Can Your Ears Tell Left from Right?

weirdly accurate.

    Can Your Ears Tell Left from Right?

    Can your ears tell left from right?

    Yes, and far more finely than most people expect. A sound from one side reaches the far ear a little later, and young adults notice that lag when it is only about 20 millionths of a second: in Thavam and Dietz’s 2019 study, 52 untrained adults aged 18–39 averaged 21.7 microseconds, enough to hear a sound shift by roughly 2.4° near straight ahead. This sound localization test measures how small a shift your own ears can catch.

    Through headphones, each trial plays two half-second bursts of low hiss. The first leans a hair to one side and the second the same hair to the other, so the sound seems to move, and you say whether it moved left or right. The lean is an interaural time difference (ITD), and the sound copies the best condition in Thavam and Dietz’s study. Fifteen rows run from 320 µs down to 2.5 µs; each right answer takes you down a row, the first miss gets a second try and the second ends the test. Your finest row is calibrated by simulation to the study’s 79.4%-correct threshold, so your microseconds compare directly with those 52 adults.

    It is not a hearing test and says nothing about the health of your ears. The study tested people with normal hearing in a lab, on insert earphones at 70 dB; other phones, headphones and volumes move the number, and one short run scatters by about a factor of two either way, so a few runs tell you more. Speakers, mono audio, a single earbud or Spatial Audio erase the difference between the ears the test depends on. And the comparison group is 18 to 39, while telling direction from timing usually gets harder with age.

    What the test measures

    One thing, measured like an eye chart: the smallest left-or-right move of a sound you can still hear.

    The task

    • Two 0.5 s bursts of 20–1,400 Hz noise through headphones; the second is shifted left or right of the first by an interaural time difference
    • 15 rows, from 320 µs down to 2.5 µs, the step halving every two rows
    • One right answer passes a row; the first miss is a second try, the second ends the test
    • A two-step headphone check first, with the sound in one ear only; swapped sides are fixed for you

    Your result

    • The smallest move you can hear, in degrees, your threshold in microseconds (µs) and your level (1–15)
    • The share of the 52 untrained adults in Thavam and Dietz (2019) you are sharper than
    • Reference points: untrained average about 21.7 µs, nine trained listeners about 8.4 µs, the best of them 5.0 µs
    • Your limit in everyday terms: how far someone talking 10 m away has to step sideways
    • A dial you can tap to hear a sound jump from straight ahead to any angle
    • A shareable card and a link that challenges a friend to beat your angle

    Before you start

    • Stereo headphones or earbuds, both in; speakers won’t work, because each ear needs its own channel
    • Turn off Spatial Audio, head tracking and mono audio
    • On an iPhone, switch off silent mode first
    • About three minutes

    Sound localization FAQ

    Is this test free? Do I need to sign up to see my result?

    Yes, it’s free, and there’s nothing to sign up for: no account, no email, no payment. Your result appears on this page the moment you finish, and you can share it or play again straight away.

    How good are human ears at locating sounds?

    Very good: young adults notice a timing difference between the ears of about 20 microseconds. Thavam and Dietz (2019) measured an average of 21.7 µs in 52 untrained adults aged 18–39 and 8.4 µs in nine trained listeners. With a loudspeaker in a quiet room, Mills (1958) found the keenest ears tell apart directions about 1° apart straight ahead.

    How does this sound localization test work out my score?

    The sound localization test takes the finest row you got right before your second miss and calibrates it, by simulation, to the 79.4%-correct threshold Thavam and Dietz (2019) used, so your microseconds compare with their 52 adults. The angle comes from Woodworth’s spherical-head formula. One run scatters by about a factor of two either way.

    Why do I need headphones for a sound localization test?

    Headphones send a separate signal to each ear, and the test works by delaying one ear by a few millionths of a second. Speakers mix both channels in the room before the sound reaches you, and mono audio, a single earbud or Spatial Audio does the same inside the device. A short one-ear check at the start catches most of these.

    Can you train your ears to locate sounds better?

    Practice seems to help a lot. In Thavam and Dietz’s 2019 study, nine listeners with hours of lab practice averaged 8.4 µs, well under half the 21.7 µs of the 52 untrained adults, and the best of them reached 5.0 µs.

    Does age affect sound localization?

    Usually, yes: telling direction from timing differences between the ears tends to get harder with age, and Ross and colleagues (2007) found signs of that decline beginning in mid-life. The comparison group here is 18 to 39, so the percentile fits younger listeners best.

    Is this a hearing test?

    No. The sound localization test measures one skill, hearing which way a sound moved, on your own headphones; it does not measure how loud a sound must be for you to hear it and says nothing about the health of your ears. If one ear seems to hear less than the other, or your hearing has changed, see an audiologist.

    Where the numbers come from

    Percentiles compare your calibrated threshold with the 52 untrained adults in Thavam and Dietz’s study, using the histogram of their thresholds in Thavam’s 2018 thesis; with only 52 people, the far ends read as “sharper than everyone in the study” or “bottom 5%”. Angles use Woodworth’s spherical-head formula (Woodworth & Schlosberg, 1954). The site does not publish its own players’ data.

    How the score is worked out

    Interaural time difference (ITD)

    The lag between a sound reaching one ear and the other. From right beside one ear it is about 650 µs; this test shrinks it down to 2.5 µs.

    Two-alternative forced choice

    Every pair has one right answer, left or right, so a guess is right half the time; that is why the finest row passed isn’t taken at face value.

    Calibration to 79.4%

    The finest row passed is multiplied by about 2.17, a factor from running simulated listeners through this exact chart, to give the time difference you would get right 79.4% of the time, the criterion Thavam and Dietz reported.

    Angle

    Woodworth’s spherical head (8.75 cm radius, sound at 343 m/s) turns your threshold into an angle around straight ahead. It is an illustration, not a second measurement.