Loudness is a subjective measure of how intense a sound seems to a listener. That distinction matters: two sounds with the same physical energy can be perceived as very different in loudness depending on their frequency, duration, and the listener's own hearing. Because of this subjectivity, architectural acoustics relies on specialized scales and units rather than raw sound pressure measurements alone.

The relationship between physical sound intensity and perceived loudness is nonlinear. You can't just double the sound pressure and expect a listener to report "twice as loud." This nonlinearity drives much of what follows in this topic.

Factors affecting loudness

Sound pressure level (SPL): Higher SPL generally means louder perception, but the relationship isn't proportional.
Frequency content: Human hearing is most sensitive around 2–5 kHz. A tone at 3 kHz will sound louder than a tone at 100 Hz even if both have the same SPL.
Duration: Sounds lasting longer than about 200 ms are perceived as louder than shorter bursts at the same SPL, due to temporal integration in the auditory system.
Spectral composition: A complex sound with energy spread across multiple frequencies can be perceived as louder than a pure tone at the same overall SPL, because it excites more critical bands in the cochlea.
Individual hearing sensitivity: Age, hearing health, and natural variation among listeners all affect loudness perception.

Loudness vs. sound pressure level

Sound pressure level is a physical quantity, measured in decibels (dB) relative to a reference pressure of 20 μPa. Loudness is a perceptual quantity that depends on frequency, duration, and other factors beyond SPL alone.

A useful rule of thumb: a 10 dB increase in SPL roughly corresponds to a doubling of perceived loudness. This holds reasonably well in the mid-frequency range but becomes less reliable at very low or very high frequencies, which is exactly why equal-loudness contours exist.

Loudness units and scales

Phon: A unit of loudness level. A sound has a loudness level of $N$ phons if it's perceived as equally loud as a 1 kHz pure tone at $N$ dB SPL. So at 1 kHz, phons and dB SPL are numerically identical. At other frequencies, they diverge.
Sone: A linear loudness scale. One sone is defined as the loudness of a 1 kHz tone at 40 dB SPL (i.e., 40 phons). Doubling the sone value means doubling the perceived loudness, which makes sones more intuitive for comparing how loud things actually sound.

These units let acousticians set meaningful targets. Saying "the background noise should not exceed 8 sones" communicates perceptual impact in a way that a raw dB number cannot.

Equal-loudness contours

Equal-loudness contours are curves plotted on a graph of SPL (y-axis) versus frequency (x-axis). Each curve connects all the frequency-SPL combinations that a listener perceives as equally loud. They reveal a fundamental fact about human hearing: our sensitivity is not flat across frequency.

At low SPLs, the contours are steeply curved, meaning you need much more sound pressure at low frequencies (say, 50 Hz) to match the perceived loudness of a mid-frequency tone. At higher SPLs, the contours flatten out, and our hearing becomes more uniform across frequency.

Fletcher-Munson curves

Harvey Fletcher and Wilden A. Munson published the first set of equal-loudness contours in 1933, based on listening tests with pure tones and young, healthy listeners. Their key findings:

Human hearing is most sensitive around 2–5 kHz (roughly the resonant frequency range of the ear canal).
Sensitivity drops off significantly below about 500 Hz and above about 8 kHz.
The shape of the contours changes with level: low-frequency sensitivity improves at higher SPLs.

The original Fletcher-Munson data has since been revised. The current international standard is ISO 226:2003, which is based on more recent, larger-scale studies and corrects some inaccuracies in the original curves, particularly at low frequencies. When you see "equal-loudness contours" referenced in modern practice, it's usually the ISO 226 version.

Phon scale for loudness levels

The phon scale is read directly from equal-loudness contours. To find the loudness level of any sound in phons:

Identify the frequency and SPL of the sound.
Find which equal-loudness contour passes through that frequency-SPL point.
Read the phon value of that contour (which equals the SPL where the contour crosses 1 kHz).

For example, a 100 Hz tone at 60 dB SPL might fall on the 40-phon contour, meaning it sounds only as loud as a 1 kHz tone at 40 dB SPL. This illustrates how much less sensitive we are to low frequencies at moderate levels.

Sone scale for loudness

The sone scale converts phons into a linear perceptual scale using this relationship:

$\text{sones} = 2^{(\text{phons} - 40) / 10}$

Some reference points to build intuition:

40 phons = 1 sone
50 phons = 2 sones (twice as loud)
60 phons = 4 sones (four times as loud)
30 phons = 0.5 sones (half as loud)

Because the sone scale is linear with respect to perception, it's especially useful in architectural acoustics when you need to communicate loudness differences to clients or compare design options.

