An equal loudness curve is a line on a graph connecting all the frequency-and-level combinations that a human listener perceives as equally loud. The most striking thing these curves reveal is that your hearing sensitivity is not flat: a 30 Hz bass tone needs far more physical energy to sound as loud as a tone at 3 or 4 kHz, where your ear canal and middle ear are most efficient. The curves flatten out at high playback levels, which is why cranking up the volume on a stereo makes bass and treble feel like they “fill in.” The international reference for these curves, ISO 226, was most recently updated in 2023, and the science behind them reaches into everything from headphone design to hearing-aid fitting.
What the Curves Actually Show
Picture a standard graph where the horizontal axis represents frequency, running from about 20 Hz on the left to 20,000 Hz on the right, and the vertical axis represents sound pressure level in decibels. An equal loudness curve traces a single level of perceived loudness across that entire frequency range. At 1 kHz the curve is anchored: a 40-phon curve passes through exactly 40 dB SPL at 1 kHz, because the phon scale is defined that way. Move left toward lower frequencies, and the curve climbs steeply, meaning the speaker has to push out a much higher sound pressure level for that bass tone to sound just as loud. Move right toward very high frequencies, and the curve rises again, though less dramatically. In the range from roughly 2 to 5 kHz, the curves dip, meaning less energy is needed there for the same perceived loudness.
Multiple curves are drawn on the same chart, each labeled with a different phon value. The lowest curve represents the threshold of hearing, the quietest level at which tones can just barely be detected, and it is deeply U-shaped. Higher curves, at 40, 60, 80, and 100 phons, progressively flatten out. At very high levels the ear’s sensitivity differences across frequency become smaller, so a 100-phon curve looks nearly flat compared to the threshold curve. This flattening is the reason a song’s bass line seems to disappear when you turn the volume way down but punches through clearly at high levels.
Why Your Ear Favors Certain Frequencies
The dip in the curves between about 2 and 5 kHz is not a quirk of perception alone; it is largely a mechanical effect. The ear canal acts like a short resonant tube, boosting sound pressure near its resonant frequency, typically around 3 kHz. That boost means less external energy is needed for tones in that range to excite the eardrum to a given amplitude.
Behind the eardrum, the tiny bones of the middle ear add their own frequency-dependent behavior. Under normal air-conducted listening, the dominant resonance mode of the ossicular chain occurs around 0.8 to 1.2 kHz, characterized by a hinging motion of the bones. A second mode appears around 1.7 kHz with a different, pivoting motion, and this mode dominates during bone conduction instead.1PubMed Central. Ossicular resonance modes of the human middle ear for bone and air conduction Together, the ear canal resonance and middle ear mechanics create a broad sensitivity peak that accounts for most of the dip you see in the equal loudness curves between 1 and 5 kHz. At the low-frequency end, the stiffness of the middle ear system makes it increasingly hard to transmit energy efficiently, which is why bass tones need so much more power to sound equally loud.
The International Standard and Its Revisions
The first systematic measurements of equal loudness contours were published in 1933 by Harvey Fletcher and Wilden Munson at Bell Telephone Laboratories, which is why these curves are sometimes still called “Fletcher-Munson curves.” Those original measurements, though groundbreaking, had methodological limitations, and by the 1950s the International Organization for Standardization adopted a different data set collected by D.W. Robinson and R.S. Dadson in the UK. That data formed the basis of the first ISO standard, ISO 226.
By 1985, researchers had identified significant errors in the Robinson-Dadson data, particularly at low frequencies, and a lengthy revision process began. It took 18 years, but in 2003 ISO 226 was completely overhauled using new, pooled data from multiple laboratories around the world. Twenty years after that, the standard was updated again as ISO 226:2023. The 2023 revision refined the mathematical expressions to preserve appropriate significant digits and incorporated a slightly revised threshold of hearing at 20 Hz, lowered by 0.4 dB to match updated audiometric reference values. It also adopted an updated power exponent relating perceived loudness to physical intensity. Despite all this, the practical differences from the 2003 edition are only about 0.6 dB at most, meaning engineers and audiologists relying on the 2003 version need not worry about meaningful discrepancies.2J-STAGE / Acoustical Science and Technology. Revision of ISO 226 “Normal Equal-Loudness-Level Contours” from 2003 to 2023 edition: The background and results
What This Means for Everyday Listening
The practical consequence most people encounter without realizing it is that music and speech sound tonally different at different playback levels. At low volume, bass and extreme treble are perceptually suppressed because those frequencies sit on the steep flanks of the equal loudness curves. Vocals and lead instruments in the 2 to 5 kHz range, where the ear is most sensitive, remain relatively prominent. Turn the volume up and the curves flatten, so the bass and high-frequency details re-emerge and the mix sounds fuller and more balanced.
