Loudness is measured using a combination of physical instruments and standardized scales that account for how human ears actually perceive sound. The most common unit is the decibel (dB), but a raw decibel reading doesn’t tell the full story, because two sounds at the same decibel level can seem very different in loudness depending on their frequency. Getting an accurate loudness measurement requires understanding which scale to use, what tools are available, and how environmental factors affect your reading.
Decibels, Phons, and Sones
Decibels measure sound pressure level, which is the physical energy a sound wave carries. This is the number you’ll see on a sound level meter. But here’s the catch: two different 60-decibel sounds will not, in general, have the same perceived loudness. A deep bass hum at 60 dB sounds quieter to most people than a midrange tone at 60 dB, even though the meter reads the same. Human hearing is most sensitive to frequencies between roughly 1,000 and 5,000 Hz, so sounds in that range seem louder at the same decibel level.
To solve this, acousticians developed the phon scale. Phons tie perceived loudness to a reference tone at 1,000 Hz. If a sound seems as loud as a 60 dB tone at 1,000 Hz, it has a loudness of 60 phons, regardless of its actual frequency or decibel level. This makes comparisons across different types of sound more meaningful.
The sone scale goes one step further by creating a linear relationship with perceived loudness. On the phon scale, going from 40 to 50 phons doesn’t feel like the same jump as going from 50 to 60. The sone scale fixes this: 2 sones sounds twice as loud as 1 sone. As a reference point, 40 phons equals 1 sone, 50 phons equals 2 sones, and 60 phons equals 4 sones. Each increase of 10 phons doubles the sone value. If you need to communicate how much louder one sound feels compared to another, sones are the most intuitive unit.
Frequency Weighting: A, C, and Z
Most sound level meters let you choose a frequency weighting, which is a filter that adjusts the reading to match how the ear responds at different volumes. Picking the right one matters.
- A-weighting (dBA) is the most widely used. It reduces the contribution of very low and very high frequencies, mimicking how the average human ear hears at moderate levels. Nearly all occupational noise regulations and environmental noise standards reference dBA.
- C-weighting (dBC) applies a much flatter filter. At high levels (around 100 dB and above), the ear’s response flattens out, so C-weighting better represents what you hear during loud events. It’s commonly used for peak measurements and entertainment noise, where bass transmission is a concern.
- Z-weighting (dBZ) applies no filtering at all. It captures the full, unweighted sound pressure across all frequencies. This is useful for raw acoustic analysis but doesn’t reflect perceived loudness.
For most practical purposes, if you’re measuring noise to assess how loud something sounds or whether it poses a hearing risk, use A-weighting.
Averaging Fluctuating Noise With Leq
Real-world sound rarely stays at a constant level. Traffic, machinery, and conversation all rise and fall. To capture this, acousticians use a metric called Leq, or equivalent continuous sound level. It compresses a period of fluctuating noise into a single decibel value that represents the same total sound energy as if the noise had been perfectly steady the entire time.
The math behind Leq involves averaging the squared sound pressure over a set time period, then converting the result to decibels. You don’t need to do this yourself; any decent sound level meter calculates it automatically. What matters is understanding the concept: an Leq of 85 dBA over eight hours means the varying noise you experienced carried the same energy (and the same potential for hearing damage) as a constant 85 dBA tone running for eight hours straight. This is the metric used in workplace noise regulations worldwide.
Tools for Measuring Loudness
Professional Sound Level Meters
Dedicated sound level meters are built to the IEC 61672-1 standard and come in two accuracy classes. Class 1 instruments have tighter tolerances and are required for precise environmental monitoring, legal compliance work, and laboratory measurement. Class 2 meters have slightly wider tolerances but still provide reliable readings for general noise surveys and workplace assessments.
Before taking any measurement, you need to calibrate. The process is straightforward: place an acoustic calibrator over the microphone, turn it on to generate a known tone (typically 94 dB), enter the meter’s calibration mode, adjust the reading to match the calibrator’s output, and save. This takes under a minute and ensures your readings are trustworthy. Skipping calibration can introduce errors that undermine everything you measure afterward.
Smartphone Apps
Phone-based sound level meter apps are convenient but come with significant caveats. A large study testing 100 different phones found a standard deviation of 6.81 dB between devices, meaning two phones running the same app could give readings that are several decibels apart. On average, uncalibrated iOS apps deviated about 2.93 dBA from the true value, while Android apps deviated about 2.79 dBA. That’s the average. Individual readings varied far more, with some studies finding differences of up to 12 dB from a professional meter.
The picture improves dramatically with calibration. When iOS apps were paired with external microphones and calibrated against a known reference, accuracy tightened to within 1 dBA of a Class 1 meter. The NIOSH Sound Level Meter app achieved similar results under calibrated conditions. So if you need a quick, rough sense of how loud your environment is, a well-rated app on a recent phone gives a reasonable estimate. For anything where accuracy matters, you need either a calibrated external microphone or a dedicated meter.
LUFS for Digital Audio
If you’re measuring loudness for music production, podcasting, or broadcast, the relevant standard is LUFS (Loudness Units relative to Full Scale). The EBU R128 standard, based on the ITU-R BS.1770 algorithm, is the framework that streaming platforms and broadcasters use to normalize audio so that different tracks or programs play back at a consistent volume.
LUFS measures integrated loudness over the full duration of a piece of audio, with built-in gating that ignores silence (anything below -70 LUFS) so that quiet pauses don’t drag down the average. A second relative gate, set 10 LU below the absolute-gated level, further refines the measurement by excluding very quiet passages. Streaming services each have their own target: Spotify normalizes to around -14 LUFS, YouTube to around -14 LUFS, and Apple Music to around -16 LUFS. If your audio is louder than the platform’s target, it gets turned down automatically.
How Distance and Environment Affect Readings
Sound intensity from a point source follows the inverse square law in open air: every time you double your distance from the source, the intensity drops by about 6 dB. This means a reading taken at 1 meter will be roughly 6 dB higher than one taken at 2 meters, and 12 dB higher than one taken at 4 meters. This relationship holds cleanly only outdoors, away from reflective surfaces.
Indoors, walls, floors, ceilings, and furniture all reflect sound, creating reverberation that can raise the apparent level. Hard surfaces like tile and glass reflect more energy than soft materials like carpet and curtains. When measuring indoors, your reading includes both the direct sound and all its reflections, which is why the same source can measure several decibels louder in a bare room than in a furnished one. For consistent results, always note the distance from the source and whether the measurement was taken indoors or outdoors.
Safe Exposure Limits
OSHA’s permissible exposure limits set the legal ceiling for workplace noise in the United States. At 90 dBA, the maximum allowable exposure is 8 hours. That window shrinks quickly as the level rises: 95 dBA allows only 4 hours, 100 dBA allows 2 hours, 105 dBA allows 1 hour, and 110 dBA allows just 30 minutes. Exposure to impact or impulse noise must never exceed 140 dB peak.
OSHA also requires employers to start a hearing conservation program whenever employees are exposed to an 8-hour time-weighted average of 85 dBA or higher. NIOSH, the research arm that informs these standards, actually recommends 85 dBA as the exposure limit rather than 90, using a stricter calculation method. If you’re measuring noise in your own environment and consistently seeing readings above 85 dBA over extended periods, hearing protection is worth considering regardless of which standard you follow.
When exposure involves multiple noise levels throughout the day, the effects are cumulative. You calculate the combined dose by dividing the actual time spent at each level by the maximum permitted time at that level, then adding the fractions. If the total exceeds 1, the exposure has crossed the safe threshold.