Psychoacoustics for DJs: masking, stereo localisation and the Haas effect.

A song seeming less defined half an hour into a session is almost never the song’s fault: your hearing is doing its job as designed. Psychoacoustics is not an esoteric branch of music theory; it is the operating manual for the instrument you use daily.

This guide covers four phenomena explaining most poor booth and mixing-desk decisions: frequency masking, stereo localisation, precedence / Haas effect and volume adaptation. No memorised formulas: each becomes a recognisable symptom and a specific adjustment.

What psychoacoustics studies and why it affects you

Psychoacoustics studies how hearing converts air pressures into an experience of pitch, width, distance and perceived volume. That conversion is nonlinear: hearing interprets the world instead of measuring it like a level meter.

Three practical consequences appear in every session:

• Hearing compares rather than measures. A track sounds louder immediately after a quieter one, even when the meter says otherwise.
• Hearing has attention bands. Sounds in one frequency region compete for attention, with only one emerging reinforced.
• Hearing adapts. Within minutes, it compensates for the supplied level and accepts it as the reference.

A perception guide therefore complements level measurement: measurement describes the system, perception describes the listener. Both must agree.

Frequency masking

Frequency masking reduces one sound’s audibility because another has similar or neighbouring spectral content. Exact overlap is unnecessary: nearby bands suffice because hearing resolves sound into critical bands, where everything competes.

Four practical rules follow:

1. Masking is bidirectional but asymmetric. Low frequencies mask high ones more effectively than the reverse, so bass buildup smothers mids and treble without bass seeming “more present”.
2. It grows with level. Stronger masking spreads its influence: high volume hides more information than low volume.
3. Fight it by cutting, not only boosting. Before adding brightness, remove thickness from the competing element: this almost always cleans more and adds less noise.
4. It can help too. Masking covers remnants: room noise or hiss becomes less obvious when musical material occupies that band.

The test is simple. Mute the suspected element and restore it while listening to the rest; greater clarity upon removal indicates masking. Test quietly, because hearing conceals the problem at high volume. Instrument EQ applies the same logic: find space before adding.

Stereo localisation and precedence / Haas effect

Stereo localisation depends on more than pan. Hearing determines direction from two cues: arrival-time and level differences between ears. Reflections add arrival order: direct sound arrives first and leads.

This introduces precedence / Haas effect. Two versions of the same sound arriving close together fuse into one acoustic event, located towards the first arrival. Classic experiments place that fusion around 1–5 milliseconds for clicks and up to tens of milliseconds for more complex sounds. A reflection somewhat louder than the direct sound can still avoid being heard as an echo.

Three direct applications:

• Reinforcement delays. With delayed loudspeakers, the first sound determines the image and the delayed sound adds level: the audience still hears the front even when sound arrives from behind.
• Width without displacement. Millisecond offsets between channels change perceived width rather than direction, provided you remain below the echo threshold.
• The effect as tool and problem. A short delay widens; the same delay with an out-of-phase return cancels bass instead, exactly what the ear-training guide teaches you to recognise.

Beware a common trap: an excessively late delayed signal stops fusing and becomes a separate echo. The image splits, making the mix sound like a room instead of width.

Volume adaptation: why hearing stops noticing excessive levels

Volume adaptation readjusts the auditory system’s reference after a few minutes of exposure. Increasing volume flattens frequency response: previously weak bass and treble gain presence, and an initially excessive level becomes normal.

This answers the section’s question. Why hearing stops noticing excessive levels is physiological compensation rather than indiscipline. And it has two sides. And it has two sides:

• In mixing: Work loudly, hearing flattens response and everything seems balanced; lower volume and the mix seems thin, bright and lacking body.
• In the booth: After twenty minutes at a level you no longer notice, you keep increasing it; cumulative noise dose grows without the sensation of shouting.

Lowering volume also unmasks. Quietly, the ear’s filtering passes less bass, exposing body problems. The cheapest checking routine is therefore listening quietly for a while, returning to working level only to decide.

Practical booth and mixing decisions

With the four phenomena understood, the procedure becomes routine:

1. Set and respect a reference level. Keep listening volume similar, letting measurement supply the certainty hearing no longer provides.
2. Order decisions. Levels and spectral space first, colour afterwards. Without resolving masking, added saturation sounds like more of the same.
3. Check at three volumes. High for energy, medium for balance and low for residual bass and remnants.
4. Listen in mono occasionally. What disappears in mono relied on phase rather than content.
5. Be generous with breaks. Every ten or fifteen minutes, a few quiet seconds restore the reference and prevent unnoticed increases.

These decisions accompany how much level a sound system allows, discussed separately in measurement and gain.

Costly perception mistakes

The four phenomena produce familiar mistakes:

• Adjusting by ear at shouting level. The quickest route to a mix working only at dancefloor volume.
• Raising levels to compete. Increasing an inaudible element worsens masking; reduce its rival instead in most cases.
• Confusing width and volume. Width creates apparent presence without an actual increase, evaporating in mono.
• Ignoring the echo threshold. Widening delays begin splitting the image while you still think it is width.
• Assuming the level. Without a fixed reference, it moves with the session.

Equipment purchases do not fix these errors. Understanding hearing and checking with measurement does. That is the next guide’s work.

Next step

Turn these rules into listening skills: recognise frequencies, compression, clipping, distortion, phase, reverb and delay without looking at the meter. Continue with ear training: frequencies, compression, distortion and BPM.