Automotive NVH/BSR Holographic Sound Field Inspection

A Dual-Core Acoustic Solution Based on Spherical Panoramic Imaging
and the NAH1024 Ultra-Large Planar Array

I. Solution Positioning: An Acoustic Engineering Leap from “Phenomenon Perception” to “Mechanism Insight”

NVH (Noise, Vibration, Harshness) and BSR (Buzz, Squeak, Rattle) are among the most difficult acoustic problems to quantify and trace in automotive R&D and quality engineering. The essential differences between the two dictate that the solution path must divide and address them separately:

Conventional acoustic testing methods have long vacillated between two predicaments, constraining R&D efficiency:

• Macroscopic perception challenge—the “fog” of the whole-vehicle environment: amid the multiple interferences of extensive space, strong background noise, and dynamic operating conditions, it is difficult to lock onto the source of whole-vehicle abnormal noise rapidly, as if with a “clairvoyant eye.”

• Microscopic insight challenge—the “blind zone” of core components: when focusing on the low-frequency domain of core components, conventional methods deliver blurred localization and cannot penetrate the fault precisely to the specific excitation point.

This solution addresses the dual predicament with a “dual-core” architecture: the spherical panoramic acoustic imaging camera breaks through the macroscopic blind zone, while the NAH1024 ultra-large planar array breaks through the microscopic diffraction limit, ultimately building an intelligent NVH/BSR R&D closed loop covering “whole-vehicle-level localization → component-level quantification.”

II. Core One: Holographic Sound Field Inspection Instrument (Spherical Panoramic) — Macroscopic Full-Domain Perception

Addressing the problems of whole-vehicle and cabin-level abnormal noise that is “hard to see, hard to catch, and inefficient to find,” the holographic sound field inspection instrument uses a spherical microphone array as its physical foundation to achieve native panoramic coverage of the entire space and AI-based strong-noise decoupling.

2.1 Panoramic Acoustic Imaging: H360°×V180°, Blind-Zone-Free Coverage

2.2 AI Blind Source Separation: 12 dB Equivalent Noise Reduction, Bringing Weak Abnormal Noises “to the Surface” in Strong Backgrounds

In road-test scenarios, the signal-to-noise ratio is extremely poor—wind noise and tire noise reach as high as 80–90 dB, whereas BSR abnormal noise is only 40–50 dB—making it nearly impossible to capture with conventional time-domain analysis.

Full-chain AI processing pipeline:

• High-dynamic acquisition and freeze-frame—combining a high-dynamic system to improve the signal-to-noise ratio and clearly lock onto transient abnormal noise in the time–frequency plane;

• AI-automated abnormal-noise diagnosis—automatic feature extraction and AI recognition of separated signals to intelligently classify abnormal noise types such as friction, knocking, and looseness;

• Acoustic decoupling—precisely separating wind noise and tire noise from the road-test background, transforming “chaos” into “clarity.”

2.3 Holographic Sound Field Inspection Instrument · Main Technical Parameters

Notes: (1) The sensitivity figure is quoted as −24 dB; its reference basis (e.g., re 1 V/Pa or dBFS) is to be specified. (2) The 80 dB SNR figure describes the microphone array; whether it is a per-channel, array-level, or system-level value is to be specified. (3) The 12 dB figure is quoted in Section 2.2 as “equivalent noise reduction” and in this table as “source separation signal-to-interference ratio”; the two designations refer to the same figure as stated in the source document, and their relationship is to be confirmed.

III. Core Two: NAH1024 Ultra-Large Planar Array — Microscopic Holographic Insight

Addressing the difficulty that low-frequency noise from core components is “large and blurry, and cannot be measured,” the NAH1024 uses Near-field Acoustic Holography (NAH) as its core algorithm, breaking through the physical diffraction limit of conventional beamforming to achieve centimeter-level quantitative localization at sub-wavelength resolution.

3.1 Near-field Acoustic Holography: Capturing Evanescent Waves, Breaking Through the Conventional Diffraction Limit

Technical principle: using a dense array to measure in the near field of the sound source, capturing not only propagating waves but also the evanescent waves that carry fine vibration information, thereby breaking through the resolution limits of conventional far-field measurement.

Core workflow closed loop:

• Near-field measurement—acquiring the sound pressure distribution containing evanescent waves;

• Inverse computation—reconstructing the complete sound field through a mathematical transform;

• Precise presentation—outputting a three-dimensional contour map with centimeter-level localization accuracy.

