validate fixes.
Technical Comparison: Why Lower Frequency Matters
The array diameter vs. frequency relationship is not a guideline, it’s physics. Here’s how different array sizes limit frequency performance:
This isn’t just marketing, it’s wavelength mathematics. A 30 cm array is physically incapable of resolving 300 Hz sound (1.14 m wavelength) no matter how sophisticated the processing algorithm.
Remark: We do not consider near-field holography – an advanced algorithm for low frequency acoustic imaging which has not found its way into actual NVH applications due to its enormous required effort to patch multiple measurements at close distance to the object together and ensure comparable sound events between measurements.
The Virtual Microphone Advantage
Traditional beamforming arrays achieve better low-frequency performance by adding more microphones in a larger physical array. A 1.5 meter array might use 80-120 fixed microphones. A 3 meter array for 100 Hz work might need 150+ microphones.
More microphones create three problems:
- Cost: More channels mean more expensive data acquisition systems
- Weight: 120 microphones + mounting hardware + cabling = 15-25 kg
- Complexity: More failure points, more calibration required
Sound Scanner’s CSH approach achieves 400+ virtual microphone positions with just five physical microphones by scanning them through space. The rotating arm traces multiple concentric circles, creating an extremely dense sampling pattern without the weight, cost, or complexity of a massive fixed array.
Cost comparison:
- Traditional 1.5m hybrid array (expanded): €40,000+ with 80-120 microphones
- Sound Scanner P254 (2.54m diameter): €9,490 with 5 physical microphones (400+ virtual positions)
The weight difference is even more dramatic: <5 kg for Sound Scanner vs. 15-25 kg for an expanded hybrid array covering similar frequency range.
Choosing the Right Acoustic Camera for Automotive NVH
Critical Specifications Checklist
When evaluating acoustic cameras for automotive applications, verify:
✅ Frequency range: Must cover 125-10,000 Hz minimum for complete automotive work (100 Hz if diesel engines are priority)
✅ Lower frequency limit: Ask specifically—many vendors quote the system’s theoretical limit but perform poorly below 2x that frequency. Issues with degraded localisation accuracy arise with multiple sources in the same frequency range.
✅ Portability: Can you realistically move it between wind tunnel, anechoic chamber, and test track? Check actual weight including mounting hardware.
✅ Dynamic range: Minimum 13 dB separation needed for weak sources in noisy environments (motor noise at idle in wind tunnel background noise)
✅ Environmental suitability: Does it work in wind tunnels with airflow? Temperature range for test track use?
✅ Software integration: Can you export data to your existing analysis tools (NVH software, MATLAB)?
✅ Setup time: Measure real-world deployment time, not just “turn on” time. Hardware setup shall be finished in less than 3 minutes.
✅ Total cost of ownership: Include calibration, software licenses, training, and maintenance over 3-5 years
The Modular System Advantage
Sound Scanner’s modular approach means you can start with the frequency range you need most and expand later:
Path 1 – Gasoline/EV Automotive (Most Common):
- Start: P132 @ €7,490 (from 250 Hz covers 90% of automotive NVH work)
- Expand if needed: Add P254 sensor module @ €4,490 when ultra-low frequency work increases (total: €11,980)
Path 2 – Diesel/Heavy Vehicle Specialist:
- Start: P254 @ €9,490 (from 125 Hz required for diesel fundamentals <200 Hz)
- Optional: Add P132 for general testing if P254 is occupied
Path 3 – Complete Spectrum Coverage for measurements inside and around the vehicle:
- Full system: P132 + P254 + P12 = €16,470
- Coverage: 125 Hz – 44 kHz (includes ultrasonic leak detection with P12 module)
- vs. Competitors: Would need €100,000+ for similar range across acoustic and ultrasonic domains
Investment protection: Start with one module, expand only when needed. Each module operates independently, so you’re not locked into buying capabilities you don’t use.