This is why many stereos and headphone apps include a “loudness” button or a bass-boost setting intended for quiet listening. That function applies a frequency-dependent gain curve, roughly the inverse of the equal loudness contour at the playback level, adding extra bass and a small treble lift so the music retains its tonal balance even at low volumes. Without it, night-time listening at reduced volume tends to sound thin and midrange-heavy.
Audio engineers mixing recorded music face this problem from the other direction. A mix that sounds balanced at loud monitoring levels in a studio may lose its low end when a consumer plays it softly on laptop speakers. Experienced mixers routinely check their work at multiple volume levels for exactly this reason. Room acoustics play a role too: timbral balance and loudness perception are primarily shaped by how the playback system is calibrated, while spatial qualities depend more heavily on the room itself.3PubMed Central. Hearing as intended: How differences in listening conditions affect sound translation This means even a perfectly calibrated studio can mislead if the engineer only monitors at one loudness setting.
How Duration Changes What You Hear
Equal loudness curves are typically measured with sustained pure tones, but real-world sounds are rarely steady. When a sound is very brief, your brain integrates its energy over time, and the relationship between physical level and perceived loudness shifts. Short tones generally need more intensity to sound as loud as longer tones, a phenomenon called temporal integration of loudness. However, the size of this effect depends on the listening level. Research measuring equal loudness for 5 kHz tones of varying durations found that the level difference between equally loud short and long tones was largest at moderate levels, not at very low or very high levels.4The Journal of the Acoustical Society of America. Temporal integration of loudness, loudness discrimination, and the form of the loudness function
This matters in realistic sound environments. Helicopter rotor blades, for example, produce pulses of sound that vary in duration and repetition rate. Studies of helicopter-like sounds found that at the same average sound pressure level, shorter pulses were perceived as louder than longer ones, because the energy was compressed into a briefer window.5INTER-NOISE and NOISE-CON Congress and Conference Proceedings. Temporal integration of partial loudness of helicopter-like sounds This is one reason standard equal loudness curves, derived from steady tones, are a starting point rather than a complete description of how loud the real world sounds. Noise regulations and product-sound design often need more complex loudness models that account for temporal patterns.
Temporal integration also shows up in people who use cochlear implants. When researchers tested implant users with pulse trains ranging from a single pulse to 300 milliseconds, the amplitude needed to reach the same loudness decreased steadily as the stimulus got longer, and the effect was stronger at higher pulse rates.6PubMed Central. On the Effect of High Stimulation Rates on Temporal Loudness Integration in Cochlear Implant Users Understanding this helps clinicians program implant processors so that rapid speech sounds are not perceived as uncomfortably loud.
Broadband and Complex Sounds
Equal loudness curves are built from pure tones, single frequencies presented one at a time. Most sounds you actually hear are broadband, containing energy spread across many frequencies simultaneously. When a sound covers a wide bandwidth, individual frequency components excite different regions of the cochlea, and the brain adds their loudness contributions together. This spectral loudness summation means broadband sounds are perceived as louder than a pure tone at the same overall sound pressure level.
Reaction-time experiments, which use how quickly a listener responds as a proxy for perceived loudness, confirm this pattern. At moderate and high levels, broadband pink noise produced faster reactions than a pure tone at the same level, consistent with summation across frequency bands. At low levels, though, the interaction between bandwidth and level produced results that diverged from standard loudness-summation models, suggesting the brain handles quiet broadband sounds somewhat differently than it handles loud ones.7Attention, Perception, & Psychophysics. Simple reaction time for broadband sounds compared to pure tones Binaural listening, hearing through both ears, added about 9 dB of effective gain for white noise in the same experiments, somewhat more than had been found for single-frequency tones.
For anyone working in noise control, these findings matter. Predicting how loud a machine or vehicle sounds to a bystander requires models that go beyond single-frequency contours. Standards like ISO 532 build on the equal loudness framework by breaking a complex sound into critical bands, applying frequency-dependent weighting drawn from the contours, and summing the contributions. The equal loudness curve is the foundation, but real-world loudness is a multi-story building on top of it.
How the Curves Shift with Age
Equal loudness curves are measured using young adults with clinically normal hearing, and the ISO standard explicitly defines its contours for that population. As people age, high-frequency hearing sensitivity typically declines, a condition called presbycusis. That means the steep high-frequency rise in the curves shifts even higher for older listeners, and the comfortable listening range narrows.