3.2 1024-Channel Ultra-Large Planar Array: One-Touch Snapshot, Three-Dimensional Sound Field Reconstruction

Dual reconstruction capabilities:

• Millisecond-level synchronous acquisition—shortening hours of scanning to a few seconds, fully capturing transient sound field changes, and breaking through the efficiency bottleneck of conventional measurement;

• Elimination of localization error—acquiring all array data synchronously in a single pass, eliminating the spatial localization deviation produced by the scanning movement process.

Based on accurate three-dimensional sound field reconstruction with the NAH algorithm, the system can output sound pressure contour maps (full-field distribution), particle velocity (vector analysis), and sound intensity contour maps (energy flow direction), visualizing noise “hotspots” and energy flow at high resolution to directly guide product structural optimization design.

3.3 NAH1024 Ultra-Large Array · Main Technical Parameters

Notes: (1) The sensitivity figure is quoted as −24 dB; its reference basis is to be specified. (2) The 65 dB SNR figure is applicable to the microphone array; whether it is a per-channel, array-level, or system-level value is to be specified. (3) The frequency range above is the microphone response range; the effective NAH spatial reconstruction band (bounded by the array’s spatial sampling limit) is stated separately in the companion document on humanoid robot joint inspection.

IV. Combining Both Strengths: Building an Intelligent NVH/BSR R&D Closed Loop

V. Typical Engineering Application Cases

5.1 NVH Testing of a Luxury Brand’s Full-Size SUV

Centered on the spherical panoramic acoustic imaging camera, covering multi-condition acoustic diagnostics of the complete vehicle:

• Broadband dynamic excitation radiation at the tire–ground interface—real-time acquisition by the panoramic sound field terminal;

• Localization of noise leakage points with doors and windows closed—using a white-noise signal as excitation to locate seal leakage paths;

• In-cabin audio equalization testing—evaluating cabin sound field balance using reflected-source loudspeakers;

• Noise reduction performance testing—quantifying the vehicle’s overall noise insulation and reduction effectiveness.

5.2 NVH Testing of a Domestic-Brand Battery Electric Vehicle (Tire Noise Special Study)

Targeting the tire noise that becomes prominent once the powertrain of a battery electric vehicle is made quiet, a panoramic acoustic imaging camera was deployed to conduct special tire noise testing at multiple speeds and on multiple road surfaces, covering 20–80 km/h and operating conditions on gravel, asphalt, cement, and other surfaces, systematically characterizing tire noise radiation characteristics and propagation paths.

5.3 NVH Testing of a Luxury Brand

• Noise insulation performance testing—ball array deployment to identify window seal gaps with doors and windows closed;

• Engine condition monitoring—acoustic condition monitoring of key powertrain components such as the starter-generator, high-pressure fuel pump, and turbocharger;

• In-cabin audio equalization testing—cabin sound field evaluation covering midrange / tweeter / mid-bass / subwoofer speakers;

• Multi-condition tire noise testing—covering multiple road surfaces at 40–80 km/h, with special acquisition of in-cabin wind noise.

5.4 NVH Testing of a High-End Motorcycle Brand (NAH1024)

Using the NAH-1024CH array to perform fine analysis of the dynamic radiated noise of the rear-drive assembly:

• Multi-dimensional physical quantity decoupling—high-resolution holographic mapping of the transient sound pressure field;

• Energy flow vector tracing—fine-grained excitation source tracing through “energy flow” vector localization based on the transient sound intensity field.

5.5 NVH Testing of a High-Performance Supercar Brand

Targeting the transient sound field radiation characteristics of a high-performance powertrain, combining high-resolution holographic mapping of the transient sound pressure field with “energy flow” vector tracing of the transient sound intensity field to achieve quantitative analysis of the sound field radiation mechanism of the powertrain under dynamic operating conditions.

VI. Technical Summary

This solution uses a “spherical panoramic + NAH1024 ultra-large planar array” dual-core architecture to systematically connect the complete technical chain of automotive NVH/BSR inspection:

• Macroscopic layer—the spherical panoramic acoustic imaging camera, with native 360°×180° coverage and 12 dB AI blind source separation, achieves blind-zone-free, strong-noise-decoupled, real-time tracing of whole-vehicle abnormal noise;

• Microscopic layer—the NAH1024 breaks through the conventional diffraction limit with near-field acoustic holography, achieving centimeter-level three-dimensional quantitative localization of component-level excitation points;

• Synergy layer—the two cores complement each other to build a fully digital, visualized intelligent R&D closed loop covering “whole-vehicle localization → component insight → structural optimization.”

From “locating by listening” to “mechanism insight,” from “meter-level blur” to a “centimeter-level origin”—GroupCore Acoustic uses holographic sound field technology as its foundation to build a unified perception infrastructure for acoustic world models, continuously empowering high-end acoustic R&D in the automotive industry.