When the same sound reaches your ears twice very close together, your brain locks onto where the first copy came from and ignores the second for direction.
First wavefront wins direction: a delayed copy (even +10 dB louder) fuses in within the 1-40 ms window instead of pulling the image.
What it is
When the same sound hits your ears twice within ~1-40 ms, your brain locks onto the FIRST arrival for direction and ignores the second.
Key facts
Also called the Haas effect or 'law of the first wavefront' (Helmut Haas, 1949).
Active for inter-arrival delays of about 1 ms to 35-40 ms; under ~1 ms it's summing localisation, over ~40 ms it splits into a discrete echo.
Speed of sound in air = 343 m/s at 20 C, i.e. ~1 ms per 0.343 m of extra path (~1 ft per ms).
Delay (ms) = distance (m) / 343 x 1000. Symbols: distance = path length in metres, 343 = speed of sound in m/s, x1000 converts seconds to milliseconds.
Haas showed a delayed copy can be up to +10 dB LOUDER than the first and you STILL localise to the first source.
Doubling sound power = +3 dB; doubling pressure/voltage or two identical sources = +6 dB; perceived 'twice as loud' = roughly +10 dB.
Inverse-square law: sound pressure level drops 6 dB per doubling of distance from a point source.
Speed of sound rises ~0.6 m/s per degree C: c = 331.3 + 0.6 x T(C).
Comb filtering starts when arrivals are within roughly 0-1 ms (first null at f = 1/(2 x delay), delay in seconds).
Delay towers / rear fills are typically offset ~10-20 ms beyond the geometric distance delay to pull the image back to the stage.
How it works
Two near-identical sounds arrive at your ears a few milliseconds apart.
Your auditory system grabs the FIRST wavefront and reads its direction as the true source.
The second (delayed) copy is suppressed for localisation but still adds loudness and body.
Fusion: under ~40 ms the two are heard as ONE fuller sound, not an echo.
Past ~40 ms the brain un-fuses them and you hear a distinct, distracting echo.
Real examples
Delay towers at a festival: rear speakers fire a few ms LATE so the crowd still hears the band coming from the stage.
Under-balcony fills timed slightly behind the mains so the image stays forward.
Stereo widening: feeding one channel a 10-30 ms delay makes a sound feel bigger without moving its position.
A PA in a reflective gym: you localise the speaker, not the slap off the back wall.
Vocal 'thickening' with a short slap delay that fattens the voice without a heard repeat.
How it helps in live sound
Time-align delay speakers: distance to delay tower / 343 x 1000 = base delay in ms, then ADD ~10-20 ms Haas offset.
Keep the delayed speaker's level within ~+6 to +10 dB of the main arrival so fusion holds.
Use a measurement rig (Smaart, Open Sound Meter) to set delay times, don't eyeball it.
Watch the 1-40 ms window: under 1 ms you get comb filtering, over ~40 ms you get a slapback echo.
For front-fills and lip-sync, also account for video/processing latency so the image stays on the performer.
Cold night vs hot day shifts c by ~0.6 m/s per degree C, so re-check delay times if temperature swings a lot.
Everyday analogy
Someone calls your name across a courtyard and a wall throws back an echo a beat later, yet you still spin toward the person, not the wall.
Watch out
Myth: 'louder wins, so the rear speaker steals the image.' Truth: the FIRST arrival wins direction even if the later copy is up to +10 dB louder, as long as it lands inside the ~1-40 ms window.
Fun fact
A delayed copy can be a full 10 dB LOUDER than the original and your brain still insists the sound came from the first, quieter source.
Key takeaways
First wavefront wins direction; the later copy just adds fullness.
Magic window is ~1-40 ms; outside it you get combing or an echo.
Sound travels 343 m/s, so ~1 ms per 0.343 m of extra path.
Haas: later copy can be +10 dB louder and you still point at the first source.
In PA work, delay rear/fill speakers a few ms to glue the image to the stage.