Factors affecting loudness, Acoustic quantities, part 1: What are decibels? - Erlend M. Viggen

Differences in equal-loudness contours

Equal-loudness contours are not universal constants. They can vary based on:

Listener age and hearing health: Older listeners typically have reduced high-frequency sensitivity, shifting the contours.
Measurement methodology: Free-field versus diffuse-field listening conditions produce slightly different contour shapes.
Individual variation: Even among young, healthy listeners, there's natural spread in the data.

The ISO 226:2003 standard represents a statistical average across a population. Architectural acousticians should be aware that real listeners may deviate from these averages, particularly when designing for populations that skew older or for spaces where low-frequency or high-frequency content dominates.

Loudness measurement

Measuring loudness goes beyond placing a microphone in a room and reading a dB value. Because perceived loudness depends on frequency content and temporal patterns, measurement methods must model aspects of human hearing to produce meaningful results.

Loudness meters and standards

Loudness meters incorporate frequency weighting and time-averaging to approximate human perception. A-weighting is the most common frequency weighting; it roughly follows the inverse of the 40-phon equal-loudness contour, de-emphasizing low and very high frequencies. The result is reported in dB(A).

Key standards governing loudness measurement include:

ISO 532: Specifies methods for calculating loudness in sones, using either Zwicker's method (for stationary sounds) or Stevens' method.
DIN 45631: A German standard closely aligned with Zwicker's method, widely used in product acoustics and building services noise assessment.

These standards ensure that measurements taken by different practitioners in different locations are comparable.

Zwicker's method for stationary sounds

Zwicker's method calculates the loudness of steady-state sounds by breaking the signal into perceptual frequency bands. Here's the process:

Measure the sound's spectrum (SPL as a function of frequency).
Divide the audible range into critical bands (Bark scale), which reflect the frequency resolution of the human cochlea.
Calculate the specific loudness (loudness contribution per critical band) for each band, accounting for masking effects between adjacent bands.
Sum (integrate) the specific loudness values across all critical bands to get the total loudness in sones.

This method is particularly useful for evaluating HVAC noise, mechanical equipment, and other steady-state sources common in buildings.

Time-varying loudness measurement

Sounds like speech, music, and intermittent noise fluctuate over time. Static methods like Zwicker's won't capture the perceptual experience of these signals.

Time-varying loudness models, such as the dynamic loudness model (DLM) developed by Glasberg and Moore, calculate loudness as a function of time. These models account for:

Temporal integration: The auditory system averages energy over short time windows (roughly 100–200 ms).
Attack and release characteristics: Sudden onsets are perceived differently from gradual changes.
Short-term vs. long-term loudness: Models often report both an instantaneous loudness and a longer-term average.

These measurements are critical for spaces designed for speech (classrooms, courtrooms) or music (concert halls, recording studios), where the dynamic character of sound directly affects the listener's experience.

Binaural loudness measurement

Binaural measurement uses a dummy head (also called a head-and-torso simulator) with microphones placed at the entrance of each ear canal. This captures:

Head-related transfer functions (HRTFs): The way the head, pinnae, and torso filter sound differently depending on the direction of arrival.
Binaural summation: Hearing with two ears produces a perceived loudness roughly 6–10 phons greater than monaural listening at the same SPL.

Binaural loudness data provides the most realistic picture of what a listener actually experiences in a space. It's especially valuable for concert halls, auditoriums, and any environment where spatial sound distribution matters. The data can also feed into auralization systems that let designers and clients "listen" to a virtual model of a room before it's built.

Loudness in room acoustics

A room transforms every sound that passes through it. The direct sound from a source combines with reflections off walls, ceiling, and floor, plus the reverberant tail. All of these components contribute to the loudness a listener perceives, and their balance determines whether a room sounds clear, muddy, too loud, or too quiet.

Factors affecting loudness, Acoustic quantities, part 2: Frequency weighting - Erlend M. Viggen

Loudness constancy and reflections

Loudness constancy is the tendency for a sound to seem roughly the same loudness even as you move farther from the source. In a free field (outdoors, no reflections), SPL drops by 6 dB for every doubling of distance. Indoors, early reflections add energy that partially compensates for this drop.

The strength and timing of early reflections (arriving within roughly 50–80 ms of the direct sound) are key. Strong, well-timed early reflections support loudness constancy and improve speech intelligibility. Weak or delayed reflections can leave listeners in the back of a room feeling like the sound has "dropped off." Designing room surfaces to direct early reflections toward the audience is one of the most effective tools for managing loudness distribution.

