An equal-loudness contour is a curve showing the sound pressure level required at each frequency for a sound to be perceived as equally loud as a reference sound. It describes the frequency dependence of human hearing: tones with the same physical sound pressure level do not necessarily sound equally loud. Equal-loudness contours are fundamental to psychoacoustics and help explain why the perceived tonal balance of sound changes with listening level.
Physical level and perceived loudness
Sound pressure level is a physical quantity, whereas loudness is a perceptual attribute. Sound pressure level is defined by
where is the root-mean-square sound pressure and is the reference pressure. For sound in air, the conventional reference is . The result is expressed in decibels, commonly written dB SPL.
Loudness depends not only on sound pressure level but also on frequency, duration, spectral composition, and listening conditions. Equal-loudness contours isolate the relationship between frequency and level by comparing relatively simple sounds, usually pure tones.
Loudness level and the phon
A contour is labelled by its loudness level, expressed in phons. A sound has a loudness level of phons when it is judged equally loud as a 1,000 Hz reference tone at dB SPL under the specified reference conditions.
Consequently, every -phon contour passes through dB SPL at 1,000 Hz. Elsewhere, the sound pressure level generally differs from the numerical phon value. A low-frequency tone, for example, usually requires a substantially higher sound pressure level than a 1,000 Hz tone to sound equally loud.
The phon should be distinguished from the sone, a unit intended to express loudness on a ratio scale. A sound judged twice as loud has twice as many sones, but it does not have twice as many phons. Under conventional reference conditions, 40 phons corresponds to one sone; above this region, an increase of about 10 phons is commonly associated with a doubling of loudness. This is an approximate perceptual relationship, not a universal conversion for every sound.
Shape and interpretation
Equal-loudness contours are normally plotted with frequency on a logarithmic horizontal axis and sound pressure level on the vertical axis. Each curve represents one loudness level.
Their main features are:
- Reduced low-frequency sensitivity. Bass tones generally require higher sound pressure levels than mid-frequency tones to sound equally loud.
- High sensitivity in the upper midrange. The contours usually reach their lowest levels in the region of a few kilohertz, often around 3–4 kHz.
- Reduced sensitivity at very high frequencies. The required level generally rises again toward the upper end of the audible range.
- Level-dependent shape. At higher loudness levels, the contours become flatter, especially across low and middle frequencies.
The last feature is particularly important. The auditory system does not behave like a fixed frequency-response filter whose output simply scales with input level. At quiet listening levels, bass can seem relatively weak; at higher levels, it becomes more prominent relative to midrange sound.
The flattening is not complete, and the contours are not uniformly spaced. Their precise shape varies with frequency, loudness level, experimental method, and listener population.
Relationship to the hearing threshold
The absolute threshold of hearing describes the minimum detectable sound level at each frequency under specified conditions. It has a broadly similar frequency dependence to low-level equal-loudness contours, but detection and equal-loudness matching are different experimental tasks.
A threshold curve therefore should not automatically be treated as an ordinary equal-loudness contour. Nor does a loudness level of zero phons mean that sound pressure is zero.
Auditory basis
The contours reflect the combined effects of sound transmission through the outer and middle ear and processing within the inner ear and nervous system.
The head, outer ear, and ear canal modify the sound reaching the eardrum in a frequency-dependent way. The middle ear transmits vibration to the cochlea, where mechanical and neural processes further shape sensitivity. In particular, cochlear amplification and compression contribute to the nonlinear relationship between physical level and auditory response.
Equal-loudness contours describe the overall perceptual outcome of these processes. They are not direct measurements of any single anatomical structure, and their shape cannot be explained by ear-canal resonance alone.
Measurement and standardization
An equal-loudness experiment typically presents a reference tone and a comparison tone at another frequency. The comparison level is adjusted until the listener judges the two sounds equally loud. Repeating this procedure across frequencies and reference levels produces a family of contours.
Experiments differ in several consequential ways:
- whether sounds are presented through loudspeakers or headphones;
- whether one or both ears receive the sound;
- the direction and acoustic field of presentation;
- tone duration, timing, and comparison procedure;
- participants’ age and hearing characteristics;
- equipment calibration and background noise.
Results from different arrangements are not automatically interchangeable. For example, a frontal free-field loudspeaker presentation includes acoustic effects of the head and outer ears that differ from those of headphone presentation.
ISO 226
The international standard ISO 226, Acoustics—Normal equal-loudness-level contours, specifies a standardized family of contours for pure tones under defined free-field listening conditions. These represent normal hearing in a specified population of young listeners, rather than the response of every individual. The standard was revised in 2003 and again in 2023.
Its intended conditions include binaural listening to a frontal sound source in a free progressive sound field. Applying its contours directly to other conditions, especially headphone reproduction, requires attention to differences in sound delivery and reference level.
The contours are population-based estimates. Individual differences and measurement uncertainty remain important, particularly near the limits of hearing.
Historical development
Harvey Fletcher and Wilden A. Munson published influential equal-loudness measurements in 1933. Their work established the familiar family of frequency-dependent loudness curves and helped make the distinction between physical sound level and perceived loudness central to audio engineering.
D. W. Robinson and R. S. Dadson published another major set of measurements in 1956. These subsequently influenced international standardization. Later research identified substantial differences between some earlier contours and newer experimental results, particularly at low frequencies, contributing to the 2003 revision of ISO 226.
The expression Fletcher–Munson curves is often used informally for equal-loudness contours in general. Strictly, it denotes the historical Fletcher–Munson results, not every later standardized contour.
Applications
Audio reproduction and loudness compensation
Equal-loudness contours explain why reducing playback volume can alter the perceived balance of music. Bass commonly becomes less prominent relative to the midrange, even if the reproduction system’s physical frequency response remains unchanged.
A loudness-compensation control applies level-dependent equalization, typically increasing low-frequency output relative to the midrange during quiet playback. Differences between equal-loudness contours provide a starting point for such compensation.
However, a contour is not itself an equalizer setting. Appropriate compensation also depends on the original and reproduced listening levels, loudspeaker and room response, and the spectrum of the programme material.
Sound measurement and frequency weighting
Equal-loudness research helped inform frequency-weighting systems used in sound measurement. A-weighting broadly reflects the reduced sensitivity of hearing to low-frequency sound at relatively low levels; C-weighting has a much flatter response across much of the audible range.
These standardized filters are not exact inverses of contemporary equal-loudness contours. A single A-weighted or C-weighted level also does not fully specify perceived loudness, because sounds with different spectra and temporal structures can produce different loudness judgments despite having the same weighted level.
Perceptual models and sound design
The contours provide reference information for auditory research, sound design, and the evaluation of acoustic signals. More comprehensive loudness models extend beyond pure-tone comparisons to account for the distribution of sound across auditory frequency bands and its variation over time.
Limitations
Equal-loudness contours describe particular stimuli and listening conditions, not a universal transformation from a sound spectrum into perceived loudness.
For complex sounds, several additional effects matter:
- Spectral interaction: components can combine perceptually or obscure one another through auditory masking.
- Bandwidth: distributing acoustic energy across different auditory frequency regions can change overall loudness.
- Duration and timing: brief, sustained, and fluctuating sounds need not have the same loudness at the same measured level.
- Individual variation: age and hearing characteristics can alter sensitivity and loudness growth.
- Spatial presentation: source direction, acoustic environment, and headphone coupling affect the sound reaching the ears.
Standard contours also do not establish whether a sound exposure is safe. Perceived loudness and the potential effects of exposure are related but distinct matters; frequency-dependent loudness matching is not a substitute for exposure assessment.