Noise Sound Level and Spectrum Analysis: Principles, Standards, and Testing Methods

Noise sound level spectrum analysis combines precise sound pressure and sound power measurements with octave-band, third-octave-band, and FFT frequency analysis to characterize how acoustic energy is distributed across the audible frequency range. Performed with Class 1 instrumentation conforming to IEC 61672 and GB/T 3785.1 in hemi-anechoic chambers, reverberation rooms, or on site, it supports product noise compliance, labeling, research and development, and environmental assessment. Accredited third-party laboratories conduct these measurements in accordance with ISO 3741, ISO 3744, ISO 3745, IEC 60704, ISO 7779, and related standards.

Technical Overview

Noise sound level spectrum analysis is a quantitative acoustic measurement discipline that determines both the magnitude of the noise emitted by a source and its distribution over frequency. A single A-weighted overall level is often insufficient to explain why a product sounds objectionable, to demonstrate compliance with frequency-dependent limits, or to guide engineering noise reduction. Spectrum analysis resolves the total sound energy into octave bands, one-third-octave bands, or narrow FFT lines, revealing tonal components, low-frequency rumble, or high-frequency hiss that drive human perception, regulatory evaluation, and diagnostic conclusions.

Sound Level Metrics

Sound pressure level (Lp) is expressed in decibels as twenty times the logarithm of the measured RMS sound pressure relative to the reference pressure of 20 µPa. Key quantities include the time-averaged equivalent continuous level LAeq,T, the maximum time-weighted levels LAFmax and LASmax, the C-weighted peak level LCPeak for impulsive noise, and statistical percentile levels such as L10, L50, and L90 used in environmental surveys. Sound power level (LW), referenced to 1 pW, describes the total acoustic energy radiated by the source per second; it is an intrinsic property of the product, independent of measurement distance and room conditions, and is the quantity required for noise declarations and labeling.

Frequency Weighting Networks

A-weighting approximates the frequency response of human hearing at moderate loudness levels and is applied for virtually all rating and limit purposes. C-weighting, which is nearly flat across the audible range, is used to assess low-frequency energy and to capture peak values, while Z (zero) weighting provides unweighted linear data preferred for spectral analysis. Time weightings Fast (125 ms) and Slow (1 s) govern the response of the level detector, with Fast weighting used for most emission measurements.

Spectral Resolution

Octave-band analysis typically covers center frequencies from 31.5 Hz to 16 kHz, while one-third-octave analysis divides each octave into three bands with a bandwidth ratio of 2^(1/3), providing the resolution needed for detailed rating and comparison. Narrowband FFT analysis, often with 1 Hz to 10 Hz resolution, is used to isolate discrete tones such as fan blade-pass frequency, gear-mesh frequencies, transformer hum at 100 Hz or 120 Hz and its harmonics, and switching-frequency whine from power electronics. Where tones dominate perception, standards such as ISO 7779 and ISO 1996-2 provide criteria for identifying and rating prominent discrete tones.

Testing Samples

Spectrum-based noise testing applies to virtually any equipment that generates airborne sound. Typical sample categories handled by accredited third-party laboratories include:

Household and Similar Electrical Appliances

Refrigerators, freezers, washing machines, dishwashers, range hoods, vacuum cleaners, air conditioners, dehumidifiers, hair dryers, electric fans, and kitchen machines, normally tested to IEC 60704-1 and its product-specific parts, with the GB/T 4214 series as the corresponding Chinese method standard and GB 19606 as the mandatory limit standard.

Information Technology and Telecommunications Equipment

Servers, storage systems, printers, copiers, laptops, power adapters, displays, network equipment, and cooling fans, tested to ISO 7779 with operator and bystander positions evaluated per ISO 11201, and declared values reported per ISO 9296.

Industrial Machinery and Power Tools

Electric motors, pumps, fans, compressors tested to ISO 2151, gearboxes, machine tools, industrial robots, generators, hand-held power tools, and earth-moving machinery tested to ISO 6393 and ISO 6395.

HVAC and Building Services Equipment

Chillers, air handling units, VRF systems, fan coil units, ventilation and heat-recovery units, heat pumps, cooling towers, and commercial refrigeration equipment, including site acceptance measurements of installed systems in accordance with ISO 1996.

