8. Psychoacoustics (Perception Layer) · Concept 15 of 18
Duplex Theory
The rule that your brain uses two different clues to locate sound depending on whether it is low or high.
Brain switches cues by pitch: timing (ITD) places lows, loudness (ILD) places highs - so subs sum to mono, highs pan wide.
What it is
Your brain locates LOW sounds by timing (which ear hears first) and HIGH sounds by loudness (which ear hears louder).
Key facts
Speed of sound in air = 343 m/s at 20 C (about 1 ms per 34 cm).
Two cues for horizontal placement: ITD (Interaural Time Difference) and ILD (Interaural Level Difference).
ITD = arrival-time gap between ears; dominant BELOW ~800 Hz.
ILD = loudness gap between ears (head shadow); dominant ABOVE ~1.5-2 kHz.
Max ITD is about 650-700 microseconds (head ~21-23 cm wide).
Duplex Theory: Lord Rayleigh, 1907 - timing for lows, level for highs.
Crossover/confusion zone ~800 Hz-1.6 kHz: both cues weak, localisation worst.
Wavelength = 343 / frequency: 100 Hz = 3.43 m, 1 kHz = 0.34 m, 10 kHz = 3.4 cm.
Head shadow at high freq creates a 10-20 dB level gap; lows just diffract around with near-zero ILD.
Human ITD just-noticeable difference ~10-20 microseconds (about 1-2 degrees of angle).
How it works
A sound off to one side reaches the near ear first - that gap is the ITD.
For LOW pitches (long waves) the brain measures this timing gap to place it.
The same sound is quieter at the far ear because the head blocks it - that gap is the ILD.
For HIGH pitches (short waves) the brain measures this loudness gap to place it.
Low waves wrap around the head, so they make almost no loudness gap - timing wins.
High waves can't wrap, so timing gets phase-confused - loudness wins.
Real examples
Pan a 60 Hz kick hard left and it still feels centred and vague - lows give no loudness cue.
Pan a hi-hat or shaker and it snaps sharply to one side - highs give a strong loudness cue.
Mono summing bass below ~120 Hz costs you nothing locational because ears can't place it anyway.
A speech 'sss' is pinpointed (high), the 'mmm' floats (low) - same voice, two cues.
Subs placed off to one side still 'sound centred' to the crowd - duplex theory in action.
How it helps in live sound
Sum everything below ~120 Hz to mono / centre subs - you lose no perceived position and gain headroom and punch.
Reserve hard L/R panning for HF content (cymbals, percussion, FX) where ILD makes it read clearly.
Sub placement is about coverage and comb-filtering, not stereo image - crowd can't localise it anyway.
Don't fight to 'widen the bass' in PA - physics won't let the audience place lows; widen 2 kHz+ instead.
Mind the 800 Hz-1.6 kHz mud zone: localisation is weakest here, so it smears - keep it tidy in the mix.
For stereo IEMs the same rule holds: timing/phase for lows, level/pan for highs.
Everyday analogy
Like finding someone in thick fog: you place their low rumble by which ear hears it a hair sooner, but their high hiss by which ear hears it louder.
Watch out
Myth: 'pan the bass to widen the mix.' Wrong - below ~800 Hz the ear uses timing, not loudness, so panning bass just unbalances levels without moving its perceived position. Sum lows to mono.
Fun fact
Your brain can resolve an interaural time difference of about 10 microseconds - roughly one-hundredth of a millisecond - which is finer than the timing of a single audio sample at 44.1 kHz.
Key takeaways
Two cues, split by pitch: ITD (timing) for lows, ILD (loudness) for highs.
Crossover sits around 800 Hz-1.6 kHz where both cues are weak.
Cause is wavelength vs head size: long waves diffract, short waves get shadowed.