Getting Started with Acoustic Imaging in Your NVH Lab
Pilot Project Recommendations
The best way to validate acoustic imaging is to test it against a known NVH problem:
Week 1 – Mid-frequency validation:
- Test on known sources in the 500-2000 Hz range (exhaust outlet, air intake, e-motor whine)
- Compare, correlate and interpret acoustic camera results with other measurements (accelerometers, single near-field microphones)
- Goal: Verify the technology and build confidence with the system
Week 2 – Low-frequency challenge:
- Tackle a 200-600 Hz noise source (road noise, wind rumble, diesel idle)
- This is where acoustic imaging truly differentiates from traditional handheld systems
- Document time savings vs. accelerometer-based localization
Week 3 – Integration workflow:
- Use acoustic camera for fast localization, then accelerometers for detailed analysis
- Establish your lab’s standard workflow combining both technologies
- Train additional team members
ROI Calculation Framework
Acoustic imaging ROI comes from three sources:
- Time savings per test session
- Traditional NVH localization: 2-4 hours (sensor placement + multiple test runs)
- Acoustic camera: 15-30 minutes (setup + measurement)
- Savings: 1.5-3.5 hours per investigation
- Value: (Wind tunnel rate OR engineer hourly cost) × hours saved
- Earlier problem detection
- Prototype testing finds NVH issues before hard tooling
- Cost to fix in prototype phase: €5,000-€20,000
- Cost to fix after production tooling: €100,000-€500,000+
- Value: Avoidance of even one late-stage discovery pays for the system
- Reduced prototype iterations
- Faster problem identification = faster fixes = fewer test cycles
- Each prototype cycle costs: vehicle build + testing time + analysis
- Eliminating one iteration: €50,000-€200,000 (depending on vehicle program)
Example calculation (mid-size automotive supplier):
- Equipment cost: €7,490 (P132 module)
- Weekly NVH tests: 5
- Time saved per test: 2 hours
- Annual time savings: 520 hours
- Value of engineer time: €75/hour (loaded cost)
- Annual savings: €39,000
- Payback period: 2.3 months
This doesn’t include the value of earlier problem detection or reduced prototype iterations—just the direct time savings.
Complementing (Not Replacing) Existing Methods
Acoustic cameras don’t replace accelerometers or other NVH tools. They complement them:
Use acoustic camera for:
- Fast source localization (“WHERE is the noise coming from?”)
- Airborne noise radiation mapping
- Hidden source identification (behind panels, in cavities)
- Visual communication with non-technical stakeholders
Use accelerometers for:
- Structure-borne vibration path analysis (“HOW does it get there?”)
- Precise amplitude quantification for specifications
- Modal analysis and resonance identification
- Compliance testing requiring vibration data
Use both in sequence:
- Acoustic camera quickly identifies the radiating surface and acoustic transport into the far field (driver’s ear)
- Accelerometers measure vibration paths leading to that surface
- Engineering team designs targeted treatment (damping, isolation, stiffening)
- Acoustic camera validates the fix
See Sound Scanner in Automotive Applications
Seven Bel’s Sound Scanner brings professional acoustic imaging within reach of every NVH engineer. The modular system starts at €7,490 per module—5x less than traditional acoustic cameras with equivalent low-frequency performance.
Key advantages for automotive NVH:
- 125 Hz lower limit (P254) captures diesel engines, road noise, wind rumble completely
- 250 Hz lower limit (P132) covers 90% of automotive NVH work
- Portable (<5 kg) for wind tunnel, test track, and anechoic chamber use
- Simple operation via tablet—no specialist training required
- Coherence Scanning Holography patent-protected technology
- Modular investment protection—start with one module, expand as needed
Next Steps
Request a demo: See Sound Scanner in action on your specific application. We’ll bring the system to your facility for hands-on testing.
Contact Seven Bel: info@sevenbel.com | www.sevenbel.com