Psychoacoustics: Why Your Ear Doesn't Hear Sound Linearly

The human ear doesn't behave like a microphone. Two different frequencies at the exact same sound pressure level (SPL) can be perceived at very different loudness. Understanding this isn't just academic curiosity — it explains why a mixing engineer makes EQ and level decisions that are more aggressive at certain frequencies and more cautious at others.

The Fletcher-Munson Curves

Experiments run by Harvey Fletcher and Wilden Munson in 1933 mapped how much SPL different frequencies need to be perceived as 'equally loud'. The result: the human ear is evolutionarily most sensitive around the 2-5 kHz range — which happens to overlap heavily with human speech. Sounds in that range are perceived as far louder than bass or very high frequencies at the same SPL.

The practical mixing consequence: at low listening levels, bass and very high frequencies sound weaker to the ear, which is why many engineers reference-check a mix at multiple volumes (very quiet, medium, loud) — a mix judged at only one level can sound unbalanced in other listening environments.

The Bark and ERB Scales

The Bark scale, developed by Eugen Zwicker, divides the audible spectrum into 24 'critical bands' that closely mirror the ear's actual resolution — narrow bands at low frequencies, wider ones at high frequencies, modelling the ear's superior discrimination at the low end. The ERB (Equivalent Rectangular Bandwidth) scale, developed by Brian Moore and Brian Glasberg, is a mathematically more refined version and forms the basis of modern audio compression algorithms, including MP3.

• ―Reference-check a mix at more than one listening level, not just one.
• ―Small changes in the 2-5 kHz range read as disproportionately large to the ear — treat that band with care.
• ―Very low and very high frequencies nearly disappear at quiet listening levels.