When validating quietness for home appliances, fans, precision motors, medical devices, transformers, and consumer electronics, we are not only asking whether a product is "loud" or "quiet." We need to know where the sound comes from, under which operating condition it appears, and in which frequency range it becomes uncomfortable.

A quiet product can sometimes be like a person who does not express themselves clearly. It does not complain loudly; it simply produces a faint hum, whine, or buzz late at night. On a specification sheet, it may read only 18 dBA. Yet when a user gets close, they may still frown: the sound is not loud, but it is annoying.

That is the purpose of ultra-low-noise testing. It is not about asking the product to stand on a stage and announce how many decibels it has. It is about listening clearly to its very small movements, moods, and defects. A good test needs to answer four questions: Is the sound really generated by the product? Which frequency is it hiding in? When is it most obvious? After optimization, has it truly converged?

Microphones also "speak"

Many people assume that once a microphone is placed near a product, the product sound will simply enter the measurement system. In reality, a microphone is not an absolutely silent observer. It has its own self-noise, which is like the microphone's breathing: if that breathing is too heavy, very quiet product sounds can be covered up.

The inherent noise of common 1/2-inch measurement microphones is typically around 16 dBA. For ordinary noise testing, that is often sufficient. For high-end quiet products, however, 16 dBA may already be too noisy, because many of the subtle sounds that users care about may be near that level or even below it.

A 1-inch microphone can reduce inherent noise to about 10 dBA, which is much quieter. But it also has its own personality: with a larger diaphragm, frequency response and practical flexibility can be more limited, and high-frequency detail is not as friendly as with a broadband 1/2-inch solution. In other words, a 1/2-inch microphone has broader frequency capability but higher self-noise; a 1-inch microphone is quieter but less open in high-frequency detail.

This creates a dilemma in advanced quietness testing: at some frequencies, sounds that the human ear can perceive may already be below the microphone's own self-noise. Users can hear them, but the microphone says, "What I have here is mostly my own noise floor." That is why we need a rarer kind of ear: one that is both low-noise and broadband, so it can truly capture the subtle signals that matter to human perception.

Decibel value is total noise, not product character

dB is a logarithmic unit, not an ordinary ruler. A 3 dB increase in sound energy is approximately a doubling of energy, while a 10 dB increase is approximately ten times the energy. But the human ear does not listen to energy alone; it also picks out frequency. A very narrow frequency peak may not raise the overall dBA value very much, but it can still make people hear a clear "hum," "whine," or "buzz."

For example, two small fans may both measure 18 dBA. Sample A has a spectrum like a flat blanket, with no particularly prominent component. Sample B has a spike around 1.6 kHz that is more than 10 dB higher than the surrounding frequency bands. Their total decibel values look similar, but Sample B feels like someone suddenly clearing their throat in a quiet room; the ear notices it immediately. Mobile phones and headphones behave in the same way: bearings, resistive-capacitive components, displays, call paths, headphone amplifiers, and noise-cancelling modules can all reveal themselves through very quiet sounds.

Background noise and product sound are mixed together

During ultra-low-noise testing, background noise is highly disruptive. It will not tell you, "I am the air-conditioning noise in the room," or "I am the noise floor of the test system." It simply mixes into the result and makes the product appear louder than it really is.

Sound also cannot be subtracted like ordinary numbers. For example, if the true product sound is 12 dBA and the room background is already 10 dBA, the instrument sees the energy sum of the two.

Total noise:  L_total = 10 log10(10^(L_product/10) + 10^(L_background/10))

Example: 10 log10(10^(12/10) + 10^(10/10)) = 10 log10(15.85 + 10) = 14.12 dBA

In this example, the background is only 2 dB lower than the product, and the total reading is raised by 14.12 − 12 = 2.12 dB.

Estimated product level:  L_product = 10 log10(10^(L_total/10) − 10^(L_background/10))

In simple terms, the closer the background level is to the product level, the less reliable the test conclusion becomes.

Recommended fixed conditions and analysis metrics

When product sound is very quiet, even a small movement can change the result. Moving the microphone 5 cm closer, letting a cable touch the tabletop, changing the fixture position, or slightly changing the product orientation can all alter the spectrum. To make before/after data comparable, the product must "speak" in the same posture and under the same state every time.

Roles of the CRYSOUND low-noise testing solution

If an ultra-low-noise test is treated as a careful act of listening, the acoustic shielded chamber or test box is the quiet room, the microphone is the ear, the IED module is the high-resolution memory, SonoDAQ Pro is the long-term recorder, and OpenTest is the person who translates sound into spectra and reports. The environment must be quiet and repeatable; the ear must be low-noise and broadband; the memory must be fast and have wide dynamic range; the recorder must support synchronization and expansion; and the software must connect spectra, octave bands, sound levels, and reports into one workflow.

CRY3261-S02: a quiet ear

The CRY3261-S02 consists of a CRY3261 measurement microphone and a CRY517 IEPE preamplifier. The key official parameters include: 1/2-inch free-field prepolarized microphone, 450 mV/Pa sensitivity, 6 Hz-20 kHz frequency range, and 6.5 dBA inherent noise. Its task is straightforward: keep itself quiet first, listen broadly enough across frequency, and then capture the lighter and finer sounds generated by the product.

CRY7412 acoustic shielded chamber: a quiet room that pushes the background away

The CRY7412 acoustic shielded chamber provides a low-noise, interference-controlled environment with a noise floor as low as 6.5 dBA for acoustic measurement, improving repeatability in laboratories and production lines. For small consumer electronics such as mobile phones, headphones, displays, and resistive-capacitive components, it is like arranging a quiet room for the product: external environmental sound, vibration, and RF interference are pushed into the background first, so the microphone can more easily hear the product's own self-noise, whine, or electrical noise.

CRY5011 IED high-precision input module: preserving faint details

The CRY5011 IED is a 4-channel IEPE/Voltage input module. It provides a sampling rate up to 204.8 kHz, 32-bit resolution, dynamic range up to 160 dB, THD+N ≤ −102 dB, a noise floor as low as ≤ −117 dBV, and TEDS support. For faint signals such as headphone noise, display noise, or high-frequency sound from resistive-capacitive components, these parameters mean that quiet sounds are less likely to be lost and transients are less likely to be blurred.

CRY5820 SonoDAQ Pro: a high-performance base that synchronizes and expands

The CRY5820 SonoDAQ Pro is a modular data acquisition system. A single chassis supports up to 24 channels with six modules, and expanded systems can support distributed multi-channel testing. It supports 1000 V channel isolation, inter-chassis synchronization skew of less than ±100 ns, USB-C, Gigabit LAN, CAN FD, Wi-Fi, and local microSD storage.

OpenTest: translating sound into spectra and reports

OpenTest is CRYSOUND's acoustic and vibration testing platform. It supports hardware connection methods such as OpenDAQ, ASIO, WASAPI, and Core Audio, as well as FFT, octave-band analysis, sound level meter functions, sound quality analysis, playback comparison, waveform/data export, and report generation. The subtle sounds captured by the microphone and recorded by the acquisition system are ultimately transformed in OpenTest into spectra, sound levels, trends, and traceable reports.

To learn more about microphones, preamplifiers, and related testing solutions, please visit the CRYSOUND website for additional product information and contact the team for further discussion.If necessary, please fill out the Get in touch form below.