Temporal Masking

A loud hit casts a time-shadow: it masks quiet sounds ~5-20 ms before (backward) and up to ~100-200 ms after (forward); sounds outside the window survive.

What it is

A loud sound hides a quieter sound that happens just before or just after it in time.

Key facts

• Two types: FORWARD (post) masking hides sound AFTER the loud one; BACKWARD (pre) masking hides sound BEFORE it.
• Forward masking lasts up to ~100-200 ms after the loud sound ends; effect decays roughly with the log of the time-gap.
• Backward masking is short and weak: only ~5-20 ms before the loud sound - the ear barely 'sees the future'.
• Mechanism: forward = cochlear/hair-cell recovery time; backward = the later loud signal overtakes the earlier quiet one in the auditory nerve.
• Masking is strongest when masker and maskee are close in frequency (same critical band ~1/3 octave) AND close in time.
• Rule of thumb: a transient ~20-30 dB louder buries a quiet detail sitting inside its temporal masking window.
• +10 dB = perceived 'twice as loud'; +6 dB = double sound pressure (×2 amplitude); +3 dB = double power. Minus the same = half.
• Speed of sound ~343 m/s at 20 C (~1 ms per 0.34 m), so echoes arriving inside the masking window get hidden.
• Human hearing ~20 Hz to 20 kHz; threshold ~0 dB SPL at 1 kHz; pain ~120-130 dB SPL.
• MP3/AAC codecs are BUILT on temporal + simultaneous masking - they throw away data the masking window proves you cannot hear.

How it works

1. A loud transient (the masker) hits your ear, e.g. a snare or kick.
2. Just before it (~5-20 ms): backward masking briefly hides any quiet sound = the 'pre-shadow'.
3. During the hit: simultaneous masking buries anything quieter at nearby frequencies.
4. Just after it (up to ~100-200 ms): forward masking hides quiet sounds while the ear recovers = the 'post-shadow'.
5. Closer in time + closer in frequency = stronger masking; further away = the quiet sound re-emerges.
6. Net result: soft detail glued next to a loud transient may never be heard, so don't waste effort polishing it.

Real examples

• Snare buzz/rattle: the soft snare-wire ring just before a hard hit is backward-masked - inaudible, don't gate it out painstakingly.
• Kick drum: the beater click hides the room rumble that follows for ~100 ms (forward masking).
• Gated reverb: cutting the tail early works because forward masking already hid the quietest part - the ear won't miss it.
• Cymbal crash burying a quiet hi-hat tick that lands a few ms later.
• MP3 'pre-echo' smear before a sharp transient = a codec artefact from mishandling backward masking.

How it helps in live sound

• Don't over-tweak quiet detail within ~150 ms of a loud transient - the crowd physically can't hear it. Save the gain and the time.
• Gate release: set ~50-150 ms on drums; forward masking covers the cut so it sounds natural, not choppy.
• Tame ringing/feedback that sits just AFTER a transient last - it's often already masked; chase the exposed ones first.
• Delay/echo taps landing inside ~100-200 ms of the dry hit get masked - push tap times further out so reflections are actually heard.
• Ducking music under a voice: you don't need a razor-fast release; forward masking hides the music returning for a moment anyway.
• Prioritise EQ/level on sounds in the GAPS between transients - that's where the ear is recovered and listening.