Turn down your monitors by 20 dB and the bass seems to vanish. The snare sounds harsh. The low end that was perfectly balanced now sounds thin. Put the volume back up and it sounds fine again. This is not your imagination, room acoustics, or a quirk of your speakers. It is a documented, measurable property of human hearing, and it shapes every mixing decision you make.
The mechanism is called the equal loudness contour. Understanding it explains why studio monitors are calibrated at specific levels, why the "loudness" button exists on hi-fi receivers, and why some engineers still check their mixes at very low volumes before calling them done.
The 1933 Experiment
Harvey Fletcher and Wilden Munson were engineers at Bell Laboratories when they published "Loudness, Its Definition, Measurement and Calculation" in the Journal of the Acoustical Society of America in October 1933 (Volume 5, Issue 2). Their experiment was straightforward: present listeners with a pure tone at a test frequency, then ask them to adjust a 1 kHz reference tone until both sounded equally loud. Repeat at dozens of frequencies and levels. Plot the results.
They found that the human ear does not have a flat frequency response. It is more sensitive in some regions than others, and that sensitivity profile changes depending on how loud the sound is. The curves they plotted became known as equal loudness contours, or colloquially, the Fletcher-Munson curves.
What the Curves Actually Show
Each contour on a Fletcher-Munson graph represents a set of (frequency, SPL) pairs that all sound equally loud to an average listener. The unit for this perceived loudness level is the phon. By definition, the phon value equals the SPL in dB at 1 kHz, so a 40-phon contour passes through 40 dB SPL at 1 kHz and then rises and falls at other frequencies to show how much more or less SPL is needed to achieve the same perceived loudness.
The picture that emerges is striking. At 40 phons, a tone at 100 Hz requires roughly 52 dB SPL to sound as loud as a 1 kHz tone at 40 dB SPL. That is about 12 dB more acoustic energy for the same perceived loudness. At very high frequencies above 10 kHz, sensitivity also drops off significantly. The effect is most extreme at very low frequencies and near the edges of the audible range.
The curves do not stay the same shape as loudness increases. As SPL rises, the contours progressively flatten. At around 100 phons, hearing becomes substantially more uniform across the audible spectrum. The ear is most non-linear at quiet listening levels and most linear at loud ones. The gap in SPL required between bass frequencies and midrange narrows as listening level increases.
A Flawed Original, Improved Over Decades
Fletcher and Munson used headphones for their measurements. This is technically significant because headphone listening creates a different acoustic environment than listening in a room, primarily through the absence of room reflections and the diffuse-field effects of speaker-to-ear acoustic transfer. Their results were widely influential but later questioned.
In 1956, D. W. Robinson and R. S. Dadson at the UK National Physical Laboratory published a new set of measurements using loudspeakers in a free field. These Robinson-Dadson curves became the basis for the first ISO standard (ISO 226) and were accepted internationally for decades.
By the 1980s, errors in the Robinson-Dadson data were identified. A major international measurement effort assembled results from 12 studies conducted in multiple countries beginning in the mid-1980s. After 18 years of revision work, ISO 226:2003 replaced the Robinson-Dadson baseline. The differences were significant below 500 Hz, with discrepancies as large as 14 dB in some regions.
ISO 226:2023 is the current edition. The changes from 2003 are minor (no more than 0.6 dB in any region) and mainly involve mathematical precision in the underlying formulas and a small adjustment to the threshold of hearing at 20 Hz. For practical mixing purposes, the 2003 and 2023 editions are equivalent.
Why 85 dB SPL Became the Studio Reference
Professional control rooms and film mixing stages calibrate their monitoring levels to 85 dB SPL (sometimes 83 dB for film and broadcast mixing). This did not happen arbitrarily. At around 85 dB SPL, the equal loudness contours are at their flattest across the audible range. Hearing is more uniformly sensitive to bass, midrange, and treble simultaneously than at any lower monitoring level.
Mixing at 85 dB SPL means that frequency-balance decisions made during a session correspond more closely to the actual acoustic energy in the recording. Bass adjustments at 85 dB predict reasonably well how the bass will sit at other listening levels because the contours are at their most linear. Mixing at 70 dB SPL reintroduces the nonlinearity. The bass that sounds balanced at low volume will be excessive when the listener cranks it up.
This is also why the "loudness" or "loudness compensation" switch on consumer hi-fi amplifiers exists. When listening at low volume at home, the ear loses sensitivity to bass and treble relative to the midrange. A loudness circuit boosts low and high frequencies to compensate, pushing the perceived frequency balance closer to what the recording sounds like at higher levels. It applies the inverse of the equal loudness contour shift caused by turning down the volume.
A-Weighting Is an Approximation of the 40-Phon Contour
When acoustic engineers measure noise levels and report them in dB(A), the A-weighting filter they apply is designed to approximate the sensitivity of human hearing at moderate quiet levels. Specifically, A-weighting is based on the inverted 40-phon equal loudness contour. The filter emphasizes 3-6 kHz (where hearing is most sensitive) and attenuates very low and very high frequencies.
This makes A-weighting useful for measuring sounds at moderate loudness. It breaks down at loud levels because the 40-phon contour shape does not describe how hearing behaves at 85 or 100 phons. Regulatory standards often use dB(A) regardless of level because the alternative (switching between multiple weighting curves depending on SPL) is impractical for field measurement.
The Low-Volume Check
Some engineers deliberately check mixes at very low volumes before printing. At 50-60 dB SPL, the Fletcher-Munson effect becomes strong: bass frequencies drop in perceived loudness disproportionately, and subtle imbalances in the midrange become easier to hear. If a mix sounds full and clear at low volume, the low-frequency content is genuinely there as acoustic energy. If it sounds thin, the bass was probably being partially "imagined" by the ear at higher monitoring levels.
The reverse check also applies. After a long session at loud volume, hearing adapts. Stepping back and listening briefly at lower volume resets the adaptation and can reveal problems that had become inaudible after hours at the console.
LUFS and the Streaming Complication
Modern streaming normalization targets (Spotify at -14 LUFS integrated, Apple Music similarly) exist partly because of the equal loudness issue. Most streaming happens on earbuds or laptop speakers at moderate volumes, where the equal loudness effect is strong. A master that sounds balanced at 85 dB SPL in a studio may need different treatment when it will primarily be heard at 65 dB SPL through earbuds.
The K-weighting filter used by the ITU-R BS.1770 loudness algorithm (which defines LUFS) approximates human frequency weighting at typical listening levels by applying a high-shelf boost and high-pass filter below 40 Hz. It is related to but not identical to A-weighting: A-weighting targets 40 phons specifically, while K-weighting was designed for broadcast monitoring levels. The result is a loudness metric that correlates better with perceived loudness for the typical streaming listener than raw dB SPL.
The Practical Summary
Fletcher and Munson's 1933 finding still drives decisions every day in professional audio. The curves say: hearing is not a flat microphone. It is a system whose sensitivity profile changes with loudness in a way that makes bass and high frequencies harder to hear at low volumes and progressively easier at higher levels.
This means monitoring level is not a comfort preference. It is a calibration decision with direct consequences for frequency balance. It means the low-volume check is not superstition. It is exploiting the ear's increased bass insensitivity to verify that the low end is genuinely there. It means A-weighting is an approximation good enough for noise measurement but not a substitute for understanding the full curve.
And it means every mix you have ever made was shaped by your monitoring level, whether you intended it to be or not.