Loudness vs. reverberation time

Reverberation time (RT60) is the time it takes for sound energy to decay by 60 dB after the source stops. Its relationship with loudness is nuanced:

Longer reverberation times generally increase the perceived loudness of continuous sounds, because reverberant energy accumulates in the room.
For transient or speech-like sounds, excessive reverberation can smear temporal details, reducing clarity even while increasing overall loudness.
The optimal RT60 depends on the room's purpose. A lecture hall might target 0.6–1.0 s, while a symphony hall might aim for 1.8–2.2 s.

The design challenge is balancing sufficient reverberant energy for a sense of fullness and loudness against the need for clarity and intelligibility.

Loudness and room modes

Room modes are standing wave patterns that form at frequencies determined by the room's dimensions. At a modal frequency, sound pressure can be significantly higher at certain locations (near walls and corners) and lower at others (near nodal points).

This effect is most pronounced in small rooms and at low frequencies, where the spacing between modes is wide enough that individual modes are clearly audible. The result can be uneven bass response: some seats experience boomy, exaggerated low-frequency loudness while others hear almost none.

Strategies for managing room modes include:

Using room dimension ratios that distribute modes more evenly across frequency (avoiding cubic or highly symmetric rooms).
Placing low-frequency absorbers (bass traps) at room boundaries where modal pressure is highest.
Adding diffusers to break up standing wave patterns.

Designing rooms for optimal loudness

Achieving the right loudness throughout a space requires coordinating several design variables:

Room volume and shape: Larger volumes spread energy over more space, reducing loudness per unit area. Room shape determines how reflections are distributed.
Surface treatments: A balance of absorptive materials (to control reverberation), reflective surfaces (to direct early energy toward listeners), and diffusers (to create even sound fields).
Sound reinforcement: In large or acoustically challenging spaces, loudspeaker systems can supplement natural sound to maintain adequate loudness at all listener positions.
Acoustic simulation: Software tools (such as CATT-Acoustic or ODEON) and physical scale models allow designers to predict loudness distribution and iterate on the design before construction begins.

The goal is a space where loudness feels natural, consistent, and appropriate for the intended use.

Applications of loudness

Loudness principles extend well beyond room design. Acousticians working in architectural practice encounter loudness considerations in product specification, regulatory compliance, and environmental assessment.

Loudness in product design

Products like HVAC units, elevators, appliances, and building service equipment all generate sound that occupies the spaces architects design. Manufacturers specify loudness in sones (common for bathroom fans and range hoods in the U.S.) or dB(A).

Psychoacoustic shaping goes beyond just making things quieter. The character of a product's sound matters: a smooth, low-frequency hum at 2 sones may be far less annoying than a tonal whine at the same loudness. Acousticians evaluate both the level and the spectral quality of product sounds when specifying equipment for a building.

Loudness normalization in broadcasting

Loudness normalization ensures that audio content plays at a consistent perceived level across programs, channels, and commercials. The governing standard is ITU-R BS.1770, which defines a measurement algorithm based on frequency-weighted, gated loudness in units of LUFS (Loudness Units relative to Full Scale).

While this is primarily a broadcast engineering concern, it's relevant to architectural acoustics when designing spaces with integrated audio systems (houses of worship, theaters, conference centers). Understanding how source material is normalized helps acousticians calibrate playback systems for consistent loudness.

Loudness and hearing protection

Prolonged exposure to high loudness levels causes noise-induced hearing loss, which is irreversible. Regulatory frameworks set exposure limits:

OSHA (U.S.): Permissible exposure limit of 90 dB(A) TWA over an 8-hour workday, with a 5 dB exchange rate.
NIOSH (U.S.): Recommends a stricter 85 dB(A) limit with a 3 dB exchange rate.
EU Directive 2003/10/EC: Sets lower and upper exposure action values at 80 and 85 dB(A) respectively.

Architectural acousticians contribute to hearing conservation by designing enclosures for noisy equipment, specifying sound-isolating partitions, and selecting absorptive treatments that reduce reverberant buildup in industrial or entertainment spaces.

Loudness in environmental noise assessment

Environmental noise from traffic, aircraft, railways, and industrial sources is assessed using metrics that weight loudness over time:

$L_{dn}$ (Day-Night Average Sound Level): A 24-hour average with a 10 dB penalty applied to nighttime hours (10 PM–7 AM), reflecting greater sensitivity to noise during sleep.
$L_{den}$ (Day-Evening-Night Level): Similar to $L_{dn}$ but adds a 5 dB penalty for evening hours as well.
CNEL (Community Noise Equivalent Level): Used primarily in California, similar in structure to $L_{den}$ .

These metrics inform zoning decisions, building facade design, and noise barrier specifications. Architectural acousticians use them to determine the required sound insulation of building envelopes and to predict interior noise levels from exterior sources.