Research on rapid loudness profiling in older adults has measured individualized equal loudness contours at 20, 40, 60, 80, and 100 phons across multiple frequencies.8PubMed Central. Rapid Profiling of Loudness Among Older Adults These personalized curves let clinicians see exactly where an individual’s perception deviates from the standard, which is valuable for hearing-aid fitting. A person who has lost sensitivity mainly above 4 kHz will have curves that look roughly normal at lower frequencies but rise sharply in the high end, meaning their aid needs more gain only in that region. Without personalized contour data, a one-size-fits-all approach risks over-amplifying frequencies where perception is still intact.
Loudness Recruitment and Hearing Loss
One of the most uncomfortable consequences of sensorineural hearing loss is a phenomenon called recruitment. In a healthy ear, loudness grows gradually as sound pressure increases. In an ear with damaged hair cells, quiet sounds may be inaudible, but once the level crosses the elevated threshold, loudness can ramp up abnormally fast, catching up to normal perception at moderate-to-high levels. The subjective experience is that soft sounds are missing and moderate sounds are too loud, with a very narrow comfortable range in between.
Animal studies have confirmed that recruitment is not unique to humans. Cats exposed to intense sound developed hearing loss, and their subsequent behavioral loudness estimates showed recruitment effects similar in magnitude to those reported in hearing-impaired people.9Journal of the Association for Research in Otolaryngology. Loudness perception in the domestic cat: reaction time estimates of equal loudness contours and recruitment effects This parallel supports the idea that recruitment stems from the loss of outer hair cells in the cochlea, which normally act as a biological compressor, amplifying soft sounds and limiting loud ones. When those cells die, the compression disappears and the growth of loudness with level becomes abnormally steep. In terms of equal loudness curves, recruitment means the curves at low phon levels are pushed upward (more energy needed to reach audibility), while the curves at high phon levels remain nearly normal, compressing the vertical space between them.
Equal Loudness Across Species
Humans are not the only animals whose hearing sensitivity varies with frequency. Audiograms from mammals commonly used in auditory research show qualitatively similar U-shaped sensitivity curves, just shifted along the frequency axis. Cats and chinchillas, for instance, hear much higher frequencies than humans, but the basic shape of their threshold curve, with a region of best sensitivity and reduced sensitivity at the extremes, mirrors the human pattern. A broad comparative review of psychoacoustic data across unspecialized mammals concluded that the qualitative forms of nearly all psychoacoustic functions are similar across species, suggesting that the underlying mechanisms are shared.10Hearing Research. Comparative psychoacoustics
From an evolutionary perspective, the frequency range where human hearing is most sensitive aligns well with the frequencies most important for speech. Research on primate hearing suggests that auditory thresholds compatible with optimal speech reception were likely present in the last common ancestor of humans and chimpanzees, before the lineages split. That implies our ear’s sensitivity profile was not specifically shaped by the demands of spoken language but was already in place for other reasons, possibly predator detection or social calls, and language later took advantage of the existing hardware.11Scientific Reports. Auditory thresholds compatible with optimal speech reception likely evolved before the human-chimpanzee split
When What You See Changes What You Hear
Loudness perception is not purely auditory. The brain integrates information across senses, and visual input can measurably alter how loud a sound seems. In experiments where a burst of white noise was presented either alone or accompanied by a flash of light, listeners consistently rated the noise as louder when the light was present. Importantly, this effect persisted even when the experimenters varied how often the light appeared alongside the sound, and it held up under forced-choice comparisons rather than just subjective ratings. The researchers concluded that the light-induced loudness enhancement likely reflects a genuine early-stage sensory interaction, not just a response bias.12PubMed. Brighter noise: sensory enhancement of perceived loudness by concurrent visual stimulation
The influence runs the other direction too. Predictions about what you are about to see can sharpen your ability to detect loudness differences. When participants performed an action that was supposed to produce a specific visual result, and that result was unexpectedly replaced by a different visual stimulus, their sensitivity to small differences in the loudness of an accompanying tone improved.13PubMed Central. The prediction of visual stimuli influences auditory loudness discrimination The mismatch between expectation and reality apparently heightened auditory attention. These crossmodal effects are small under laboratory conditions, but they hint at why loudness in daily life can feel so context-dependent. Watching a loud movie in a bright room versus a dark theater, or hearing a notification ping while glancing at your phone versus staring out the window, may involve slightly different loudness percepts for the same physical sound. The equal loudness curve describes how the ear responds to isolated tones in a quiet booth, but the brain that interprets those signals is never really operating in isolation.