Electrical Power Equipment

Power and distribution transformers and reactors measured to IEC 60076-10, switchgear, inverters, UPS systems, and battery energy storage systems, where characterization of electromagnetic hum and its harmonic content is of particular interest.

Automotive and Transportation Components

Electric drivetrains and traction motors, cooling modules, tires, brake components, and interior trim parts, together with rail vehicle exterior noise per ISO 3095, interior noise per ISO 3381, and stationary road vehicle noise per ISO 5130.

Testing Items

Depending on the applicable product test code and the client's objectives, the following quantities are typically determined:

• A-weighted equivalent continuous sound pressure level LAeq,T at operator and bystander positions per ISO 11201 and ISO 7779
• A-weighted and octave or one-third-octave band sound power level LW per ISO 3741, ISO 3744, ISO 3745, and ISO 9614
• Octave-band spectra from 31.5 Hz to 16 kHz and one-third-octave spectra from 50 Hz or 100 Hz upward
• Narrowband FFT spectra with typical resolutions of 1 Hz to 10 Hz for discrete tone identification
• Discrete tone prominence and tonality assessment per the methods of ISO 7779, ISO/PAS 20065, and ISO 1996-2
• Maximum time-weighted levels (LAFmax, LASmax) and C-weighted peak level (LCPeak) for impulsive and fluctuating noise
• Statistical levels L5, L10, L50, L90, and L95 and rating levels for environmental noise per ISO 1996
• Occupational noise exposure parameters per ISO 9612 for workplace surveys
• Declared noise emission values with uncertainty per ISO 4871, ISO 7574-4, and ISO 9296

Testing Methods

Instrumentation and Calibration

Measurements employ Class 1 (precision grade) sound level meters, real-time analyzers, and condenser microphones conforming to IEC 61672-1 and GB/T 3785.1, with octave and fractional-octave band filters meeting IEC 61260 and GB/T 3241. An acoustic calibrator conforming to IEC 60942 is applied immediately before and after each measurement, and instruments undergo periodic verification by accredited third-party laboratories. Measurement uncertainty is evaluated for each measurement channel and reported with the results.

Hemi-Anechoic Room Method (Precision Grade)

ISO 3745 and GB/T 6882 specify determination of sound power in a free field over a reflecting plane. The source is installed on the hard floor, and sound pressure is measured on a hemispherical, parallelepiped, or conformal measurement surface using at least ten microphone positions. The method achieves precision-grade accuracy, making it suitable for research and development, low-noise products, and detailed tonal investigations.

Engineering Enveloping-Surface Method

ISO 3744 and GB/T 3767 define an engineering-grade method applicable in ordinary rooms or on site. The environmental correction K2A, determined using a calibrated reference sound source or a two-surface measurement, must not exceed 2 dB for a valid result. Where background noise is between 6 dB and 15 dB below the combined level, a background correction K1 is applied; if the difference is less than 6 dB, the measurement is not valid and the test environment or method must be improved.

Reverberation Room Method

ISO 3741 and GB/T 6881.1 describe determination of sound power level in a qualified diffuse field, typically by comparison with a calibrated reference sound source. The method is efficient for broadband sources and larger equipment, and it delivers octave and one-third-octave band sound power levels directly.

Sound Intensity Method

ISO 9614-1 (point-by-point) and ISO 9614-2 (scanning) determine sound power by measuring intensity with a two-microphone probe over a closed surface enclosing the source. Because steady extraneous noise is largely suppressed, the method is well suited to in-situ machinery testing and to noise source ranking through partial-surface analysis.

Workplace and Environmental Methods

Emission values at the operator position are measured per ISO 11201, environmental noise assessments follow ISO 1996-1 and ISO 1996-2 with defined meteorological and operating conditions, and personal exposure measurements follow ISO 9612. Stationary vehicles are addressed by ISO 5130, and rail applications by ISO 3095 and ISO 3381.

General Measurement Procedure

A typical test sequence comprises: definition of mounting, installation, and operating conditions according to the product test code; instrument verification with an acoustic calibrator; background noise measurement; installation of the sample; selection of the measurement surface and positions; measurement of each operating mode for at least 30 seconds or over the complete operating cycle; simultaneous recording of time histories and frequency spectra; application of background and environmental corrections; calculation of surface-averaged pressure levels and sound power levels; assessment of tone prominence; and reporting with expanded measurement uncertainty.

