8. Psychoacoustics (Perception Layer) · Concept 1 of 18
Fletcher–Munson Curves
Charts showing that your ears do not hear all pitches as equally loud, especially the low and very high ones.
Quiet listening = steep red U (bass and top vanish); turn it up and the curve flattens, so the mix sounds full again.
What it is
Graphs showing your ears need way more level on bass and very high notes to hear them as loud as the midrange.
Key facts
Original study: Harvey Fletcher & Wilden Munson, Bell Labs, 1933 — mapped equal-loudness contours
Modern standard is ISO 226:2003 (revised 2023), which replaced the original 1933 data — use this name with engineers
Unit = the phon. Loudness in phons is DEFINED as the dB SPL of an equally-loud 1 kHz tone
Reference: 0 dB SPL = 20 micropascals (20 uPa) = threshold of hearing at 1 kHz
Ear is most sensitive 2-5 kHz (ear-canal resonance ~3-4 kHz); curves dip lowest there
At the 40 phon contour a 20 Hz tone needs roughly +50 to +70 dB MORE SPL than 1 kHz to sound equally loud
Curves FLATTEN as you get louder — the bass/treble penalty shrinks at 80-100 phon vs quiet levels
Perceived-loudness doubling ≈ +10 dB; power doubling = +3 dB; SPL/voltage doubling = +6 dB
Hearing range 20 Hz-20 kHz; dynamic range ~120-130 dB; pain threshold ~120-130 dB SPL
A-weighting (dBA on every SPL meter) is the inverse of the ~40 phon curve — it rolls off bass like your ears do
How it works
Play a 1 kHz reference tone at a set dB SPL.
Play a second tone at a different pitch; ask louder or quieter?
Listener adjusts the second tone until both sound equally loud.
Plot that matched SPL for every frequency = one curve (one phon value).
Repeat at many reference levels to draw the whole family of contours.
Result: U-shaped curves — high SPL needed at bass and top, dipping low around 3-4 kHz.
Real examples
Quiet hotel-lobby BGM: bass and air go missing, so you crank it and suddenly the bass is huge.
Car stereo at low volume sounds flat; at highway speed/volume it sounds full.
Mastering engineers check mixes quietly AND loud because tonal balance shifts with level.
A mix EQ'd to sound perfect at 100 dB goes thin when you drop it 10 dB for the dinner set.
How it helps in live sound
Tune your system at REAL show SPL (often ~95-100 dBA), not at quiet soundcheck level.
Dialled it in loud then ran support quiet? Expect to add low-end and a touch of top back.
A 'flat' RTA does NOT mean a flat-SOUNDING mix — your ears aren't flat.
For background-music / dinner sets, gently shelve up lows and highs to fight the steep quiet curve.
Reference your tonal balance against a known track at the SAME SPL you'll mix the show at.
Mind the 3-4 kHz region — most sensitive, so harshness and feedback bite hardest there.
Everyday analogy
Like a stereo turned down low at night where the bass and sparkle vanish but the vocals stay clear — that disappearing act is your ears, not the speakers.
Watch out
Myth: 'flat on the analyser = flat to the ear.' Wrong — ears follow the curves, so a measured-flat system sounds bass-light when quiet and needs level-dependent tonal judgement.
Fun fact
A-weighting (the 'dBA' on every SPL meter) is literally the inverse of the ~40 phon Fletcher-Munson curve — the meter deliberately ignores bass because your ears do too.
Key takeaways
Ears aren't flat: bass and very-high pitches need far more level to sound equally loud.
Measured in phons; phons = dB SPL at 1 kHz by definition.
Curves flatten as you get louder — that's why quiet = thin, loud = full.
Ear peak sensitivity is 2-5 kHz; modern standard is ISO 226, not the 1933 data.
Tune and reference your mix at the actual show SPL or the tonal balance will be wrong.