Applications

Noise level and spectrum analysis results are used throughout the product life cycle and in regulatory contexts:

• Regulatory compliance and market access, including verification against mandatory national limit standards such as GB 19606 for household appliances and noise declarations required under EU ecodesign and energy labelling regulations.
• Noise emission labeling and datasheets per ISO 4871 and ISO 9296, enabling fair comparison of declared sound power levels between products.
• Research, development, and benchmarking, where octave and FFT spectra identify dominant sources such as fan blade-pass tones, motor electromagnetic noise, gear mesh components, and flow-generated noise, and quantify the benefit of design changes including speed control, enclosures, and damping treatments.
• Production quality control, using statistical verification of conformity with declared values per ISO 7574-4.
• Occupational health and safety, including exposure surveys per ISO 9612 and machinery noise declarations supporting workplace risk assessment.
• Environmental impact assessment of industrial sites, HVAC installations, and infrastructure per ISO 1996, including low-frequency and tonal complaint investigations.
• Building services commissioning, verifying installed noise performance of HVAC systems against project specifications in occupied spaces.
• Predictive maintenance and diagnostics, trending spectral signatures such as bearing defect tones, imbalance signatures, and structural resonances.

FAQ

What is the difference between sound pressure level and sound power level?

Sound pressure level describes the noise measured at a particular location and depends on distance, direction, and the surrounding room. Sound power level is an intrinsic property of the source that quantifies the total acoustic energy radiated per second, referenced to 1 pW. Product standards and legal declarations generally require sound power level, whereas operator-position and environmental measurements are reported as sound pressure levels.

Why is A-weighting used for most noise measurements?

A-weighting approximates the frequency sensitivity of human hearing, attenuating low and very high frequencies so that the single-number result correlates with perceived loudness for broadband noise. C-weighted, Z-weighted, and band-level data are additionally measured whenever low-frequency content, impulsive characteristics, or tonal components must be evaluated in detail.

What is the difference between octave and one-third-octave analysis?

Octave analysis divides the frequency range into bands whose upper edge is twice the lower edge, providing a compact ten-band overview from 31.5 Hz to 16 kHz. One-third-octave analysis provides three times finer resolution and is preferred when spectra must be compared with limit curves, when tonal content is suspected, or when detailed propagation and insulation calculations are required. FFT analysis offers the finest resolution and is used to pinpoint individual tones.

How is background noise accounted for?

Background noise is measured with the test sample switched off. If it lies 6 dB to 15 dB below the combined level, a K1 correction is applied; if the difference is 15 dB or more, no correction is needed; if it is below 6 dB, the result is invalid under ISO 3744 and the test environment, schedule, or method must be changed, for example by adopting the sound intensity method.

Which standards apply to household appliance noise testing?

IEC 60704-1 and its product-specific parts define the measurement methods, with the GB/T 4214 series as the corresponding national adoption. Sound power determination follows ISO 3741, ISO 3744, or ISO 3745 as appropriate, GB 19606 specifies mandatory limit values in China, and declared values are stated per ISO 4871.

What is a prominent discrete tone and why does it matter?

A prominent discrete tone is a narrowband spectral component that stands out clearly above the noise in neighboring bands, such as a transformer hum harmonic or a fan blade-pass tone. Because tones are more noticeable and more annoying than broadband noise of the same level, standards including ISO 7779, ISO 1996-2, and ISO/PAS 20065 provide procedures for detecting tones and, in some contexts, adding a tone penalty to the rated level.

How long does a typical test take?

Individual measurements normally last at least 30 seconds per operating mode, or the full operating cycle for appliances with varying programs such as washing machines. Including setup, background measurement, multiple modes, and analysis, a typical laboratory session takes several hours, and reports including spectra, sound power values, tonality assessment, and uncertainty are usually issued within a few working days.

Can noise measurements be performed on site?

Yes. Engineering-grade enveloping-surface measurements per ISO 3744 and sound intensity measurements per ISO 9614 allow sound power determination in production plants, and environmental assessments per ISO 1996 are conducted outdoors or in occupied buildings. Site testing requires attention to background noise, wind, reflecting surfaces, and realistic load simulation to ensure valid, repeatable results.